G.T.

Markov

ANTENNAS

Progress Publishers Moscow

nt en na s

,,

EU
Progress Pu bllshers
First jni/JlisMd 1958
 CONTENTS

Introduction • • • • • • • • • • • • • • • • • • • • • • • 9
1-1. Purpose . . . . . . . ... ... . • • • • • • • • • • 9
1-2. £~axwe11 's Equations • • • • • • • • • • • • • • • • 10

Part One

GENERAL THEORY OF RADIO \\'AVE RADIATION AND

RECEPTION

Chapter One. Radiation from Elementary Sources . • • • • • • 15

l-1. Radiation of an Electric Dipole • • . . . . . • • • 15
1-2. Radiation of a Magnetic Dipole . • . . • . • • • • 18
1-3. Radiation of an Elementary Electric Loop • • . • • 19
1-4. Radiation of an Ele1nentary Magnetic Loop • • • • • 22
1-5. Effects of a Perfectly Conducting lnf1nite Plane on
the Radiation fro1n Elen1entary Sources . . . . . . . 23
1-6. Unidirectional Radiation . . . • • . • • • . • . • • 27
1-7. l{adiation of a Rotating Field • • • • • • • • • • .. 29
Chapter Two. Radiation of Dipoles of Finite Length . • • • . 31
2-1. The Electric Dipole . . . . . . . . . . . . • • • • 31
2-2. The Integral Equation Method . • . . • • • • • • . 32
2-3. Current Distribution in a Symmetrical Dipole in the
First Approximation . . . . . . • . . . . • . . 34
2-4. Field Intensity of the Dipole in the Radiation Zone 38
2·5. Calculating Power Radiated by an Antenna by the
Poynting Vector Method • . . • . . . • . • • • • 41
2-6. Field Intensity in the Vicinity of the Antenna • • • . 44
2-7. Calculating the Power Radiated by an Antenna by the
Induced E.~\\F Method. Input Resistance of a Symmetrical
Antenna . • . . . . . • . . . . . . . • . . . . . . 48
2-8. Calculating the Input Resistance of an Antenna by
Reducing It to a Homogeneous Long L inc \Vith Losses 53
2-9. Radiation of a Sy1nmetrical Magnetlc Antenna. Sym-
metrical Slot in Screen • • • , • • • • • • • • • • • 57

'
 2-10. Additional Remarks Concerning the Current Dis-
tribution in an Antenna . . . . . . . • • • . . • . 62
2-11. The E igenfunctions Method • . . • . . • • • • • • 68
Chapter Three. Radiation of Two Coupled Antennas • . • • . 77
3-1. Directional Diagrams of Antennas . . . . . • • • . 77
3-2. Mutual and Total Impedances of Antennas . . . • • . 83
3-3. Active Reflector . . . . . . . • . . . • . • • • . . 86
3-4. Passive Antenna . . . . . . . . . . . • • • • • . . 87
3-5. Calculating Mutual Resistance of Antennas . . . . . 89
Chapter Four. Radiator Systems and Antenna Parameters • . . 93
4-1. General • . . . . . . . . . . . . . . . . . • • . 93
4-2. Radiation of a Linear eo-Phased Dipole Array . . . 93
4-3. Radiation of a linear Dipole Array with a Variable
Phase . • . . . . . . . . . . . . . . . • . . . . . 99
4-4. Radiation Resistance of a Dipole System . . . . • • 102
4-5. Radiation of Continuous System of Sources with a
Travelling Wave . . . . . . . . . . . . . . . . . . 103
4-6. Application of the Equivalent Surface Currents Theorem
to the Calculation of the Radiation Characteristics of
Antennas . . . . . . . . • . . . . . . . . . • . . I 08
4-7. Radiation of an Ideal Plane Antenna . . . . . . . 112
4-8. Effect of Changes of Field Antplitude and Phase in a
Plane Antenna on the Directional Diagram . . . . . . J 16
4-9. Antenna Parameters . . . . . . . • • . • . . • . . 122
4-10. Definition of the Current Distribution in an Antenna in
Accordance with a Prescribed Directional Diagram . . 130
Chapter Five. The lnftuence of the Earth and Metal Bodies on
Antenna Radiation . . • . . . . . . . . . . . . . • 157
5-l. Radiation of a Symmetrical Dipole Located over Plane
Earth . . . . . . . . . . . . . . . . . . . . . . 157
5-2. Radiation of an J\!;ymmetrical Dipole ~ Disposed on
the Surface of Plane Earth . . . . . . • • . . . . . 162
6-3. The Influence of the Earth on the Radiation of
Multiple Antennas . • . . . • • • •• • • • . . • • • 164
5·4. The Influence of an Infinite Circular Cylinder on the
~. Radiation of an Electric Dipole . . . . . • • • • • . . 166
5-5. The Influence of an Infinite Circular Cylinder on the
Radiation of a Longitudinal Slot • • • • . . • • . • 171
Chapter Six. Receiving Antenna Theory . • • • . . . . . . . 177
6-1. A Symmetrical Dipole in the Field of a Plane Electra·
magnetic Wave . . . . . . . . • • • . • . . • . . 177
6-2. Power Dissipated in the Load of a Receiving Sym·
metrical Dipole. , • • • • • • • • • • • • • • • • • 180
4
 6-3. Current Distribution in a Receiving Dipole . . . . • • 182
6-4. Application of the Principle of" Reciprocity to the
Study of the Properties of Receiving Antennas . . . . 183
6-5. Eft'ective Area of an Antenna • • • • . . . . . . . . 187
6-6. Mutual Resistance of Receiving and Transmitting
Antennas . • • . . . . • . . • • . • . . . • . • . 188

Part Two
TRANSMISSION LINES OF I~ADJ() WAVES
Chapter Set.ten. Transmission Line Theory • • • • • • • • • • 190
7-J. Genera I . . . . . . • • . . . • • • • • • • • • • 190
7-2. Rectangular Waveguide Theory . . . . • . . • . • . HJ2
7-3. Circular Waveguide. Theory . . . . . . . . . . . . . 207
7-4. Brief Information R~garding Coaxial Lines . . . . •. 219
7-5. Single-Wire Line Theory . . . . . . . . . . • • . . 221
7-6. ()ther Transmission Lines . . . . . . . . . . . . . . 228

Chapter Eight. Methods for Matching the Line to the Load . . 234
8-1. Twin-Wire Line Equivalent to a \V.aveguidt! • . . • . 2~J4
8-2. Transmission Lines of Finite LPngth . . . . . . . . 2~i5
8-3. Narrow-Band Matching of the Line to the Load . . . 239
8-4. Broad-Band Matching of the Line to the Load . . . 244
8-5. Matching the Line to the Load in a Broad Frequency
Range • . • . . . • . . . . . • • . • . . • • . . . 250

Chapter N uze. Parameters and Elements of Transmission Lines 254

9-1. 1\vo- Wire and Four- Wire Lines . . . . . . . • • • . 254
9-2. Coa,,ial Lines . . . . . . . . . • . • • . . • • • • 260
9·3. W2·vegulde Lines . . . . . . . . . • . • • • • • • • 264
9-4. Switc~ing Devices and Duplexers . . . . • • • • • • 280
9-5. Devices E n1 ploy ing Ferrites . . • . . . • . . • • • • 288

Part Three
TYPES OF ANTENNA DEVICE~

Chapter Ten. Ultrashort-Wave Antennas • • • • • • • • • • • 306

10-1. Antenna Types . • . . . · • • • • • • • • • • • 306
10·2. Simple Antennas of the Wire Type • • • • • • • • 306
10-3. Simple Antennas of the Slot Type . . . . . • • • • 313
10-4. Construction of Simple Wideband Antennas ••• 320
6
 l0a5. Multi-Unit Antennas . • • • • . . . . . . . . . . . 334
10-6. Slot Waveguide Antennas . . . . • . . • • • • • • • 340
10-7. Director Antennas . . . • • • • • • • • • • • • • 349
10-8. Helical Antennas • . . . • • • • • • • • . • • • • 355
10-9. Dielectric Rod Antennas • . • . . . • • • • • • • 361
I Q.. J0. Horn Antennas . . . . . . • • . . • . • • • • • • 368
10-11. Slow Phase Velocity Impedance Antennas • • . . . • 377
I 0-12. Lens Antennas . . . . . . . . . . . . . . • . • 384
10-13. Parabolic Antennas . . . . . . . . . . • • • • • • 402
J0-14. Other Reflertor Antennas . . . . • • • • • • • • • 416
Chapter ElPven. Short-Wave Antennas • • • • • • • • • • • • 422
11-1. Classification of Antennas . . . . . • • • • • • . • 422
J 1-2. Simple Dipoles of the Tuned Type . . . . . • • • • 423
11-3. Simple Dipoles of the Multiple-Tuned Type . . . . • • 424
11-4. Corner Antenna . . . . . . . . . . • • . • • . • • 426
11-5. Multiple Short-Wave Antennas of the Tuned Type . 427
11-6. Multiple Short-Wave Antennas of the Multiple-Tuned
1~ype . . . . . . . . . . . . • • • . . . . . . • 431
11-7. Rhombic Antenna . . . . • . . . . • . • • • . . 433
11-8. Travelling-Wave Antenna • . . • . • • . . • • • • 441
J1 ..9. Diversity Radio Reception . . . . . . • • • • • • 450
Chapter Twelve. Me.dium- and Long-Wave Antennas . . • • • 451
12-l. Classification of Antennas • . • . . . . . . . . . . . 451
12-2. T- and L-Antennas . . . . . . . . . . • . . - . . . 451
12-3. U-Antennas and Antennas with Multiple Downleads 462
12-4. Mast Antennas . . . . . . . . • • • • • . • • • • 464
12-5. Slot Antennas on Low Supports . . . • • • • . • . 472
12-6. Wave Antennas • . . • . . . • . • • • • • . • • 474
12-7. Frame Antennas • • • . • . . • • • • . • . • . • 478
J2-8. Rad iogonion1ctcr3 ., .. . . . • • • • • • • • • . • 481
Appendixes • • . • . • • • • • • • • • • • • • • • • • • 484
Bibliography • • • • • • • • • • • • • • .. • • • • • • • • 509
 SYMBOLS INDEX

A-vector potential. /..--length, inductance.

a-radius, trans\'er~e dirncnsion /-length of dipole arn1, length
of wavPguide. of line segn1ent, light.
B- reactive conductivity, nutg· M-mutual inductance, rnagnct-
netic induction. isation.
b-tran~;ver.se din1ension of t\1---nlagnetontotivc force.
wavegu1dt'. tu-~- ratio of current n1oduli,
C- electric capacitance. nurnber of dipoles in array
t• 1-- velocity of light in vacuun1. row, .rna~nctic JJJon1ent:
!•--directive gain, electric n- nun-1ber of rows in array,
induction, distance be- nurnb~r of windings. coef-
t wecn wires, dimension of Jicient of refraction.
antenna (out put) opening. P- power, polarisation of
d- dian1eter, distance bet\vcen 1nediun1.
antennas, distance between p- reflect ion coefficient, electric
dipoles. n1on1ent oi
dipole.
F.- electric field intensity. Q- Jinear charge, quality.
fi- electromotive force. q-surface density of electric
e-electric field intensity, charge.
electric. R- active resistance, distance,
F-focus, fr~quency. radius.
{-frequency, focal distance, r-distance, radius.
reflection coefficient. S- Poynting vector.
G- active conductance. s-area.
H-magnetic field intensity, T-period.
height. t-time.
h-height, propagation con- V- voltage.
stant, magnetic field V- volume.
intensity, effective length. v- phase velocity.
/-linear current. W-wave impedance.
i-imaginary unit. X- reactance.
J-current surface density. Y-total (complex) con·
j-current volumetric density. ductance.
K-travelling-wave ratio. z- (complex) impedance.
k-wave number (phase coef- a- coefficient of phase, angle,
ficient) for free space. electric (magnetic) polar·
k1-surface capacity factor. isabiJity of field.

1
Ji- atten uatio n constant, angle. 6-- angle between wire axis
y- propagation constant, angle. and direction towards point
A- angle. of observation.
6- angle of losses. 6- angle.
8 - perm ittivi ty or medium, A.- wave -leng th.
anten na amplification J.L- perm eabil ity of medium.
coefficient. ~-velocity of light to phase
q- efficiency factor. velocity ratio.
Q- volumetric density of charg e, <p- angle .
resis tivity . ..p- phase shift angle.
a-sp ecifi c cond uctan ce. eo-·-~ angular frequency.
 Introduction

1-1 Purpose

Antennas are essential components of transmitting and

receiving systen1s. They are coupled v.rith the transn1itter
or the receiver through feeders or waveguides. Sornetirnes
this coupling is direct, requiring no feeder devices.
1·he high-frequency electromagnetic ·osci Jlations excited
by the transmitter are conveyed along the feeder system as
plane waves and on reaching the antenna they are trans-
formed by the latter into free-space spheric a I waves. The
plane \vaves associated \Vi th the feeder line transfer electro-
nlagnetic cnc.'rgy from the transmitter to the antenna. The
free spherical waves generated by the antenna carry electro·
magnetic energy frorn the antenna into the infinity, i.e.,
their waveguide is free space. The transrnitting antenna is
said to radiate e]ectromagnetic energy.
The electron1agnetic waves travelling in space excite in
the receiving antenna oscirlations of high-frequency currents.
Plane waves are excited in the f~eder system coupling the
antenna with the receiver and are conveyed to the receiver.
These waves carry high-frequency energy exciting the receiv-
ing set. The receiving antenna is said to receive electro-
tnagnetic energy. Note that only part of the power reaching
the receiving antenna is transmitted to the receiver.
The rest of the energy is radiated back into surrounding
space.
The main purpose of a transmitting antenna .. feeder system
is to transmit electromagnetic energy from the transmitter
to the antenna and then into surrounding space with maxi-
9
mum efficiency and without noticeable distortions of the
modulating signal. The other purpose is to distribute the
radiated energy in space in a definite pattern. In one case,
the energy is to be radiated uniformly in aiJ directions;
in another case, it is to be radiated in a sufficiently narro\\'
beam. The receiving antenna and the feeder line coupling
it with the receiver should also operate at a maximum
efficiency and possess a sufficiently wide pass-band. The
response of the antenna to the waves arriving from various
directions should follow a prc&cribed law, i.e., it will just
like a transmitting antenna, possess a definite directivi ty.

1-2. Maxwell 's Equations

Modern antenna theory is based entirely on the basic
equations of electromagnetic theory, viz., Maxwell 's equa-
tions. These equations are the generalisation of data gained
by experience and their validity has been confirmed by
practice.
Throughout this course we shall always be concerned
with such electromagnetic processes as are harmonic in time,
or vary in time as sin rot; or as e,wt in complex form.
Furtherrnore, we shaJ I write the vector of the instantaneous
•

v a Jue of the eJec tr ic fie Id intensity as e= I m (Eeiwt),

where E is the fieJd con1plex amplitude; for sitnplicity,
complex amplitudes will be written further on without dots
above the letters. Complex electromagnetic waves may be
regarded as sums of harmonic asci 11ations, i.e., they may
be expressed in the form of a Fourier expansion.
We shall keep to the rationalised practical metre-kilogranl-
second-couloinb system of units. \\1c sha1J consider a horno-
geneous and isotropic medium assuming the existence of
a given distribution of the exciting (external) electric and
magnetic currents for some of the regions.
Under thesf conditions Maxwell's equations may be
written in di fferentiaJ form as:
rot H = iroe'E ~- je;
rot E = - iroJl H- jM, (I-I)

where E is the complex amplitude of the electrical field

intensity. in volts per metre;
10
 H, the complex amplitude _of the magnetic field
intensity, in amperes per metre:
s' =s( 1-i~), the complex dielectric constant of the medi urn;
e, the dielectric constant of the n1ediun1, in farads
per metre; in vacuum:

fl, the permeability of the medium, in henries per

n1etre; in vacuum ..-,0 =4n-I0- 7 h/m;
a, the conductivity of the medium, in mhos per metre;
je , the complex amplitude of the external electric
current density in amperes per square metre;
j~, the complex amplitude of the external magnetic
current density, in volts per square metre.
The external magnetic current is a fictitious quantity,
since magnetic charges are non-existent in nature. Ho\\'ever,
the introduction of this concept makes for a relatively simple
analysis of the radiation of slots in conducting screens.
The relations (1-1) are usually supplemented with the
following equations:

divE=~~
8

divH=~, (1-2)
I'
where Qand m are space densities of the electric and of the
magnetic charge respectively.
The equations (1-2) are derived from the equations (1-1),
for the continuity equations are always valid
div je + iiDQ=OJ
div JM+ irom='O. (1-3)
Note that for free space, the quantity
1 1
Vl= Jf =3X 10
Eof'o

11
Is expressed in metres per second representing the velocity
of light in vacuum; and then the quantity

w = .. /~= 1203t~376.6
0 r 'eo

is expressed in ohms representing the wave impedance of

free space. Note also that 1>.

1_1 m 2n
ko=CJ.J Y 8 J.t = -= i l
o o vl "'o

is the wavt! number (phase coefficient) and A.0 is the wave-

length (in metres) of the waves travel ling in free space.
IncidentaJiy, it follows fron1 the equati 1ns (1-3) that in
defining fields one may proceed frotn the current distribution
alone, since charges can be readily defined provided the
distribution of the currents in the radiator syste1n has been
given. This is precisely the way we shall proceed trom
now on.
Two auxiliary vectors are usually introduced to solve
MaxweJI 's equations (I-I): the vector potential of the
electric currents Ae and the vector potential of the magnetic
currents AM. Furtbermore, the electromagnetic field vectors.
E and H, are determined by means of the auxiliary vectors
through these equations:
1
E=-ictlJ.tAe+.l(J)£ , grad divAe-rotAM;

(1- 4) •

Substituting (1-4) tnto (J-1), we obtain the following

vector heterogeneous wave equations for the auxiliary
potentials •
t,.Ae + kiAe=- .iel
flAM + k 2 AM= jM, (1-5)
where
4A=grad divA -rot rot A;
k = ro Ve'JJ..

In defining the radiations from antenna systetns, the inte-

gration of MaxweiJ's equations may thus be reduced to the
12
integration of the vector heterogeneous wave equations
(I -5).
It will be reminded that the solutions of MaxweJJ's equa-
tions are unique, when:
a) they satisfy the boundary conditions on the surface of
discontinuity;
b) they satisfy the infinite boundary conditions; in other
words, they follow the law of radiation at infinity;
c) they are finite throughout, provided they refer to either
a space or a surface distribution of exciting currents. If the
electromagnetic field is excited by a linear distribution of
currents, the solutions of MaxweJI's equations will possess a
dipole singularity, i.e., the field must tend towards infinity
as the point of observation is brought nearer to a linear
radiating current.
The boundary conditions on the surface of discontinuity
of two media are expressed as the continuity of the tangential
con1pone~ts of the intensity of the electric and n1agnetic
fieJds between tnediun1 1 and 1nediuru 2
( 1-6)

On a perfectly conducting surface the boundary conditions

assume different form. On the surface of an ideal conductor
the tangential component of th, electric field intensity
vanishes altogether while the normal component is equal to
the ratio of the electric charge surface density to the die-
lectric constant of the medium surrounding the conductor:
q
~ - e.
En -- (1-7)

As for the magnetic field intensity in the same case, its

normal component vanishes while the tangential component
is equal to the surface density of the electric current:

The boundary conditions at infinity amount to this:

the electromagnetic perturbation wiJJ move away from the
exciting sources into infinity as traveiJing waves. There
can be no waves travelling from infinity to the exciting
sources.
13
 Let us mention also the complex form of Poynting"s
theorem:
{- S[EH*]" ds+ iro S("'~ 1•- sJ~t•)dV +
I V

+ saf:fl dV = ~ s(-jMH*- je*E) dV, (1-9)

V V
where n is the external normal to the surface s enclosing
the volume V which contains the exciting sources.
The right-hand side of this equation defines the complex
power output which is taken from the generators. The
first term in the left-hand side defines the complex power
as leaving the volume V, the second term defines the reactive
power stored in the volume V and the third term defines the
power dissipated as heat in the volume V. The power leaving
the volume V is described by the complex Poynting vector
S=~[E, H*]. (l-10)
 PART ONE

General Theory of Radio

Wave Radiation and Reception

CHAPTER ONE
Radiation from Elementary Sources

1-1. Radiation of an Electric Dipole

We shaJJ begin this course of "Antenna Theory" by
investigating the radiation from elementary sources, i.e.,
sources whose geometrical dimensions are small in comparison
with the wave-length of the oscillations which they excite.
Furthermore, we shall assume that the reader is famiJiar
with the Hertz dipole theory.
Let us assume that we have a boundless homogeneous
space. Let the vector distribution of the density of the ex-
citing electric or magnetic current j(x', y', z'), be given for
a certain limited volurne V of this space. The vector potential
of these currents at a certain spatial point P(x, y, z) is
kno,vn as: ,
A(x. y,
I
z)= 4n
sj(x', y', z') e-illl'
r dV, (1-1)
V
where
r= V(x-x') 2
+ (y-y' ) + (z-z')•.
1

The expression (1-1) is the solution of the vector wave

heterogeneous equations (1-5). The solution is unique since
it satisfies the boundary condition at infinity and has a
finite va Iue throughout.
By substituting this expression into the equations (1-4),
we can define the vector of the electric field intensity E
and the vector of the magnetic fieJd intensity H for any point
of the space.
15
 Let the exciting current be electric and directed along
the z-axis of Cartesian coordinates. Let us suppose that the
current occupies an infini tesimaJ volume of space at the
origin of the coordjna tes, i.e., Jet there be an electric Hertz
dipole with a momentum lel, where /e is the current of the
dipole in amperes and lis its length
z A" 'Z in metres (Fig. 1-1 ). In this case
the vector potential will have only
the z-axis component and we shall
write the relation (1-1) as:
11 e _1~1
_e-ilu
__
Az- 411 , , (1-2)

where
r= VX + 1J
2 1
-~ Z1 •
1-1. An electric We shal1 now turn our attention
dipole. to the field of the dipole in the zone
of radiation (r~ A) and, by substi-
tuting (1-2) into (1-4), we shaiJ describe the zone of radiation
in the spherical coordinates
.. r, 6, cp.·

(1-3)

where 6 is the angle between the axis of the dipole and the
direction to the point of observation of the field.
lt follows from (1-3) that a) the Hertz dipole radiates
travelling waves moving away into infinity at the velocity of
light in the given medium; b) the surfaces.of equal phases of
these waves are spheres with their centres coinciding with
the position of the dipole; c) the amplitudes of electric and
magnetic fields intensities vary inversely proportional to
the distance between the point of observation and the
dipole; d) the vectors of the electric and magnetic fields are
mutually perpendicular, being perpendicufar to the direc-
tion of propagation of the wave and oscillating in phase;
e) the vector E lies in a plane passing through the dipole
(in a meridian plane), the vector H lying in an azimuth
plane; furthermore, the rat iD oi. J.he.ir....maa.itudes is equal
M

16
to the wave Impedance of the space
Ell k=
-11 ~ = -
we
Y" ohrns.'
-
e
'

In contradistinction to the vector potential, the magni-

tudes of the electric and magnetic fields intensities depend on
the angle of observation a. Because of axial symmetry, the
field does not depend on the angle of observation <p. The
graphs showing the field intensity magnitude plotted as a
function of the angle of observation are known as directional
s=o· ,=o·

8=270

Fig. 1-2. Directional diagrams of the dipole.

diagrams (characteristics). We see that in the meridian

plane (in the plane of the vector E), the dipole directional
diagran1 is a sinusoid plotted in polar coordinates. 'fh is
curve may conventionaJiy be termed a "figure of eight".
In the azimuth plane (in the plane of the vector H), the
directional diagram of the dipole represents a circle
(Fig.l-2). Thus, the Hertz dipole radiates the maximum energy
in a direction perpendicular to its axis, radiating none
along the axis.
The power radiated by the Hertz dipole can be defined
using the Poynting vector procedure as follows
(1-4)

where Rr.d is the dipole radiation resistance, expressed

for a lussless medium as:

R>:d == ¥ v~ (~ r. (1-5)
where A. is the wave-length for a given medium.
2-.2122 17
 In the case of vacuum, the expression becomes:
R:td=20 (k 0 i)I=80n• (~Y. (1-6)
where A0 is the wave-length for vacuum.

1-2. Radiation of a Magnetic Dipole

Now, let the exciting current be a magnetic dipole with
a moment JML, Jying at the origin of the coordinates and
directed along the z-axis (Fig. 1-3). The vector potential
of the dipole is given by:
(1-7)

In accordance with the expressions (l-4), the components

of electric and rnagnetic fields in the radiation zone are
. /M lk . e-ikr
E =-t sin 6-
' 4n r
• /Mlk2 • e-ik,.
H9=t stn 0 · (1-8)
4nrofl

We see that a magnetic dipole, just like an electric one,

radiates spherical waves which move away into infinity
z with the velocity of light.
The directional diagrams of
a magnetic dipole coincide
with those of an electric one,
i.e., its radiation is maximal
in the equatorial pJane and
zero along the axis. We see
~---t--+----.-.u that in the meridian plane
there are electric I ines of force
in the case of an electric
dipole, and there are 1nagnetic
1ines of force in the case of
a magnetic dipole.
Fig. 1-3. A n1agnelic dipole. Con1paring the radiations of
the electric dipole with those
of the 1nagnetic one we can easily bring out the following
rule: if the moment of the electric dipole current is equal to
that of the magnetic dipole current, the magnetic field of the
electric dipole is equal to the electric field of the magnetic
 dipole and of reverse sign. Furthermore, the electric field
of the electric dipole differs from the magnetic field of the
magnetic dipole by lf··=~ E
times and of the same sign. This
constitutes the interchangeability of fields produced by
elementary dipoles of electric and of magnetic types.
The po\ver radiated by the magnetic dipole, calculated
after Poynting is evidently expressed as:
J
pI= 2JM 2Gtd, ( 1-9)

where Gl:d is the radiation conductivity of the n1agnetic

dipole expressed as
2n , /'8 ( 1 ) •
Gl:d=3 V ~ X" • ( 1-1 0)

By comparing (1-5) and (I- !0) \Ve shall find that

Rr.d
al:d = w• , <I- 11>
where W= ·{~is the wave impedance ~f the medium.
The expression (1-11) enables us to define the radiation
conductivity of a magnetic dipole provided the radiation
resistance of an electric dipole of the sarne length is known.

1-3. Radiation of an Elementary Electric Loop

Let there be in an unbounded space a circular electric
loop of radius a with a uniformly distributed current If •
Let thr centre of the loop coincide with the origin of the
coordinates and the axis of the loop be directed along the
z-axis. In spherical coordinates, the loop current will then
only have the <p-component. Let u.s calculate the field of the
loop in the zone of radiation. Referring to Fig.l-4, let us
choose on the loop two elements of current If adcp symmetri-
cally relative to the x-axis and, considering these current
elements as electric dipoles, let us write the expression of the
vector potential caused by them at point M of the xz-plane.
Furthermore, it should be noted that the potential caused
at point M by the x-components of these current elements
is zero, as the difference in distance between the elements
and point of observation is zero, with the current element
2* 19
components being of opposite signs. The radiation may be
due only to the y-axis components of the current elements;
their vector potential will be doubled and have the com-
ponent
e _ 2/j cos cp ad 4p e-iler
dAcp--~4n-- r • ( 1-12)

It should be noted here that, since the point of observation

of the field is at infinity the beams from all the loop elements
z N
-;1\~
I
t
0
I
z Ho I
-----L

-
----'
Fig. 1·4. Explaining the calculation
of the field of an electric loop.

can be considered parallel and the field phase will be calcu-

lated from the loop centre as the basis. We may then assume
with a high degree of accuracy that in the expression (1-12)
•
I I
-~-·
r r0 '
'~'o -a cos cp sin a.
Substitutini th~se expressions into (1-12) and sumn1ing
up the radiations of all the loop elements, we shall find
that

( 1~ 13)

20
 Now, Jet us assume the Joop radius to be s1naJI in comparison
with the wave-length ka~I. Then we may state that in
( 1-13)
etka cos cp stn $~ 1 + ika cos <p sin e.

Integrating (1-13) under those conditions we shall find

that
e . Jjka 2 • e-ikr0
Acp~t 4 stn e 'o .
Further, taking into account that fron1 (1-4) div Ae=O we
have
Jjc.oJ..&.ka 3 • e- ikf0
E, = 4 Sln e 'o . ( 1-14)
1"he rnagnetic field intensity is then expressed as
E,_
H8 -
- - ,__ • ( 1-15)
y~
We see that the radiation of an electric loop is sin1i Jar
to that of a magnetic dipole. On cotnparing· (1-14) and (J-8),
\\'C obtain:
I M l = iCJ.)~ Ifs' ( 1- 16)
\vhere s=na2 is the area of the loop.
An electric loop of sn1aJI radius may thus he regarded as
a n1agnetic dipole whose axis coincides with that of the loop
and whose moment is defined by (1-16).
The loop gives rise to maximal radiation in its plane
and radiates no electromagnetic energy along its axis,
besides the field produced by this loop has the same polari-
sation as the magnetic dipole field.
The radiation resistance of the electric loop is defined
as the ratio of the radiated power to the square of the
effective value of the current in the loop
2Pz
Rl:l =I'~ 11 .

But according to (1-16) 1n1=~:!. From (1-9), we shall

write that
(1-17)

21
or, substituting here the expression (1-10), we obtain
8 ,/f! n 1s1 (1-18)
Rr-, = 3 V 8 A.• • I

For vacuum, this formula becomes

320n4s1
R El= ) • • ( 1-19)
·o

The expression above is usuaiJy applied to the calculation

of the radiation resistance of loop antennas. It should be
pointed out that whereas the electric dipole radiation
resistance is inversely proportional to the square of the wave-
length, the radiation resistance of an electric loop of small
radius is inversely proportional to the fourth power of the
wave-length.

1-4. Radiation of an Elen1entary Magnetic Loop

Now, let us consider an elementary magnetic loop with a
uniformly distributed current /~. Let the centre of the loop
coincide with the origin of the coordinates and let us
direct the loop- axis along the z-axis. By analogy with the
electric loop, we may immediately write the expression for
the magnetic loop field intensity in the zone of radiation

(1-20)

Comparing (1-20) with (1-3), we obtain

· JM s. •
l el = - trot ( 1-21)
1

In this case the elementary magnetic loop rnay be regarded

as an electric dipole with the current moment defined by
(1-21). The loop radiates a field which has the same polari-
sation as the field of an electric dipole and its directional
diagram coincides with that of a dipole.
We may introduce the concept of radiation conductivity
of the loop, which is defined as the ratio of its radiated power
to the square of the effective magnetic current. From
22
(1-21) we can write that the conductivity of the loop
2P l: ooze•sz rozs.Js•
GI 1 = I/ e JZ [I = Rl:d - ll • (1-22)
Substituting here expression (1-5), \\'e obtain
8 , le n 1 s2
G:~:, =3 V ~ A.• • ( 1-23)

Comparing (1-23) with (1-18), we rnay \\Tite

R~,
GI.t = \W2• (1-24)
where Rl:I is the radiJtion resistance of an electric loop
of the same radius.
We see that the radiation conductivity of a magnetic loop,
ts related to the radiation resistance of an electric loop in
the same way as the radiation conductivity of a magnetic
dipole is related to the radiation resistance of an electric
dipole. ·

1-5. Efl'ects of a Perfectly Conducting Infinite

Plane on the Radiation from Elernentary Sources
We have been so far considering the radiation of dipoles
and loops in an unbounded space.
Let these antenna dipoles and loops now lie at a certain
height above a perfectly conducting infinite plane h (Fig.l-5).
Due to the electromagnetic field of the antenna dipoles
and loops surface electric currents are induced on the plane.
Let us consider the direction of these induced currents. In
the case of a horizontal electric antenna the surface currents
flow parallel to the current in the antenna dipoles. In the
case of a vertical electric antenna and a horizontal magnetic
loop, they flow in radial directions. In the case of a hori-
zontal magnetic antenna dipoles, the induced currents have
components parallel to the antenna axis as well as those
perpendicular to it. FinaJJy, in the case of a vertical magnetic
dipoles and a horizontal electric loop, the induced surface
currents have azimuth components.
In the upper and lower half-spaces of the plane the induced
surface currents give rise to secondary fields which (once
the plane is considered infinitely thin) are symmetrical in
relation to this plane. Since the plane is a perfect conductor
23
and infinite, it screens the lower half-space, and the induced
currents wilJ be so distributed that the secondary fieJd at
any point of the lower half-space will be equal in magnitude
to the primary field produced by the dipoles and loops
in that haff.. space and of opposite sign.
This leads us to what is known as a mirror method, in
which the upper half-space secondary field will be assun1ed
to remain the sarne if we remcv~ the conducting plane and
place in a mirror point a source of such current as is equal
in rnagnitude to that of the real source and is so directed
that the tangential component of the total electric field
vanishes on the surface of the plane under consideration.
In the case of a horizontal electric dipole, a horizontal
electric loop and a vertical magnetic dipole, the current of
the mirror image, as shown in Fig. 1-5, a, b and c, flows in
a direction opposite to that of the current of the real source.
In the case of a vertical electric dipole, a horizontalrnagnetic
loop and a horizontal magnetic dipole, as shown in

111 l'
T • '..
,ll
Hone, coinciding
-----------------~~
/1
l __ _ -·
(0) (/JJ {C)

if If'
• • •
Plane,
------~-~--~-----~~
h
j_
'I
I
-·- -~-

(d) (I) (f)

Fig. 1·5. Direction of current in mirror image.

Z4
Fig. 1-5, d, e and f, the current in the mirror image flows
in the same direction as that of the real source.
When the source is just above the plane, in the case of a
horizontal electric dipole, a horizontal electric loop and a
vertical magnetic dipole, the primary and secondary fields
are of equal magnitude and of opposite signs, the total field
is equal to zero and the radiation vanishes. And inversely,
in the case of a vertical electric dipole, a horizontal magnetic
loop and a horizontal magnetic dipole, the primary and
secondary fields are equal in magnitude and sign and the
total field is double that of the sarne source in free space.
As for the resistance and conductivity of the radiation of
the sources, they are ~qual to zero in the first case and
doubled in the second case. The doubling of the radiation
resistance and conductivity is connected with the fact that, tn
that case, the density of the radiated energy at every point
of the space is quadrupled, whereas the power is radiated only
in the upp~r half-space.
Thus, the radiation resistance of a vertical electric dipole
of length l lying on the surface of a perfectly conducting
plane is expressed as:
4n , /Ji ( f"l ) .
1

Rid = 3 Jl 8 ( l-25)
The radiation conductivity of a horizontal n1agnetic
loop lying on the plane is expressed as:
16 , le n 3s2
Gl:t = 3 J1 -~ 1.,4 ( l-26)
and that of a horizontal rnagnetic dipole lying on the plane
as:
/e
4n. , ( I )
G:~:d= 3- V Ji ;: .
1
(1-27)
The directional diagrams in a vertical plane of a vertical
electric dipole, a horizontal magnetic loop and a horizontal
magnetic dipole lying on the surface of the plane under con-
sideration (h=O) are shown in Fig. 1-6. As can be seen, the
vertical electric dipole and the horizontal magnetic loop
radiate maximum fnergy along the surface of the plane and
give no radiation in a direction perpendicular to the pJane.
The horizontal magnetic dipole has an undirected radiation
in its equatorial pJane and a radiation in the form of half
a figure of eight in the meridian plane.
25
 The horizontal magnetic loop and horizontal magnetic
dipole lying on the surface of an infinite perfectly conducting

•
(D) (b)

(C) (d)
Fig . 1·6. Directional di agran1s in vertical plane:
of a hori zont al
a) of a vert ical elec tric dipo le; b) of a hori zont al mag neti c loop; c)zont
mag neti c dipo le in the equa toria l plan e of the dipo le; d) of a hori al magne1 ic
dipo le tn Its mer idia n plan e.

plane examined above represent, in fact, radiating slots

cut in a conducting screen. In practical work these slots
may be excited through a coaxial line and a rectangular
waveguide, as shown in Figs. 1-7 and 1-8.
The magnetic- current of these antennas is but the
difference of potential (voltage) between the edges of the slot.

•
Fig. 1-7. Exc itat ion of an ann ula r Fig . 1-8. Exc itat ion of a slot
slot by a coa xial line : by a wav egu ide:
J-sl ot In the screen; 2-e lect rlc lines of J-sl ot In the screen; 2-el ectr lc lines
force; 8-co ndu ctln g screen; 4-co axla J of force; 3-co ndu ctln g screen;
line . 4-re ctan ,ula r wavegulde.

The ratio of the power radiated through the slot to the square
of the effective value of the voltage of the slot determines
the radiation conductivity of a slot antenna. Thus, the radi-
ation conductivity of an elen1entary annular slot antenna
may be determined from (1-26) and the radiation conduc-
tivity of an elementary linear slot antenna from (1-27).
The slots cut in a conducting screen may also be excited
by a twin feeder (Fig.l-9). Slot antennas of this kind radiate

,2, ~--

Fig. 1-9. Excitation of bilateral slots:

1-stot Jn the screen. 2-twln feeder.

in both half-spaces and are ea lied bi latera.l slot antennas,

whereas the slot antennas sh·own in Figs. 1-7 and 1-8 radiate
in one half-space only and
are known as unilateral.
Evidently, the radiation
conductivity of bilateral
slot antennas is double that
of their unilateral counter-
parts.
/ It should be noted that
/, thesimpleexpressionsgiven
/l'L above for calculating direc-
-- - - - ~--~~--~---..-.11 tional diagrams as well as
the radiation resistance and
conductivity of elementary
wire and slot antennas are
valid only in the case of
r a perfectly conducting plane
f(g. 1-10. Explaining the calcula- screen of infinite extension.
tion of a unidirectional radiation. The mirror method can
be used and simple expres-
sions for the radiated field can be obtained only in that
case. In all other cases, the influence of conducting screens
on the radiation of antennas is more ,complex.

1-6. Unidirectional Radiation

Let us consider the radiation in free space of electric and
magnetic elementary dipoles in superposition. Let the
electric dipole be directed along they-axis and the maenetic
J7
dipole along the z-axis (Fig. 1-10). Let the currents in the
d i pole.s osci Ilate with the same frequency ro. Let us consider
the radiation fields of the dipoles in the xz-plane. This
plane is the equatorial plane of the electric dipole and
therefore from (1-3), its field is given by

In accordance with (1-8), the field intensity of the magnetic

dipole at the same point of observation is:
1r: lk e-ikr
EM=-i
cp 4n
sin 8--.
r

Since the directions of the field vectors of the electric and

magnetic dipoles coincide, the magnitude of the tota J fie) d
at the point of observation is given by the arithmetic sum
of the magnitudes of the field vectors

. I~ lk" e-ikr
E f4> =t-4nooe
·
I
-
/~roe .
- stn
e • ( 1-28)
r le k
- u
Let the excitation intensity of the dipoles be such that
the atnpli tu de ratio of their currents is defined by the ratio

-'~--
le -
_!!__-- , /~
WE - V 8 '
11

i.e., it is equal to the wave impedance of the surrounding

medium. -
Then, the radiation of superposed elen1entary dipoles,
electric and magnetic, for the xz-plane i.s given by
le lk 1 - ikr
EttJ = l 4nCil8
T
11
e r (1 +· sin 8). (1-29)
We have obtained an expression to define the unilateral
radiation of such a system. In fact, in th.e positive x-direc-
tion, \Vhen 8=90°, the radiation of the dipoles is double
that of one of the dipoles in that direction. In the negative
x-direction, when 8=-90°, the radiation of the dipoles
equals zero.
28
 The directional diagram of so superposed electric and
magnetic elementary dipoles is shown in .Fig. 1-11, has the
form of a cardioid. In the xy-plane, the directional diagram
z of the dipoles will likewise
have the form of a cardioid.
A practical exan1ple of
such a ~ystent rnay be

Fig. 1·11. lJtrectional dia- Fig. 1~12. Cardioid

gram of an electric dipole antenna:
and a magnetic dipole in /-dipole~ 2-Joop
su perposition.

afforded by an antenna arrangement consisting of an el£~ctr:c

dipole and an electric loop superposed as shown in Fig. 1-12.
Antennas of this kind, known as cardioid antennas, are
utilised in direction finding.

1-7. Radiation of a Rotating. Field

Let us consider the radiation of two electric (or n1agnctic)
dipoles lying in space at right angles to each other and
excited with their currents 90° out of phase. Furthern1ore, the
an1plitudes of the currents in the dipoles are assumed equal.
To begin with, let us consider the radiation field in the
plane of the dipoles in the yz-pJane (Fig. 1-13). In that plane,
the field vectors of both electric dipoles coincide in direction
and in accordance with (1-3) they are ex pressed as:

But 1:= 1: e-19oo, hence the total field will be

l~lk
1
E . e-ilv -l'l
a= t iiiws , e • ( 1-30)

29
 Thus, in the yz-pJane, the amplitude of the field does
not depend on the angle e (the directional diagram is a
circle) and the phase variation of the field in that plane
follows th.e travelling-wave law, which is indicated by the
factor e-;a.
In the x-direction, the field vectors of both dipoles are
at right angles to each other and have a 90° phase difference.
Consequently, the end of the vector of the total field rotates
z

,
Ea
,
£'

Fig. 1·13. Illustrating the calculation of the radia·

tion of a rotating Held.

in a plane perpendicular to the direction of wave propagation,

i.e., a rotating field \vith a circular polarisation is radiated
in the x-direction.
In all other cases, the crosslike arrangement under
consideration radiates a rotating field with an elliptical
polarisation; in particular, in the xy-plane, the radiation
with a circular polarisation in the x-direction turns into
a radiation with an elliptical polarisation in a direction form-
ing a certain angle q> < 90° with the x-axis and then, into
a radiation with a linear polarisation in they-direction.
We have therefore an elementary arrangement radiating
continuously in all directions; there is no direction in which
the radiation would equal zero.
 CHAPTER TWO
Radiation of Dipoles of Finite Length

2-1. The Electric Dipole

Let us consider the radiation of what is known as a
symmetrical electric dipole, which is the simplest antenna and
very often an essential element of various complex antenna
systems.
A symmetrical dipole is a rectilinear cylindrical conductor
of length 21 and radius a, fed in the centre by a high-frequen-
cy oscillator (Fig. 2-1). ThP high-frequency oscillator n1ay
be connected int.> the antenna by means of, for example,
a twin feeder.
t

.re--
~7
Fig. 2-1. A symmetrical dipole.

The oscUiator emf will then induce in the dipole electric

currents distributing them so that the electromagnetic
field set up in the surrounding space may satisfy both Max-
\\'ell 's equations and the boundary conditions. If, as we shall
henceforv1ard assume, the dipole is a perfect conductor
(conductivity a= oo), the boundary conditions on the dipole
surface are such that the tangential component of the
electric field intensity is equal to zero everywhere but in the
region of the external ernf. In the region of the external emf,
81
I.e., where the high-frequency oscillator exerts its effect,
the sum of the tangential components of the external emf
and of electric field intensity is zero. ,
The problem of defining the electromagnetic radiation
consists, first of all, in relating the current in the dipole
with the electromagnetic field set up by this current. If
the distribution of the current in the antenna is given, the
electromagnetic field can be readily defined from the
equations (1-1) and (1-4). However, the distribution of the
current along the dipole is not known beforehand, and
therefore the problem of defining the field in space becomes
considerably more complex.
From a strict electrodynamic approach, there are two
methods for solving the problem of the dipole excitation, viz.,
the integral equation method and the eigenfunctions method.
We shall give a brief description of the former.

2-2. The Integral Equation Method

Let there be, in free space, an infinitely thin magnetic
loop in the form of a ring of radius a and width b.ln accordance
with (1-1), the vector potential of this loop is defined as:
..
A
M 5
1
= 4n J
M e-ikR
/? ds, (2-1)
s
where the integration is done over the loop surface and JM
is the vector of the surface magnetic current density.
In the problen1 under consideration, the magnetic
current in the cylindrical coordinates r, q>, z has only an
azimuth component JM=J:. Calculating the x- and
1
y-components of vector Af\1 -from (2-1) and taking into
account that J~=-,/~sincp' and J~=JMcoscp', and
then that A~=- A~ sin q> +A~ cos q>, we obtain the
following expression for the azimuth component of the
vector potential of the magnetic loop:
b
+-
1 tn

A:(r, 'fl, z)=ln. 5 5 J:(a, 'P· z')cos(cp'-cp)x

b Cl>' =0
1'=--
2
-ikR
X eR adz' dq>', (2-2)

82
where
R= V (z-z')1 + r* +· a•- 2ta cos(cp' -q>)
Is the distance between the point of observation (r, q>, z)
and the point of the sources (a, <p', z').
Since the magnetic current J:
does not depend on the
angle q>, the vector potential has only an azitnuth corn·
ponent, which, likewise, does not depend on the angle cp.

l l

Fig. 2-2. Illustrating the integral equation

method:
1-magnetlc loop; 2-conductlng tube

The intensity of the loop electric field is found from

the expression EM=- rot AM having both a radial corn·
ponent and a longitudinal component.
M oA:
E, == iJ
z
;
M 1 0 M
E, == - 7 a,(rAcp);
E:=o. (2-3)
Let us place a perfectly conducting cylinder of radius
a and length ,21 symmetrically in the magnetic loop field.
Under the effect of the magnetic loop field there wiiJ be
induced in the conductor such surface electric currents as
will have longitudinal components 1:
on the cylindrical
conductor surface and radial components on the conductor J:
end plates.
To simplify the calculations and avoid taking account of
the existence of electric currents on the end plates, we may
regard the conductor as a tube with Infinitely thin walls
(Fig. 2-2). The induced currents will have then only longi-
tudinal components. ftowing along the outer as well as the
inner surface of the tube. The vector potential of the
induced currents has only a loneitudinal component and is
3--2122 88
expressed as: ,
A: (r, q>, t) = -Jn
+I 1ft
1:na(z') e-lkR , ,
R adz dcp , (2·4)
2
1 1 =-ltp':O

where 1:
is the total current in the z-section of the con-
ductor, equal to zero at its ends, (:I: l)-0. 1:
The intensity of the electric field caused by the induced
currents is expressed as Ee == 1 (kiA e+ grad div A•) having
'BQ)
both a radial component and a longitudinal component
e_ 1 a•A•z.
8 , - ime araz'
e I (
Ez=.1(1)8 1 A=)
e
kAz+a• ;
iJI
z
E~=O. (2-5)
The electric current induced in the dipole is so distributed
that the longitudinal component of the total electric field,
i.e., the field of the current of the loop and that of the induced
currents, is equal in magnitude and opposed in sign to
the external emf _on the conductor surface. We obtain then
the fol Jowi ng equations:
a• A: + k. A ez
0Z 1
r=a
== •
tms 11- ra (rA cpM)
r ur
I r=a+o
-
. r.teX t
tc.oe.c, • (2-6)
Here, the external emf E:xt is, in fact, the magnetic
current (with the reverse sign) of the exciting loop J:.
The vector potentials A: and A~ are expressed by (2-2)
and (2-4).
Since the unknown function, i.e., the electric current
/~ is the integrand in the expression .(2-4), the equation
(2·6) wi 11 be, to put it precisely, an integro-differential
equation. We shall examine below the solution of the equa·
tion (2 .. 6) in the first approximation.

2-3. Current Distribution In a Symmetrical Dipole

In the First Approximation
. -
In many practical cases, thin dipoles are utilised, i.e.,
dipoles of such thickness as is small in comparison with
their length and that of the transmittina wave. We shall
~
•

"
therefore assume the radius of the dipo}e to be infinitely
small: !f<l and1:<1. The magnitude of the vector potential
A: at point z' -=z on the surface of the conductor will there-
fore be determined in essence by electric currents flowing
in the vicinity of point z'-z, and for~-..0, the influence .
of the currents flowing on the remaining portions of the
dipole., can be disregarded, unless at point z'=z the current
is zero. We may approximate the expression (2-4), for r=a,
as follows:

(2-7)

where
R, = V<z-z') 1
+ a•,
fl is a constant, smaJJ in comparison with the wave-length.
But since for z- L1 < z' < z + £\, we may assume that
e-tkR,~ 1 and that the current within this interval is
constant and equal to that at point z' =z, instead of
(2·7) we have:
(2-8)
where
r+4

Assuming that a< L\, we obtain:

1 2A
Q= 2i In a • (2-9)
It follows from (2-9) that with vanishing a the quantity Q
will tend towards infinity, the expression (2·8) becoming
more and more accurate as the remaining part of the (2·4)
integral becomes finite. .
We shall further suppose that the ~xternal emf is applied
to an infinitely small section of wire In the centre of the
dipole. And since the radius of the wire a is small in com-
parison with the wave-length, the effect of the magnetic loop
may be repJaeed by that of an equivalent electric dipole,
assuming the dipole moment, in accordance with (1-21), to
1• 81
 be equal to J:l=-iweJ: bna1• The vector potential of this
dipole has a finite magnitude small in comparison with that
defined by (2-8) at all points of the wire, except those in the
region of the external emf.
Thus, for all points of the antenna, except those at which
an external emf is applied and where the current may equal
·zero, the integro-differential equation (2-6) may be approxi·
mated to the following differential equation:

(2-10)

This is a so-called telegraph equation, valid for a long

lossless line. Taking into account that the current at the
ends of the antenna equals zero and designating the current
at the feed points of the antenna by I 0 , the solution of the
equation (2-10) may be expressed as: ~
/e =z I sin k (l-l z I) (2-11)
z o sin kl •

We may also obtain with the same approximation the

distribution of the electric charge in the antenna. Making
use of the continuity equation (1-3), which, for the linear
current, is written as:
die
~ +iroQ=O,

where Q is the charge per unit length of dipole, we obtain:

k cos k (1-z) k cos k (l +z)
Qz> o = iW I 0 sin kl ' Q.r<~o ==-lOO I o sin kl • ( 2·12)

Thus, in a thin antenna, the current and eharge are approxi-

mately distributed in accordance with the sine law. As
follows, however, from the very analysis which led to the
expression (2-11), with the radius of the wire tending towards
zero the current distribution only nears a sinusoidal distri-
bution, without ever actuaiJy becoming sinusoidal. The
expression (2-11), in particular, is not valid for current
nodes, in which the current neither obeys the sine law, nor
is equal to zero.
In the majority of cases met with in ante~na theory, the
current in a symmetrical dipole is assu,med distributed in
accordance with (2-11). This assumption considerably
simplifies all the calculations and in most cases the results
obtained are close to those real. '
It follows from (2-11) and (2-12) that:
a) current nodes (zeros) and charge antinodes are set up at
the ends of the antenna;
b) current antinodes and charge nodes (zeros) are set up at
the distance of a quarter of a wave-length from the dipole
Pbds. Then, yet another quarter of a wave-length farther,
current nodes and charge antinodes are set up in the same
way, etc.;
Q

.,.., """" / /
-- {0)
-"' ..,.,/
(/J)
I
'
I

_, _,., /
(C)
Fig. 2-3. Current and charge distribution in a sy•nrnctrical
dipole:
).. A 6A
a-l=T: b-l=~; c-l=a

c) the current and charge at every point of the dipole are

90° out of phase;
d) a current antinode, node or some intermediate value,
as set up at the feed points of the dipole, depending on the
relation of the dipole length to the wave-length;
e) along the dipole the phases of current and charge change
by 180° when passing through zero.
A number of characteristic current and charge distributions
along a diAole is given in Fig. 2-3. It should be noted that
a symmetrical dipole with an overall length of 2l equal to
half a wave-length is known as a half-wave dipole. A dipole
of 21 length equal to a full wave-length is known as a wave
dipole.
Throughout this course we shall assume that the current
unless otherwise specified is distributed in accordance with
the sine law.
·81
 2-4. Field Intensity of the Dipole
In the Radiation Zone
Let the current distribution in the dipole be given. We
can then determine the field of the antenna at any point
of space, including very distant ones, i.e., in the radiation
zone. Furthermore, since the antenna is thin, we shall
neglect the radiation of the magnetic current caused by the
discontinuity in the centre of the antenna. •

ZCIJS(J
\
z \
/,

,
I.'
Fig. 2-4. Explaining the calculation
of the field intensity of a symmetrical
antenna.

Turning to Fig. 2-4, let us mark out at the point z of the

wire an element of length dz. In accordance with (1-3),
the intensity of the electric field caused by the element dz
of the antenna in the zone of radiation equals:
/e tUkt -
dE = i z sin ee-tk,., . . (2-13)
• 4nwer 1
The field intensity caused by another· element dz lying
symmetricaHy to the first one relatively to the centre of the
antenna, at the same point in space equals: ,
le dzk 1
dE = i z sin ae-tk'•. (2-14)
• 4nW8r 1
We assume directions r 1 and r2 to be parallel, since the
point of observation is infinitely distant and at any rate lies
at a distance considerably greater than the length of the
antenna.
88
 Substituting the expression of the current at the point
z...:...f:= I, sln ~~',;;-') into the expressions'.(2-13) and (2-14)
and adding up these expressions, we obtain:
dE ... dE +dE = i I 0 dzkl sin a sin k (1- 2) [,-~. +
I I 4 21Cil8 SiO k/ TI
,-ilv•]
fa •

Taking further into account that

1 1 I I
• -~-~ ;
'• '• 'o
',-= r 0-1 cos a,
f I = f0 -f- Z COS 6
and that •

sink (l-z) = J [elk et- z> -e-lk Cl-z)],

21
we obtain I

dE== 1o dzk• sin 9 e-tkro [elk ( l - z) - e - tk a- '>]X

Bnwe sin klr 0
X [elkz eo• & +e-tkz coa ']. ·

Integrating this expression along the length of the antenna

from z=O to z= 1, we have:
£._.i 601 0 e-tllra cos (kl cos 9)-cos kl (2-15)
r0 stnkl sin& ~

where we have substituted for vacuum 2no.l£

k .... 60.
The expression (2-15) defines the electric field intensity
of a· syn1metrical antenna in the radiation zone. In par-
ticular, for a half-wave antenna ( l-{) , this expression
becomes:

E=i
60/
o e-tkra
cos (;cos e) (2 .. J6)
r0 sin 6 •
The curves described by (2-15) and giving the dependence
of the values of the intensity of field E on the angle of
observation e. are known as the directional diagrams
(characteristics) of the antenna in tts meridional plane.
Directional diagrams are usually erected in polar
coordinates (sometimes in Cartesian ones).
-
•
 It should be pointed out that, owing to axial symmetry,
the radiation in the equatorial plane of the antenna is
uniform and therefore, the directional diagram of the an-
tenna in that plane represents a circle, in polar coordinates.
It should be further pointed out that the field intensity
phase does not depend on the angle of observation 8 and
to an observer in the zone of radiation, the waves seem to
e·o·

tao·
(0)

tao•
I

(C)
Fig. 2-5. Directional diagrams of a sym1netricaJ antenna:
A
o-1-
4 : angle ot aperture of soo-
A
b-l- T• angle of aperture of 44°i
5
c-l- A, angle of aperture of 31°.
8

originate from a point coinciding with the centre of the

antenna. This point is usually referred .to as the antenna
phase centre.
Fig. 2-5 gives the directional diagran1s of the antenna in
its meridional plane for three particular cases: l=}. l=}
and l= ~A. The same figure gives the values of the angles
of aperture of the directional diagrams, i.e., of the angles
within the limits of which the field intensity does not fall
by more than V2 times below the field intensity in the
direction of the maximum radiation. This aperture angle is
often referred to as the half-power width _oi the directional
diagram. '
An analysis of the expression (2-15) and of the curves in
Fig. 2-5 shows that, when 2l < A <4l, the directional
diagram has only one lobe, with its maximum in the direction
which forms with the axis of the antenna an angle 9=90°.
As the wave-length diminishes, this lobe becomes narrower
(the angle of aperture of the directional diagram becomes
smaller). When L>i, the major (main) lobe of the direction-
al dtagram becomes still narrower and there appear, in
addition, minor (silte) lobes. With a further decrease of the
wave-length, the major lobe begins to decrease, whereas the
additional minor lobes increase. Thus, when, for example,
l= A, the radiation vanishes in the direction 0=90°.
It should be noted that in the direction of the antenna axis,
just as in the direction of the dipole axis, the radiation is
always zero.

2-5. Calculating Power Radiated by an Antenna

by the Poynting Vector Method

Let us consider the power radiated by a symmetrical

antenna. Using the Poynting vector method, let us calculate
the power radiated through a sphere whose centre coincides
with that of the antenna. Furthermore, the sphere radius
is taken such as to ensure that the surface of the sphere is in
the zone of radiation.
The average power flow conveyed through a unit area per·
pendicular tc the direction of the power motion is expressed
by the Poynting vector

S == f [E, H*J. (1·10)

Since in the zone of radiation, the vector E is In phase

with the vector H (in the case of a lossle.-;s medium) and
their relation is equal to the intrinsic r~sistance ·or space, we
have
EE*
JS I=- 240n • (2·17)
furthermore, the only component of vectorS is perpendicular
to the surface of the sphere.
41
 The power radiated by the antenna 1s e~tpressed as:
Pz- 24~:rt sEE* ds, (2-18)
•
where ds is defined in spherical coordinates by the expres-
sion ds=r: sin 8 d8 dcp (Fig. 2-6).
Substitutang into (2-18) the exprrsion for E from
(2-15), we obtain:
n
p _ 30/: [cos (kl cos 8)-cos k1) 1 d6
(2-19)
z - sln1 kl 810 a .
'= 0

If we relate the radiated power to the square of the

current at the antmode of the antenna 1. == •• (here and :·kl
further on we have in m1nd the current amplitude values),
we obta1n a quantaty expressed in ohms and known as
the radiation res1stance of the antenna:
2Pr,
Rr.a-= fl • (2-20)
a

The integration of the expression (2-19) leads to the

following expression for the radiation resistance of the
antenna (related to the current antinode), first obtained by
Ballantyne in 1924 [1]:
Rza==30{2[C +In 2kl-Ci 2kl] +cos 2kl [C +In kl + Ci 4kl-
- 2C1 2kl) +sin 2kl [S1 4kl-2Si 2k/]}, (2-21)
where C==0.5772 is the Eiler's constant;
%

S
Si x == sin u du, the integral sine;
X •
0
ID

Ci x=- S eo: u du, the integral cosine.

• X

The integral sines and cosines can be expressed by the

following series (2):
. l.t' lx'
St X==X- 3 Si +-g- 5i- ... ,
Ci x==C+ lnx--- ~
2 21
1
+.!.4 .41~- •••
 As can be seen from (2-21), the radaatJon reststance of a
symmetrtcaJ antenna depends on the ratto of the length
of the antenna to the wave-length. Thus, for example, the

z f60

120

80

40

0.2 04 06
Fig 2 6 Expldtntng the calcu- Fag 2 7 Rad 1at aon resistance
lation of the rddldted power of symrnPtrtcal antenna

radrahon resrstance of a half-wave antenna ( f=0.25)

ts 73.1 ohm~ and the rad1at1on reststance of a wave
l
antenna (~= 0.5) ts 199 ohms
Table 2·1

1/'A Rza lf'A Rza 1/l Rza. ,,,., Rza lf'A RJ.a 1/1 N.l.a,
ohm" ohms ohms ohms ohms ohms

0 125 6 4 0 225 54 0 325 144

I
0 425 209 0 525 185 0 625 105
0 150 13 0250 73 1 0 350 168 0450 212 0650 166 0650 93
0 175 23 0 275 96 0 375 187 0 475 210 0 575 145 0 675 87
0 200 36 0 300 120 0 400 200 0500 199 0 600 127 0 700 86

The values of the radiatton res1stances of a symmetrtcal

antenna sttuated in free space, related to the current anti·
l
node, for a number of values of;: are gaven In Table 2·1
and Fig. 2-7.
 2-6. Field Intensity In the Vicinity of the Antenna
We have been considering the field intensity of the antenna
In the zone of radiation. Knowing this field has enabled us
to define the radiation intensity of a symmetrical antenna
in various directions, as well as the power radiated by this
antenna. Now, let us investigate the field intensity in the
vicinity of the symmetrical antenna. Here again, since the
ante..11na has a small radius, we shaJJ
' 1 neglect the radiation of the magnetic
r M field in the gap of the antenna.
l ~d----:.t-----7 Let us, to begin with, find the com-
ponent of the electric field intensity
paraJJeJ to the axis of the antenna.
In accordance with (I -4). this com-
ponent is expressed as:
J d2 A:
Ez= iooe ( k 1 A~+ dz• ) • (2-22)

The vector potential at point M

(Fig. 2-8) equals:
Fig. 2-8. Explaining
the calculation of the
field In the vicinity
of the antenna.

The integration in (2-23) extends from z' =0 to z'=l, since

we have taken account of the currents symn1etricall}' distant
from the centre of the antenna and situated at distances
r~.=Vx•+(z-z') and r.==Jfx• +(z ;-z') from the point
1 1

ot observation M. Substituting (2-23) into (2-22) and taking

into account that .

we obtain:
•
E, =;w:4n J
1'51
Twice integrating this last expression by parts, we find:
'

Taking into account that the current at the ends of the

antenna equals zero and that, on the other hand,

when z' = 0, also equals zero we have:

E == _. I· ldlz\ ( e-lkR, +e-lkRs) +
1 ICUB4n dz z=l R R 1 1
2 ld/ 1 e-tkRo 8
+ i(J)8 4n: di z=o R + 0

. I
+ twe 4n s (d
I

dz
z
2
+ k•J
/ )1
(e-tkr, + e-tk'')d
' rI r2
, (2-24)
Z •
z' =o

where R, = Vxz -t (z-1) 11 is the distance from the upper

end of the antenna to the point of observation ·M;
R.=Vx• +(z+l) 1 , the distance from the lower end of
the antenna to the point of observation M:
R0 =)I'"x• + z•, the distance frotn the centre of the
antenna to the same point M.
Assuming the sine distribution of the current along the
antenna
I =I sink (l-z)
z o sin kl '
we obtain:
•

Jdlzl __ k/ 6 ldl,l _ -k/ coskl

1 dz z = l - sin kl ' dz z =0- 0 sin kl ·
Thus, in. the vicinity of a thin antenna, the longitu-
dinal component of the electric field intensity, i.e., the coni·
ponent parallel to the axis of the wire, may be represented
45
as follows:
_ i 301, { ,-tkR,+e-fkR,_
E, - - ~•n kl kle-tkRo ( _ )
R1 R1 2 COS R, • 2 26
where account is taken of the fact that for vacuum, oo!n=30.
It follows from the expression (2-25) that the •ntensity
of the electric field E, on the surface of the antenna (when
x==a) does not equal zero. As an example, Fig. 2-9 shows
the distribution of E~(a) along a half-wave antenna (I:;=}).
We see that the resistive component, t.e., the component of
the electric field intensity, which
is in phase with the current, remains
, almost constant along the whole
length of the antenna, whereas the
reactive component, i.e., the com-
ponent of the electric field intensity
Z•-t Z•O Z•l which is in phase with the current,
Fig. 2·9. Distribution of tends towards infinity at the ends
reststive and reactive of the antenna when a-+0. Actual-
components of field ln-
tenstty along antennas ly, however, the tangential com-
/-resistive component, J- ponent of the electrtc field intensity
reacllve component E,(a) must be equal to zero on the
surface of the antenna. But we
assumed a sinusoidal, i.e., an approximate distribution
of the current along the antenna and we have, further-
more, neglected the magnt'tic field current in the gap of the
antenna; that is why the component E,(a) should not vanish
on the surface of the antenna even though the latter is
assumed to be an ideal conductor.
Let us now find (in the c.ylindricdl coordinates %, tp, z)
the other components of the electromagnetic field intensity,
viz., Ex, t.e., the normal component of the elt'Ctric field
intensity and H", the azimuth component of the magnetic
field intensity.
It follows from the first Maxwell's equation in cylindri·
cal coordinates that
iooeE ~= ~ ~ (xH .,),
an,
imsEa == - iJz • (2-26)
Writing (2-25) as:
E-== 301, .!.~ (s•lleRa+ 1 -tkR, __ 2roslel••liRo)
• k stn kl ~ u~
and compartng wtth (2-26), we obtain:
1
H =l
•
o
4nx su1 kl
{ e-lkR, + e-tkR, 2 cos kle-t•R, }, (2-27)
•

where an equivalent quantity

(1)8
12~n has been substituted
for the factor T .
Substttut1ng now (2-27) into the expresston forE" (2-26),
we have:
E _ {e-tkR,( -l)+e-lkR~( +l)- 2
30/ 0 klp-lkRo }
"'-' xsin kl R1 z R1 z cos R, Z •
(2-28)
Equattons (2-27) and (2-28) express the magnettc
field fntenstty and the normal component of the electric
field 1ntenstty 1n the vicintty of a symmetracal antenna.
It ~hould be noted that tn the case of an antenna whose
radtus tends towards Zt'ro, the normal component of the
electrtc field intenstty on the surface of the antenna, expressed
by (2-28), tends towards anftnt ty, whereas the tangent tal
component of the eJectr1c fteld tntenst ty, expressed by
(2-25), remains f1n1te everywhere except at the ends of the
antenna. Thus the vector of the electric f1eld Intensity 11
almost perpendicular to the ax1s of the antenna. Thts enables
to a~sert that the dpproxlntate s1nusoadal current dtstrt·
button along the antenna provtdes a d1strtbutton of the
elertromagnettc. fteld 1n the space surrounding the antenna
(tncJudtng the wave zone) close to reality ..

~·--------~~~~--~------
D
0
Ftg 2 10 Methods of m~dsunng t.urrent
and charge tn an antenna

It should be noted that the magnetic fteld tntenstty on

the antenna surface determan~s the current in the antenna
while the normal component of the electric field determines
its charge. Th1s results from the followtng equattons:
I.==2:r.aH, (a),
Q.= 2naaE. (a), (2-29)
41
where H, (a) and EJt(a) are the intensities of the magnetic
and electric fields on the surface of the antenna at the point
of observation.
The expressions (2-29) enable the experimental definition
of the current and charge distribution along the antenna.
A loop may be used to measure the current and a dipole to
measure the distribution of the charge; their position is
shown in Fig. 2-10. The dimensions of the loop and dipole
should be small in comparison \\'ith the length of the antenna
and the wave-length, otherwise the field will be distorted
and the measurement inaccurate.
Incidentally, the expressions (2-29) show that for a given
magnitude of the current(and consequently also of the charge),
the magnetic field intensity and the normaf component
of the electric field intensity on the surface of the antenna
increase as the radius of the antenna decreases.

2-7. Calculating the Power Radiated

by an Antenna by the Induced EMF Method.
Input Resistance of a Symmetrical Antenna
In the Poynting vector method, the integration was per-
formed over a spherical surface of an infinitely large radiu~.
However, since the space surrounding the antenna is free,
we may perform the integration over any surface '-nclosing
the antenna in order to calculate the power radiated by the
antenna.
The surface is a cylinder with height 2L and radius x;
a symmetrical antenna is placed in the centre of the cylinder
along Its axis (Fig. 2-11 ). Let us examine the Poynting
vector on the surface of this cylinder. The 11ormal components
of the Poynting vector in cylindrical coordinates are ex·
pressed as: •
I •
Ss== -2EzH..,;
1 •
s.= 2 E1&H..,. (2-30)

Evidently, the integral of the normal component of the

Poynting vector over the surface of the cylinder defines the
power fed to the antenna and radiated by it. Moreover, since
E and H are no longer in phase, as was the case in the radi-
ation zone [see equations (2-2:i), (2-27) and (2-28) J, the
power is expressed in complex form. i.e .• it has an active
component (radiated power) as well a~ a ("eactive one (power
oscillating in the vicinity of the antenna).
Let the cylindrical surface under consideration coincide
with the antenna surface, i.e., let us assume that x=a.
L=l. Then, when the radius of the antenna a tends towards
zero, the integrals over the upper and lower bases of the
cylinder (antenna) wiiJ tend towards zero and the power is
defined by integrating only over
the la'teral surface of the cylinder
(antenna):
l an
P=2 ~ S Sxadzdcp. (2-31)

Substituting into (2-31) the ex-

pression S:c from (2-30) and taking
account of the fact that the field
does not depend on the coordinate
q>, we obtain:
l

P= - ~ E~ (a) 21taH~z. (2-32) ...

t=O
Taking further into account that
21taH;= 1:. incstead of(2-32) we have
I
P=- ~ E~(a)J;dz. (2-33) Fig. 2-11. Explaining the
1.=0 calculation of the radf·
Thus, in o,rder to define the power ated power.
radiated by the antenna, we must
take the product of the current by the tangential com-
ponent of the electric field intensity on the surface of the
antenna and integrate this product along the antenna. Since
the electric field intensity on the surface of the antenna
Ez(a) is, in fact, the emf per unit length of the antenna in-
duced by the current in the antenna, this method of computing
the power is known as the induced enlf . method. The induced
emf method was proposed simultaneously by D. A. Rozhansky
[3 J and L. Brillouin [4 ], and later applied to antennas by
I. G. Klyatskin [5 ].
It should be noted that, actually, on the surface of the
antenna, the tangential component of the electric field
4-2122 49
intensity 'is, as was mentioned above, equal to zero every-
where except In the region of the external emf's. For- this
reason, the integration (2-33) should, in fact, be reduced
to an integration within the region of the external emf's.
We are considering the external emf (the generator's emf)
impressed in the centre of the antenna on a small section of
its length 2111, where the current may be regarded as constant
and equal to the current at the feed points of the antenna
I:. For this reason, the expression (2-33) is found to equal
J•u
P= ; (2-34)
Al Al
where V.=-2 SE,(a)dz=2 S E'!x dz1

r=o reo

Is the voltage applied by the oscillator to the antenna.

But let us go back to the expression (2-33) relating the
power to the square uf the current in the anti node as was
done above. We may write:

... (2-35)
where

(2-36)

represents the complex resistance of the antenna in relation

to the current in the antinode. The active component of this
resistance represents the radiation· resistance; as for the
reactive component, it represents the reactance of the antenna
in relation to the current at the antinoda.
Let us calculate the complex resistance of a symmetrical
antenna of arbitrary length and vanishingJy small radius.
Substituting into (2-36) the expression for E,(a} from (2-25)
and I ,=1 .sink (l-z), we obtain:
l

Z,;a==30 s
As a result of the integration of this expression, we obtain
for the limit case i •0, T •0:
ZEa=R'Ea + iXE••
where
REa =30 { 2 (C + 1n2kl-Ci2kl) +cos 2kl
(C + tnkl +
+Ci4kl- 2Ci2kl1 +sin 2kl [Si4kl-2Si2klJ}, (2-38)
X:t1 =30 {sin 2kl [c-tn~+Cl4kl-2Cl2kll-
-cos 2kl [Si 4kl-2Si 2kl] + 2Si2kl }• (2-39)
The active component of the antenna reststance thus obta1ned
Is the same as when calculated by the Poynting vector
method (2-21 ). This is fully justifiE'd since in both cases,
we integrated the Poynting vector over closed surfaces
enclo~ing a symmetrical antenna with a sinusoidal current
distribution. The shape and dimensions of the surface of inte-
gration do not matter in this case since we are considering
an antenna in a lossless space and the whole of it! radiated
power is lost into infinity. As for the reactive component
of the antenna resistanct'~ it determines the power osctl-
lating tn the vicinity of the antenna. Const'quently, the
magnitude of the reactance of the antenna depends on the
surface of integration of the Poynting vector. When the
radius of the wtre a tends towards zero, an dccordance with
(2-39), the reactance of the antenna tends towards infinity,
with the exreption of the half-wave antenna, in which case
it ts found to equal 42.5 ohms.'
If we relate the cornplex reststance of the antenna to the
current at the feed points (in the centre of the antenna),
we must take account of the fact that 1. = san \, .
Then

(2-40)
' Thus, for resonant tunlna. balf -wave antennas ~bould be
somewhat shortened.
I/
The resistances Rr.o and Xzo are called the antenna radiation
and reactance, related to the current at the feed points.
These resistances are also called the resistive (active) and
reactive components of the input resistance of the antenna
(in the absence of losses in the antenna), since they represent
the resistance between the input terminals of the antenna,
to which the oscil1ator is connected.
When an antenna is short (l/A~l), the trigonometrical and
Integral sines and cosines may be expanded in series
(kl) 1 I
aln kl = kl- 31 , • • •;
n-ro I.. I - - 1 (k/)l I •
\..V~ n." - .I - 2! T . . ••

. (k/)'
S1kl = kl- 3 •31 + · ··,
• (kJ)I
Ctkl=C+Inkl- 2 •21 + ··
Substituting these ex pans ions into (2-38) and (2-39) and
taking into consideration on1y the main terms of the expan-
sion, we obtain the following rather simple formulas for the
antenna input resi§tance:
1
Rx.o = 20 (kl) ; (2-41)
X1:o = - W w cot kl (2-42)
where
(2-43)

represents the antenna wave impedance.

Thus, it appears that the actjve resistance of a short
antenna may be calculated from the same expression as for
a dipole antenna [see (1-6) ], and the reactance, from the
expression for the long lossless line disconnected at the end.
The expressions (2-41) and (2-42) are used for antennas
of a length smaller than a quarter of a wave-length.
When the antenna is fed in the potential antinode <f-
-=0.5, 1.0, etc.), the calculation of the magnitude of the
input resistance of the antennas by means of (2-40) gives
infinitely large values. This is understandable because,
according to the assumptions \Ve made regarding the current
distribution along the antenna, the current at the feed points
is found to equal zero. Actually, in this ca.se, the current at
D
the potential anti nodes is never·equaJ to zero and even though
the input resistance does become large, it nevertheless
ren1ains finite.
For this reason, the expressions (2-40) are not valid when
the antennas are fed in the vicinity of the potential anti·
node. They still give satisfactory results for comparatively
thin antennas, whose length is smaller than l/A-=0.4.

. 2-8. Calculating the Input- Resistance

of an Antenna by Reducing lt to a Homogeneous
Long Line with Losses
As we saw above, the expressions for calculating the input
resistance of a symmetrical antenna as obtained by the
induced emf method fail to give accurate results when the
antennas are fed in the vicinity of current nodes.

Fig. 2-12. Distribution of current ampli·

tude along a sytnmetrical antenna:
/-circular sine lawr 2- hyperbolic 'sine law.

To avoid the arising difficulties, in technical calculations,

the current is assumed to be distributed along the antenna
in accordance with the hyperbolic sine Jaw (Fig. 2-.12), i.e.,
it is assurned that, as a result of the radiation, each element
of the antenna possesses a certain active resistance.
The actual distribution of the radiated power is compli-
cated and from the standpoint of physics, the assumption
mentioned just now is unjustified. However, in engineering
practice, it is quite in order to suppose the radiation
resistance to be distributed uniformly along the antenna.
The method examined below for calculating the antenna
input resistance consists of three distinct stages. To begin
with, uc;ing the induced emf method, we determine the an·
tenna radiation resistance in relation to the current anti-
node, the current distribution being assumed sinusoidal.
Then, taking into account the conservation of the radiated
power condition, the resistance is distributed uniformly

"
along the entire antenna. After that, applying to such a
homogeneous line with losses the equation for long lines, the
input resistance of the antenna is found.
In accordance with the conservation of the radiated power
condition, the integral along the whole length of the antenna
of the product of the square of the current by the active
resistance of the antenna inf1ni tely small element is equal to
the product of the square Lf the current in the antinode by
the radiation resistance in relation to the current ant1node
computed under the assumption that the current distribution
is sinusoid~tl,

(2-44)

Substituting here I.= I a sink (l-z), we obtain:

l
Rr.a=2R 1 ~ sin• k (l-z) dz.
r=o

But once sin' k (l-z) =} (1-cos 2 k (1-z)], the inte-

gration yieJds
- R:ta
RI- l ( 1-··;;:') • (2-45)

Equation (2-45) defines the d1strtbuted radiation resistance

of a symmetraca1 antenna per un1t of length.
Now, we may constder the antenna as a long line whose
propagation constant is:
v=P +ik, (2-46)
where
•
k == ¥ is the wave number;
P== ::, the attenuation constant.
It should be noted that the distributed resistance causes
a slight change in the wave impedance of the antenna.
This should be written as:
_ -v2R,+uilL1 ==
W.,-
IO>C I
yz:;c- yl -•--
I
2R ~m {I ; Rt)
•L·
1
~ L w• -,
(I) l
or, since

we have:
w Wl ~ wW ( 1-;{) , (2-47)
Consequently, the antenna wave impedance should be re·
garded as a complex valut.
Using the expressions for a long line with losses, the expres·
sion for the input impedance of a syrnmetrical antenna may
be written as:
Zlo = W w ( 1- if )cth CPL +- lkl). (2-48)
Transforming (2-48) In accordance wtth the expres~1on

cth(~/ + ikl) = sh 2 Pl- t sin 2kl

eh 2Pl-cos kl
we obtain:
. sh 2Pl- {stn 2kl
Z1;o = R1;o + tX};o = W w eh 2~1 cos 2kl -
. sin 2kl + {sh 2Pl
- ' W w eh 2Pl-cos 2kl • 2 49
<• )
Here, the wav~ impedance is usually defined in accordance
with (2-43). The expression (2-49) together with (2-43) and
(2-45) are used in many technical calculations of real
antennas.
To simplify the calculations, Table 2-2 supplies the values
of the distributed radiation impedance R1l of a symmetrical
antenna, dllpending on the rat1o of the length of the
antenna to the wave-length.
Table 2-R

R1l, R1l. R 1l, R 11, R 1l,

1/'A ohms 1/'A ohms l/A ohms l{A c,hms l(i. ohm a

0.125 17 6 0.250 73.0 0.376 155.0 0.600 199 0 0626 121.0

0.160 26.3 0 275 88.0 0400 J68.0 0 525 194 0 0650 105.0
0.175 36.5 0.300 104 0 0.425 181.0 0 550 181.0 0.675 93.0
0.200 47.0 0 325 120 0 0 450 192.0 0 575 163.0 0.700 91.0
0.225 60 5 0 350 138.0 0 476 199 0 0600 J5J .0
8:
H.,
A"g

?1/110

W=560.
W=460

01 -
1
~
W=
-500
t500
·"
f/100

0 Ill D.2 113 ll4 116 IJ6 tl7

Fig. 2-13. Input resistance of a symmetric:al Fig. 2-14. Input resistance of a symmetrical
antenna (resistive component). antenna {reactive component) .

...
 In Figs. 2-13 and 2-14, the resistive Rl:.o and reactive Xt,
components of the input resistance of a symmetrical antenna
are plotted against the ratio of the length of the antenna to
the wave-length l/A, for a series of values of wave impedances
of the antenna Ww· An examination of the curves shows that:
I) the active part of the antenna input resistance has its
maximum value in the presence of an antiresonance (l/'A~
~0.5). This maximum valu~ depends on the antenna wave
impedance: the larger the radius of the antenna, i.e., the
smaller its wave impedance, the smaller the maximun1;
2) the maximum of the reactive part of the input resistance
is about half the maximum of the active part (XEomax~
~ Rl:~max). This maximum value of the reactive part of the re·
sistance depends also on the· antenna wave impedance and
is all the smaller as the wave impedance is smaller;
3) for small values of the length of the antenna (l/'A <
< 0.25 the ·reactive part of the input resistance has a nega-
tive sign. When 0.25 < l/A < 0.5, the reactive part of the
input resistance has a positive sign, then negative again
and so on. It should be noted that the input resistance of
the antenna in the presence of an antiresonance (l/A,~0.5)
may be approximately defined from the simplified expression
w•
Rl:o= R w ' (2-50)
Ea
which is derived from (2-49).
If the antenna is short, i.e., li'A < 0.2, the input re·
slstance of the antenna may be expressed as
Rl:.a
R l:.o = stn 1 kl ; (2-40)
X Eo= -W w cot kl. (2-42)

2-9. Radiation of a Symmetrical Magnetic Antenna.

Symmetrical Slot in Screen
Although magnetic currents do not exist in nature, we
may examine the radiation of a magnetic antenna in free
space by analogy with an electric antenna.
Let us imagine an ideally conducting cylindrical magnetic
antenna of length 21 and radius a symmetrically excited
in the centre by an external mmf (Fig. 2-15). There will
67
arise In the conductor a magnetic current which will be so
distributed that the tangential component of the magnetic
field intensity will be equal to zero everywhere except In
the region of the external mm f. In the region of the externa]
mmf, the sum of the tangential component of the magnetic
field intensity and the external mmf equals zero..
Since the vector potential of magnetic currents obeys ·
the same equation as the vector potentia) of electric cur-
rents and the boundary conditions with respect to the mag-
netic field in the case of a magnetic antenna coincide with
--
...._
I
______ /,. --------·
I
•
:;: it.rz •- I/' 1 ~
~------ ~------
• 2'
Fig. 2·15. A symmetrical magnetic antenna:
/-electric loop; 2-magnetlc conductoJ tube.

the boundary conditions with respect to the electric field

in the case of an electric antenna, the distribution of the
magnetic current in a magnetic antenna coincides with the
distribution of the electric current in an electric antenna.
Thus, in a vanlshingly thin symmetrical magnetic an-
tenna, the current in the first approximation is distributed
in accordance with the sine law
/M- /M sin k ( l - r z I ) (2·51)
1 0
- sin .'d '
where 1: is the magnetic current at the feed po1nts of the
antenna.
Now, we may use the principle of the interchangeability
of the fields of the electric and magnetic antennas formu-
lated in Chapte One and write the expressions for the electro-
tnagnetic field set up by the magnetic antenna corresponding
to those of· the field for the electric antenna (2-15), (2-25),
(2-27) and (2-28). .
In the radiation zone, the electron1agnetic field of the
magnetic antenna is defined as

(2-52)
 In the vicinity of a symmetrical magnetic antenna.
the electromagnetic field is I

k . /~ { e- tkR, e- illRa
H• - = - t 41t(I)Jt sin kl R1 + R1 -
e-tkRa
-2 coskl Ro · ,
H . k t': e-tkR, l
~=I 4nmtJ. x sin kl -R, (Z- >+ (2-53)
e-tkR, e-tkRo }
+ R. (z+l)-2coskl Ro z;
/M
E =-i ~ {e-'"R•+e-i1tR,_2coskle-tkRo}
' 4nX SIO kl • J
To define the radiation conductivity of a syn1n1etrical
magnetic antenna, \Ve may use the expression (1·11) and
write:
(2-54)

where R1:a is the radiation resistance of the synin1etr1cal

electric antenna.

W= V~ , the wave impedance of space,

Gz., the radiation conductivity of a symmetrical
magnetic antenna of the same Jength as the
electric antenna
Thus, for example. for a half-wave electric antenna
Rza=73.1 ohms. In accordance with (2-54), for a half-wave
magnetic antenna, we obtain:
Gta ~ (a~~--~)1- 0.000514 ohm-
1
;

for a wave electric antenna Rr.a= 199 ohms. In accordance

with (2-54), for a wave magnetic antenna, we obtain:

Gza==(a:;~6>1 =0.0014 onm-•.

Generalising (2-54), we may obviously write:

Zzo,
Yze= w• (2-55
69
where is the complex Input resistance of the electric
Z!o
antenna;
Yto. the complex input conductivity of the magnet it
antenna with the same geometrical dimensions as
the electric antenna.
Let the magnetic antenna have a rectangular cross section
of length b and width 't, b being ~'t. When this antenna is
placed on an idea11y conducting infinite plane, a unilateral
slot is forrr1cd in a screen of length 21 and width b, exci~ed In
the centre (Fig. 2-16). The distribution of the magnetic

I I ---
~I
,~·
ill ~"' I

1 rr £ b
_I l
f 2

-
Fig. 2-16. Symmetrical -slot:
/-slot In screen; 2-clcctrlc generator; 8-screen: 4-magnetlc antenna.

current (of the electric field intensity) in the slot is, evi-
dently, defined by (2-51). Using the mirror method, \Ve find
that the electromagnetic field of the slot in the half-space is.
doubled in comparison with the field of the magnetic antenna
in free space. On the shadow stde of the screen, the field
equals zero everywhere. . ·
As for the radiation conductivity and the input conduc-
tivity of the slot, they are also doubled and expressed as:
2Rl:a
Gta= w• ; (2-54a)
2Z:to
YEo= w• , (2·55a)
where Rr.a and Zr.o are the radiation resistance and the input
resistance of the electric antenna of cro~~-section
dimensions b X t in the case of free space1;
60
 Gl:a and Yl:o, the radiation conductivity and the
input conductivity of a slot in a screen, whose
dimensions correspond to those of an electric
antenna.
If the same slot is made bilateral (Fig. 2-17), the radi-
ation conductivity and the input conductivity will again be
doubled and will be expressed as:

(2-54b)

(2-55b)

In accordance with these ~xpressions, the radiation con-

ductivity of, for example, a bilateral half-wave slot equa Is
1
Gl:a = 0. 002052 ohrn - •
Above, we have, by introducing the concept of the mag-
netic antenna, calculated the radiation of a slot in a plane
screen of infinite ex tension, using the ex press ions for the
radiation of the corresponding electric antenna in free space.
The direct connection
between the radiation
of a slot in a screen and
that of the correspond-
ing electric antenna in /J
free space was est ab- Fig. 2-17. Bilateral slot in conducting
lished by A. A. PistoJkors screen;
(6, 7J who formulated /-screen: 2-osclllator .
the so-called duality
principle, which states that the probletn regarding the radi-
ation of a slot in a screen corresponds to that regarding
the radiation of a metal strip of the same length and width
located at the place of the slot. In the plane in which
the slot was cut, the empty regions and those filled with
an ideal conductor change places. Accordingly, E and H
change places in the boundary conditions in that plane.
Moreover, they do so in the whole of. the external space.
A. A. Pistolkors has established that from the point of view
of setting up of the field intensity, lA of electric current
In a metal antenna is equivalent to 60n= 188.4 V in a slot
1 According to M. A. Leontovich (JETP, 1946, No. 6), a strip of
width b may be regarded as equivalent to a cylinder of radius a=-b/4.
61
 radiator. As can be seen from (2-55b), the Input condu~­
tlvity of a slot in a screen differs from the input resistance
of an electric antenna likewise by (60n)1 times.
The duality principle is fairly simple to generalise for
the case of a source-free space. Let this space be bounded by
a surface s and let us suppose that an electric field E, is
prescribed on one part of the surface and a magnetic fteld
H, on the other one. ·
In accordance with the uniqueness theorem, the boundary
conditions on the surface s under consideration completely
determine the field inside the bounded region. Let us intro-
duce into Maxwell 's equations
rot H = iiDeE:
rot E = - iro~J. H
the new variables
E8 = V iws E and H.,= V- iCilJI H.
They are related as follows:
rot H.,=kE.;
rot E8 ==kH,..
This means that the solution for E, in the existence of
boundary conditions on the surface ·
E80 = f, (x, y, z), H 80 = f. (x, y, z),
will also be valid for H1 , if we interchange the boundary
conditions, i.e., if we assume:
H,. 0 =f, (x, y, z), E1 ,={1 (x, y, z).
Thus, the duality principle can be stated as follows:
"The solution of Maxwel1 'c; Pquations for a magnetic field
for the given boundary conditions is also the solution of the
same equations for an electric field in the boundary con-
ditions, where theE and H fields are interchanged. Moreover,
,to the same boundary values of H, in the first case, and E.
in the second one, will correspond the same values of these
fields at the point of observation In both cases." ·
2-10. Additional Remarks Conceralng
the Current Distribution In an Antenna
...
We have shown above that the current In a symmetrical
electric antenna and the intensity in a symmetrical bi}ateral
slot in an infinite screen are, in the ~rst··approximation,
•

f2 .
distributed according to the sinusoidal Jaw. This result was
derived from the approximate solution of t9e equation (2-6).
Let us now examine a more strict solution of this equation.
To this end, let us return to the expression of the vector
potential of the electric currents (2-4) and, conforming to
the work of M. Leontovich and M. Levin (8 ), let us con-
sider the integral
l
T ==- S I: (z')e-t;R dz', (2-56)
Z'=-l
where
R= V(z-z') + Q ; Q -== ,.• + a•- 2ar cos (cpl -<p).
1 1 1

If we take into account that when z' < z,

VR· -Q I dZ = - R dR
-::-;::-_=-==
ZI = Z -
,
I

v~·-Q·
and when z~ > z,
~
J.l
== z + VR· - Q•, dz = I
+ .rR dR 1
,
r R 1
-Q

the expression {2-56) may be written thus:

,

z"= -1
Let us further take into consider at ion that
dR -d 1 R+ VR•-o•
YR•-Q·- n 2A '
where A is constant and assumed small in comparison
with the length of the antenna and that of the wave.
Integrating by parts in the expression T and taking
Into account that I:(± l)=O, we obtain:
•

T = - 2fz (z) e-tko In 2~ +

,
+ In R+ (R1 -Q1
~A
e' z'-z]
1z (z -·&klz(Z )X R
1 • e ,
X
•'=-1 J

'
•
z'=z
iki: (z' / ; z]
•

• •
X [ I:' (z') - e-llcR dz'•
61
 In the vicinity of the antenna surface, when ka< I, we
may approximately assume:
R~Jz'-zJ, R+VR -Q 1 ~2Jz'-zl.
1
e-t•o~l
and noting that, when r >a
ln:+q>

S ln 2~ dq>' = 21n ;A,

(f)' =cp

we obtain the following expression for the vector poten·

tial on the surface of the wire:

(2-57)
where
J

V [I~·, z] = 5 In 2 (z2~z'>u:· (z') + ikl~· (z')] e-tkr•-•''dz'-

z'= -l
l
- \ In (z~:;z>
~·
2
[1:· (z')- iki: (z')]e-ikw-• dz' ....
1

z'=z
l -
~:::: 5 In 1z;;' 1[ik/: (z') +I:=:: 1 1~ (z')] e-lk
z'= -l
2
I z- z' I dz'.

In (2-57), the first term coincides with the expression

(2·8) obtained at the beginning of the present chapter, and
the second term is the additional one, which we had earlier
discarded.
Substituting now lhe expression (2-57) into the equation
(2-6), we arrive at the following integro-differential equation,
investigated by Hall en l9] and later, by Leontovich and
Levin [8 J,
dale
dz~ +k 1;=iffie4n x{K+O[l~, z]}
2
(2-58)
where
I . 4 G diV k. V
X= a ; twe n == d za + '
21n~

K Eext
== , - ll, a ( AM)
or r ' t••+•·
64
 Note that the second terrn in the expression K is usuaJJy
neglected, which is permissible in the case of thin antennas
( 10].
Since x is a small quantity (a~.1), let the solution of the
expression (2-58) be represented in the forn1 of a series:
Jz= Joz + X/,z -t-
2
X /1z + •••
By s~:~bstituting this series into the equation (2-58) and
the boundary condition for the current and considering the
coefficients of the terms of the sarne power as equal, we obtain
the following system of differential equations:
., I
foz+k 10 z=0, I Ol (±l) = 0~
l~z
,
+ k'l 1 z = uoe4n {K + G ("/ oz' z}}, /1,(±1)=0;
+
11 : k 1 I 1 z= iCJ.le4nG [/,z, z], I •z (±l)= 0;
• e • I • e • • • • • • • . . . . (2-59)
The equations thus obtained are solved by the method of
successive approximations.
Let us first exan1ine a half-wave antenna. Frot~l the first
equation and the first boundary condition of (2-59), we see
that, irrespective of the distribution of the external emf
in the antenna, the current in it has a sinusoidal distribution
sin k ( l -I z l) I O
I oz= I o sin kl = o cos kz. (2-6 )

Multiplying the second equation of (2-59) by /,~ and

integrating in the interval from -1 to +l. we obtain:
l

•
iroe 4n ~ {K + G [/ oz• z]} I oz dz=O .
-l

Substituting here (2-60) and noting that

G [/ oz• z] = I o0 [cos kz, z],
we find: •

l
~ K cosltzdz
-I
Jo=- ~~------- •
~ a (cos kz, z) cos kz dz
_,
5.-2122 66
 •
If we ntgl~ct the second item in K and consider that the
external emf is applied in the centre of the antenna on an
infinitely small area, so that Ee:t = U~ (z-0), where
U is the voltage at the gap of the antenna and ~(z-0),
the delta-function which is equal to zero everywhere with
the exception of the point z=O, where it is equal to infinity.
and in addition

~ 6 (z-O)dz= I,
-QD

then
l l

~ K cos kz dz= ~ V cos kzb (z-O)dz= V.

-l -1

Consequently

where
'
Ztro = ~ G [cos kz, zl cos kz dz,
-I

evidently represents the input resh,tance of a half-wave

syn1metrical antenna. The calculation of this integral leads
to the quantity
Z 1n = 73.1 + i42.5 ohms.
Thus~
the input resistance of a half-\vave antenna is found
to be cotnplex and its reactive component is of an inductive
nature. For resonanl tuuing the antenna has to be shortened.
A more detailed investigation of the solutions of the systen1
(2-59) shows that the amount by whicb a haJf-wave antenna
should be shortened to reach the resonant length may be
defined as:
~~ 0.225
y=- l.
In --
a
Now let us examine an antenna of length other than half
a wave. From the equations (2-59), we see that when the
length of the antenna is not divisible by half the wave-
length, the first equation has only a zero solution. so that
l,z=O and G(/ 0, , zJ=O.
66
In the case of an arbitrary distribution of the external emf,
the solution of the second equation (2-59) may be written as:
z
4:t
I ~~=- ksin 2hl
J(l)E { ·
stn k <1-z ) Jr K (z ,>stn
· k (L +z ,) dz, -t-
z' = -l
l

+ sin k (l + z) ~ K (z') sin k (l -· z') dz' } •

z' c z

Let a 1un1pecl e1nf be appli{'d in thr ct'ntre of th£' an-

tenna, so that K ~ E~xt = U~ (z' -0). Then, the distri-
bution of the current :n the antenna \\'ill, in the first
approxitnation, be expressed as:
I ='XI = - H•>l!'2nt.rx sink (l-f21). (2-61)
z lZ k COS k/

The ratiq between the voltag(\ and the current at the

gap of the antenna represents the antenna input resistance
and, in this approx i n1at ion, it is
found to be pure I y reactive t . t -
u .60 ;7
Xtn= lz(O)=tycot kl. - Zo -
Fig. 2-18. An asynltnl"tri ...
To calculate the inphasc con1- cally excited antenna.
ponent of the input resistance,
the next approximation for the current shouJd be calculated.
However, this will nut Le done here.
Let now the en1f be Jun1pcd not in the centre of the antenna
but at a certa.~n point z(l (Fig.2-18). Then K ~E~v.t == U6 (z'-
-Z 0 ) and the distribution of the current in the antenna is,
in the first approxirnation, expressed:
\Vhen z<zo
I _ -., 1 __ tme4n!!I- . k ( 1 _ ) c:. k (l- ).
z- ,.~~ 'z - k sin 2k l s1n .. +z . ., 1n zo '
when z> Z 0

ime 4ltUX • k (l -ZrSln

\_ · k (l .,.-2 {2 62)
I z= XI tz=-kstn2k7Sln 1
0
)
• -

Thus, in that case, the distribution of the current in the

antenna, which retnains sinusoidal to the left and right
of the point where the generator is connected to the antenna
depends on the location of this point. It can be seen from
s• 67
(2-62) that the currents amplitudes in the arms of the antenna
are unequal and depend on the ratio sink (l+z0 )/sin k (l-z0 ).
Fig. 2-19 shows a number of current distributi on curves which j

differ substantially from those in Fig. 2-3. Thus, when

l2=l+zo= 0.251v and 11 =l-zo=0. 5A, the current in the
left arm of the antenna is vanishingly small relatively to
the current in the right arm of the antenna. When 12 =
=l-J-zo=0 .25 A and l, =l-z0 =0. 75 A, the direction of the
current in the antenna is different from that in a wave an-
tenna in the case of a symmetrical excitation. It is only in the

Fig. 2-19. Current distribution In an asymmetrical antenna.

casE' of a half-\V8ve (resonant) antenna that the distributi on

of the current does not depend on the feeding point of the
antenna.
The san1e kind of electric field distributi on as in Fig.
2-19 takes also place in a bilateral slot, if it is excited asym-
metrically, i.e., when the slot is fed as shown in Fig. 2-16,
but not in its centre.
Next, we shall turn to the calculation of the current
distribution in the antenna by- another method, viz., thE'
eigenfunctions rr1ethod.
•
2-11. The Eigenfunc tions Method
In the eigenfunctions method, the symmetrical antenna is
considered as an ideally conducting prolate spheroid excited
by a circular magnetic loop lying directly on the surface of
the spheroid (Fig. 2·20).
Let us introduce the coordinates of the prolate spheroid
obtained by the rotation of the ellipses around the major
axis. We get a system of spheroids the orthogonal surfaces
of which are a system of two-sheets hyperbolo ids. Let us
68
direct the z-axts of a rectangular systen1 of C\1ordinates along
the n1ajor axis of the ellipse and let d be lhe focal distance
(Fig. 2-21) and cp the azimuth angle, the azimuth of the x·
axis being zero. The s1gn of <p should be determined by the
1 2
z

Fig. 2-!lU. Explaiuin;.: the elgenlunctions method:

1-magnetlc loop. 2-conductlng spht·rold.

right-hand screw rule. The coordinate surfaces of the

spheroid are defined by the equations
' +r r -r
eh "V= ' d 2 , cos()= ..!..d '

and the connection between the rectangular coordinates and

z
D=O

O=const

Fig. 2-21. Spheroid coordinates.

the spheroidal ones ts found from the equations

...

X= f sh 'lj1 sin 6 COS q>;

d
y= 2 sh 11' sin 0 sin cp;
d
z= 9 eh 1J' cos e.
69
 If we take ,, 0, cp as the spheroidal coordinates, the
relation between the spheroidal and rectangular com-
ponents of a certain vector F is expressed as:
F..p=F x~feh 'ljl Sin 0 eoscp + F1 ~~eh 'ljl sine sin cp +
1 d
1 2 + Fz h
sh ..p cos 0;
1 d I d .
F, = F" h 2 sh "P rose cos rr ·-1- FV h 2 sh 1p cos eSin q>-
1 1
1 d . .
- F z h 2 eh \j) stn 0; (2-63)
I

Frp=- F" sin <p F1 cos <p, +

2
where h1 =h1 =fVeh '1jl-cos•o are the scale factors.
Let an infinitely thin magnetic loop of elliptical form lie
in an infinite space. The vector potential of this loop is
ex pressed by (2-1). In the pro bIern under consideration, the
magnetic current has only an azimuth co1nponent JM= 1:,
so that when the ratios (2-63) are applied to the expression
(2-1) for the azimuth cotnponent of the vector potential we
get the expression
A:('ljl, 0, cp) = 5
~ J~ (,P', 6', cp') cos (cp' -cp) e-~kR ds, (2-64)
s
02 231
where
) ... ds = ~ ) ... h~h·3 d6' dcp';
s O'=a, q>'=-o
h ,1 = 2d 1I
V
h2
C 1J'
I
-COS
20 I
;

d ~ h 1'I ,_sJ.n O' .

.h,.; -=-= :l
To solve the problen1 by the eigenfunctions method,
. I e-lkR . • .
the functton G = 4n R IS rcprt'sented tn the form of a
spheroidal functions expansion, i.e., in the form of an in-
finite double sum of space harmon1cs:
(» IJ)

1 e-ikR k ~ ~ Em ( ' )
G= 4n R = 2n.i ~ ~ N cos m <p - cp X
n=o m~o mn
X S~J (c,
cos 0') S~], (c, cos 8) X
I R~~ (c, eh 11'') R;:~ (c, eh 11') for 11' >'I'';
X ). R1n'J (c, eh 'l1') R!:~ (c, eh 1Jl) for--t-
2 65
< ..p', <• >
VO
\\'here
em= I \VhCI1 tn=O, £,n=2 \\'hentn== 1, 2, 3 ..•• k=
2:t d
=T' c=k2;
S!,~,!, the angular spheroidal function of the first kind
sin1ilar to the associated Legendre functions;
R:,:~. the radial spheroidal function of the first l<ind
sin1ilar to the Bessrl function of the half-entire
orJer;
R~:~, the radial spheroidal function of the fourth
kind siruilar to the HankeJ function of the
second kind of the half-entire order,
~~
CO (
· dkr.)2 (k + 2tn)~
N mn = 2 ' t k!(21< + '21lt + J) •
k=o 1

In the Jast expression, a dash after the syn1bol of the sum

indicates that the su1nn1ing up occurs for even values of
k if n is even, and for odd values of k if 11 is odd.
The expr(\Ssion (2-65) is the solution of the wave equation
for the a function in spheroidal coordinates.
Note that because of axial synnnetry, the rnagnetic cur-
rent of the loop does not depend on the angle cp. rlence,
substituting (2-65) into (2-64) and taking into account that
2:'1 ( 0 when m ~-F 1;
~
\ cos (q/ -q:') cos tn ((~' -qJ) cl<p' = ·\ -wen
2n h m= 1,
qJ = 0 ~tn

\\·e obtain:
'1:a

A~ (~1, 0) = ·- .f J: (lp', 0') h, h~ dO' ik x

rv = "•
~
L
1
X N 111 s(u (
tn c, cos
O' ) s(l)
tn ( c, cos o X a>
n=o
x J R~~~ (c, eh'!'') R~~' (c, eh 1p) for,~> 1jl';
(2·66)
l R~~ (c, eh 1p') R~~, (c, eh \f')Jor 1p < 11>'.
Thus, the azimuth component of the vector potential does
not depend on the angle <p. As for the radial and A:
meridionalA: cornponents of the vector potential, due to
axial symmetry. they equal zero.
11
 If an fdeaJJy conducting spheroid, the radial coordinate
of the surface of which wi I I be designated by 1f'o < 1J7',
is now introduced symmetricaJJy into the field of the mag-.
netic Joop under consideration, electric currents will be
induced on the surface of the spheroid, the currents whi.clJ
have only a longitudinal comp nent Je=J8. Under the In-
fluence of the surface electric currents, a secondary fieJd
is formed of such a va Iue that the tot a I fie Id, i.e., the fieJ d
of the currents of the magnetic loop plus the field of the
currents of the spheroid, has no tangrntiaJ component of
the electric field intensity on the surface of the spheroid.
The secondary field rnay be regarded as the field of the
primary wave, reflected by the spheroid and rnoving away
towards infinity, i.e., it may be represented in the form of
the ex press ion (2-66) with unknown re fleet ion coefficients
of the space harn1onics
61
A;· reil ('ljl, 9) = - ~ J; {'ljl', 9') h 1 ha d9' ik X
01 =81
ao

X~: ;n s~t,! (c, cos 9') s~~) (c, cos 9) Rit,! (c, eh 'P') X
n=o '"
- x R~t,/ (c. eh 1p), (2-67)
where an are the reflection coefficients of the space harmonics,
determined from the boundary conditions on the surface
of the spheroid. In accordance with (I -4), the vector of the
electric field intensity is expressed as:
E=rot AM •

and since AM =A;, the corr1ponents of the E vector in

prolate spheroid coordinates are:
e.=- hlha !o (h,A;); 1
1 a ( M'
E8 = h h a;r, h,Aq~);
1 I 't'

E~=O. (2-68)
Thus, the boundary condition on the surface of the
spheroid for the total vector potential may be written in
the following form:
a (h A
I0. I
M.
cp
tot) :
it=••
= 0. (2-69)

12
 Adding up (2-67) and (2-66) and substi· \V hen 'I'< "l.,'
tuting into (2-69), \ve obtain for the coefficients of expan-
sion a,.: ,
d
c}tJ,,o r2d sh 'l'oR (1)
tn (c, eh 'l,o) l R~~· (c, eh "'')
an=- d (1) • (2-70)
!: sh .. ,., R<• l (c eh ..., ) R 1n (c, eh'$')
d~'o ') 'to 1n ' 'tO

The solution thus obtained satisfies Maxwe ll's equa-

tions, .the boundarv conditions on the surface of the ~1-
~

lipsoicl and at infinity and is therefore the unique solution

No\v substit ute (2-70) into (2-67) and add it to (2-66) Then,
placing the rnagnet ic loop on the surface of the spheroid.
i.e.. assutning 11~' = ~-,o and taking account, furthermore,
that
d R~ ',1 (c, c h 'i'oJ
1
(1) cl R 1:/ ( c, ch 'i'o) 141 h
R (c,
1 ,1 eh~\) 0 ~, 0 - Rtn (c. c '1' 0 ) cJ'I-, ==
0
l
=-·~-.

ik ~ sh 'i'o
\Ve C'btain
~~ OD

A;· 101
(ljJ, 0)=- S J:h 10 h"3 odll'_L ~ nX
1
r~~ n~o

X S\',: (c. cos B') s'.'~: (c' cos 0) X

R~~ (c, eh 'I')
X~-~~------ (2-71)
0 [ -d s11 'l'uR (t)
all'o tn (c, eh 'i'o) ]
2
where
h10 == f V ch•w,.- cos• 0'; h~o == ~ sh 11'. sin a·.
·rhus, the expression (2-71) provide s the solutio n of
the problern of the excita tion of an ideaJJy conducting sphe-
roid of a circular magnetic loop with a uniform azimuth
distrib ution of the magnetic current The magnetic
loop located on the surface of an idea~ ly conducting sphe-
J:.
roid forms a circula r slot on the spheroid and the magne tic
curren t J:
is equal in magnitude and opposite in sign to
the electric field intens ity £ 8 in the slot.
The electric field intens ity in the space surrounding the
spheroid is determined by substituting (2· 71) into (2-68) •
•

71
lt should be further noted that since A~=A:=o and
A: does not depend on the angle cp, divAM=O. Hence, in
accordance \vith (1-4), the magnetic field intensity wi 11 be·
expressed as:
_ - , (;0 8 AM.
Hq>- cp
tot
•

We may now also write the expression for the electric

current induced on the spheroid. The surface density of the
electric currPnt is expressed as:

Thus, the eigcnfunctions method, which is the strict

classical method for solving the problcrn under consider-
ation, enables to determine the exact value of the electro-
magnetic field in the space surrounding the antenna and
the exact value of the electric current distribution in the
antenna. __
The difficulty in calculating the field according to the
expressions obtained consists in that the terms in the series
do not always decrease sufficiently fast. However, the
angular and radial spheroidal functions are partly tabu-
lated [11 1 and this mak('s the calculations easier.
The strict solution of the problen1 regarding the exci-
tation of a spheroid by an infinitely narrow slot located in the
cenjre of the antenna. was first given by Chu and Stratton
[ 12]. Designating the voltage at the slot in the centre
of the antenna as U, we obtain in that case in (2-72):
•
b2 £i
SJi S~! 1
(c, cos 8') h;o h~o d6' =US~!' (c,O) 2 sh 1p0 •
6' -:,: QI

Further, if \\'e take account of the fact that the full

current flowing through the cross section of the spheroid
is expressed as:
231.

I:= 5J~h .. dq>,

q>=O

74
\VC obtain the following expression for the current at the
feed points of the antenna 1: ,
1
d sh 1t' ) X
-- U ( ·-
l eo-- - 60icl)k 2 °
~-, ~
R~~!_)(-::-c,~c_h_,~_o)_I_S..;..;.;;!_,
1
X ___ <_c_,_O)._)- - - - : - - •
(2-73)
~o N tn f
a~. ( ~h ~'•R~! 1
(c, eh ~'~) J
Q0/5 aot5
Q012
yta= 'flU
!b aotz /ftr-707
~
~ aooo ~c,
~ o.oou ~~
~ 0.006 4 ~ 0.006
Jl\
~
~ aooJ
8,.~ ~
aooa \.
' ~
- -~ ~
\.1
~
;:)
aooo ' ~ 0.000
~-QOOJ
~-Q003
6 A
-aoo a2 aJ a4 a5 aB 0.1 aa
"""""""
21/A.
6
er-
•
~ll.l
-o.oo a2 oJ a4 o.5 aa 0.7 o.a
QOf5 QOM~~~~-----­
,J
Q012 -- O.OQ~+-~~~_.~
Cb va=7Q7 !b
~ QOIJD ~ QOOI 1----+---' --t---11 +--....,.--...-.t
~alK16
~-,
I \ 1 Q006 ..,____
& QOOJ
~ ~
~
~

~·
~ ..... - § QOOJ t--~-+-1-+---+-~---1
~atw ~ 0.000 ~+--~ -+---+--1
!§.
Sl.
~-0.003 - ~-o.ooa~~___.,; ~~~
-aoo60.2 0.3 \...-- V ~11A
0.4 0.5 06 Q? 0.8 --QOOI(J2 0.3 114 0.5 116 ll7 118
Fig. 2-22. Curves of the input conductance of an antenna .

.I

Evidently, the ratio of the current to the voltage repre-

sents the input conductance of the antenna
Cl)

Y:to= ~ (0,. +iB,.), (2-74)

n=o
15
where
1
• - imll ( d )
Gn + tBn= 60 2 sh 'i'o X
R';.l
[S'!: (c, 0)] 1 (c, eh 'l'o)
X ----::-"a~l--=-d---=-,.-l--~]~ •
N1n O'i'o 2 sh "i'o RIn (c, eh ,o)

In Fig. 2-22 the active Gt and reactive 8 1 parts of the

input conductance are plotted as a function of the ratio
of the length of the antenna to the wave-length 2//A for
a series of ratios of the length of the antenna to its largest
diameter 2l/2a. The values of the conductance when n=3.5 ...
for the given dimensions of the antenna are relatively
small and are not given here. It is seen from the curves that
the maximums of the active part of the conductance 0 1
occur in the vicinity of 2 l/'A=0.5. The reactive parts of the
conductance pass through zero when 2 li'A is somewhat
smaller than 0.5. Thus, the resonant length of the antenna
is found to be somewhat smaller than half a wave-length;
moreover ft is the smaller the thicker the antenna is. When
the diameter of the antenna changes, the active part of the
conductance in !he vicinity of the half-wave length rerrJains
practically unchanged, whereas the reactive part of the con-
ductance increases when the diameter diminishes. If we
pass from the conductance to the resistances, we obtain
the foJiowing resistance values for the half-wave length:
ZEo = 73 + i 30 ohnts when lfa = 7,070;
Z1:o= 73 + i 24 ohms when lfa=707;
Zl!o=73+ i 12 ohms _when l/a=70,7.
Ca lcuJation! show that the current distribution curves
along the antenna are close to the siQusoid.
 CHAPTER THREE
Radiation of Two Coupled Antennas

.
3-1. Directional Diagrams of Antennas
We have discussed the radiation of a single syn1metrical
antenna in~ df!tail. Let us now examine the radiation of two
synunetrical antennas in free space.
Two parallel antennas of the same length lie at a distance
d frorn each other (Fig. 3-1 ). Voltage U, (of frequency
w) is applied to the first antenna
f and voltage U 2 (of the ~a me fre-
quency) to the second antenna.
Electric currents wi 11 arise in
the antennas the cornp1ex arnpfi-
tudes of which wi 11 he designated
- ~

--
Fig. 3-1. Coupled an- Fig. 3-2. Explainin~ the calcu-
tennas. lation of radiation characteristics.
by / 01 at the feed points of the first antenna and by / 02
at the feed points of the second antenpa. Their ratio will
be as follows:
l o t - me'"'
I 01 - ' (3-1)
where m is the ratio of currents amplitudes;
1J'1 phase shift between currents.
77
 The points of interest, in the first place, are the fields
which are set up by this pair of antennas in the radiation
zone. Let us first examine the equatorial plane of the an-
tennas (Fig. 3-2). In that plane, the radiation of each of
the antennas is non-directionaL However, due to mutual
radiation, their fields interfere with each other: in some
directions they are added up, in others deducted one from
the other.
If we assumP their current distribution to be sinusoidal,
the field of each of the antennas is defined by (2-15). In
the equatorial plane, we have for the first antenna
60 101
E1= i . - ( 1- cos kl) e-ikr, (3-2)
r, san kl
and for the second one
60
E2 = i r /
stnkl
02
( 1 -cos kl) e- ikr,. (3-3)
2

The field vectors at the point of observation are parallel,

so that (3-2) and (3-3) can be added up arithmetically.
It should be noted in addition that the point of observation
of the field is at infinity, so that, in adding up, we may,
with a high degree of accuracy, assurne:
..
I I I d d
~ ~ , r,-r 0 = -2 coscp, r,-r 0 = - - coscp,
2
r.. '· fn
where cp is the angle between the direction to\vards the
point of observation and the line joining the centres
of the antennas;
r 0 is the distance from the centre of the plane
formed by the antennas to the point of observation.
Thus, we obtain for the total field:
. kd . kd
60 I 01 , -1 cos cp ·• - , -1 cos cp
E=i r stn
. k,• (1-cos kl) e- ikro (e
•
+me' e ].
0
(3-4)
As is seen from (3-4), the amplitude and phase of the total
field depend on the angle of observation cp and on the ratio
of the currents. Let us examine some cases of excitation
when the antennas are excited in eo-phase, i.e., let m= 1
and 'l'=O. Then the expression (3-4) wil1 be reduced to the
following one:
E = t. 120. I 01
- ( 1- cos kl) e- tkro cos (kd
- cos cp ) • (3-5)
T 0 SIO kl 2
78
 It can be seen from (3-5) that the shape of the direction·
aJ diagram of eo-phased antennas dep~nds on the ratio
between the distance of the antennas and the wave-length.
There are directions in which, by con1parison with the
field intensity of one single antenna, the field intensity of
the antennas is doubled. These directions are determined
frorn the condition

COS \2 COS C('m:iX = ±

/ kd )
},
or
ll'A.
COS q)max = d, 1l = 0, 1, 2 ... (3-6)

Thus, the first radiation maxin1um occurs for any values

off and coincides with the direction <rmax=±-~. This
is understandable since in that direction, the difference of
the ray paths equals zero. The second radiation n1axirnurn
4
occurs when = 1 and coincides with the direction (jim;1x =
=0°, 180°
At the same tirne, there arc directions 1n \Vhich 1he
radiation of the antennas equals zero. These directions are
deterrr1ined frotn the condition
kd ..
cus (.2 cosrp 0 )=0,
or
nA.
coscpo=u, n= 1, 3... (3-7)

Thus, when 4< 0.5, there are no directions in which the

radiation equa Is zero. The first zero of radiation occurs
when ;=-= 0.5 and coincides with the directions q> 0 =0°, 180°.
Fig. 3-3 shows the an1pl itude directional characteristics
of eo-phased antennas for a series of values of as f,
ea lcu lated from equation (3-5). .
As regards the phase directional characteristic, it can
be seen from (3-5) that for a value of f
< 0.5, it repre-
sents a circle, i.e., when the angle cp changes, the phase
of the field remains unchanged. For values of ~ 0.5, f
19
the phase of the field changes by 180• when passing through
the zero of radiation.
Now, et the antennas be in anti phase, i.e., let m= I and
'1'=180°. The expressiou (3-4) will then be reduced to the
following one:
E=- 120. I 01 (1-coskl)e-tkro s1n
. - coscp . (kd ) (3-8)
rdstn kl 2
In that case, when q>=90° and cp=270°, the radiation is
always equal to zero, Irrespective of the ratio { . The
direction-s tn which the field is doubled are determined from
the condition
nA
cosrrmax =u, n== I, 3 .•. (3-9)

lt can be 'ieen that when } <0.5, there is no direction with

a double value of the field intensity. When {=0.5, the
field is doubled in the direct ion CVmax =0°, 180°. The zeros
of radiation are detern1ined frorn the condition
.. n'A.
cos <p o = d , n = 0. l, 2 ... (3-10)

and as noted a hove, the first zero of radiation occurs for

any values of {-in the dirrctioncp 0 =90°, 270°. The second
zero of radiation occurs when ~=I and coincides with
the dtrections c:p0 =0°, 180°.
Fig. 3-3 shows the amplitude directional characteristics
of anti phase antennas for a series of Vi! lues of lt should f.
d
be noted that for any values of A' the phase character-
istics of anti phase antennas change by 180° when passing
through the zero of radiation.
The excitation of antennas in quadrature, i.e., when m=l
and 'i'=90°. is aJso of interest. In that case (3-4) is written
as:
12 101
E=i
r0
~SlD kl (l-coskl)e-tkro+t••ocos(45°-~l:OSQJ'.
2 j
(3-1 J)
BD
r .!L-o.25 d
-f~o.5 f=0.75 d -l
-
~

~
~
ADO• iT=0.375
uo• goo 10"
1•1
w•
I

f80° ( I ~ q;=0° M'lt++-l- (/)= o· rooo ~ (/)= 0

0
180
0
I : tao'!- I T -l--,=o·
;=o· I
270° 270. 270° 270° 270.
•
ao• go• 10.

rao·--1\::;t(:---l~...l_J ·' • •rp=e· .eu"l I I f!J=o· tao" ) I -1-~=o· rso-"____..;;::~ . 9=0·

f/=180°
270° gz.o•
NO" uo· ..11o·
.. ·---....
I

I
rao· >I J 9J=o· reo• (:=] I r,=Ow 180. } y ") f/)=0·

I 1/J=g(}"
270
11
I
270°
I

270°
Fig. 3-3. Directional characteristics of t\vo coupled antennas for variou~ distance4ii bet\\·een them (d)

~
,_____________ :md :rinus pha~e ~hifts ('f). ---------- _ __ _j
 When the distance between the antennas is equal to a
quarter of a wave-length ( k?d =45°) the radiation in th&
direction rp= 180° equals zero and, in the direction <p=O,
it is doubled in comparison with the radiation of a single
antenna with the same current. The a1nplitude directional
characteristic has the shape of a cardioid and the second
antenna, which reflects energy towards the first antenna is
known as a reflector. The same figure shows the amplitude
directional characteristics for a number of values of {when
the antennas are excited in quadrature.
It is easy to define also the expression for the field in-
tensity in the radiation zone, in the plane of the axes of the
antennas. In that plane, the field of each of the antennas
is defined by (2-15). Designating once again by r 1 the dis-
tance from the centre of antenna 1 to the point of obser·
vation and by r1 , the distance from the centre of antenna 2
to the same point and adding up the fields, we obtain the
following expression for the total field:
_- . 60 I 01 _ 'kru cos (kl cos 8)- COS kl
E- t r
0 sinkl
-e ' - sine
x
kd kd .
x e i - sin 6
1 +me' e
''tj' - i - sm ''
2 (3-12)
'

where 0 is the angle bet,veen the direction to\\rarrls the point

of observation and the antenna axis.
An examination of (3- I 2) shows that the directional
characteristic of two symmetrical antennas in th~ir com-
mon meridional plane represents the product of two fac-
tors. The first factor, which is defiJaed by the expression
sin0)-coskl represen t s th e d"1rec t•Jona 1 ch arac ter1s
cos(kl cos · t•tc of
0
a single antenna; the second factor, which is defined by the
kd . ~ kd 5 J D
ex press ion e i 2 ~•n · + me;..p e-; 2 n , represents the factor

ot the combined radiation of the two antennas.

82
 3-2. Mutual and Total Impedances of Antennas
[~et us consider the interaction between the antennas .
The longitudinal cornponent of the electric field intensity
set up by each of the antfnnas in the system under con~ider­
ation (Fig. 3-1), is expressed by (2-25). The con1ponent
Ez, set up on the surface of antenna I will consist of two
! teJTIS
EZ1 = E Z11 -~ E z 1 2.
\Vhere Ez 11 is the component of the electric field intensity
induced by the currents of antenna I;
Ez 12 , the con1ponent of the electric field intcn~i ty,
induced by the currents of antenna 2.
In just the sam~ way, the cotnponent of the electric field
intensity Er,_ set up on the surface of antenna 2 wi 11
consist of t\\'O i terns
EZ2 =--- EZtl +E f!1
where Ez 21 is the cornponent of the electric field intensity
induced by the currents of antenna 2;
Ezu• the con1ponent of the electric field intensity,
induced by the currents of antenna I.
Using the induced emf method, in conforn1i ty with (2-33),
the po,ver fed fron1 the generator to antenna J is:
: 1 =1 z 1 =l
P, =- ~ Em/;1 dz,- ~ Em I; 1dz •• (3-13)
:r 1 =o z,::.:o

and the powe! fed from the generator to antenna 2,

Z2 ::.1 z"l =I

pI = - ~ Em I;, dzt- ~ Ettt J;. dzt, (3-14)

z2 =0 z.~.:o

where I :u is the current at the point z. of antenna /;

I z• is the current at the point z,_ of ttntenna 2.
On the other hand, the power fed to the antennas tnay
be expressed as: _,
•
I OJU1
PI= 2 ;
(3- J5)
1•02 U1
P. = 2 •

6*
On comparing (3-15) with (3-13) and (3-14), for two coupled
antennas, we obtain equations similar to Kirchhoff's
equations for coupled contours with lumped constants,
U~=/oJZu+/ozzll' t (3-16)
V,.=lozz2 a·+ /o1Zu, J
where
(3-17)

* dz ;
E zs• I za (3-18)
1

(3-19)

(3-20)

Here, Zu and Z 22 represent the separate resistance s of an-

tennas 1 and 2, -i.e., the resistance of each of the antennas in
the absence of the
I(Jf other, and andz.2 z2h
theso-calledmutual or
coup) ing resistances.
Moreover, they are
a 11 related to the
2
currents at the feed
f
points of the an-
Flg. 3-4 Equtvalent circuit of two tennas.
coupled antennas. It is seen from (3-19)
•
and (3-20) that the
mutual resistances are equal Z12=Z 2 1. A more detailed
anaJysis shows that this equality ts likewise maintained
even when the antennas are of different lengths. From now
on, we shaJI therefore speak of the mutual resistance as of
a quantity which is defined according either to (3-19) or
(3-20}. The mutual resistances of the antennas are complex
quanti ties.
On the basis of (3-16), \Ve may reduc~ twu coupled
antennas to the equivalent circuit shown in Fig. 3-4.
84
 Dividing the first equation (3-16) by I 01 and the second
one by I o2 , we obtain the ex press ions for the full input
impedances of antennas I and 2:
Z01 = u, z
r= 11 +me 11;
i'i' z
Zot-- u: -
- z~~+-e
1oz
I - i'i' z11
tn
(3-21)

In these equations, the second terms in the right-hand side

represent the resistances introduced by the second antenna
into the first one and by the first antenna into the second
one.
In the case of arbitrary current amplitudes and phases,
the introduced rcsistances are not of equal magnitude and
depend on the ratio of the currents in the antennas as \Veil
as on the geometrical din1ensions of the radiating system,
which is seen from the expressions given in Paragraph 3-5.
It is only in t~e case of co-phasaJ oscillations with equal
amplitudes (me' 4 =-==I) that the introduced resistances become
equal to each other and to the mutual resistance. Inci·
dentaiJy, this explains why mutual resistances are sometimes
taken as the resistances introduced in the presence of eo·
phased osci Jlations.
If we assunte that Z11 = R11 +iX, 1 , Z,,=R,,4-iX,, and
Z 12 =R 12 +iX 1 ~ the equations (3-21) may be given the fol-
)o\ving form convenient for ea lcuJa tion purposes:
Z 0 , = R11 + m (R 11 cos "' ·- X 11 sin '\f') + i [X 11 + m (R 11 sin 1IJ ~t­
-t- X,, cos"'')];
Z01 = R11 +7i (R 11 cos 'lj:1 +X11 sin '1,) + i [xu- ~ (R 11 sin'lj:1-
-X11 cos 'lj:1)]. (3-22)

The equations enable to calculate the full input impedances

of the two coupled antennas if the ratios of the currents.
i.e., m and 1J' are known beforehand.
~Vhen reactances are connected at the. antenna input with
the object of tuning them up, the quantities X 11 +Xu and
X 22 + X at, where X H and X 2 t are reactances inserted into
the antennas at their feed points and not associated with
the radiation process, should be introduced into all the above
given expressions, instead of the separate reactances Xta
and x22•

85
 Let us now define the power fed to the system and radi-
ated by each of the antennas.
The power supplied from the generator to antenna 1 ·
and rad ia1ed by the system is:
I o• I o* t .
Pr.1 = 2 [R 11 -t-tn(R, 1 cos1J'-X 11 stn ~')1, (3-23)

and the power suppl ieci to ;tntenna 2 and radiated by the

~ysten1 is:
I o2 I *o 2 1
Pr., = 2 R•• + nz (R 11 cos 11' +X,. sin 'I') . (3-24)
The total power radiated by the systen1 is equal to the
sun1 of (3-23) and (3-24):
*
2 R cos1J'].
P1:= I OJ 2I 0 1 [R 1 ,+m IR •• +m (3-25)
12

Whence we detertnine the radiation resistance of the whole

sys1en1 with respect to the curJ ent at 1he feed points of
antenna 1:
(3-26)

3-3. Active Reflector

As we saw in Paragraph 3-1, if nz=l, '\f'=9o• and
d l
A. = , antenna 2 reflects energy towards antenna I. In that
4
case, the field in the direction of antenna I in the radiation
zone is doubled in comparison with the field of one antenna.
To obtain the stated ratio of the currents (m= 1, ,J, =90°),
both antennas should be fed at a definite relation between
the feeding voltages U, and U2 • In .that case, antenna 2
is referred to as the active reflector.
Therefore, in accordance with the ex press ions (3-22),
the fuJJ input impedances of the antennas are:
ZOI =R •• -XII+ i (XIJ + R•• ),
zo.=RI! +XII+ i (XII-RI!). (3-27)
It fo11ows from this equation as well as from (3-26) that
the radiation resistance of an antenna with an active
re fleet or is:
(3-28)
86
 3-4. Passive Antenna
Let antenna 2 be passive, i.e., such that it is not frd
frorn the generator (U 2 =0) and is excited by the field of
antenna /.
Let a reactance X 2 t be connected to antenna 2 \Vith the
ohject of tuning it up (Fig. 3-5). The equivalent circuit
of such a systern is sin1iJar to that represented in Fig. 3-4
anrl differ~ frorn it only in that, instead
of the· genera tor of voltage U 2 , the 2 f
reactance X ::!t is connectPd to contour 2.
For this circuit, Kirt. hhotT's equation~ I,
are written as:
u,
ul =I o,z" --t- I 02z,2;
0=/o2(l22-4-iX2t) t /o,Ziz· (3-29)
The left-hand side of the second equa- •
tion is zero. becausr U 2 -=- 0. Incidt.\ntal lv, J
Fi~. :_J-.3. ~y~teru
it fol10\\IS that Z 02 -=-=.l~'.!--:. 0, i.e., since consisting of one
1 11 ~ active and on(' pas-
it is passive, the full itnpcdancc of . sive nntenn:l.'\.
ant('nna 2 equals Zl\ro. The st~cond equa-
tion (3-29) ena hlcs to dett·nuinc the currcni induced in the
passive antenna,
(a-JO)

Substituting (3-30) into the first equation (3-29), we obtain

the ("'X press ion for the itnpedance of the active antenna, i.e.,
of antenna 1,
(3-3 J)

On comparing (3-30) with (3-1), Wfl obtain the following

expressions convenient for ea lcu!ation purposes
R 212 ·t X 211 •
m= (3-32)
R ~22 + (X 22 + X 2 t) 2'
1Jl = n +arc tan ~ 11 -arc tan
11
Xu: X
22
21
• (3-33)

The expressions (3-32) and (3-33) are generally used for

calculating the current amplitudes and phases in a passive
antenna.
87
 Since Zo2=0, the radiation resistance of a system con-
sisting of an active and a passive antennas is expressed as:

R!.o=R 11 +m (R 11 cos ..p-X 12 sin 'i'). (3-341

When X at =0, the current in the passive antenna is always
smaller than that in the active antenna (m<l), and di-
minishes continuously as the distance between the antennas
increases, owing to the fact that the n1agni tu de of the
coupling resistance is always srnaJler than that of the
resistance of the passive antenna.
When the distance between the antenna is very smaJJ,
outing to the fact that Z. 2 - ,.Z 22 , we have: nz~ I, 1J'~n and
R~(~~ 0. In that case, the two antennas system ceases to
radiate, because the energy supplied to the system by the
oscillator does not take up the entire space between the
antt\nnas.
\\'hen the passive antenna is tuned in resonance and
X 22 + x· 2 t =0, the current in it reaches its rnaxin1um value
at all distances; at small ones, d may become larger than the
current in the active antenna. In the case of resonance,
(3-32) and (3-33) take the following fortn:

R 212 + x~11 .
m= ~-___;:_;;;

Rzz '
(3-32a)

(3-33a)

When the value and phase (depending on distance d and

tuning X 2 t) of the curreut excited in thr passive antenna are
such that the n1aximum field is set up in the direction of
the active antenna and the minin1um field, in the direction
or the passive antenna, the latter is referred to as a passive
re Hector. }-lo\\'ever, the passive antenna n1ay be tuned up
in such a \\'ay that the max irnurn energy wi 11 be radiated
in the direction of the passive antenna and the minimum
energy, in the direction of the active antenna. In that case,
the passive antenna directs energy in its own direction and
is consequently referred to as a director.
ln the case of a passive reflector, the optimum dtstance
between the antennas lies between d=0.2A and d=0.25A..
Hence the difference between the directional diagram and
that of the antenna with an active reflector (Fig. 3-3) is
that in the passive reflector, there is al\\·ays a lohe in the
direction of the retlector, whereas no such lobe is observed
in the case of the active reflector. This is due to the fact that
it is in1possible in the passive antenna to fulfil sinlultaneous-
ly the conditions m= 1 and 'IJ=90° when =0.25, since f
the operator can only change the rnagnitude and sign of
the resistance X 2 t.

3-5. Calculating Mutual Resistance of Antennas

To calculate the n1utual resistance of antennas, the gt'ner-
al expression (3-19) obtained by the induced ernfrnethod
is used. This n1ethod was first applied to the calculation
of the rr1utual resistances of antennas for1ning part of an-
tenna systems by A. A. Pistolkors in 1928. Applying the
expr(~ssion (3-19) to the calculation of the coupling of two
parallel antennas of equal length, as shown in r:i~. 3-J,
one needs to substitute into it the corresponding vahJ£:as of
Ez 12 and /z,· If we assun1e the distribution of the currt~nt
to be sinusoidal, In =J~/-iil sink (1-z), .the expression for
Ezu. will be obtained in the form of (2-25). The calculation
proceeds in just the san1e way as shown in Paragraph 2-7.
As a result, \Ve obtain the following expression:

Z 1 2 = s1~~-
2
n kl
{ [1 -t- 2e- ikl cos kl] [Ci kd-

-Ci k (l/ dz -~- l - l ) - i Si kd + i Si k <Vti•

2
+ 1 --l)] + 1

+[I-!'- 2eikt cos kl] [Ci kd-Ci k <Vtr + t• + l)-

- i Si- kd 4- i Si k ("Vd 2
+l 1
-~ /)] -+ eiskl [Ci k <Vd +41 -r 2/) --
1 2

1 1
-Cik(V-d*+l2 +l)-iSi k (J/.d +4l +2l)+
+ i Si k(V d + l + l)l + e- ilk/ [Ci k <V d + 41 --2/)-
2 1 1 2

-Ci k (J/ d + l -l)- i Si k (Vd* -+- 47 2/) +

1 2 2
-

+ i Si k (Jfd + l -l)l}. 1 1
(3-35)

In the limit case, when the antennas 1 and 2 coincide

(d=O), this expression defines also the natural resistance
of the symmetrical antenna Zs1=Z22·
89
 .. -
+70'" H,3
·- -
- -
.

+60
-
-
+50 ,....

+40
-
- ..
- -- -
+JO
f-~

-
-+20
-
~~ ,_
+10 loo.
r ~
1 ~
- I 'N l
r/
1\.
'-~
-
rti/,A.
0
0 ~
2 ., .':I
J
~a
- \ '"'- J
- . - ~- - ifJ ~
-to -
~
V
- --
1---
-
11
-20 l
J
- -
1\.
I

Fig. 3-6. Cu!ve of the resistive con1ponent or the mutual resistance

of half ·wave antennas.

Figs. 3-6 and 3-7 give the curves of the resistive Rt 2 and
reactive X s2 components of the mutual resistance of half-
\vave antennas, calculated by means of (3-35). As can be
seen from these curves, in the limit case d=O, for infi-
nitely thin antennas, the mutual resistance pa~ses into
the natural resistance of the half-wave antenna and ac·
so
quires the values Rt2=R 11 =73.1 ohms and Xt2= . Xu=-=42.5
ohms.
Appendix I contains a table of the values ot R1 2 , as cal-
culated by A. A. Pistolkors and first pubJished by him in
1928 [13 ]. Appendix I I contains the tables of the values
of R: 2 and X1 2 , as composed by V. V. Tatarinov and pub-
lished by him in 1936 12]. The tables concern half-\vave
antennas spaced at a distance d apart and displaced along
the axis by a distance h in wave-lengths.
To illustrate the pcrforrnance of t\\'O coupled antennas,
one of which is passive, Fig. 3-8 shows the curves of R~o,
tp, I 02 , lo E-:;=o and E,... as a function of the n1ag-

+41)
.

+30 •

~20

1- ~ l.
11
-
+10
~ ~
,j,..

1-- - I
, , ,J ~

~.
~-

0
0 r • 'tt
I
'
-~'- ~
~

~
~

~
I
~ ~

-to ~
,
J
5M
-20

-30
=~
-40 M
fig. 3-7. Curve of the reactive component of the mutual resistance of
half-wave antennas.
91
nitude of the passive antenna reactive impedance X22+X2t•
In the same figure, the currents I 02 and I 01 and the field
360" t8
•
1/1
--..ol
t'fL//
£,.
_.. ----- - V - 1.6 llllo.

"" ~-o·
280~ ~ ~ ~,_ t4
" ........... .....
'-lln•
·- .- ....,.,.
- k l2
~--
-, i"-
200° ~ I.OI 1......
~~
- -- - ....... !
.. Ill ""' i\.
I

- •
-"'"'"-- -- 100

f60a '7 lllllliii .~, ~ ,_ - ~---· ~--

80
- ,
~> ~ ..-r:. , ..............
;f'.:JJ'J"
..... _ ,,:_ pa-: 60
~

I~V r-._, Ill"

t'tEII
BOa ~
~~ ~- - ~---- .
0.4
,... ___
lo-_....
r--...:
~

-.... 40
f=D.25 1=0.25 -
4d'J 0.2 20
.__
KJJ +Kit
-tOO -80 -~~ -4{! -20 0 20 4D 60 80 100
Fig. 3-8. Type of operation of a system consisting of an active and a
passive antennas plotted as a function of the tuning of the passive
antenna.

intensity in the direction of the active antenna E,=o and

in the direction of the passive antenna E~= 180 o are given
in relation to the current and field intensity of one single
antenna on condition that the radiated power remains the
same in both cases. As is seen from this figure, the passive
antenna plays the role of a refte_ctor when its reactive im·
pedance X22 +Xt is positive, i.e., is of an inductive nature;
this same antenna plays the role of a director when Xa 2 + Xt
is negative, i.e., has a capacitive nature.
 CHAPTER FOUR
Radiator Systems and Antenna Parameters
.
4-1. General
In the preceding chapters, we saw how, by con1bining
dipoles or antennas (electric as well as rnagnetic), one could
obtain various directional characteristics.
In the present chapter, we shall discuss n1ethods of ob·
taining highly directional characteristics or radiation charac-
teristics by means of systerns consisting of a large nun1bcr
of relatively simple radiators discretely or continuously
distributed in space. Methods of obtaining highly direction-
al characteristics deserve special attention due to the
practical importance of devices with characteristics of
this particular kind.
The systems of radiators are made up of wire, sJot, dielectric
or other types of radiators. In aJJ cases, use is rnade of the
radio \vave interference phenomenon according to which, the
radiation of all the radiators are added up in some directions
and subtracted in others, due to differences in path length
of the rays from SP,para te elements of the systen1 and
differences in the phases of the currents in these elements.
We shall examine separate systems of radiators with differ-
ent current amplitude and phase distributions, without going
into details as to how these distributions can be obtained.

4-2. Radiation of a Linear Co-Phased Dipole Array

Many antenna systems utilised in practice are rnade up of
half-wave dipoles suitably disposed in space and excited
at definite ratios of current amplitudes and phases. We shall
98
Investigate the radiation of a system of dipoles lying on
one plane and excited in eo-phase.
Let nm half-wave dipoles, forming a linear array as shown
in Fig. 4-1, lie at a distance d, apart bet\\'een the centres
of neighbouring dipoles
in different ro\vs, and at
...... m•B a distance d 2 bet\\1een
d, the centres of neighbour-
• - ~--.....v ing di poJes in one ro\v
(n being the number of
-n•l rows of dipoles and m,
Fig. 4-1. ;\rray of he11f -wave dipoles.
the number of dipoles
in one row).
In accordance with (2-16), the field intensity of each
of the dipoles in the radiation zone is:

_ ros ( :;- cos B)

E- Eo . u
SUI ' (4-1)

where 6 is the angle between the direction towards the

point of observation and the axis of the dipole ;
601
E0 =i a e-ikr., .the field intensity of the dipole in its
'o
equatorial plane at the distance r 0 •
/
/
/
/
z 10
ktt,sina
i
11, ktl,sina
T z /J
nkd,slna

miEOI lrd,sin \
p jJ \
11
c '
Fig. 4-2. Explain ing Fig. 4-3. Vector diagratn of the composi-
the calculat ion of the tion of the fields of a system of dipoles.
radiatio n of a dipole
array in the magnetic
vector plane.

To begin with, let us investigate the field in the xz-

plane, i.e., in the equatorial plane of the dipoles or in the
9~
rnagnetic vector plane (Fig. 4-2). Since at the point of
observation, the directions of the field v-ectors from the
separate dipoles coincide, the resultant field intensity of
the system in that plane is defined as
E!~t= mEO [1 -t- e-lkd, sin + e-i'J.kd, ~In a+ (1

••
+ • • •
-L e-I ( n -
I
1) k d I sIn a]
•

I
We have taken into clccount that, bccau~e - R.:l

I I
'o
~ -- ~ ... :::::: the atnplitudes of the field E" 0 of all
'• 'n-t
the dipoles are the same and the difference in path length of
the rays from neighbouring rows equals rr- 'r-• =d 1 sin a,
where a is the angle between the direction to\vards the point
of observation and the z-axis:
This expression correspond~ to the vector d iagrarn sho\VII
in Fig. 4-3. Let Q be the radius of the circurnference de·
scribed near the polygon forn1ed by the vectors m 11:~ 1 e-lktl.--lnt.c.
Then, from the triangle aho, we find

m l£ 0 1=2{! sin (~sin 1.1),

and from the triangle aco, we find
IE tot
2 . liil~d, . )
.tz 1 = QSJll \
2 - stn a .
Hence
nlld 1 )
tot sin ( 2 sin a
(Exz l=nzf£of (kd 1
~-. (4-2)
sin sin a)
2
We have used the factor m because from the point of view
of the formation of the directional diagran1 in the equa .
torial plane, a11 the dipoles of one r£'w hehave as one dipole
with current l=mlt. where / 1 is the current amplitude in
one dipole.
Let us analyse the expression (4-2). When a=O E~~t =
=nmj£ 0 1, i.e., in a direction perpendicular to the plane of the
array, the field intensity is equal to that of one dipole in the
system, multiplied by the total number of dipoles. l"he
radiation zero occurs in the direction defined from the
condition
nkd. s1na
2
·
0
= N n, N = 1, 2, 3
from which we derive
- N)~
stn ao = d • (4-3)
n '
Thus, for example, when n=B and d 1 ={ sinao=~ and there
is no radiation in the direction of the angles a 0 = 14.5°,
ao=30° and cx 0 =48.6°.
Fig. 4-4 gives the
directional diagram
for that case, calcu-
lated from (4-2).
As can be seen, the
n=8 directional diagratn
has a major lobe and
a nu1nber of minor
side ones. The nunl-
·900 bcr of minor lobes
is the 1arger, the
larger the number of
rows of dipoles. In
...
the example under
consideration, the
width of the major lobe
at the zeros of radia-
tion equals 29°. In the
general case, when
there is a large nu1nber
Fig. 4-4. Directional diagram of a eo- of dipoles and they are
phased dipole array. spaced sufficiently
wide apart, the width
of the major lobe at the zeros may be defined by the fol-
lowing simplified expression derived from (4-3)
A.
2aft = 2 L [rad]. (4-4)
where L~nd1.
Thus, the directivity of the eo-phased array is all the
larger as the width (or length) of the system L is larger
relative to the wave-length A.
Let us now define the minor radiation maximums. For
small values of d 1 in comparison with the wave-JPngth, the
variation of denominator in (4-2) can be neglected in the
region of the minor maximums, so that fhe position of the
B6
 minor maximums can be found from the condition
'
. ( rtkdl Stn
Slfl . CX.max ) =
2 ± 1.
Hence
. 2N + 1 A. N I
SJn amax = 2 nd• , == ' 2•
'"->
\."1 ... • (4-5)

Subs tituti ng this value into (4-2), we obtain:

tot I- .. --I ( 4.. 6)

s1n -- ---
tl 2
We see that the magnitudes of the rnaxin1un1s of the ntinor
lobes of the directionai diagram of a eo-phased array cannot
be smaller than _!_ of the rnagnitude of the rnaxin1um of
n
the rnajo r lo.be, because in (4-6) the denon1inator is always
sn1aller than unity.
It can be further seen from (4-6) that for a large value
of n, the magnitude of the tnaxh num of the first rninor lobe
equals approxintately
tot 2
IExz (ama x)l=n ml£o f3n,

i. e., constitute:; approximately -K of the magnitude of the

maxi mum of the major lobe.
Let us also note that it fo] Jows frorn (4-2) that the nutn her
of dipoles in a row of the eo-phased array does not in-
fluence the forrn of the directional diagrarn in the magnetic
vector plane and deterrnines only the absolute value of the
field intensity.
Let us now examine the directional diagram of the array
in the yz-pJane, i.e., in the electric vector plane. Let P
be the angle between the direction towards the point of
observation and the z-axis in that plane (Fig. 4-5). Then,
similarly to the above, we obtai n for the total field in the
radiation zone the expression ·
tot . t'A)
cos ( -n Sill
Evz = n£ 0 2 ll -+- e-lkd. , ,,"~ +
cos~
+ C -llkd 1 sin p -r ••• + e-1 (m-1) kda Slfl fiJ,

7-212 2 97
or
tot cos ( -f sin ~ sin ( mk:• sin ~
(4-7)
sin • sin p
2

liere, the first factor on the right-hand side of (4·7) (dis-

regarding the factors n and~ Eo I which do not depend on the
direction) represents the directiona l character istic of a
V baJf-wave dipole and tbe second
factor is that of the combination of
the radiation .of the non·directional
radiators. All the deductions which
were drawn with regard to (4-2)
T
llj
apply also to this second factor
in (4-7). It is clear that the form
I of the directional diagram in the
electric vector plane depehds on
"'----,'-------~~~ the number of dipoles in one row
, m and does not deplhd on the
number of rows n. Only the radiation
intensity depends on the nuJnber
- of ro\vs in that plane.
It should be further noted that
when the array is square, i. e., when
Fig. 4-5. Ex J)l aining the
calculation of the radi· n=m and dt=d 2 , the directivity
ation of a dipole array in of the radiation in the plane of the
the electric vector pi an e. electric vector of the array is some-
what larger than the radiation
directivity in the plane of the magnetic vector, due to the
presence of the first factor in (4-7), -Le., due to the directivity
of the radiation of the dipole in its meridional plane.
The expressio ns (4·2) and (4-7) define the directional
diagrarns of the eo-phased array in the two principal planes,
One could also determine the radiation in the other planes,
but the picture of the radiation in these two planes is found
to be perfectly adequate for technical purposes.
In antenna theory, one often meets the concept of normal-
ised radiation characteristics, by which one understands
the ratio of the magnitude of the f1eld intensity in the radi-
ation zone 'fn a given direction to the magnitude of the field
intensity in the radiation zone in the direction of the maxi-
98
mum radiatio,n. Thus, for the magnetjc vector pJane,
the normalised radiation characteristic of a eo-phased array
is expresSed as

F (a)=
1E~~t I 1 sin (
=-
sin a) nk:, (4·8)
nm I E0 f n sn1 ( kd, sna}
1 \ '
2 I

and for. the electric vector plane, the normalised radiation

characteristic of the array is given as the product of two
normalised radiation characteristics
...

(4-9)
where
cos (-;sin p)
•
FI ( p) =.: cos p (4- J0)

is the norQ;Jalised characteristic of a half-wave dipole, and

1
sin (m~da sin p)
(4-11)
sin 1
sin p
2
the normalised characteristic of m non-directional radiators.

4-3. Radiation of a Linear Dipole Array with Variable

Phase
Let us exarnine an array consisting of one row of half-
wave dipoles (Fig. 4-6) excited by currents of equal
amplitude but with a
!I
phase shift 1J' between the
currents of neighbouring z ,
dipoles, so that
I ,- I e+tcr-1>• -
- 1 ' *"
where r= I, 2, 3, ... , n.
Designating by q> the ..Jd .......
angle between the direction Fig. 4-6. Dipole arrpy with vari
towards the point of obser- able current phase.
vation and the x-axis in
7* 99
the magnetic vector plane (Fig. 4-7), we obtain for the
total field in the radiation zone:
E~~~ = £ 0 [1 -1- e+' <"'-kd cos (f)) -t- ... -t- e+t <n-1> <11>-k.t ros q)>],
where

Hence the directional diagran1 of the array is expres~ed as

tot sin ( i {'Ill- kd coscp)]

(4-12)
sin (1p- kdcos cp)
2
Let us exan1 ine particular cases of excitation of the array.
When 11-'=0, we get the case of the eo-phased dipole~
discussed earlier. Assun1ing that in (4-12) 11'= 180°, we
obtain the expression for the antiphase dipole array:

tot sin ri (180~ -- /.·d cos cp >]

I Exz /= J Eo I 1 - --- (4-13)
- sin [ (180°-kd cos!p)J
2
It follows from this expression that in the direction q> = ± 90°,
when n is even, the radiation vanishes. If we assume that
d/'A=0.5, then, in the direction of
the axis of the array (cp=0° and
cp= 180°), the radiation wi!l be at its
1 maximum and equal to n I E 0
Assuming _that in (4-12) 'I'= kd,
2
we obtain the expression for the dipole
array with a phase varying in ac-
J cordance with the travelling-wave law:
z
sin rn;d (I - coscp) 1 (4-14)
sin (1-coscp)
2
n
\Ve see that in the direction cp=O,
Fig. 4· 7. Explaining such a system always gives a maxi·
the calculation of the mum of radiation equal to n 1E.\,
radiation of a dipole
array with variable irrespective of the value of d/A. When
phase. di'A ::.;;0.25 and n is ~n odd nurnber,
I

lOO
we get a zero of radiation in the direction ff= 180°. Thus,
in this last case, the radiation is found to be unidirectional:
the energy is radiated in the direction of 'the rnot ion of the
wave in the array, none being radiated in the reverse di-
rection. Fig. 4-8 shows the directional diagran1 of an array of
eight dipoles spaced at a distance A/4 apart and oscillating
with a phase shift 1jJ=90°.
10. .

n=8

180"------..::: ~----"-··- f!J=o·

270.
Fig 4-8. Directional diagram of an ·array with
a travelling wave.

If the antennas are arranged in one line at a distance d

apart (in accordance with Fig. 4-5, but taking into account
that the angle e is complen1cntary relative to the angle
p in that figure) and excited \Vith a phase shift of the cur·
rents '~>· the directional diagram is expressed as

101
cos ( i ~os 9) sin [ i (1JJ-kd cos 9) 1. (4-15)
~in (11-,- kd ens 0)
2
It is also of interest to note that from (4-12) as well
as from (4-15), it follows that the principal max \mum ot
radiation in the general case lies in the direction which
is determined fron1 the condition .

COS
- kd19
amax- • (4-16)

Thus, for example, when d/A=0.5 and 11'= 10°, we have

Umax = ± 86.8°.
101
 4-4. Radiation Resistance of a Dipole System
In the two preceding paragraphs of the present chapter
we have examined characteristic examples of the formation
of highly directional radiation characteristics by means
of a linear array of haJf-wave dipoles. To obtain a definite
distribution of the amplitudes and phases of the currents
in the dipoles, we have to knour the resistance at the termi-
nals of the dipoles of the given system.
Extending to the dipole system the relations for two
coupled dipoles discussed in Chapter Three, let us write for
an array of n dipoles, the following system of Kirchhoff's
equations:

• • • • • • • • • • • • • • (4-17)
U, =I ,zn, -t-1 1 Zn• -f- ..• + I nznn•
where Ule and I 1c are the voltage and current at the
term i na Is of the k- th dipole;
Z pll• the 1nutual resistance between the p-th and
k-th dipoles;
zltk• the resi~tance of the k-th dipole proper.
Dividing V k by '"J k' we obtain the resistance of the k-th
antenna, with due account to the effect of all the other
dipoles in the system
z, = ~ z,, + ~ Z-_~t.t + ... +Zu + ... +~ Zl!ra· (4-18)
Thus, to detern1ine the impedance of a dipole we have
to know the resistance of the dipole, the mutual resistance
between that di poJe and all the other ones, as well as the
magnitudes and phases of the currents in all the dipoles
relatively to the current amplitude and phase in that dipole.
To give an example, let us calculate ~he active part of
the hnpedance of one dipole in the eo-phased array of half·
wave dipoles shown in Fig. 4-1. whenm=2, n=2, dt=iand
d2 =-}. Note that in this example, the impedance of all the
separate dipoles are equal due to their symmetrical arrange-
ment. Since ~-=1, in accordance with (4-18), theactive
part of the impedance of the first dipole will be:
R,=R,,+ R,.+ R,.+ .R,,.
102
 l~rom A. Pistolkors's table (see Appendix I) we find:
Ru=73.1 ohms, Rt2=-12.36 ohms, ,Rta=26.4 ohms,
Ru=-11.8 ohms. C:Onsequently, the impedance of each of
the four dipoles will be Rt=R 2 =Ra=R.=75.34 ohrns.
The computation of the active as well as the reactive
part of the impedance of the antenna in any other system
is performed as in the example quoted.

4-5. Radiation of Continuous Systems

of Sources with a Travelling Wave
Let us investigate the radiation of a thin rectilinear con-
ductor of length L, through which there flows an electric
current varying along
the conductor in ac-
cordance with the travel-
ling-wave law:
I = 1oe- i k X ( 4-19)
J(

where I 0 is the current ....._-z-~I~

at the origin of the con- .._,__ _ _ L _ ____,...,.
ductor.
Let us choose on the Fig. 4-9. Explaining the calculation
conductor an element of of the radiation of a conductor with
a travelling wave of the current.
current I xdx (Fig. 4-9).
In accordance with (1-3), the field oi this ele.ment in the
radiation zone is:
. I xtJxk 1 • e- ikr
dE=-t 4nwe s1n8 r •
Substituting· here the expression (4-19) and taking into
account that for the radiation zone, ..!. ~ ..!_ andr=r.-xcos9,
'o r
we have for the total field:
L
E = i I ok• sin 8 e-lkro
4nm£ r,
5 e-lkx (I-COS 8) dx.
z=o
The stated integration yields
. I 0k1L . 8 e-tkro -tez» sin Cl» (4-20)
E =J 4 stn e "' ,
2T,(l)8 '• • w
where
kL
cp-= 2
(1-cos 9).
103
To begin with, note that in the direction of the conductor
axis, the radiation equals zero, because when 0 =0, sin a=
=0. The main maxitnum of radiation is at an angle to the
axis of the conductor that satisfies the condition 0<0max <90°,

uo·
'
'

170" 270°
Fig. 4-10. Directional diagran1 of Fig. 4-l I. Directional diagram
a conductor with a travelling wave of a conductor with a travelling
of the current U'A= 1.0. wave of the current LI"J...= 1.5.

i.e., the radiation of a conductor with a traveJiing wave

of the current is directed at a certain angle to the conductor
in the direction of travel of the wave. Furthermore, the
longer the conductor rei a-
tively to the wave-length,
11 the cJoser is the rnain
1 -tnaxitnum of radiation
·1-U..J..L.u.L.LUU+-!:-L-'..&...£..~..~,..~..~'-t--rl_z against the conductor
.-'...,.__-:& -~~ axis and the larger the
------ i nutnber of minor lobes.
Fig. 4-1 :l. Layer of transverse electric This is illustrated in
currents. Figs. 4-I 0 and 4- 1I
showing the directional
diagram of the conductor for LIA=l.O and LIA-=1.5,
calculated according to formula (4-20).
We have exatnined the radiation of a conductor with a
travelling wave of the longitudinal current. In practice one
nlfets travelling-wave antennas in which the ~l~ct.-ic
currents have components perpendicular to the axis of the
antenna. Let us therefore exatnine the radiation of a thin
layer of transverse electric currents propagated along the
layer in accordance with the travel ling-wave law (Fig. 4-12).
Let Jo be the density of the surface current at the beginning
of the layer. Then, the density of the current at the point
x of the systE:'tn \Vi 11 be:
JY = Jue- iki.x, ( 4-21 )

where ~=~ is the ratio of the velocity of light to the phase

v~Jocity of the current wave in the layE'r.
Let us select at the point x of the layer, an elerncnt of
current J~ldx. The intensity of the field set up by this ele·
n1ent in the radiation zone in the xy-pJane wi JJ be:
. J tdxk 2 e-ikr
dE=t 1 4nwt.
cosa r ,

where a is the angle between the antenna axis and the

direction towards the point of observation.
Integrating this ex press ion along the \\'hole length of
the layer, we obtain:
(4-22)
where
kL
'l' = 2 (~-cos a).

If we eliminate the factor cosa from (4-22), we obtain

an expression which defines the field in the xz-pJane of the
'
antenna.
Note that the main radiation maximun1 occurs in the
direction which is determined from the condition lJI ==0 or
COS a,max = ;.
This condition is fulfilled when 6~1, Le., when the phase
velocity of the wave in the layer is e~qual to or higher than
that of light. When ;=0, the velocity of the wave in the
layer equals infinity (eo-phased osci Jlat ion of currents)
and the angle of the radiation maxin1utn a,max is found to
equal 90°. When ;-.1, the velocity of the wave approaches
that of light and the direction of the radiation maximum
approaches the axis of the layer (amax 0). --i>

105
 Let us determine the width of the directional diagram
when ;= 1, i.e., when the radiation maximum coincides
with the axis of the layer (amax=O). Assuming that in (4-22)
'Y=rc, the following expression is obtained for the deter-
mination of the width of the directional diagram at the
zeros of radiation:
).,
cosao= 1- r.
When i·~ 1, the angle a is sufficiently small and for
the width of the directional diagram at the zeros of radi-
ation, we obtain the approximate expression:

2a0 =2 Y 2 ~. (4-23)

For the determination of the width of the directional dia-

gram at half power, we assume that in (4-22)
sin (k;(l-cosa. 1,)] =0.707[~(1-cosa.,,)],
from which we obtain the approximate relation

-2a•,,=--=2 Yo.886~. (4-24)

When ~>1, the wave phase velocity in the layer is inferi-
or to that of light. In that case, in accordance with (4-22),
the maximun1 of the major lobe of the directional diagram
coincides with the layer axis. At the same time, the size
of this lobe relatively to that of the minor lobes or the
radiation intensity in the princip~l direction depends in a
complex way on tht' value of L/'A, wher£las when £==1 the
radiation in the principal direction is the larger the larger
is L/'J... •
As we shall see below, to each value of the relative
velocity of the wave in the layer (!..<1) there corresponds
VJ
a definite electric length of the layer (L/A) for which the
maximum radiation of energy occurs in the main direction.
This relation is found from the condition lf'==-j a.nd is
defined by the ratio
(4·25)
106
Antennas with a low phase velocity, the dimensions of
,,,.hich satisfy the condition (4-25) are usually referred to as
antennas of axial radiation of optimum length.
The width of the major lobe of the directional diagram
at the zeros of radiation is determined from the expression
A.
cos a 0 =~-L,

and if we substitute in it the value of ~ taken from the

condition (4-25), the width of the major lobe at the zeros
is determined fron1 the expression
A
coscxo= 1-u.
For a highly directional radiation we may assurne that
a•
cos a 0 ~ I-i. Then

2tt,=2Y~. (4-26)

As for the width of the directional diagram of an antenna

of optimum length at half power, it is obtained from the
expression
sin [k2L(;- cose~, 1.)] =0.707 ~ [k; (6- cosa, 1.>]
L
and ior A~ 1, it is approximately defined as

2Gt't.=2 J!o.28 ~ . (4-27)

Comparing (4-23) and (4-24) with (4-26) and (4-27), we
note that an antenna with a low phase velocity of optimum
length has a narrower directional diagrarn than an antenna
the phase velocity of which is equal to the velocity of light.
A comparison of (4-23) and (4-26) with (4-4) shows that
antennas of axial radiation possess a considerabl) lower
directivity than antennas of the sa~ length consisting or
a series of linear dipoles with a co-phasaJ excitation. This
is due to the fact that antennas with an axial radiation
concentrate energy in a narrow bunch of waves in two planes
whereas eo-phased linear antennas concentrate energy in a
narrow bunch of waves in one pJane only.
101
 As for the minor Jobes of the directional diagram, it is
clear that for an antenna of optimum length with an axial
radiation, the rnagnitude of the maximum of the first n1inor
lobe amounts to 11, of the magnitude of the n1aximum of
the major lobe.
Note in addition that if the length of the antenna is
doubled in con1parison with the optimum one, i.e., the length
determined fron1 the condition (4-~5), then in the principal
direction (a=O), the radiation is found to equal zero.
Let us consider the example of an optimun1 antenna of
axial radiation. Let ~-=-= 1.1; then L/A= 5 and the width of the
directional diagram at the zeros is 2a 0 =51.5° and at half
power, 2a., . ~ =- 27. 2° .
.

4-6. Application of the Equivalent Surface

Currents Theorem to the Calculation
of the Radiation Characteristics of Antennas
Before we continue to discuss the radiation of other possi-
ble systems, let us dwell on the equivalent surface electric
and magnetic currents theorem, known to the reader
frorn the course on clrctromagnetic fif:'Jd theory [ 14 and 15 ].

Regionf Hegion~

E-
p
4nfenno
-
En
,_ Hn

•
Fig. 4-13. Explaining the equ•valent surface cur-
rents theorem.

Here is a brief explanation of this theorem. Let us examine

the radiation of an antenna lying in a free space. Let E
and H be the vectors of the electromagnetic field set up by
the antenna. Let us divide the whole of the space with the
arbitrary closed surface s enveloping the antenna into two
regions: I and 2 (Fig. 4-13).
108
 {

Let us expand the E and H vectors en the surface ~ 111

two component vectors, one tangential and the other nor1nal
to the surface:
E= E 1 -t-En:
H=Ht+H,,.
Using the linearity of MaX\\'ell 's equations, Jet us mcnta11y
examine two pairs of vtctors separately:
1) the pair of vectors H1 and E," assun1ing that E1 =-0
and Hn=O;
2) the pair of vectors E, and H, assun1ing that H,=O
and E,1 =0.
In the first case, the surface s behaves as an ideal electric.
conductor, since the n1agne.tic field vector on it has only
a tangential component and the electric field vector, only
a norrnal con1ponent. That is why, in that case, a surface
electric current appears to flow on the surface s, the
density of which is expressed as
(4-28)
where n is the external norrnal to the surface under consider ..
ation (region 2).
The vector potential of this fictitious surface electric
current at a certain point P of the region 2 n1ay be
determined frt,nl the ex press ion
1 C e-ikr
Ae=iitJJe r ds, (4-29)
s
and the intensity of the electric and magnetic fields at the
same point · P, from the expressions
Hc=rotAe;
Ee =z - trof.1
. Ac -f- . I , gra d d.1v Ae . (4-30)
l.(j)£

In the second case, the surface s behaves as an ideal nlag-

netic conductor, since the electric field vector on it has only
a tangential component and the magnetic field vector, only
a norn1al component. 'That is why, in that second case,
a surface magnetic current appears to How on the surfact! s,
its density being equal to the tangential component of the
electric field intensity and determined fron1 the ex pressi"n
JM =[n, E,]=[n, E]. (4-31)
109
The vector potential of the fictitious surface magnetic
current at the point P. is determined from the expression
4n
s
A M = _!_ J,. e-;1,. ds
r '
(4 ·32)
s
and the electric and magnetic fields at the po;nt P are
determined from the expressions
EM= -rot AM,
H M =-· troE
• 'AM -,-
_l
.I
gra d d'IV AM • (4·33)
tOOf.l
The total field at the point of observation P is thus de-
fined by the sum
E= Ee+ E.. ,
H=He+ H... (4-34)
Thus, the field in the region 2, which is free from
sources and bounded by the surface s, may be set up by the
electric and magnetic currents distributed on that surface
and, in that sense, the real sources in the region 1 may
be replaced by the "equivalent" surface electric and
magnetic currents.
The choice of the surface s enveloping the true sources of
the field is absolutely arbitrary. In the case of, for example,
a symmetrical dipole, the surface s can be coincided with
the surface of an ideal conductor in the centre of which there
is a gap for connecting a high-frequency oscillator. In this
case, the fictitious surface electric currents are identical
to the real surface electric currents flowing on the surface
of the conductor. As for the gap in the conductor, within
this gap, the surface electric current as well as the surface
magnetic current are fictitious and provide an equivalent
substitute for the real electric currents flowing in the
internal region of the gap in the conductor.
It is frequently found more convenient to calculate the
radiation characteristics of antennas by performing the
integration not over the volume which contains the real
sources of the field but over a certain closed surface enclosing
the real sources. This is the way to calculate, for example,
the directional characteristics of horn-type antennas, cover in&;
the horn with the surface s and integrating the radiation
of the fictitious sources lying at the output opening of the
horn (neglecting the radiation of the sources disposed on the
remaining part of the surface s).
110
 Now that we have given the necessary explanations re·
garding the equivalence theorem, let us select on the surface
s under consideration an elementary area ds and define the
field created by this area at the point P. It is evident that
we must regard this area as a Huygens elernent. Let us
coincide the area with the xy-plane of a (',artesian system
of coordinates and Jet the origin of this system lie in this
area (Fig. 4-14).
1/ :&------ £:&, -----7
/
/
/
/
-r~ /
/
/
z/

t.L'
17
Fig. 4-J4. Explaining the radiation of a HU'ygens elen1ent.

Let the vector of the electric field intensity coincide

with the y-axis, the vector of the magnetic field intensity
with the negative direction of the x-axis, and the normal to
the area, with the negative direction of the z-axis. Then,
in .accordance with (4-28) and (4-31), the densities of the
surfac~ electric and magnetic currents will be:
J~=HJC, J~=Ey. (4-35)
Let us inveSttigate the field in the radiation zone in the xz-
pJane.
Since the electric moment of the elementary area equals
J:CU and the magnetic moment equals J~ds, in accordance
with (1-3), the electric field intensity caused by the electric
dipole at the point P will be:
J~sk 1 e-1/cr
dE(xz) = i 4n(l)8 r .. '

and, in accordance with (1-8), the electric field intensity

set up by the magnetic dipole will be:
.. J"'dsk
. % '-lkr
dE (.rz) : 4-n- cos e&
= -' -

Ill
 The total field will be found to equal:

dEcxz>
Jedsk'
•11
= t 4 we
e- ikr
1-
JM
X
c
CO£
J
-;;-cos a . (4-36)
'
1t r Jy r.. J
When the surface s is in· the radiation zone (relatively
to the real sources),
J;e =-yr;=-roe
Ey k
an
d J~
u=-k
Wf. E
y·
..
1y

Hence (4-36) becorncs:

( 1 -I
dE.o~(xz) = - t. Eyds
2~r - cos a ) e - ik r . (4-37)
For the radiation in theyz-plane, we obtain in a similar \vay:
E,.ds
dE(!Jz) = - i i~r (I · f- cos~) e- ikr, ( 4-38)

where p is the angle between the z-axis and the direction

towards the point of observation in the yz-plane.
Thus, the expressions (4-37) and (4-38) define the radi-
ation of the Huygens eJernent. The radiation of such a system
has already been investigated in Chapter One. The di-
rectional diagram -of a 1-1 uygcns el~n1ent has the form of a
cardioid; it is sho\\'n in Fig. 1-11.

4-7. Radiation of an Ideal Plane Antenna

Let us discuss the radiation of an ideal plane antenna of
dimensions a and b. Let the plane of the antenna coinc1de
with the xy-pJane (Fig. 4-15) and let the field of the antenna
be prescribed as:
E" - k

Evbdx
dE<xz) = -- i 2Mo (1 +cos~) e-lllto+lkx •iu u,

112
where 0 r 0 is the distance frorr1 the centre of the antenna to the
point of observation. The total field of the whole of the
antenna at the point P is detcrrrlined as the intr~ral of the
11
/I ..

b r-rg =-XSill a

' Fig. 4-15. An ideal pi ane antenna.

preceding expression taken along the x-ax is fron1

a 2

-
E(xz)-
Evb
-- t" 2A.r
• ( 1 -+-cos a) e-
0
lkro s ei~v •in n dx.
x=-a/2

After periorming the stated integration and reducing

the similar terms, we obtain:
. E,,ab sin 'I' __
E<xz> ==- t 21"'o (I -1- cos a.) 'Y e lkro, (4-39)
where
ur ka .
T =-2 s1n ex.
Sin1ilarly, for the yz-plane, \\'e obtain the expression

(-1·4U)
where
lcb . R
<11 ~ -~, s1n 1.~ •

Let us now draw the necessary conclusions fron1 (4-39) and

(4-40). First of all, note that the maximurn value of the
B-2122 113
field Intensity occurs in a direction perpendicular to the
plane of the antenna ((1=0) and is equal to: r
I

lE I= Eyab
Ar
I
(4·41)
0

i.e., it is equal to the field intensity in the antenna

multiplied by the area of the antenna and divided by the
product of the wave-length by the distance from the antenna
to the point of observation.
The zero radiation directions in the xz-plane are de-
termined from the condition sin'1'0 =0or~ sina0 =nN. where
N = 1, 2, 3, ...
Hence
(4-42)

From this expression we see that in the xz-plane, the number

of directions in which the radiation vanishes wi IJ be all
the larger as the ratio a/ 'A is larger. In just the same way,
the zero radiation directions in the yz-plane are determined
from the condition
. -- N'A
SlnP 0 = b , N=l, 2, 3 ... (4-43)

and the number of the zero radiation directions in that

plane will be all the larger as the ratio bi'A is larger.
When the dimensions of the antenna are large in
comparison with the wave-length, the width of the n1ajor
lobe of the directional diagram at the zeros of radiation is
determined from the expressions-
A. A
2ao~ 115° a 0
and 2~,~ 11~ b . (4-44)

Obviously, as can be inferred from (4-42) and (4-43), there

will be no radiation zeros if A>a and A>b.
Let us also stress that in the xz-pJane, the directional
diagram of the antenna does not depend on the dimension b
of the antenna, and in the yz-plane, it does not depend on
the dimension a of the antenna.
Fig. 4-16 shows the directional diagram of an ideal plane
antenna where the value of the function ~" is plotted
51

114
on the y-axis and the argument 'I'= na sin ° on the x-axis.
We see that the magnitude of the maximum of the first
minor lobe amounts to approximately 1/ 1 of the major
lobe. Let it be reminded that in a linear eo-phased array of
half-wave dipoles, the first minor lobe has the same relative
intensity.
The same diagram shows the width of the major lobe et
half power 2a. 12 , i.e., the size of the angle a, ·within which

Q6~~~~~~~~~~~+-~--~~

Q4.....,_-+---+--+--~..._+-- -+--+--~-+---1

Q2~~~--~~~+-~--~~--~~

- -5n--4?r-3R"-2~-:n- ~ 2w 3:71' 4'H 5~

(J=:na6inajA
Fig. 4-16. Directional diagram of an ideal plane
antenna.

the field intensity does not fall below 0. 707 of the 1nagnitude
t>f the field intensity in the main direction. We see that
the width of the major lobe at half power of an 1deal plane
antenna is . expressed as:
2a. 1.=51°~ or 2P•,.=51°f. (4-45)

The power radiated by an ideal plane ~ntenna is fairly

simple to calculate. Indeed, the power radiated by 1 rn1 of
I E 11
the antenna equals S== 240n and, consequently, the
power radiated by the whole of the 'antenna equals
1
P 1 E l ab {4-46)
z = 240.n
.
Here, E represents the amplitude value of the electric field
intensity on the surface of an ideal plane antenna.
115
 4-8. Eft'ect of Changes of Field Amplitude
and Phase in a Plane Antenna on the Directional
Diagram
In the preceding paragraph, we examined the radiation
of an ideal plane antenna with a uniform distribution of
its field amplitudes and phases. Let us now consider a plane
antenna of the san1e dimensions a and b in which the field
phase wi 11 be assumed constant in the whole of the antenna,
the field amplitude being assumed constant only in the
direction of the y-axis and changing along the x-axis as

C, + C1 cos(¥)· i.e.,
Ev=E 0 [ C,-/- C1 COS(¥)]. (4-47)

Since the distribution of the field amplitude and phase

along they-axis has not changed, the forrn of the directional
diagran1 in the yz-planc ren1ains the san1e as for the eo ..
phased plane antenna. For this reason, the radiation will
be considered only in thr xz-pJane of the antenna.
Referring to Fig. 4-15 and substituting into (4-37) the
expression (4-47), the integration along the x-axis yields:
E cxr> = - t· 2Eob
-x-;:-
0
(I ,
1
cos a ) ><
ll

a
x=--
2

Performing the stated integration and reducing the sin1ilar

tern1s, \\'C arrive at the following expression:
•
£(xz)= -i :A:: (1 +COS r.t) e-lkro X
sin 'V 2 cos 'l'
X c I llf
'I
-t - cI -:;:;-
., " 1-
( ')
--=- 'I'
) 2 ,
(4-48)
j[

where
nr ka .
T =-
2 s1n a..

Let us now consider special cases of the distribution ul

the field amplitude in the antenna. f.Pt C1=l and C2-==0.
116
This gives us the case of the co-phasal surface alr(~ady con·
sidered in Paragraph 4-7. -
Let Ct =0 and C 2 ~-= I. This is the case of a eo-sinusoidal
distribution of the field arnplitude in the direction of the
x-axis, when the field at the edges of the antenna fa) Is down
to zero. We have:
. E0ab 2 cos •y ..
ro n I -· ~ J1 'I'

The directions in which the radiation vanishes are

determined from the condition
cos '1'0 =0 or ~sin a.=~ N, N = 3,5 ...
Hence
~ina 0 =~:, N=3,5... (4·50)
In the case of a high directivity, the \vidth of the 1najor
lobe bet\\·een the zeros of radiation is approximately
expr€ssed a~
2a 0 -::::: 172° -} • (4-51)
i.e., the n1ajor lobe of the directional diagrarn between the
zeros of radiation will, in that case, be 1.5 tirnes wider
than the n1ajor lohe of an ideal eo-phased antenna provided
the size of tht! antenna is the san1e. The ha If- power \vidth
of the major lobe of the directional diagrarn is, in that case,
deteimincd from the expression
o A
2aa, 2 = 67 -(I . (4-52)
At the same time, the relative magnitude of the n1inor lobes
decreases and the maximum of the first lobe constitutes now
only 1/ 11 of the magnitude of the maxirnum of the major
lobe.
Now Jet C1 = / 1 and C2 = 11 • Here, although Wi! do get
1 2

a maximum of the field amplitude in the middle of the an·

tenna, the field decreases towards the edges down to only
1
/
1
of the maximum value. In that case, for a highly
directional antenna, the width of the major lobe between
the zeros may be defined from the expres~ion
{4-53)

117
 The half-power width of the directional diagram may be
defined from the expression
2a. 1, =57 .3° ~ . (4-54)
The value of the maximum of the first minor lobe constitutes
0.095 of the value of the maximum of the major lobe.
Thus, the faster the amplitude of the field in the antenna
decreases towards its edges, the wider the major lobe and
the smaller the relative auaplitudes of the minor lobes of
the directional diagram. ·
Now, let us consider the influence on the directional
diagrams of a linear change of the field phase in the
antenna. Let the field intensity in the antenna change as
E = E 0 e-ik,Jt. (4-55)
1
In accordance with (4-37), at the point of observation in
the xz-plane, the field intensity caused by the element on
the antenna surface is:
dE<xzl = - i ;-:::x (I +cos a) e-tkr.+t Ck •In a-k,) "'·
0

Integrating this expression over the antenna surface. we

obtain: ..
. E.ab (I ) sln 'I'' _11,
E <xz) = - t 2A,,• + cos a. 'I'' e o, (4-56)
ka . k a
\\·here 'I''= 2 stn a- 2 .
Note that k2a = 'IJ>, represents the difference of phase
between the field in the centre and at the antenna edges.
lt can be seen from (4-56) lhat- lhe rnaxilnum radiation
occurs for the condition 'I'' =0 and its direction is determined
from the ex press ion •
sinamax= !: . (4-57)
As for the zero radiation directions, they are determined
from the condition '1''-=±Nn and may be calculated from
the ex press ion
sin a,=± N !=_
a
+ '~'
na
11
, .V== 1, 2, 3, . . . (4·58)
Thus, we see that in the case of a linear change of phase
in the antenna. the form of the directional ·diagram remains
•
IJIJ
identical to that of the directional diagram of a eo-phased
antenna, i.e., is described by the function but the s;v.
direction of the maximum radiation differs from the normal
to the plane of the antenna by an angle determined in ac-
cordance with (4 .. 57).
Such a linear change of phase at the output opening of
antennas is utilised for di~gram scanning (hunting) in
radar a.ntennas. A detailed analysis of (4-56) shows that
when k.~k the fortn of the directional diagram of the
antenna undergoes but an insignificant distortion in the
course of the hunting process.
Let us consider the directional propertie~ of an antenna
the field phase of which changes as k 2x2 • Such a phase
distribution occurs. ior exam.ple, in the case of horn-type
antennas.
In the present case, the field intensity in the antenna
is ex pressed as
Ev=
•
E0 e-lk~JC • 2
(4-59)
In accordance with (4-37), the field intensity at the point
of observation lying in the xz-plane of the element of the
radiating surface equals:
dEixz) = - i ;tdx. (I +cos a)
0
e-tkr. H (k •in a-lf.x) "'·

To determine the intensity of the field set up by the whole

antenna at any point of the xz-pJane, this expression should
be integrated over the radiating surface. As a result, we
obtain a rather complex expression containing Fresnel 's
integrals (see [ 161 for example).
Fig. 4-17 shows the directional diagram of a plane an-
tenna, calculated in accordance with the expression just men·
tioned. In this figure, ljl 1 =k,lT)
I \I
jS the phase shift between
the fields at the edges and in the centre of the radiating
surface. As shown in the figure, in the case of a quadratic
change of phase, the maximu1n of radiation corresponds
to the direction a=O. The major lobe of the directional
diagram widens and, in the case of sufficiently large phase
shifts, it is divided into two. The level of the minor lobes
increases; furthermore, the zero radiation directions vanish.
For large values of '1' 2 , the minor lobes are entirely absorbt'd
by the widening major lobe.
119
....
~ .f... ·~_

1
EmtJx , Emux T

I" 1/!.=0
'2 I~ «-
z---
2
gr

' 1/ I

r a~
·' D.5 Coplloso/ I
excilution
t
r - 7
- t\
'\.: I
I
•

7·si ,...- - m ina

~/~~
,_
........ 7
\; V"\ "-./~ ...
'F!a ,_./
-
i"-. T·J
a a o o a m
-

~ 4 2 2 4 1086420246810

f
-Emaz
- -E
I £111111 f
1'\
•

~=:11" ,V\ J7\. §l=t5~

y ~ / ~ '
.
, / .........

1151 '\ "- I Q5

'\
.If ' ~ II \ fiT'a

•
rJ
- • 1 1 2 o z
L

4
1 L
6
!.
a m
"' 9ra.~ •'na
t A
m a 6 4 2 o z 4 6 a
'm ,sina
.;t"

Fig. 4-17. Influence of a quadratic change of phase on the directional diagram of a plane antenna. I
,,
 It is also of interest to consider the directional diagratn
of a plane antenna in case the phase of 'the field changes
r.
as k 8 In practice, this change of phase occurs in lens and
parabolic antennas when the radiator is taken out of the
focus.
Let the field intensity on the radiating surface change as
(4-60)
The field intensity at the point of obser\' at ion in the
xz-plane caused by the antenna element is:

d E(lcz> = - i E;::;.x (I + cos a) e-'"'•+1 (~t sin a-~t.x'> s.

The integration of this expression over the radiating
surface is rather complicated. Usually, the directional
diagrarns are calculated by rneans of approximate expressions
[16], which 'are also fairly cornplicated and are not given
here. On the whole, the properties of the directional diagran1
of a plane antenna with the cubic change of phase under
consideration can be surnmariscd as follows.

E
Ema.r

Q5

~a sina
T
-8 -6 -4 -2 2 10

Fig. 4-18. Influence of a cubic change of phase

on the directional diagram of a plane antenna.
'

Just as in the case of a linear change of phase, the direc-

tion of the maximum of radiation deviates fron1 the normal
to the plane of the antenna, but this deviation is at-
tended by a distortion of the directional diagram (Fig. 4-18).
When a phase changes but little at the edge of the surface
1:11
('i'a), the direction of the maximum radiation may be de·
termined from the expression
.
SIn Clma X =-=
o.sw.A.
na . (4-61)
When 1J? 3 ~A.. the error in the deterrnination of the di-
rection of the radiation maximum does not exceed a few per
cent. A comparison of (4-61) and (4-57) shows that the
deviation of the radiation maximum direction is small in
the case of the cuhic change of phase than in that of the
linear one (by approximately 1. 7 times).
The. distortion of the forrn of the directional diagrarn
consists in that the major lobe widens and becomes asym-
metrical; the level of the minor lobes increas es on one side
of the major lobe and decreases on the uther one; further-
more, the decrease in the level of the minor lobes occurs on
the opposite side to the direction of deviat ion of the major
lobe.
If, with the field phase changes on the radiating surface,
the field amplitude drops towards the edges, the influence
of a change of phase on the directional diagram decreases.

4"-9. Antenna Parameters

Antenna systern s may be characterised by a numbe r of
parameters which enable us to appraise and compare the
antennas. In particular, the width of the directional
diagram between the zeros 2ao and 2a~ 12 at half power can be
considered as belonging to parameters of this kind.
In the present paragraph, we shaJJ consider other useful
parameters. The field intcn,i ty in -the radiation zone of any
antenna can be represented in the following form:
30kh I •
E= i r
eff e- 111 '"F (8, cp) e1•, (4-62)

where F ( 9, q>) is the amplitude normalised directional

characteristic of the antenna;
'J', the phase directional characteristic of the
antenna;
I, the current at the point of the antenna to
which the phase directional characteristic
applies ;
hert, the so-called effective length of the anter1na.
The concept of the amplitude normalised directional charac-
teristic of antennas has already been giVen in Paragraph
4-2. The phase characteristics of antennas were also men-
tioned earlier. The effective length is mentioned here for
the first time. To make things clear, Jet us exarnine the
application of (4-62) to some radiating systems. For an
electric dipole lying in free space, the electric field intensity
is ex pressed as
. 30klle . e lk
E8 =t r stn e- '·

On comparing this expression with (4-62), we obtain:

F (9)=sin 9, '1'=0, hett=l, (4-63)
•
i.e., the amplitude normalised directional characteristic
of the dipole represents a sinusoid ("figure of eight") in
a polar system of coordinates; the phase characteristic does
not depend' on the angle of observation 8 and represents a
sphere in space and a circle on a plane; the effective length
of the dipole is equal to its geometrical length and measured
in metres.
For a symrnetricaJ electric dipole in free space , the
electric field intensity is expressed as
- . 601 0 e- lkr o cos (kl cos 9)- cos kl
t
F ·- r -.;_._~:-"---=---
sin k l sin 8 ·
0

The amplitude normalised directional characteristic of

a symmetrical electric dipole is expressed as
F (8) =cos (kl cos 9) -cos Id
Sill 9 ( I - cos kl) .

Then, on comparing this expressions with (4-62), we obtain:

_ 0 h -~ 1-coskl (4-64)
'I'- ' eff- n sin kl ·
As we see, the effective length of a symmetrical dipole is
not equal to its actual length and is a function of the length
of the dipole as weiJ as of the wave-length. The effective
length of a half-wave dipole ( 1/'A ={)
for example, equals
heff =!::. =.! 2l, i.e., constitutes 0.635 of the overall length"
n n
of the dipole.
123
 The effective length of a symmetrica I dipole 1nay be de-
termined by integrating the function of the current along the
whole length of the dipole and relating this integral to
the current at the feed points of the dipole:
l
-~\I sink(l-z)d _!::._1-coskl
h''ff - I 0 • kl
2- . kl
0 ·~· s1n n stn
Z=O

Hence, the ter1n effective length of an antenna may be

understood as the length of a dipole with a uniform current
distribution, which sets up the san1e rnagnitude of field
intensity in the direction of the maximum radiation as the
antenna under discussion \Vith the san1e current at the feed
points.
As for the phase directional characteristic, in the case
of a syrnmetrical dipole, it does not depend on the angle
of observation e either and, in polar coordinates, repre-
sents a circle. Note that the calculation of the field phase
in the above expression is performed relatively to the radius-
vector applied to the centre of the dipole. If the radius-
vector is related to another point of the antenna, the phase
directional characteristic relatively to this point will no
longer be spherical.
In the case of a eo-phased array of nm half .. wave dipoles,
the electric field intensity in the rr1agnetic vector plane
is ex pressed as
. ( nkd, s1n
s1n
. a. )
E -
--~>
. !iOI :tm e-lkr o \ 2 -
r0 • ( kd 1 •
SID SUI Q ) '
2
where r0 is the distance frotn the centre of the array to the
point of observation of the field.
The amplitude normalised directional. characteristic is
defined by the expression
nkd 1 \
sin ( sin a}
1 2
F (a.) :;~A- ( kd
1
) •
n sin sin a
2
The phase directional characteristic is spherical (11'=0)
and, consequently. the effective length of the antenna is
expressed as:
A
heff=-nm.
n
(4-65)

124
Note that, in practice, the concept of the effective Jen~th
of an antenna is applied only when discussing the radiation
of simple wire antennas. That is why the expression of
hprc wiJJ not be applied to other radiating systern~. although
this could be done, if desired.
The concept of the dirccti\'e gain, first put for\vard by 1\.
Pistolkors in 1929 [ 17 ], is current today in antenna theory
and practice. By directive gain one undC'rstands the ratio
of the square of the IJJagni tu de of the intensity of the field
set up by the antenna in a given direction to the rncan (along
all directions) value of the square of thr fk~ld intensity
I f 12
D= L"""' .,
-- . (4-66)
; I~W'lrl
•

Let us exprC'ss the rnean value of the square of the field

intensity through the power radiated by the antenna. The
power rnay pe expressed as the product of thr surface of a
sphere of radius r 0 by the rnean value of the Poynting
vector on the surface of the sphere:
'

But the na~an value of the Poynting vector is

•.
E~l'31l
Smean = 2 4on ,
hence
2 60Pl:
EmeLtn = 2 • (4-67)
'o
•

Substituting (4-67) into (4-66), \Ve obtain:

I E 1~,:
D=oop . (4-68)
:E

It was in 1his forn1 that the expression for the directive

gain was put forward by M. S. Neun1an.
Let us examine the application of th~s expression to a
number of antennas. For a Hertzian dipole, the tnagnitude
of the field intensity in the direction of the maximum
radiation is 1E I= aokll , and the power it radiates,
'o
Pl: =-f. 20 (kW. Consequently, the directive gain is D = 1.6.
/26
 In the case of a symmetrical dipole in free space, the mag-
nitude of the field intensity in the direction perpendicular
to the dipole axis and the power radiated by the dipole are:
£=601 0 1-cosld
r0 sin kl •
Jl
p ~,=2 sin: kl Rr.a·
Consequently. for a symmetrical dipole. the directive. gain
is ex pressed as:
120
D= R (l-coskl) 1 • (4-69)
r.a
For a half-wave dipole, for example, kl=90° and Rza=73.1
ohms and, therefore, D=l.64and for a wave dipole, kl=180°
and Rz.= 199 ohms and, therefore, D=2.41.
As we saw earlier, for a eo-phased array of half-wa\'e
dipoles, the magnitude of the field intensity in the direction
of the maximum radiation is:
1EI=60la nm
- 'o
and the power radiated by that dipole array is:
Jl
Pr.= 2
a Rr.anm,
where Rzl4 is the mean ntagnitude of the radiation re-
sistance of a dipole in the system ..
Consequently, the directive gain is expressed as
120
D=R nm. (4-70)
Za •
Thus, for example, for an array consisting of four dipoles
(n=2, m=2), the radiation resistance of a dipole in the
system is Rr.. = 75.34 ohms and the directive gain is therefore
D=6.37.
In the case of an ideal plane antenna, the magnitude of the
field intensity in the direction of the maximum radiation
equals:

126
and the power radiated by the antenna is:
E's
p - 11
z- 240n •
consequently, the directive gain is expressed as
D --~ns
A,• t
(4-71)
where s=ab is the area of the antenna.
Let, for example, the area of an ideal plane antenna equal
1 m1 and let this antenna be excited by a wave A= 10 cn1.
The directive gain of the antenna is then D = 1,256. If we
excite the same antenna by a wave A=3 cm, the directive
gain will be D = 13,956. In the case of a cosinusoidaJ distri-
bution of the field amplitude· along one of the sides of an
ideal plane antenna, the field intensity in the principal
direction is 1E I= Eys..!, and the radiated power,
'Ar0 n
E1 s 1
- y
Pz- 240n 2·
Hence, the directive gain is determined from the expression
4ns
D=O.Bl V·
Let us quote one more expression for the directive gain,
which is often found useful. In accordance with (4-62), the
magnitude of the square of the antenna field intensity in a
given direction is:
301 kh 1 / 1
J E JI == ,.
eff F I ( 8 • cp).
0

The mean value of the square of the field intensity of the

antenna along all directions is expressed as
ft 1:1
1
E'mean === zv1
4 0
S SIE 1•r.• sin 6dfJ dq>.
w=o- 9•t ,.
•
Substituting these expressions into the initial express1on
(4-66), we obtain:
D= 4nF1 (8. f) • (4-72)
ft 1ft '
~ ~ F1 (8, cp) sin 8 dO dtp
· - · 41)=1
121
where F (0, cp) is the amp Iitude normalised directional
characteristic of the antenna.
l''he expression (4-72) enables us, in particular, to calcu-,
late the directive gain by means of the graphic integration
of a known space directional characteristic and to obtain
an approxin·1ate estimation of the magnitude of the directive
gain. Indeed, assuming that \Vithin the limits of a certain
solid angle ~w=l\ 8l\cp F ( 8,cp)= 1, and that outside its
limits F ( 9, cp)=O, we obt~in the following approximate
•
expression
(4-73)

Thus, in the case of an ideal plane antenna, for example,

the width of the directional diagram at half power equals
0.89~ and 0.89 i- in the two principal planes. Substituting
these values into (4-73), we obtain:
4ns
D=l.27 A- 2 ,

although a rnore accurate definition of the directive gain

of an ideal plane antenna is given by (4-71).
In the case ef an axial radiation antenna with a phase
velocity equal to the velocity of light, the \vidth of the
directional diagram at half power is defined hy (4-24).
If we consider the directional diagran1 to be syn1rnetrical
relatively to the antenna axis, the substitution of (4-24) into
(4-73) yields the following expression for the directive gain:
L
D=3.55 T I

\vhereas the use of (4-72) Jeads to the more accurate expres-

•
s1on:
•
(4-74)

According to this expression, the directive gain is the larger

the larger is L/A.
In a slow-wave axial radiation antenna, the directive
gain depends on the ratio between the electric length L/'A
and the relative phase velocity ~. Taking (4-22) and (4-72)
into account, the maximum directive gain is obtained \vhen
'l' =i-·1.e., when the fields in the 1nain direction from the

J21J
first and the last elements of the antenna are in antiphase.
The substitution of (4-27) into (4-73) yields a somewhat higher
valueofD=lt.2i whereas the use of (4-72) yields the more
accurate expression:
L
D=7.2 1: · (4-75)

Apart from the directive gain, the concept of the antenna

power gain is also used. By antenna power gain s, one under-
stands the ratio of the square of the intensity of the field
set up by the antenna under discussion in a given direction
to the square of the intensity of the field set up by a ha If-
wave dipole in its equatorial plane. lt is furthermore
assumed that the power fed to. the antenna and dipole under
investigation is the same and that the half-wave dipole lies
in free space. Since the directive gain of a half-wave dipole
equals 1.64,, the expressions quoted yield
DT)
£=1.64. (4-76)

where TJ is the efficiency of the given antenna.

Note that, sometimes, the power gain of the antenna is
defined not in relation to a half-wave dipole but in relation
to a hypothetical (omnidirectional) radiator. In that case,
the coefficient 1.64 in (4-76) is omitted.
The antenna efficiency is defined as the ratio of the power
radiated by the antenna to the power fed to the antenna,
consisting of the radiation power and the po\\ter of the losses
in the antenna,
P'E.o
T)= . (4-77)
P%o + Plosses
If we relate the radiation power and the power of the losses
to the square of the current at the feed points of the antenna,
the ex press ion for the efficiency wi II be:
Rzo .
T)= • (4-78)
Rto +· Rlosses
where Rtosses is the resistance of the losses of the antennas
related to the current at the antenna feed points.
Apart from the parameters h:e1t. D~ s, TJ, the con~ept of the
effective area of the antenna Self ts also used tn antenna
9-2122 129
theory and practice. This concept will be dealt with later
during our study of the receiving antenna theory.
The parameters of some types of antennas will be
discussed in the course of our study of the corresponding types
of antennas.

4-10. Deflnit ion of the Current Distribution

in an Antenna in Accordance
with a Prescribed Directional Diagram
a) Stating the Problem

Until now we have been defining the directional diagran1s

of antennas knowing their current distribution. Thus, we
discussed two ways of achieving highly directional antennas:
by way of the uniform distribution on a plane of the current
in arnplitude and eo-phase, and by way of the travelling-
wave current distribution along a certain straight
line in space. rfhe question arises as to \\'hether these are
the unique ways of obtaining directional antennas. Is it
not possible to establish other distributions of the current
in an antenna, such as would ensure sti11 narrower direction-
al diagrams than- in the cases stated above while permitting
the use of antennas of reduced size? Apparently these and
sitni1ar probletns may arise when designing antennas with
any form of directional diagram and not only highly
directional ones.
The answers to these questions can be obtained by solving
the problen1 in the following manner: having prescribed the
forrn of the directional diagran1, we look for the distribution
of the current on a certain straight line or on a plane in
space that will ensure the prescribed directional diagran1.
Stated in this manner, the problen1 will. not have just one
unique solutton. In other words, we shall obtain several
possible current distributions, each of which will ensure
the given form of the directional diagram. This is confirmed,
for exa1nple, by the results obtained during the investi-
gation of the co-phasal distribution and of the current
distrFbution in accordance \\'ith the travelling-wave law.
As we saw earlier, the directional diagrams obtained in
both cases have an identical shape. A comparison of the
current distributions obtained in this way gives a unique
answer to the question as to which of them is the optimum
130
 one from· the point of view of, say, the minimum size of the
antenna or of the sirnplicity of its design.
'The problems of antenna technique where the forrn of the
directional diagram is considered pr~scri bed as well as
where the current distribution corresponding to that
directional diagram is being sought, are referred to as
the reverse problen1s of electrodynan1ics. The importance of
stating and solving these problems is not limited to the
purpose of finding out the optimum current distributions for
frequently utilised forrns of directional diagrarns for which
the design princi pies of the corresponding antenna systerns
are known, as in the example exan1ined above. When the
necessary directional diagrams are of cornplex forrn, the
deterrnination of the current distributions \Vhich will ensure
these diagrams, i.e .• the synthesis of antenna systents \Vith
directional diagrams of special forrn rnay have an inde-
pendent aQd, frequently, pritne irnportance for technical
purposes. Sontetimes, the object of the reverse problen1s of
electrodynatnics is not to find the current distribution in the
antenna but a certain parameter, directly connected with
the geometry of the antenna and determining its diagrarn.
Such a parameter may be the distance between the elernenta-
ry radiators of which the antenna is made up, for exan1plc,
the depth of the grooves of a ridged antenna, or the length
of the dipoles fn an array of equaJly spaced dipoles, etc.
We shall be concerned with the problerns of the first type
wherr the object is to find the function of the current distri-
bution on the antenna. The problems of the second type
may be reduced to those of the first type, since, eventually,
the above-mentioned parameters of these problen1s define
the current distribution in the antenna.
Let us examine the various mathematical methods \Vhich
wi 11 enable to establish the connection between the direc-
tional diagram and the function of the current distribution.
To begin with, we shall investigate the Fourier integral
method, which enables to establish many general princi-
ples and to arrive at estimates which have to be taken into
account for solving the reverse problems of electrodynamics.
b) The Fourier Integral Method
Let us examine the reverse electrodynamic problem stated
as follows: radiating elements (elementary electric currents,
for exatnple) are distributed along the x-axis, within the
9* 181
interval -oo~x.-+ oo (Fig. 4-19), in accordance with
the expression
I (x) = u (x) e1117 (x)
where u (x) is the amplitude distribution of the currents
of the radiating elements;
'I' (x), the phase distribution of the currents of the
radiating elements along the x-axis.
The field of each of the radiating elements in the f~r zone
(r-+ oo) may be written as:
,-tier
dE=AI (x)dx ,

where r is the distance from the radiating element to the

point of observation P;
A is a certain constant number (the radiating ele-
ments are considered as non-directional).
Let the function of the directional diagram of the linear
system under consideration F ( 8), which. in the general
case, is complex, be prescribed. We have to define the distri-
bution of the amplitude and phase of the currents of the
radiating elemeQts on the x-axis that wiJ1 ensure this di·
rectional diagram of the linear antenna under consideration.
~fhe total field set up by all the elements lying on the
x .. ax is is expressed as:

E (P) =A e -tkro s I (x) et

CD

lex cos~ dx.

'o _.,
In this expression. the directional diagram is described
by the function
QD

f (6) = sI (X) eJIIxCOI. d". (4 .. 79)

-CID

In accordance with the conditions of the problem, we

need to find the distribution of the sources for which the
directional diagram described by (4-79) should approach
the prescribed function F (8) with the required accuracy,
•
1.e ••
CID

F (6)= S
_.,
I (x) ellutcoat lU. (4-80)

132
Let us • introduce a new variable detern1ined from the
expression
~=k cos a. (4-81)
When 0 changes in the whole range of the values under
consideration from 0 to 11:, ~ changes within the interval
-k~s~k. Taking account
of (4-81 ), the expression (4-80)
____,p
n1ay be written as I

CID I
F (')==SI (x) e1h dx. (4-82)
-QD

But on the right-hand si·de

of this expression, we have
the Fourier transformation of L
the function I (x). This means
that the function I (x) can
be found from the expression z•O
aD

I (x)=dn sF (~) e-~d~. (4-83)

-aD
t

Thus, the expression (4-83)

yields the current distribution
along the x-axis needed to ob-
tain the prescribed directional Fig. 4-19. Systetn of coordi .
diagram F (8). Note that, in nates utilised.
accordance with (4-81). the
function F (;) coincides with the directional diagram of
the linear system under consideration only in the interval
- k <;; ~ k. This means that, while leaving the function
F (;) unchanged in the interval - k ~' ~k. its value
may be arbitrarily prescribed outside of the interval.
Furthermore, it may happen that various distributions of
the radiating elements I (x) along the ~-axis will correspond
to the same form of the directional diagram of the antenna
which is described by the function F (;)when- k<i<k
because, in accordance with (4-83), I (.r) is defined by the
function F (;) on the whole of the axis -oo..;;.;; oo.
Taking account of the above, the current correspnnding
to the prescribed function of the directional diagram
181
F (6) = F (arc tos}) can be written in the form of two items
k -k
I (x) = -dn Sf(6) e-i~xd6 +Jn SF (6) e- ~ 4 +
-k -ao
Cl)

+ SF me-t~x dS = 11 (x) +I. (x). (4-84)

k

Consequently, the whole straight line - oo ~ ~ ~ oo serves

as the line of integration for the item / 2 (x), with theex-
ccption of the section- k ~ ~ ~ k.. The definition of the cur-
rent in accordance with the prescribed directional diagrarn
F (s) is arbitrary because the i ten1 of the current I 2 (x) tnay
be prescribed arbi trariJy. Let us investigate the physical
1neaning of each of the i tcn1s in (4-84).
Since the directional diagram of the antenna (i.e., the
values of the function F (~) on the section -k ~ ~ ~ kJ,
corrPsponds to tht· field in the far zone, and the fie]d in the
far zo11e itself defines the power radiated by the antenna,
the first item / 1 (x) corr1pletely defines the power radiated
by the antenna and is therefore referred to as the active or
radiating currenf.
Thus, the addition of the item / 2 (x) changes only the
near fil'ld of the antenna and, consequently, changes only
the oscillating reactive power accurnulated near the antenna.
Hence, the itcn1 /~ (x) is referred to as the reactive current
of the antenna.
It rnay be shown that, if we define the quality of the an-
tenna system Qas the ratio of the reactive power accumulated
near the antenna (coupled wtth the electrtc as well as the
n1agnetic field) to the radiated power, we may obtain for
it the ex press ion •
-k CIO

~ I F (') (1 lit + ~ I F <;> J• 4

Q=-m k k • (4-85)
~ IF(~) (1 d;
_,
The structure of this expression is easy to explain on the
basis of the preceding arguments. The exact proof is asso-
134
ciated with complicated n1athematicaJ calcuJations and is
not given here [ 18}.
As will be shown in the folJowing section, the \vidth of
the directionaJ diagrarn which may be obtained fron1 a
linear antenna of finite length 21, is defined by the quantity
Q. Let us pass on to the esti1nations which t.~nable to
establish such a connection.

c) Antennas with Fast and Slo~~· (~hange of Currrnt J:11zase

In accordance with the above, the velocity with \\'hi eh

the current phase changes in an antenna is d(~terrnincd by the
magnitude of the derivative of the function ~-: (x). Let us
agree that the name "antennas with a s)o\v chang(' of phase"
119) wiJJ be used to designate antennas for \\hi eh the fol-
lowing inequalities are satisfied
rnax I11'' (x) I < k \\'hen -l~x~..:;l,
max I 'I'" (x) I < k~ \\'hl"·n -L~:::::x<l
---- -.;;;:: ' (4-86)

21t •
\\·here k = T ts the wave number. These inequa I it ie
mean that the current phase in the antenna chang('S n1ore
slowly than the phase of a plane wave propagated in vacuurn.
Antennas for which the current phase changes faster wi 11
be designated as ultradirectional. The meaning of this ter1n
wi 11 he expJained later.
Let us estirnate the width of the directi,)nal diagrant be-
tween the zeros for antennas with a slow change of current
phase. In accordance with (4-79), the norn1alised directional
diagram of a linear antenna of dhnension 2l (Fig. 4-19)
may be written as:
l
~ u (x) 1 ilkx cos 9+ t 1x1 1 dx

fs (6)=----'":"""I-------- • (4·87)
max ~ u (x) e1 lkx co• &-tt 1-'ll dx
_,
Let us note that, for antennas with a slow change of the
current phase, it may be approximately assumed that in
the direction of the maximum of the directional diagram,
the fields of the antenna separate elements are added up in
135
 •
eo-phase, 1. e.,
I I

max
~
\ u (x) e1 lb ca. '+"' fx>J dx ~ Ju (x) dx. (4-88)
-l -I

The correctness of .such an assumption is easily verified

for the case 11>' (x) =k 1 where k 1 is a constant magnitude
satisfying the inequality O~k,~k.
In this particular case, one may always find such an angle
8max for which kcos9max =k 1 and the fields of all the
elcn1ents at the point P disposed at such an angle are added
up in eo-phase. Naturally, this angle 8max corresponds to
the direction of the 1naxirnun1 radiation of the given
linear antenna. In the general case, since the velocity with
which the current phase changes on the antenna is Jimi ted,
one may always find a direction of e for which the fields of
all the radiating elements distributed along the x-axis are
added up almost in eo-phase, i.e., the (Jhase displacenaents
due to the difference in path length of the rays arrivingat
the point P aln1ost entirely compensate the phase displace-
rnents between tke fields of the separate elements due to
tht' function 'i' (x).
Taking account of (4-88). we rnay rewrite (4-87) as:

I
~ U (X) f'l(kx Coli+ 11'(Xl)cf.x

/.(0)=-1 I • (4-89)
S u (Jt)dx
-I

Let us estimate in 111odulus the n1agnitude of the deriva·

tive of the function / 1 (8), defined by (4-89).

t:lJ sin 015 '1

·;I u (x) dx

Ia'aJO) I
E;; I _, •

\ tl (X) d~
_,•
136
Let us designate
i
5_,I ; I u (x) dx

•
~ u (x) clx
-L
then \Ve shall obtain
(4-90)

Let us note that the coefficient Pu is not larger than unity~

furthermore, it reaches this 1naxin1un1 value if the currPnt
in the antenna is concentrated at the points x= :1: I (i.e.,
m that case, the value of the reducing item LJ-' which
in!luences the magnitude of the integra) in the nurneralor
of the expression for p, is at its rnax in1un1).
In the other cases, fhe factor ~I rcduct•:. the magnitude of
the integral in the nurnerator ot the l)xprcssion for Pu
relatively to the n1agnitude of the int()g1aJ in the dcuon1i·
na tor, since at a 11 points except x = ±l. 1. j'<
Substituting for Pn into (4-90) the n1axirr1un1 value of
this coefficient, which is equal to unity. we find~
•

Iill 'a~~, I~ kl.

Thus, the velocity of change of the modulus of the normal-
ised directional diagrarn of a linear antenna cannot exceed
the electric length of the antenna in the case of a slow
change of current phase in the antenna. But the velocity of
change of a normalised directiona! diagran1 determines the
\Vidth of the antenna darcct1onal diagran1.
Indeed, if the velocity of change of the normalised di-
rPctional diagram for a change of the ang!e H does not excetd a
certain magnitude f' rnax (0), the angle ~0 within the
lirnits of which the directional diagram change~ fron1 its
n1aximum magnitude, equal to unity, to the magnitude
1-u
a<l. cannot be smaller than the quantity f~u(O) lil we
assurne that at all vaJues of 6 within AB, the velocity of
181
change is maximum and equals f' max (8)1. On the strength
of this, we may write the evident inequality for the width
of the directional diagrarn between the zeros
2 2 A \.
20o ~ , = = · (4 91)
f max(O) kip,~, ~lPa -

r:or a eo-phased and uniform di~tribution of the radiating

sources along a Hnear antenna, we have established in
Paragraph 4-2 that the w&dth of the directional diagram
between the zeros equals:
A.
26 0 = (rad].
7
Consequently, for antennas with a slow change of current
phase, the width of the directional diagram may be ob-
tained :np, times smaJJer than in the case of a co-phasal
current distribution, through the corresponding choice of the
current distribution I (x). However, one should take
into account that tf we are to obtain a suffici£'ntly srnaJJ
level for the minor lobes of the directional diagram, the
quantity Pu should be smaller than 0.5 (p,,=0.5 for a uniform
distribution of the amplitude). Consequently, the co-phasal
and uniform distribution of the radiating sources is near
enough to the optimum from the point of view of obtcn-
ing the minintum width of the directional diagrarn (for
prescribed dimensions of the antenna) if we lin1it ourselves
to the investigation of antennas with a slow change of cur-
rent phase.
In the case of antennas with a rapid change of current
phase, the inequality (4-88) is not valid. Hence, instead
of (4-90), on the streugth uf (4-89). one should write

Idf~O) I~ P-tPakl, • (4-92)

where
I

~ u (x) dx
-l
P•== ----,~-------------- • (4-93)
max ~ U (%) ei (k.¥ COI~+1j) (.¥)1 iJ%
-l

The coefficient p~ shows the number by which the ampli·

tudes of the currents in an antenna wi1h a rapid change of
13H
 phase need to be multiplied so as to obtain the sam~ value
of the field in the direction of the n1ax inltlnl of the direction-
al diagrarn as for a eo-phased antenna in which the distri-
bution of the current amplitude corr~sponds to that in the
antenna with a rapid change of current phase under invt·sti-
gation. Since the coefficient p~ n1ay be as large as drsircd,
the width of the directional diagram bet\\'een the zt'ros
for antennas with a rapid change of current phase for the
prescribed dimension of the antenna 2! rnay be obtained
as small as desired. 'fhat is why antennas \Vith a rapid change
of phase are often ca11ed uJtradirectional antennas.
The following conclusions tnay be dra\vn frotn the above.
For the prescribed dir11ension 21 of an antenna, \\'e rnay
obtain a directional diagra!n of any fornt. lio\Vt\vcr, an
increase in the velocity of change of the directional dia·
gram with a chanf,!e of the angle, at the prescribed dinJtlnsion
of the antenna, requires an increase of the velocity of change
of current phase in the antenna, and this, in accordance
with (4-93) requires, for the prescribed value of the ficJd
in the far zone of the antenna, an increas(\ of th(~ current
arn pJ i tu de in the antenna.
From the point of view of the practical design of the an·
tenna, an increase of the velocity of change of the phase of
its current will render it rnore complex, since a larger
nurnber of discrete sources will be needed for approxi-
mating with suffici~nt accuracy the continuous current
distribution theoretically obtained. Furtht.'rtnore, an increase
of the an1plitude of the currents in the antenna leads, for
the sarne radiated power (for the same field in the far zone),
to a decrease of the efficiency of the antenna, if the surface
along which the currents flow, possesses a finite conductivity.
It should be noted that, in practice, it is nevertheless
found possible to achieve ultradirectional antennas with
even higher velocities of change of current phase than
'i''(x)=k, so that, from the point of view of the practical
design of antennas for the current distribution obtained, the
boundary between antennas with a slow change of current
phase and ultradirectional antennas is somewhat relative.
This applies to antennas in which the ultradirectivity
effect is achieved through a slow phase velocity.
In the general case, a slow phase velocity is not neces-
sarily associated with obtening ultradirectional dia-
grams. For this reason, all the estimations quoted do not
139
apply to just any antenna with a slow phase velocity but
only to antennas in which the distribution of the ampli-
tude and phase of currents is prescribed specifically with
the object of obtaining the ultradirectional effect. In par·
ticuJar, aJJ that was said above does not apply to travelling-
wave antennas with a slow phase velocity, for \vhich the
\Vidth of the directional diagrarn i~ even larger than in the
case of a eo-phased antenna of the same length.
Let us also note that if we do not aJter the form o. the
directional diagrarn but make the antenna each time. with
a decreasing ditnension of I, then beginning from a certain
dimension of l, we shall eventually arrive at ultradirectionaJ
antennas. At the sarne tirne, a decrease of I will raise the
quality factor of the antenna Q, since when I decreases,
the active current rerrtains unchanged (conventionally,
the directional diagram does not change in the case of a
change of l) and an increase of the current amplitude in the
antenna can only be connected with an increase of the reac-
tive currents and, therefore, with an increase of the reac-
tive fields in the vicinity of the antenna. Thus, an increase
of the degree of ultradirectivity of the antenna leads to
an irnprovement of its quality factor and this leads to the
narrowing of the pass-band of the antenna from the point
of view of its input resistance, since the input resistance
of antennas of high quality factor are very closely de-
pendent on the frequency.
The direct connection between the velocity with which
the current phase changes in an antenna and its quality
factor enables to clas~ify antennas into ordinary and ultra-
directional antennas, not only from the point of view of the
velocity of the change of the curr~nt phase in the antenna
but also from the point of view of the so-called ultradirec-
t i vi t y coe ffi c i en t , w h ieh eq ua Is
•
v==l +Q,
where Q is the quality factor of the antenna, determined in
accordance with (4-85).
As can be seen from the preceding arguments, the ultra·
directivity coefficient reflects the ratio of the sum of the
radiated and reactive powers accumulated near the antenna
to the radiated power. The boundary between ordinary and
ultradirectional antennas is, in that case, a certain_ al~o
relative, quantity for which the practical design of an
J40
antenna systern is still possible. This boundary value v ;s
usua JJy considered to be of the order of sever a I tens.
As follows frorn the definition of y, this parameter is
closely connected with the phase velocity of the currents
in the antenna and it is possihle to establish a simple cor·
respondence between the n1agnitudc of the paran1eter y
and that of the phase velocity of the current~ in the
antenna detern1ining the boundary between ordinary and
ul tradirect ional antennas.
In concJusion, note that the Fourier integral rnethod is
seldom applied to the practical ea lculation of antennas in
accordance with a prescribed diagran1, because by subst i-
tuting into (4-83) an arbitrary function of the directional
diagram, we obtain, in the general case, a current \\'hich
differs frorn zero along the whole of the axis- oo~ x ~ Of..J. In
practice, the dirnension of the antenna or the range \Vi thin
\Vhich thi& dimension rnay he varied is usually specified in
advance, so that other rnethods are used for practical calcu-
lations, in particular the partial directional diagrams r11ethod.
l"rue, such rnathen1atical restrictions rn'ay be imposed on
thP function F(~) that, in caJculating the current distri·
but ion in accordance with (4-83), the value of the current \ViiJ
diiTer fron1 zero only \\'ithin a certain finite interval on the
x-axis (see, for exarnple, 1181). However, for practical
calculations, this n1ethod is too complicated .

d) The Partial Directional Diagranzs Method

Let the function of the elen1entary radiators distribution
/(x) in the interval -I~ x ~I be prescribed as a sum tn
of functions
m
I (x) = L An/ n (x). (4·94)
n=o

Then, the function of the directional diagran1 f (6) \Vi 11

be represented as a sum of partial diagra1ns fn (6) to each
of which corresponds, in the int~?rvaJ -I-s;;, x ~I, its own
function of current distribution 1,. (x).
Indeed,
I I m
f (0)= sI (X}
-1
etb cot I dx = sLA.I,. (X)
_,n::o
etb COl 'dx.

141
 Changing the order of the integration and summation
and introducing the designations
I

fn (6) = sIn (x) elk% COl 8 dx,

-l
we find
m
f (0) = r. A"f (0).
n=o
n

Then, the solution of the reverse electrodynan1ic problem

is reduced to the expansion of the prescribed directional
diagram F (H) in accordance with the systen1 of functions
In (6). The coefficients An of this expansion enable, in
accordance with (4-94), to construct in the interval -l~x~l
the current distribution which corresponds to the prescribed
diagram F (8). Note that the current distribution thus
obtained wi IJ be different from zero in the prescribed inter ..
vaJ - l ~ x ~~ inasmuch as the current I n(x) corresponding
to each partial diagram fn (6) differs frorn zero only on that
section. Furthermore, from the mathematical point of view,
the systen1 of functions I n(x) can always be prescribed in
such a way that the velocity of the phase change will, for
each of them, be less than k. Then, an ultradirectional
antenna will not be obtained by just any combination of the
currents I n(x). However, in all these cases, in order to
obtain a sufficiently accurate approxirnation of the
prescribed diagram by a systenl of functions fn (0) the t

dirnension chosen for the antenna should satisfy the condition

(4-91) and the system of functions of the partial directional
diagrams should be sufficiently complete.
The partial directional diagrams rnethod ~nables the quanti-
tative estin1ation of the influence of the degree of ultra-
directivity of the antenna on its current arnplitude. The
graph of Fig. 4-20 shows how the ratio of the an1plitude
(maximum) of the current umax in an ultradirectional
antenna of length 2/ to the current amp!itudc in a eo-phased
antenna U 0 of length 21 varies with the quantity 2/jy,
the directional diagran1 of which is represented as:
6) =~Ill (kl cos 8)
fu ( kl cos a •
142
The calculation was effected by the partial diagrarns
method; moreover, for the ultradirectional antenna, the
directional diagram was prescribed as:
f (S) =_sin (akl ~o~~l ,
akl ens U

where a is the coefficient, lar~er than unity, character-

ising the degree of ultradircctivity.
The functions taken as the functions In (x) \\'ere:

(x)=--~··--}.:-~==-xT ,/t-(~.)
1
1
I
tn k;;r, V l~-x~ tn V ;

/2n + 1 (x) =
(-) )"- 't
· ------:-::--- X T
I( :r. V [2 - xz
1r1 +1
II
-/
-
X
'
•

where T 1 n and T•n• 1 are the even and ocld C~hebyshcv

polynomials of th(~ corresponding argurnents.
Then, .the corrrsponding partial diagran1s are repre-
sented as:
fn (0) = ./" ( kl COS 0),

where .I" is the Bessel function of n order.

The expansion of l (fJ) in a fn (0) series is easy to prrforn1
o\ving to the complrteness of the Bessel functions syst~rn.
The calculations ha\'~
also sho\\'n that the cur- diJ
rent phase in the antenna 80~----I---+---
changes frotn point to +--

point. 60~----~--~~
An examination of the
graph in r.:ig. 4-20 en-
ables us to conclude that
the practical efTect of
the uJtradirectivi ty can 0 f 2 J
be uti Iised to raise the
directivity of antennas Fag. 4-20. Graph illustrating the
of smal1 length, \\'here an1plitude of the currents in an ultra-
directional antenna.
the needed increase of
current amplitude is not
high. An attempt to utiJise the ultradirectivity effect
for antennas of great length (¥->1) will cause the power
gain to fa) I as a increases due to the sharp increase of the
143
amplitudes of the currents and the attending sharp decrease
of the antenna efficiency caused by an increase of its losses.
Thus, the present investigation of the methods for solving
the reverse problems of electrodynamics, which do not,
or course, cover the whole variety of methods of this kind
used in antenna technique, shows that the solution of the
reverse problem regarding the calculation of the current
distribution in an antenna in accordance with a prescribed
function of the directional diagram does not present any
difficulties of principle. Furthermore, it follows from the
expressions (4-83) and (4-84) cited above, that prescribing
the modulus of the directional diagram, which is usually
adequate for practical purposes, does not yet provide a unique
definition of the current distribution along a linear con-
ductor. due to the fact that, in accordance with (4-84 ), when
prescribing the n1odulus of the directional diagram,
the deterrnination of even the active current of the antenna
/ 1 (x) is not unique (with an accuracy of up to the arbi·
trary phase charact~ristic of the antenna). This means that
when thE:' directional diagram is prescribed in this way, with
the phase of the function F (;) varying in the interval
-k ~ ~ ~ k and the function F (~) assuming different
forms for ~ > k, the current distribution may vary within
a wide range. In the absence of a unique solution to the
reverse electrodynan1ic problem, there arises the possi bi Ii ty
of imposing on the prescribed function of the directional
diagram a whole series of conditions ensuring current
distributions which, apart from the prescribed directional
diagram, would also confer to the antenna a series of de-
sirable characteristics, such as efficiency, a wide pass-band,
simplicity of construction. etc. The mathP-matiral exprps-
sion of these conditions may usually be found only in the
case we know a concrete way of obtaining the current distri-
bution. So far, the general approach to the· problem regarding
the obtention of optin1u1n current distributions frotn the
point of view of efficiency, \\'ide pass-band, etcq has not
been worked out in antenna technique.

e) The Chebyshev Polynomials

Apart from the possibtlities indicated above for obtaining
optimum current distributions, the form of the direction
al diagram is itself often prescribed non-unique. Indeed,
144
the technical characteristics usually requared of a direction·
al diagram by the radio-technical system to \Vhich the
given antenna is connected, are Jirnited to the width of the
directional diagram (at the zeros or at half power) and to
the level of the minor lobes. Moreover, frorn the point of
view of the characteristics of the radio-technical systern
as a whole, the forrn of the minor lobes and their distri-
bution in the non-operating sector of ~pace are rnore or less
indifferent. However, due to constructional consideraUons
(dhnensions and si1nplicity of the antenna system) this
question is by no means indifferent inasrnuch as the current
distribution depends on the values of the function of the
directional diagram F (~) in the whole range of angles.
We are thus confronted with the problern of finding the
optimum forrns for the directional diagrams of highly
directional antennas which, for prescribed width of the
directional diagrarn and level of the 1ninor lobes. would
reduce the dimenssons of the antennas do\\'n to a rninimun1.
The problem may be formulated in two ways:
1. The dimensions of the antenna systen1 and the level
of the minor lobes being prescribed, determine the current
distribution in the antenna that will ensure the rninirnurn
width of the major lobe.
2. The width of the dircrtionaJ diagran1 being prescribed,
determine the current distribution in the antenna that
will ensure the minimum level of the rninor lobes for the
dimensions chosen for the antenna system.
Stated in this way, the reverse problems of electrody-
namics are solved with the help of the Chebyshev polynon1ials;
we shall give~ a brief outline of their properties. A more system-
atic account of the theory and properties of these poly·
nomials may be found in the literature (see, for exarnple, (21 f).
The Chebyshev polynorrtia Is have tht following form:
T m (X)= cos (m arc cos .:r). (4-95)
It is easy to prove that the function cited here is the
polynomial of power m with respect to #the variable x. Let
us rnake the substitution x=cos 6 and utilise the expression
ror the tx pans ion of cos m6 in a cos 6 power series:
cosmu=cos
.£. ,. .£.
u-
( m )
cos •-• u~ san
. z u~ +
2
+ ( m4 ) cos,..m-' u s1n
R. • ' -'-
u t .••• (4-96)

10-2122 145
where

(:) = kl(:~k)l .

If we replace sin2l>= l-cos 2 ~ and, in turn, cos ~=x we

obtain the follo\\·ing expressions for the first seven
Chebyshev polynomials:
T 0 (X)=I;
T 1 (x)=x;
T 2 (x)=2x 2 -1;
T. (x)=4x 1 -ax~ (4-97)
T'~~, (x) == Bx - - Bx - t- 1;
4 2

1,, (x) ~- 16.x5 - - 20x 8 -I- 5x;

T 6 (x)=32x 6 -48x" -i- 18x2 - l ;
T 7 (x)= 64x 1 - 112x1 1- 56x 1 ·-- 7x.
Fig. 4·21 shows the graph of the function T7 (x) for
O~x~~I.3. In the interval -1.3Gx~O, the form of the
graph of the function T, (x) is similar to the one shown.
As can be seen fron1 the graph, a11 the rr1aximun1s of the
Chebyshcv polynomial on the segment-l~x~l are equiva-
lent and equa I to unity and the nun1 ber of the zeros of
the function on that segment is equal to the power of the
polynon1ial. Outside the segment-l~x~l, whenlxl> I
T1, on the 111odulus grows steadily without restrictions. Ac-
cording to the polynon1ials theory used here, the functions T,1
constructed within the segment -l~x~l in agreernent
\Vi th the a bove·mentioncd rule, are the polynomials of least
deviation from zero. Tl1i~ Htt!an~ thal lhe deviation of the
rnodulus of these polynomials from zero at the points of the
anaximum values of the polynomials on the segment -J~x~
~1 is at a minimun1 in con1parison with any other poJy-
nonlial of the sarne power with real coefficients and with a
highest-term coefficient equal to unity.
Let us now consider the sante Chebyshev polynomial
T1 (x) on the segment -I~ x~I but with a changed scale
T 1 (a 0x) where a(l> I. In that case, the graph of the function
1''1 (aox) on the segment -lE;; x~ 1 (Fig. 4-22) acquires
th~ characteristic form of a multilobe directional diagram,
the maximum of which occurs for x=+l. The level of the
-
minor lobes of this muJtilobe diagram is determined by the
146
ratio r,:a.o) . because, in accordance wit11 the graph in Fig.
4-22, the amplitude of the n1ajor lobe equals T7 (a0) and
the amplitude of the minor lobe equals unity. In accordance
with (4-95), the distance along the x-axis bt't\veen the point
with the maxin1um value and the first zero of thl~ function
(~x in Fig. 4-22), is writ-
ten as: 30 ;;&}
A l 3l 20
L.\..\ = 1- -- COS , ( 4-98)
a0 2tn
10
'vhere m is the po\\·er of g
the polynomial. re:
For tn=7, this yields: j .a
t
4
6.x = ( 1- _!__cos _:t_) . 3
u0 14
2
The positfon of the zeros X 0
and the n1axin1utns Xm of
f .--.I
I·
the function on the x-axis
is deterrnin~d by the re- '
* l
~
•

4-. •
1a dons
a 0 x 0 -= cos ~ + ll_~ ·
2nz '
prt
a ox m=cos _.1:1:__ , (4-99) -
~ .. ~·
- l-t
t141.r
where pis the ordinal nutn- 0 02 Q4 Q6 08 lt07t2
l>er of the zero or max itnutn Fig. 4-21. <"Jra ph of tht' Chel>y shev
and "' is the power of the poJynonJial T7(x).
polynotnial. For the graph
shown in Fig. 4-22, it should be assurned that tn=7. Sin1ilar
expressions will apply to the polynon1ials T m (x) of any
order m.
Now assume that for an antenna of a certain definite
design, the directional diagram n1ay be repr(;•sented in the
form of a polynomial of power m at· the variable x, pro-
portional to the cosine or sine of the angJe of obsrrvation
and that, in addition, the coefficients of this polynornial
may be given any value through varyin2 the elcn1ents of
the antenna.
If we consider that the power of tt-e polynomial is pre-
scribed, the coefficients of the polynorni als should be cho~.en
10* 147
so as to correspond to the coefficients of the Chebyshev
polynomial T,. (a 0x) of the same order m. Then, for the
prescribed level of the minor lobes T ..~«e)' the width of
the directional diagrarn at the scale of the variable x will
be at its minimum and, conversely, for a prescribed distance
on the x-axis between the maximutn of the diagram and

·30 l7 !"';/"')
·20 I ~
a1 =tf4
'iriB-
6
5
~4
3
2 •
T,fa,)
.., 1ft.
• 1 11 r-,

..

- ft- • ... • ~
- +

V 06 ,I l.z'J
t2 tO aa a6 ().2 0 Q2 114Xm x, 08 f.O t2
AzJ-.
Fig. 4-22. Graph of the Chcbyshcv polynornial T., (a0x).

the zero (L\x), the level of the minor lt>bes will be at its
minimum. The last assertion is directly derived from the
properties of the Chebyshev polynomial inasmuch as, among
all the polynornials of the same power, the maximums of
the n1inor lobes (on the segment -l~x-=-1) of this poly-
nomial differ the least from zero. The first statement is
easily proved if we take into account that outside the
segment -l~x<l, an1ong all the polynomials with real
coefficients of the same power, the Chebyshev polynomial
possesses the highest velocity of increase when JxJ becomes
larger than unity. ·
148
 Taking advantage of the above favourable properties of
the Chebyshev polynomials, we can construct antennas with
a directional diagram of optimum form. Let us illustrate
this on the example of a eo .. phased rnuJti .. unit linear antenna.

f) Designing an Antenna rt'i th

a Directional Diagra1n of Optimum Form
Let us examine a concrete antenna sy~ten1 which naay
be optimised by the Chebyshev p(\lynomials. Suppose an
antenna system ts prescribed as an array of radiators equally
%

d
•· .... ' ................. . •··· ..... •· ... • ••••• ~
I, , I~ I., Z:, I, I, /11 .z;,

.1, •. ..... ... ..1i

1,., 1, zr,
.'
I, 11
... ... .. .. .. .•r,., r,. ...
(/J)
Fig. 4-23. D'.strlbution of the dipoles along a linear antenna:
a-even number of dipoles: b-odd number oi dipoles
I

spaced at a distance d apart along the x-axis (Fig. 4-23).

All the radiators have a co-phasal feed and the directional
diagram of each radiator in free -;pace represents a circle
in the ¥Z·plane, i.e., we are examining only the
con1bination factor and taking no account of the directional
properties of each radiator.
The field of such a system of sourceS in the far zone in
the case of an even number of radiating elements in line is
(without taking account of the phase factor):
N

lEaN@ >I= L / cos 2k;l (~d) sin P

4 (4-100)
t•J
149
and, correspondingly, for an odd number of sources:

• (4-101)

If we introduce the ne\\' variable u equal to:

(4-102)
we obtain:
N

I E 1 N(8)1= L_Jd cos(2k--l)u ;

k=l
N

Jl::zN+ 1 (8) J =-= ,L. Id cos 2ku . (4-1 03)

ll=O

On substituting x=cos u, each of these expressions

can he reduced to the expression of the polynotnial of 2N -1
and 2N po\vrrs respectively. To this end, we n1ake use of
(4-H()) and replace sin 2 u by (l-cos 2 u). Following a
nuJHul'r of transforn1ations, we obtain instead of (4-103):

N N
~~ ~ I A 2 k- 1 zq _ 1.
IE 2N( e) I=: ~ ~ d 2Q-1 X '
q:.l k=q
N

IE z.lV+t (0) 1--- ~..,

- 6 c x'q (4-104)
q=-o
wherr
k-1
p 2k-1
A:::~ 11 9
=(--1) - L_ p-k+q 2p
•
'
p=k-q
k

A::= (-l)k-q L_
o=k-q
p-k+q
p 2k
2p .

Thus, the directional diagratn of an array of eo-phased di-

poles is expressed by a polynomial of 2N-l power or 2N of
the variab]e X, changing from --1 to +1. rfhe power of the
polynornial is determined by the number of dipoles in the
ISO
array (it is one less than the overall nunlbt~r of dipolt.;)s),
i.e., for the prescribed distance bet\veen the dipoles. the
power of the polynornial is closely connected \\'ith the di-
rnensions of the linear antenna. The coefficients of the poly-
nomials (4·1 04) are ex pressed in the forn1 of polynornia Is
of various powers of the currents in the dipoles and they ntay
be given any value through th(~ adequ::atc choice of th('Sc
currents.
Thus, we have obtained a practicable antenna systt•rtt
which enables us to n1akc use of the properties of the
Chcbyshev polynomials and construct an antenna with a
directional diagram of optirnurn forrn if the currents in the
dipoles Id are chosen in such a \V ay that the coefficients of
the polynornials in (4·104) \Vould correspond to the cof'f-
ficients of the Chebyshcv polynornials of the satnc orrlt'rs.
The condition of the equa Ji ty of the cot-lf fie il~nts ruay he
\\'ri t ten as:
N
~I A2k-1 _ Q2N- I (U )2q- 1. (4-l 05)
~ d IQ-1 - 2Q-I u •
'
k=q
N

LidA::
k=q -
=a:~(u 0 ) q; 2

N--·l is the coefficient in the 2N -1 ordrr Che-

where a 22 q-J
byshev polynomial of x2 q-• in (4-97);
a:~, the coefficient in the 2N order Chcbyshev
polynomial of x 2 q in (4-97).
The order of re-evaluation of the paratneters of the di-
rectional diagram represented by the Chebyshev polynornial,
from the variable x into real paratneters (level of the minor
lobes and \\·idth of the directional diagrarn) of the antenna
under consideration (i.e., the re-evaluation frorn the
variable x to the variable p) wiJJ be given in the foJJowing
section.
Let us now consider how we should choose the distance d
bet\\'ecn the dipoles in order to obtain a 'directional diagram
of optimurn form. To this end, let us analyse the ex press ion
which connects the variable x in the expressions (4-104) to
the variable p corresponding to the angle of the point of
observation with respect to the linear antenna under
consideration, and the nature of the changes of the graphs
151
of the same functions of the directional diagrams but
constructed once in the x, F(x) coordinates and then in the
~. F(p) coordinates for various values of d.
In accordance with the above,

X=COSU=COS f~sinp].

As ran be seen f1 ntn this expression, the range of change of

the variable is lirni ted to the segment -1 ~ x ~ J for any
range of change of the variable u. Let us investigate how the
range of change of the variab]e u influences the character
or the directional diagrarns. Since the range of change of u
is deterrnined by the rnagnitudc of d, let us take several
particulur values of d.

20 F(/J)
m
16 ~-

12
-
8 ·- .... . -
-
4 --
J

I ;.c: ~~----~-~:SW s-;,8

-10 -ro -so -Jo -to o ffJ 3fl 50 ro 10
Fig. 4-24 C>ptilnurn clircctional diagran1 for 2N-;8.

then u =-f s1n ~ and when the angle

changes from --;. to + T, the v.ariable u changes
in the · intf'rva I fron1 --;- to T. At the same time,
the graphs of one and the same function f(x) (let us take,
for exatnple, the function represented in Fig. 4-22) and
F (p) have the same character (the function F (p) is ~epre­
sented in Fig. 4-24).
2. If d== 'A, the graph for F (x) remains the same (Fig.
4-22), and the graph for F <Pl acquires quite a diflerent
form (Fig. 4-25), since, when d==A., u=nsin ~ and when
162
p changes from -f to f, u changes. from -n to n,
which corresponds to the change of x frotn -1 to I
through 0 and back to -1.
d J
3. When A< on the graphs F(p) constructed for
2
one and the same function F (x), as the ratio decreases, *
so does the number of lobes in the directional
diagrarn of a linear antenna.
Having investigated all these particular cases we can
draw the oJ Iowing conclusions.

20
fl
•
18
16
f4
IZ
10
8
'
6
4
2
- [/J
Pgo -70 -
I'W'

-fO 0 fO 30 50 70 !J()
Fig 4-25. Directional diagram for d--=A..

When d={-. the directional diagram of a linear antenna

is described by the graph of the c:hebyshev poJynlnniaJ pro-
vided the col1dition (4-105) is fulfilled \Vhen its argurnent
changes from 0 to 1.
When d < i··the directional diagram is described by
only part of this graph ami when d> } , there occurs a
larger or lesser repetition of this graph at the angle p.
When d= 'A., an additional maximwn is obtained along
the linP.ar antenna ( ~ =i)
equal to the principal maxi·
mum when P=O, and in the case of the utilisation of
non-directional elements in an array, placing the radiator
at such a distance gives rise to a directional diagram of
unsatisfactory form.
153
 If we leave aside the values of d close to A., then, for any
d<'A, we obtain the same form of directional diagram, in
which a change of d only leads to a change of the number
of rninor lobes. Thus. it is possible to obtain n directional
diagran1 with the same level of the n1inor lobes and the
sarne width of the n1ajor lobe between the zeros from sys-
terns with any distance between the dipoles smaller than
/. but with the san1e nutnber of elements. But, as follo\vs
from the above, ultradirectional antennas can be obtained
in the case of small{-; this will manifest itself in that the
currents in the elernents of the I inear antenna wi 11 sharply
change fro1n E'lement to elernent. Frorn this point of view,
it is preferable to choose d within the range of from -~ to
J~-t\ without reaching the values of d for which a second
nJajor lobe begins to appear when ~= -~- .

g) Computation of a Linear Antenna with

a Directional Diagram of Optimum Fornz
-
1. Given are the width of the directional diagram bet\\·reJI
the zeros 2~ 0 and the distance d between the dipoles.
In accordance with (4-99), the location of the zeros of the
directional diagram represented by the Chebyshev poly-
nonlial of the m order on the x-axis is defined as:

a t = cos (2P + 1) n ·
o"' o 2m •

the first zero of the diagran1, which detern1ines the width

of the tnajor !obe, corresponds to the ca.se p=O,
Then
X0 = COS ( r
nd . A )
Sln t'o ==COS 2m a; •
n I (4-106)

We select m and a 0 in such a way that the condition (4-106)

should be fulfilled; whereupon one should take into account
that the larger the value chosen for m, the smaller the level
of the 1ninor lobes of the dipole array.
The values found for ao and m as well as that of f which
154
we know from the conditions of the probien1, are substi-
tuted into the expressions (4- 105) \Vhich establish the re-
lation between the coefficients of the Chebyshev polynornial
of the m order and the coefficients of the poJynorniaJs of the
21V or 2N-l order £•xpressing the directional diagranJ of a
systern consisting of 2N -t-1 and 2N dipoles (ex press ion
4-104). At the san1e tirne, one should take into account that
tn=2N-1 for a systetn conc;isting of 2N dipoles and tn=·-=
=2N for a system consisting of 2lw' -t-1 dipoi<:'S. \Ve obta1n
a system of equations for definin~ the currents in the
dipoles Id.
Example I. ''[he giv~n width of the uirC'ctional c.Jia .
gram is 2P 0 =35;). The distance cho~cn bet\\'(\~n the
"
dipoles is d = ~-. We prescribe m= 7 (!Ill' nu mher of the
dipoles 2N = 8) and deternline the parant<.~tcr u0 :
n
cos-~-

a = - 2t1l gr.:
== 1.0 o.
o
COS
(n-:r ·R)
SI 11 t'O
.
Hence, in accordance with (4-97)
T., ( 1.09 5) = 9 .I.

Consequently, the Jevel of the rninor lobe~ is:

R=20log\~f.l, ( J )
=--19.2 db.

In accordance with (4-105), the systcn1 of equations for

defining the currents will becorne:
64/ 4 =64 (a 0 ) 7 • I 4 = 1.89;
16/ 1 -1121,=-112(a 0 ) 1 • 1,=2.21;
41 1 -20/,-+ 561,=56(a 0 ) 1 • 11 ==2.88;
11 -3/ 1 + 5/ 1 - 7/ 4 = 7 (a 0 ) ·/I 1 ~ 3.17.
Example 2. The given level of the minor lobes is 30 db.
The number of dipoles in the antenna is 2N = 8, the
A
distance between then1 is d = 2 .
155
 We 20 log T, (a 0 )= -3 0d b. Hrnce u 0 =1 .18 .
find
According to (4-105), the system of equations for defining
the currents becomes:
64 I • = 64 ( 1.18)' -• / 4 = 3. 18;
J 6/ ,-1 12 /, = -1 12 (I. 18)' ~I,= 6.3;
4I I - 20 I, + 66 I 4 == 56 (I .18) • I. = 12;
1

1, - 31, + 5/ I - - 71. = 7 • I . 18 • 1, = 12 .15.

At the san1e time, the wid th of the directional diagram
2P 0 is:
n .
Sltl Po \ =
cos(ln) = 0.826;
COS ( -
2 / ao
sin ~.=0.382, 2~,==45°.
 CHAPTER FIVE
The Influence of the Earth
and Metal Bodies on Antenna Radiation

5-l. Radiation of a Symmetrical Dipole Located

over Plane Earth
So far, we have been considering antenna systen1s located
in free space. Under actual conditions, antennas are installed
near the surface of the earth or near various n1etal bodies.
In n1any cases, due to its high conductivity, the Parth
is regarded as an infinite ideally conducting plane surface.
Hence the estimation of the influence of the earth on the
radiation of antennas can be considerably sirnpli fied and
the n1irror method appJied.
As for the estimation of the influence of n1etaJ bodies on
the radiation of antennas, here too \Ve are forced to idealise
the conditions of the problem and consider bodies of the
simplest forrn. To begin with, Jet us discuss the problern
of the radiation of a symmetrical dipole situated above
the earth.
To simplify the problem, let us assumt! that the earth is
an ideal conductor and plane. Let us place a symn1etrical
dipole (Fig. 5-1) at a distance h above the earth and let us
construct its mirror image relatively to# the earth. Then, in
the case of a vertical dipole, the current in the mirror image
will have the same direction as the current in the dipole
itself, and in the case of a horizontal dipole, the current
in the mirror image will have the opposite direction.
Let us calculate the intensity of the field set up by the dipole
in the radiation zone in the presence of the earth, assuming
IS1
that /\ is the angle between the horizon and the direction
towards the point of observation in the vertical plane (Fig.
5-2). Let r 1 be the distance from the dipole to the point of
observation, r 2 , the distance from the tnirror image of the
dipole and r 0 , the distance from the intersection 'Ni th the
plane of the line joining the dipole to its mirror image.

Th e

11 6)
v1
l t~2
''
'" e \
\

Fig. 5-l. Constructing the mirror images of dipoles:

1-dlpole: 2-mlrror image of the dipole.
-
Since r1 =r0 -h sin fl, r1 =r 0 +h sin~' the total field
of the dipole and of its mirror image is expressed as:
E = i 1201 a e -llu. cos (kl sin A~- cos kl cos (kh sin~) (S-I)
'o cos
in the case of a vertical dipole (in its tneridional plane)
and
120/ -tl:r,)
E=- a e ( 1-cos k/) sin (kh sin 11) (5-2)
'o
in the case of a horizontal dipole (in its •equatorial plane).
An exarnination of the expressions (5-l) and (5-2) reveals
that in the direction of the horizon (on the surface of the
earth), the vertical dipole sets up the maximum field
intensity, whereas the horizontaJ dipoJe does not radiate at
all in that direction.
Fig. 5-3 shows a number of directional diagrams of a
dipole installed horizontally. It can be seen that when the
dipole is low (h<'A/4), the maximum of radiation is di-
rected upwards at an angle t\=90°. Raising the dipole leads
to a decrease of the raditttion upwards ~nd when h= 'A/2,
158
the maxin1um of radiation is at an angle L\ =--=30°. A further
increase of h leads to the appearance of additional Jobes in
the directional diagrams, their number increasing as h
increases, and the maximum of the first lobe frorn the
horizon pressing itself closer and closer to the horizon.
If we take into account the finite conductivity of the
earth, the tnirror tnethod becornes invalid and the problern
of the radiation of the dipole brcornrs considerably rnore
complicated.
Even though the exact solution of the proble1n concerning
the influence of the earth of finite conductivity on the radi-
ation of antennas is possible,
it leads to very con1 pI ica ted
resuJts. That is why \\'e sha11
not d\\'ell on the rr1ethods
of the exact solution of the
problem and shaJJ give h
f
the approx irnate rnethod
applied in practice. ,
'
I
F~or the firld in the radi- I . , /
ation zone, the influence of h
the earth 1nay be esti1nated 1 '
I (~"'
~..L- - -
by rncans of the reflection Fig. 5-2 Call.: u l at i ng 1he direct i v i t y
coefficients. Here again, d i agra 1n of the dipole.
we assume the surface of
the earth to be plane and construct the rnirror irnage of the
antenna relatively to this surface, as was done in Fig. 5-2.
'The current amplitude and phase in the tnirror i1nage are
supposed to depend on the angle of observation of the field.
The ratio of t~e complex amplitude of the current in the rnir-
ror hnagc to Lhe cornplex amplitude of the current in the
antenna will be considered as equal to the reflection coef-
ficient. Thus, two rays will reach the reception point: one
coming directly from the antenna and the other, reflected
from the surface of the earth in accordance with optical
laws.
Disregarding the influence of the earth, Jet the field
intensity of a symmetricaJ dipo]e be: ·
(5-3)

where Emax is the field intensity of the dipole in the

direction of the n1aximurr1 radiation.
159
 F (~)
is the normalised directional characteristic
of the dipole.
In that case, the field intensity of the wave reftected
from the earth at the same point of observation is:
E.= EmaxF (d) If I e-llkh sfn ~+t• • (5-4)
Consequently. thP total field intensity is defined by the
,,.
.....--.-......;..:_11• O.IIA

Fig. 5·3 Directional diagrams of a horizontal dipole in the

vertical pi an e.
•
expression
E=E,+ E.=Emaxf(A)(I +·t/fe-ttkllstn6+t•], (5-5)
where the reflection coefficient for a vertical dipole is defined as:

(6-6)

160
and for a horizontal dipole, by the expression

In (5-6) and (5-7), e; is the complex relative pern1ittivity

tO E
QD-
£,
08 m
Q ...........--+--.j .....-+-- 0.7~~~~4-+-~~

Q6 ..........---+-t ~ Q6 ~ ~-+-~~--+--+--4
Q6~~~-+~~~~ IJ5..___. ;.._~--+--..:! ..............-~~

Q/it---+'111 04 \
aa~ --+-+--+-~~--+-_. Q3~~~LJ~~~
e=3;6=fltKJ(J5
02 .......,..__~~~~....,;: ....=*-"1&1
Q2
Qf af 11-+--+-~-+--+--+--~
o to 20 .10 40 5o BD 70 tJO uo 0 10 20 30 40 50 60 70 80g}
11=0.25A . h=fl5A.

tOE
~--+- ......
aa~- ~........_ ~~
- ...........
Q7t-++- ._._,. .....~......_..._...-t
aB~~6~~~~~+-t

Q5~~rH~,~~~
ll4 ................ ~,__,..__..,_~ --+--t

Q3~+-+---+-~~ ~~---+-~ ......~

112 --+-~
O.f -+---+--+ --+--~~--li-----t Fi~. 5-4. Directional diagrams of
a horizontal dipole in the vertical
0 1030 40 5060 pi ane taki n~ into account the finite
h•A conductivity of the earth.

of the earth, equal to:

E ,' =e-e(l - t.o)
- ,
(1)8
•
(5-8)
0

Fig. 5-4 shows the directional diagrams of a horizontal

dipole plotted in Cartesian coordinates in accordance with
the above-mentioned expressions [22] when A=30m for
11-2122 161
ideally conducting earth (a= oo), wet earth (s,=8, a==
=O.oos-s m) and dry earth (s,=3, o=0.0005- 1 m).
As can be seen from this figure, the influence of the earth on
the directional diagram amounts, on the whole, to a certain
widening of the diagram and the replacement of the zero
radiation directions by minimum radiation directions.
As for the influence of the earth on the radiation
resistance and the input resistance of a symmetrical dipole,
in order to simplify the problem, the earth is considered as
an ideal conductor, in which case, the radiation resistance
is defined as the sum of the natural resistance of the dipole
and the resistance introduced into the dipole by its mirror
image. At the san1e time, the input resistance of the dipole
is determined from the engineering expressions (2-49).

5-2. Radiation of an Asymmetrical Dipole

Disposed on the Surface of Plane Earth
Let us investigate once again the electromagnetic field
in the vicinity of a symmetrical dipole.· We saw that the
field intensity in the vicinity of the dipole is expressed by
(2-25), (2-27) and (2-28). As follows from these expressions.
in the equatoria~ plane of a symmetrical dipole, the com-
ponent of the electric field
intensity Ex normal to the di-
pole axis or, which is the same
thing, tangential to the equa-
,
I
/
/ l torial plane, equals zero. Conse-
quently, the equatoriai plane
I J.
of a symmetrical dipole is an
~ equipotential plane with a zero
potential.
Fig. 5·5. An as}manetrical It is obvious that if we neg-
dipole. Ject the Jo\(rer half of the dipole
and align its equatorial plane
with an ideally conducting plane surface of infinite extension,
the electromagnetic field obtained in the upper half-space of
the equatorial plane of the dipole is the same as for a sym ..
metrical dipole (Fig. 5-5). On the surface of the conducting
plane, the electric lines of force running normally to it Jiill
terminate on electric charges and there will arise surface
currents caused by the change of the charges in time. The
density of the surface charges and currents i$ determined~ by
\
~

161
the normal components of the electric field Intensity E, and
the magnetic field intensity H, from the expressions (2-25)
and (2-27).
The density of the electric charges at the point of the
plane is defined by the expression
. 60/ 0 { e-ikV x2 +I• e- ill~ }
q~= B0 E,=- z . tl r -cos kl • (5·9)
(z::o) san . J xi+ l• x
and the• density of the electric surface current by the
expression
Js=H r; =i 1.• _ {e-lk V .c•+l'- cos kle- 1klt} . (S-10)
Cz=O) 2nx su1 kl
It follows from the abov~ that the radiation of an asy•n-
metricaJ dipole is determined by the currents flowing in
the dipole itself and by the currents flowing on the surface
of the con~ucting plane; furthermore, the radiation of the
currents of the plane is equivalent to the radiation of the
mirror image of the dipole relatively to this plane.

8·10·

,..,............_. 6·10·
Fig. 5-6. f>irectional diagrams of an asymmetrical dipole in the
vertical plane.

To excite in an asymmetrical dipole a current Of the same

amplitude as in a symmetrical dipole, it is, apparently,
necessary to reduce by half the voltage fed to the dipole.
Hence, th~ input resistance of an asymmetrical dipole is
haff that of the corresponding symmetrical dipole
ZI;taymm
Zzoasymm = 2 (6-1 I)
168
The sarne goes. of course. for the radiation resistance

(5-12)

In particular, the radiation resistance of a quarter-wave

di poJe (/ : : : : A/4) is found to equal 36.6 ohrns.
The directional diagrarns of an asyn1n1etrical dipole in
the vert ica I p Jan(• arc defined by the ~an1e ex press ion (2-15)
as for a symmetrical dipole.
Fi~. 5-6 shows several directional diagrarns of an asym-
rn('trical dipole in the vertical plane in the case of ideally
conducting earth.

5-3. The Influence of the Earth

on the Radiation of Multiple Antennas
The m(~thod for estin1ating the influence of the earth on
the radiation of syrnrnetrical dipoles set forth above rnay be
applied to any rnultiplc antenna. In this case, the expres-
sion (5-5) rernains unchanged, F (!\) designating the
nortnaliscd dirt'ctional charactf'ristic of the antf'nna. taking

'.r V
r t&___ z
h ~, h •

, , , , , .. ,,,,/,'///1 '.'1." P PI
'
t
'
lj
r:l•' ''''lrr-7
h
j_ --- -~- --- --- h
l.
•

Fig. 5·7. Calculating the fddlataon of a rCO·ph~lSl'd etrray

taktng into account the influence o the earth.

no account of the influence of the earth and Emax' the field

intensity of the antenna in the direction of the maximum
radiation. I~"urtherrnore, the phase direct ionaJ characteristic
is detertnined relatively to the centre of the antenna and
h designates the distance from the surface of the earth to the
centre of the antenna.
Thus, for exatnple, in the case of a eo-phased plane array
of half-wave horizontal dipoles situated at a height h above
164
the earth (Fig. 5-7), \\'e ha\'e in the plane of the magnetic
\'ector /
_ . 60/ antra - ikr,,.
Ema'll-' , e '
0

\\·here rn is the distance frorn the centre of thr array to the

point of obsPrvation of the field.
Substituting these ex pression5 into I
(5·5), \Vhere lz is detcrnlint'd frotn
Fig. 5-7 and f.._ in accordance \Vi th .L ....L
(5-6),\\'£& rnay construct the directional
characteristic of the antenna in the ,,,,,,,,,,,,,,,,,,
vertical plane taking approx in1ate ac .
count of the in nuencP o the real earth.
Z'
"'
--- ---
When calculating the radiation re-
sistance of a co-p ha t.~d array, one shou fd
I'
--- ---
3'
-~ E x pi a i n i n~
take into account that th(• currt·nts in tPihcg . c(i!)IcuI et ti un of the
the mirror irnage of horizontal dipoles n~s1 ~t a nee of ;4 dipole
have a negative sign r(llatively to the array t.:tking «Jccnunt
currents in the dipoles of the array, of t h~ In f1ueiH'C of the
so that the signs of the values of the earth.
rnutual resistauces taken frorr1 A. Pistol·
kors's table should be reversed. For exatnple, when n=2
and m =2, d 2 = 1./2, d, = A/2, h=~ A (Fig. 5-8), we have:
for the upper dipoles
RI= RI== R 11 + R 11 + R 11 -t- R •• - R 11, - - R,.· - R t : t ' -
- R 1 .. · = 7:i.l - 12.36 ·+ 26.4- 11.8- 1.18 + l. 77 - 3. 76 +
+ 5. 75= 77.92 ohms;
for the lo\\'er dipoles
I-?.= R4 = R11 +Rat +R •• + R1 4 - R21 '-:-Rtt'-Rta' -R~,· =
=-12.36-t- 73.1-11.8-t- 26.4 +I. 77-4.08 + 5. 75-
-8.83=67.95 ohms.
The average magnitude of the radiation resistance of a
dipole of the system will then be Rr-a= 72.93 ohms and in
accordance \\'ith (4-70), the directional gain is D==6.58.
165
 5-4. The Jnftuence of an Infinite Circular
Cylinder on the Radiation of an Electric Dipole
The probleJn concerning the calculation of the influence
of metal bodies of complex shape on the rac!iation of an-
tennas situated in their vicinity has no exact solution. That
is why the conditions of the problem have to be considerably
idealised. UsuaJJy, the body is regarded as an ideal conductor
and its surface as coinciding with the coordinate surface of
the corresponding system
z of coordinates. ·
\Ve shal1 examine here
the radiation of an
electric dipole situated
near a circular, infinitely
• long and ideaiJy con-
ducting cylinder. We
sha I1 make use of the reci-
~ procity principle. This
~ examp Ie wi 11 serve to
~~~t».._....,,on
il1ustrate the main laws
governing the influence of
--111 metal bodies on the
radiation of antennas.
Fig. 5-9. L>itT~action of a plane wave Let the cylinder be
on a c1rcular cylinder· excited by a plane electro-
magnetic wave propa-
gated in the direction of the x-axis from infinity to the origin
of the coordinates (Fig. 5-9). Let the vector E of the eJectro-
nlagnetic wave have only a z-th component. Then, assuming
the magnitude of the field intensity in the plane x=O
to equaJ E0 , we put down:
E'z = E oetk% = E'oetkr cos fP ~ (5-13)
2:t
h
werP k =-r·
Longitudinal eJectric currents are induced in the cy Iinder
under the inftuence of the field of the electromagnetic wave
acting upon it. These currents radiate a secondary field,
i.e., set up waves reflected (diffracted) from the cylinder
and moving away into infinity. The secondary field has
also onJy the z-th component of the electric field intensity.
This secondary fieJd E; is such that the to!al electric field
166
 I

on the surface of the cylinder equals zero:

E;+ E;=o when r=a. (5-14)
In order to solve the problem, let us represent a plane
electromagnetic wave in th~ form of the superposi tion of a
spectrun1 of cylindrical waves, i.e.. let us represent (5-13)
in the form of the followin~ expansion=
aD

E'z == E oeikr COl (11 ==: ~~ E i,. J (kr)

£o ~ cos nm (5 .. 15)
" n T'
n=o

u·here en= 1, \\'hen n === 0 and s,. = 2, \V hen n ==- 1, 2, 3...

It is easy to Sl~e that the expression (5-15) satisfies
the wave equation ~E -t-k 1 E=0, which, in the cylindrical
systern of coordinates, when E=Ez, will be written as:

(5-16)

Indeed, after substituting (5-15) into (5-16), we obtain the

Bessel equation

! (r iJJ,dr(kr))1 + (k
.!..r or 1
- ~)
1
r 1
J (llr)=-0.
,.
(5-17)

Note that the other cylindrical functions 1ikewise satisfy

the equation, (5-17). in particular the Neuman function
and, consequently, the Hankel function
H~•, (kr)= J n (kr)- iNn (kr),

which has the following asymptotic expression:

n n'
... / 2 -l ( ler-n-;--;)
H~'> (kr) ~ V nkr e " ' (5-18)
Jtt -t> CID

i.e., this function represents a travelling wave rntlvin&

away from the origin of the coordinates towards infinity.
For this reason, let us expand the secondary (reftected
from the cylinder) field in a series of Hankel 's functions
tD

E;=E.l:a,.e,.i"H~, (kr) cosncp, (5-19)

·-=· 167
where a" are as yet unknown coefficients, which do not
depend on the coordinates.
The expansion (5-19) 1s sirnilar to the expansion (5-15)~
it satisfies the wave equation (5-16) and the prtnciple of
radiation at infinity.
To determine the coefficients of the expansion an. we add
the ex press ions (5-15) and (5-19) and substitute (5-14) into
the bounde~ry conditions, thus oLtaining:
J, (ka)
a ==- --''---- (5-20)
,. fi~~l (ka)

·rhus. the total field is expressed as:

«.

E = E }: e in costHp J (kr)- J '11 ~l!!_~ //C2l (kr) (5-21)

z o
fi':.O
" n lf,<12) (ka) "

T'he expression (5 . . 21) \Vill b<:' highly accurate 1f \\'e regard

the fit'ld S<'l up by the t:lectric dipole a1 the point q(r «fl, 0)
in tb~ \'icin1ty of the cy1in-
z der as parallel to the z-axis
and ~ituatf'd at the di~·tant
t point (r', 0. 0) 111 the di-
rect ion of the x . . ax is. In that
case, in (5-21)
[zl r ItA' I k'l - tkr'
• E=--=-t-~-.e --·
o -l ..lwf r'

No\\', let us transf<.'f the dipole

,,·hit:h fl..lnuuns parallel to it.se]f
frorTJ the po1nt p to thr po1nt
q, 111 a 1n t a i n 1ng 1t s e Iec t r i c
•
tnornent /~/unchanged. Theu,
Fig. fl·llJ. Cah.:ulatin~ the radi- in accordance \\'tth the :-eci-
atton of an ele"·tric uipole. procity theorenl (forntulated
. in the next chapter). tht' field
of that dipole at tl1e point pas equal to the f1eJd \\'luch \Ve had
at the point q \\rhen the dipole \vas at the point p I.e .•
it is ex pressed by (5-21 ).
Thus, if \\'e place an <.. lectric dipole at the point (r', (f, O)
in the vicinity of the cylinder (Fig. 5-1 0). then the fl~ld
set up by the dipole at a distant point (r, ,, 0), ts def1ned
J61J
by the ex press ion
- t llr ...,
~ enin cos n (<p -( Jl ') X
u- u

x J n (k r') -! ..! !( A a} H tt) (k r') (5-22)

H( 2,
ka)
rr
(n ' •

at u'h en th e dipo le is situatf.'d on the- su rfac{\ of

It iC\ rlcar th
;-; t~) , the field \\'hich it se ts up eq ual~ Z('ro
the cy.linder (r'
re. Th e ex pr es sion (5·2 2) rnay be generali~t~d ror
everyw he nut
'rtnina t ion of the field of the di po le
the casP of t hP ciE)tt

' -r''
~-
r
I

..·
-+ -- -- ~--. _L__ .. .._._ .. ,..

z'...._
L_. __ z ______..,.
I
­
Fig. 5- il Explmntn~ th t' ritdiatHln of '' un~~Jtudi
nal dipolt.' 111 thtt rnrradional plane

ria l plan e (tt =9 0' "). bu t al so in an y

on ly in the eq ua to is
p~ an e, I.e., 1n the di re ct io n to rra in g \\'ith the ax
meridional tt < J8 0u
of the cylinde r an an gle 6 ly tn g lfl th e in te rv al 0<
ed. if a SU lJilar re a~ on in g ts adop ted in th e
(Fig. 5- ll ). Inde angl e tt,
ve ar rt vi ng at th e cy lin de r at an
cas'"· of a plane wa io n
fie ld of tht ~ di po le in th e ra di at
then, for defining the
zone, Wf: ob ta in :
/el k 2 sin 2 it - tkH ~~
to
.,. '
: e
~ ,cos n p -
E.= -l - 4n we R t'.n l (l CJ1 ) X
- ..
, ..

61 Hll• (k r' sin 0) I (5·23)

-J (k r' sin 6) - J n (k a sin
,. H~1 • (.~a sin (tj n

where
R= l' r• +(z - z' )
1

169
 Furthermore, in the spherical system of coordinates (R,
6, cp) the meridional component of the electric field inten-
sity is defined from the expression
EA-- Ez. (5-24)
v- sin{} '
wher~as Efl' =0 and ER=O.
The magnetic field intensity has only an azimuth com-
ponent, which is defined from the expression
(5-25)

For a symmetrical dipole situated in the vicinity of

a cylinder and lying in a longitudinal direction, the radiatea
field wiJJ be detern1ined through the integration of (5-23)
along the entire length of the dipole. Let the centre of the
dipole coincide with the plane z=O. Then, assuming the
distribution of the currt'nt in the dipole to be sinusoidal
I =I sin k (I- 1z I)
z o ~in bl

_ e-tkR e-lkRo
and taking into account that R ~ Ro e tkz' cos o
(Fig. 5-11), the expression for the total field will be:
CID

Eo= E ~ enin cos n (<p- cp') X

n=o
X J (kr' sin tl) - J n (ka stn ".L H<•> (kr' sin t)) (5-26)
n H~1 )(ka sin '6') n '
where
E-._ l,k cos (kl cos -6)-cos kl ,-tkR,
- ' 2:twe sin kl sin t · R0 '
R 0 = Vr 1 +z•.
Thus, the field of a symmetrical dipole is defined as the
product of the field E of the dipole, situated in free space,
by a factor which takes account of the existence of an
infinite cylinder.
In the cross section of the cylinder (6=90°), the di-
rectional characteristic of a longitudinal dipole does nut
depend on its length and is determined. only . .by the radius
170
of the cylinder and the distance between the cylinder axis
and the dipole. Fig. 5-12 shows several directional character-
istics of a longitudinal dipole in the transverse plane as cal-
culated by Carter {23). All these characteristics concern a dis-
tance between the dipole and the cy Iinder axis equal to r' =
=0.24A; the variable para meter is the radius of the cylinder.
We see that,when the radius of the cylinder is sn1all and equals
a=0.0016A, the directional characteristic differs insignifi-
cantly.from a circle (Fig. 5-12, a). When the radius of the

no• •
po•
(D) (C)

Fig. 5-12. Directional characteristics of a longitudinal dipole In the

transverse plane:
a-r'=-0.24A, a=0.0016).; b-r'=0.24A,
a== c-,..
0.08A.; == 0 241; Q= 0.16,_,

cylinder becon1es a=0.08A, there is a considerable decrease

of the radiation in the blind region of the cylinder
(Fig. 5-12,b) .. An additional increase of the radius of the cylin-
der leads to a further increase of its reflecting action and when
the radius of the cylinder reaches a=0.16 A the characteristic
becomes, in fact, unidirectional.

5-5. The Influence of an Infinite Circular Cylinder

on the Radiation of a Longitudinal Slot
••
•
Let us examine the radiation of a longitudinal slot on an
infinite ideally conducting circular cylinder excited from
within the cylinder. To this end. we shall again make use
of the diffraction of a plane electromagnetic wave on the
cylinder, with the subsequent application of the reciprocity
principle.
111
 Let the plane electromagnetic wave be, just as before,
propagated from infinity towards the x-axis (Fig. 5-9), bui
now let the magnetic field H~ be longitudinal. Designating
by 11 0 the magnitude of the field intensity ;n the plane
x=O, we may evidently write by analogy with the above:
CID

H; =floe- tkrco• cp=H 0 ~ EninJ n (kr) cos n<p. (5-26)

. n=o

Just as in the preceding case, we sha IJ write for the re fl~cted

field:
QD

H;= Ho ~hneninH~2 , (kr) cos ncp, (5-27)

n=o
where hnare unknown coefficients.
The electric field intensity in a cylindrical system of
coordinates (r, cp, z) is determined from the expressions
E = 1 .! aHz E = - _1 aHz E.=O. (5_28 )
r iO)(~ r dq> 1
rp LroE ar ' .
Consequently, the boundary conditions on the surface of
the cylinder are written as:
-~(H: + H:),=a=O. (5-29)
Substituting into (5-25) the expressions (5-26) and (5-27),
we obtain for the coefficients hn:
,
J n (ka)
bn =--~- (5-30)
2 H~ )' (ka) '

where the dashes in the cy I indrical functions designate the

derivative of these functions. _
If we now take it that the plane electromagnetic wave is
set up by the distant longitudinal magnetic dipole situated
at the point p (r'. 0) 0), then •
IMlkZ
. z e -lkr'
H =---1
0 4nrof.L
~~-
r'
and the field ~et up by this dipole at the point q(r, q>, 0) is:
I''Vllk2 e-lkr'
If.=- i 4 nw~
z r' Jf: e t
CID

11
·n cos ncp X
n=o
J,'1(ka)
x J,. (kr)- (IJ' • H~',(kr) . (5-31)
Hn (ka)

112
Making use of the reciprocity princi pJe transfer the dipole
from the point p to the point q. Then, the' field of this dipole
at the point p wi 11 equal the field at the point q when the
dipole was at the point p, i.e., it is defined by (5-31 ).
Thus, if we place the rnagnetic dipole on the surface of
the cylinder at the point (a, cp', 0), the field set up by it at
the distant point (r, cp, 0) is defined by the expression

J' k 1
X J (k )
n a -
n<
(2 )'
H n (ka)
a) H( 2 '(k
n a J.
) (5-32)

Let us further note that

J" (ka) /1,(,2)' (ka) -- /{<2)
n
(ka) Jn' (ka) = nt~ a , (5-33)

and fleneralise the ex press ion (5-32) for the case when the
direction frorn the antenna to the point of observation

(H,IJ,rp)
z

z-z'

,_,,
•

Fig. 5-13. Explaining the radi-

ation of a longitudinal slot.

forms with the axis of the cylinder an angle ft (Fig. 5-13).

We thus obtain:
/~lksin6 e-tkR~ ·n cosn(<p-cp')
H.=· 2a•wp.a R ~ n H~tl' (kasin t)
8 1
'
(5-34)

173
where
R== Yr 1
+ (z-z') 1
•

In spherical coordinates, the magnetic field Intensity

has
•
only a meridional component defined from the expres-
ston
Hz
He=-sino&' (5-35)

and the electric field intensity, only an azimuth component

(5-36)

A magnetic dipole situated on the surface of an ideally con-

ducting cylinder forms a slot with a voltage Uz=l~ between
its edges.
Let the slot be narrow and have a length equal to half the
wave. Then, the field of this slot at the distant point will be
determined through the integration of the expression (5-34)

...

Fig. ~-14.Directional characteristics of a longi•

tudinal slot in the transverse plane.

along the slot. Coinciding the centre of the slot with the
plane z=O and assuming the voltage in the slot to be distri-
buted in accordance with the sine law U.=U, cos kz,
we obtain:
H H 2
•
~ .,. cosn(«p-q>')
•= nllla sin 0 t <d
11•0
s,.& H< 1 l~. (ka 1111 t) '
Ill .
'
(5-37)

114
where
. u.k cos(.!}cos")e-~lrRo
H = t 2 :tCtlf.L sin it' R, '
Ro= Jfr•+za.
Fig. 5-14 shows the directional characteristics in the trans-
verse plane, as calculated by A. A. Pistolkors [24) for cylin-
ders V{ith radius a=~ A and a=:n
A. Fig. 5-15 shows the
ratio of the field intensity in the blind side (cp=n) to the
field intensity in the illumi- tO
nated side (cp=O) depending on
the radius of the cylinder. The
curve shows that for cylinders
which are thin in relation to as ~
the wave-length, the direction- Ill
al diagra111 of the slot differs
but little from a circular one. kA
An increase of the radius of the Q'o Ill tO tl z.o 1.5 J.IJ
cylinder leads to an increase Fig. 5-15. • Ratio of the Ueld
of its screening effect. intensity in the blind side to
Let us also determine the that in the illurninated side.
power radiated by the half-
wave slot in the cylinder. Since the field in the radiation zone
is defined by the two components EfP and Hft, then, in ac·
cordance with (1-10), the Poynting vector has only a radial
component and, taking account of (5-36) is written as:
Sr= 2 :on EcpE;, (5-38)
where
U0 cos 2 cos6)
(~ -lkRu ~
E -- e ~ e in cos n<p " (5-39)
,- 1
n•ka sin 'it Ro n=o "' H~2 )' (ka sin ft)
lt is supposed here that the azimuth coordinateof theslot<p' ::;:0.
Substituting (5-39) into (5-38) and integrating over the surface
of a sphere of radius R,, we obtain for the radiated power
"'
" '" U'0 cos1 ~ cos t)
24~ n S S
(

p=
1)=0 Q)=O
n&k'IJi !n• " X

~ ~ e,.e,.;a-• cos ncp cos mtp R:sfn" dftd

XL. ~ [n<•)'(kaslnt)]lH( 1 )'(kasint)]•
5 40
q>, <• )
•=• m=o 11 ,.

118
but
tn 0 when m =I= n
cos nq> cos mq> dcp ==
S 2n
-E,. when 1n=n
so that (5-40) will be \Vritten as:

0
P-= 120n•k
U' r
J
cos• (-]-- cos ft) ~'"' - ln dt•. (5-41)
sln' ft n~ A) 11
2a 1 IH(I)' (..If a s1n u
_
it - o =0 1
n

Now, if we relate the radiated power to the square of th~

voltage in the antinode of the slot, we obtain the expression
for the external radiation con-
110 ••
ductivity of the slot in the
tOO - ~
cy I inder
O.fKJ f 0=-
U'I. .
2P
(5-42)
~ Q&? ._ ()

·~
~ {270
~ Thus, the radiation conductivity
~ QBO of a narrow half-wave slot depends
I on the ratio of the radius of the
050
cylinder to the wave-length. This
kA
0.4/JO I 2 ~ dependence, calculated from
6 10 20oo
(5-41) and (5-42) for the va Jue
Fig. 5-16. Radiation cnnduc-
livi ty oi a narrow half -wave of ka, which changes in the inter-
~lot on a cylinder val fron1 0 to oo, is represented
in Fig. 5-16. We see that, for
infinitely thin cylinders, the calculated conductivity agrees
with (2-54) and equals

Gka -•o = cr:o~)' = 0.514 milliohm,

•
and, for infinitely thtck cylinders, the conductivity agrees
\Vith {2-54a) and equals
2 73 1
Gka -.«)= (l ' 0n)2
2
• = 1.028 nu"I)"10hms.
Jt can be seen frotn the curve that if the circumference of
the cylinder exceeds 10 wave-lengths, the conductivity of a
slot on a cylinder differs from the conductivity of a slot on
an infinite plane screen by not more than 3 per cent.

176
 CHAPTE R SIX
Receiving Antenna Theory

6-l. A Symmetri cal Dipole in the Field

. of a Plane Electromagnetic Wave
Let us examine a rectilinear electric dipole of length
21 and radius a. Assume that the radius of'the dipole is very
small in comparison with the wave-length.
Let this dipole lie in free space in the field of a plane electro·
magnetic wave propagated in a direction which forn1s with
the z-ax is of the dipole an angle 8 (Fig. 6-1 ). Let the vector of
the electric fiP!d intensity lie in a plane passing through the
axis of the dipole. Since
the vector E forms a
right angle with the di-
rection of propagation of
the wave, it has only the
component E 9 in polar •~
coordinates and its pro- 1 1 ~...
jection on the axis of the
d i po Ie is E z = - E sin e. Fig. 6-I. E X pI~ in i ng t ~ e ca'· cuI a ti on
Designati ng by ~o the of the emf mduced m a d•pole.
field intensity in °the
centre of the dipole, the component along the axis of the dipole
at the point with the z-coordinate is obtained from Fig. 6-1:
(6-1)
The component of the electric field intensity along the
dipole axis represents the emf per unit length of the dipole.
12 -2122 177
 Thus, a plane electromagnetic wave induces in the dipole
a distributed emf with a constant amplitude, i.e., inde-
pendent of the z-coordinate, and with phase kz cos echanging
in accordance with a linear law.
We shall now turn our atten tion to the load current con-
nected in the' centre of a receiving dipole. In accordance with
the superposition principle and due to the simultaneous
action of several emf's distributed randomly along the entire
circuit, the current flowing at auy point of the circuit con-
sisting of active and reactive resistances of invariable mag-
nitude equals the sum of the currents which would arise
~d£. l dz at that point in case each of the
~: i, i:: given emf's acted independently.
Let us suppo se that an
....--z
~

I __... ....,_ I - . . . . - t elementary emf Ezdz exerts its

effect at the point with the
Fig. 6·2. Expla ining the calcu-
lation of the current in the z-coordinate of the dipole and that
load of a dipole. at all the other points of the
dipole, the emf's equal zero.
Under the influence of this emf, an elementary current
io arises in the centre of the dipole (in the load) (Fig. 6-2).
The magnitude of this current will be defined in the fol-
lowing way: Jet us mentally switch off the ernf Ezdz and,
instead, switch on in the centre of the dipole an emf d! 0
of such a magnitude as to leave the value of the current i0
unchanged. Current iz arises at the point z of the dipole
under the influence of the ~mf df 0 •
Let us now n1ake use of the reciprocity principle, well
known in tetra pole theory, which states that if an emf of
any kind, localised at any point of an electric circuit devoid
of nonlinear resistances, gives rise to a current at any other
point of this circuit, when transferred to the second point,
the satne emf gives rise at the first point .to an electric cur-
rent of the S3me strength as in the first case. The valid ity
of this principle in antenna theory was proved by Sommer-
feld and Sveshnikova [25]. According to this principle

(6-2)
or
de,= E.f(z )dz. (6-3)
Here f (z)=~- represents the function of ~e current dis·
118
tribution for an emf concentrated in the centre of the dipole.
Note that this is not the function of the current distribution
in a receiving dipole since it is the effect of the distributed
em f.
The total current arising at the dipole centre is determined
by the emf obtained through the integration of (6-3) along
tbe entire length of the dipole
+l

£,== S l:.f(z)dl. (6-4)

Z'= -l

Thus, in order to determine the electric current arising

in the centre of the dipole under the inHuence of the emf
distributed along it, we may replace the effect of this dis-
tributed emf by that of an equivalent emf concentrated in
the centre of the dipole, i.e., at the same point of the dipole
at which the current is determined. This emf is defined by
(6-4).
Let Z 0 be the resistance of the receiver connected in the
gap in the centre of the dipole. Determine current I • Howing
through this resistance using the Tevenin's theorem which,
in this case, states that if resistance Z0 is connected to the
two points of a gap in a dipole, current I o which wi 11 be
set up in this resistance equals the ratio of the difference of
potential betvleen these points of the dipole, before con-
necting the resistance to the sum of this resistance Z 0 and
resistance Z~o measured between the points taken.
The difference of potential set up between the two points
of a gap in a dipole is, evidently, equal to the equivalent
emf determined above. Thus, the current in the resistance
of the load is defined as
+I
S Ezf (z) dz (6-5)
I
o
= -'=~-~~~--
Zo+ Zto •

The expression (6-5) shows that a rec&iving dipole may be

regarded as a generator with an emf f and an internal
resistance Z10• The equivalent circuit of the dipole is shown
in Fig. 6-3.
The resistance Zzo is measured between the points of the
gap in the wire and represents the input resistance of the
dipole when dealing with a transmitting dipole.
t2• 17f
 Let us now calculate the current of the load of the dipole
I 0 from (6-5). The function of the current distribution f (z)
is ex pressed as:
z -
) _slnk(l-z)
kl
h
w en
O< <l
z ,
'( san
0

z) _sink(l-1-z) when O> z >-l.

-
'( stn ,.: 1
0 ...
(6-6)

Substituting into (6-4) the value f (z) from (6-6) and

the value Ez from (6-1) and performing the stated inte-
gration, WP obtain:
£ 01/~~ cos (kl cos 8)- cos kl
1o=- n(Z0 -t-Zl: 0 )sinkl sinO • (6-7)

(6-7) expresses the relation of the load current in a receiving

symrnetrical dipole to the angle of arrival of the electromag-
netic wave. This relation detern1ines the
directional characteristic of the receiving
symn1etrica I dipole .
.~;.. We see that the relation between the
current and the angle of arrival of the
~ electromagnetic wave in a receiving
~ • ·symmetrical dipole is the same as the
relation between the field intensity and
the direction towards the point of obser-
Fig. 6-3. Equiva- vation in a transmitting symmetrical
lent schen1e of a dipole.
receiving dipole. l''hus, the directional characteristics of
a receiving syrnn1etrical dipole coin-
cide with those of a transmitting symmetrical dipole.
The rnagnitude of the current in the load for a prescribed
electric length of the dipole and a prescribed angle of arrival
of the electrotnagnetic wave is proportional to the product
of the electric field intensity by the wave-length; further-
more, the n1aximun1 value of the current takes place when
the dipole is tuned in resonance.

6-2. Power Dissipated in the Load

of a Receiving Symmetrical Dipole
Let us determine the power dissipated in the load of a
receiving symmetrical dipole. Note that Zl:o=RI{)+i.A,;,
and Z0 = Ro + iX 0 • Then, the power dissipated in the load
180
 evidently equals:
'

p = / /:Ro
0 Eo~E:,A 1 R0
o 2 = n2[(R:. + Rto)t-t· CXo +X .to>!J X
x[cos(klcos9)-coskl)l .
sln1 kl sin• e (6-8)

The rnaximum power in the load is dissipated under the

following conditions:
1) when the reactive part of the load resistance X 0 equals
the reactive part of the dipole input resistance X to and is of
opposite sign, i.e., when
X0 -=- XI 0 ;
2) when the active parts of the load and the input
resistance equal:
Ro =R>:.o·
Consequently, the maximum power dissipated in the load
equals:
P = E 0 ~E~,p (cos (k/ cos 0)--cos k/) 1 •
o max 4n 2 Rr.o sin1 kl sin 2 0 (6-9)

For a wave arriving at the dipole at an angle 6=90(\

the maximum power \vill be
•
• ,. I
p = E o~E o,,~ (I- cos k/) 1 •
o rnax 4n 2 R l:o sin 2 kl (6-1 0)

Note that part of the power absorbed by the dipole from

the passing electromagnetic wave is spent on the reverse
radiation of the dipole. In the absence of an active resistance
of the load (R 0 =0), i.e .• \\'hen the dipole is shortened or
when a purely reactive resistance is connected to it (X 0 #)),
the whole of the power absorbed by the dipole is spent on
reverse radiation.
As a result of the dipole reverse radiation, the field of
the antenna currents (secondary field) is superimposed on
the field of the passing plane electromagnetic wave. Due to
interference between these fields, the distribution of the total
field in the vicinity of the dipole presents a complicated
picture. In particular, on the surface of the dipole (if it
consists of an ideal conductor), the tangential component
of the total electric field intensity and the normal component
of the total magnetic field intensity equal zero.
181
 8-3. Current Distribution in a Receiving Dipole
Note that the distribution of the electric current Induced
in a receiving dipole differs from the current di,tribution
of the same dipole when it is used as a transmitter. This
will become clear if we rentember that the current distri-
bution in a dipole depends on the distribution in it of the
exciting emf.
Indeed, in the case of a rer.eiving dipole. the exciting
emf is distributed in accordance with the law expressed by
(6-1), whereas in the case of a transmitting dipole, the
exciting emf is concentrated; in particular, it is concentrated
in the centre of the dipole.
Let us find the current distribution in a receiving
dipole. To simplify the problem, we assume the dipole
to be unloaded (Z 0 =0) and infinitely thin. Turning to
Fig. 6-1, let 1: be the current at the point z of the dipole
and A~, the vector potential set up by the current of the
dipole on its surface. As we saw in Chapter Two, the vector
potential on the surface of a dipoJe obeys the equation (2-6).
In the case of an unloaded receiving dipole, this equation
will be written as:
dl A~ z e •
dzl +k Az = - tcoeE,. (6-11)
Substituting into the right-hand side of this equation the
value E, from (6-1) and making use of the approximate
expression (2-8) which relates the current to the vector
potentia) at the point z of the dipole, we obtain:
dzl~ ' e l(tU~ . (6 12)
dz• + /(, Iz= Q Er,~ Sin 0 etkz eo• 6• •

Let us limit ourseJf to the case of a normal arrival of the

electromagnetic wave on the wire ( e:::ooo). Then (6-12)
will be written as:
d•J: I e tCtl8 E
dz• +k lz= Q o8• (6-13)
lt is evident that, in this ca~e, the current is an even
function relatively to the origin of the coordinates and
the boundary conditions for it are written as:
die
1:==0 when z= ±l and a'==~ WhfTP z==O. (6-14)
IBJ
It may be shown that the solution of the equation (6-13)
in the boundary conditions (6-14) is written as:
e iCJ.ll! E 1
I z==- , 2 o~ k 2 cos kl (cos kz-co s kl). (6-15)

Observe that for a half-wave dipole, kl=90 ° and the cur-

rent distrib ution is sinusoidal, i.e., it coincides with the
current distribution in the case of a transm itting dipole.
It may be said, as was pointed out in Chapter Two, that
for a half-wave dipole, the current distribution does not
depend on the distribution of the excited emf in the dipole.
The same applies to dipoles excited on odd harn1onics
(kl=n-!-, where n=3, 5, 7).
For other lengths of the dipole, in accordance with (6-15)
the current distrib ution differs from the sinusoidal one.
Since the question of the secondary radiati on of antennas
wi 11 no longer interest us, we shaJ I not dweJ I any longer
on the current distribution laws in a receiving dipole.

6-4. Applic ation of the Princip le of Reciprocity

to the Study of the Proper ties of Receiving Antennas
Above we discussed the t'ffect which a plane electromag-
netic wave has on a receiving symmetrical dipole. To define
the current in the antenna load, we made use of the reci·
procity principle. It was
found that, to define the
current in the antenna
load. the input resistance
and the directional •
charac teristic of the --~-~- -_.._
dipole when it is utilised
as a vibrator should be
known.
From a broader aspect,
it appears that there
, J

is no necessity to work Fig. 6-4. receivin

Explaining the theory of
g antennas.
out a special theory for
receiving antennas and
that it is sufficient to make use of the transmitting properties
of the antenna under investigation and apply the reciprocity
principle. -This is what M. S. Neuman did in 1935 (26}.

'"
 Let us examine two arbitrary antennas 1 and 2, situated
far apart. Let Z1 be the resistance connected to the terminals
of the first antenna and Z2, the resistance connected to the
terminals of the second antenna (Fig. 6-4).
Let the antenna 1 be a transmitting antenna and the
antenna ·2 a receiving one and Jet f t be the emf induced at
the terminals of the antenna /. Then, under the influence of
this ernf, a current will appear in the antenna 1 the n1agnitude
of which at the tern1inals wiJI lJc defined as
8,
I (6-16)
•=z +Z· ·
1 ln1

where Z 1111 is the input resistance of the antenna 1.

Under the in tlucnce of the curr('nt of the first antenna, a
field intensity is set up in the vicinity of the second antenna,
which, in accordance with (4-62), is expressed as

E!1 ==.: l· ?,----

0 k h ,1. e- t'll r"f I (f) m) e
1 " 1'
't'l (6-17)
r0 t T '

and current / 21 wi II be induced in the load of the second

antenna.
Substituting inlo (6-17) the expression of the current
from (6-16) and ex pressing C1 through the rest of the
rnagni tu des:
(6-18)

When the emf f, in the antenna 1 and the emf f 2 in the

antenna 2 arc switched off, the current arising at the
terminals of the second antenna \ViJl be:

• (6-19)

where Z1n 2 is the input resistance of the second antenna.

The intensity of the field set up by the second antenna
in the vicinity of the first one will, in accordance with
(4-62) be expressed as:
. 30kh I
E 11 =l z •e-lkrop (8 I m)el'tl'2 (6-20)
r0 I T I

and current /1 2 wi 11 be induced in the load of the first antenna.

184
 Substituting into (6-20) the expression .for / 2 fron1 (6-19)
and defining f 2 , we obtain: '
(6-21)

According to the principle of reciprocity. the· emf

switched on at the tern1inaJs of the antenna 1 is rflated to
'!
th~ current 12 t induced under the effect of this ernf at the
terminals of th~ antenna 2, as the f'fl1f f 2 switched on at the
terminals of the antenna 2 is related to the current lt 2
induced under the effect of this etnf at the tern1i naJs of the
antenna /:
(6-22)

Substituting here the ~xpressions f 1 and f 2 and transferring

to the left-hand side of the equation thus obtained, tht.
n1agni tu des relevant to the antenna I and to the right-hand
side, the rnagnitudes relevant to the antenna 2 we obtain

(6-23)

Assun1e that the pararnetcrs of the second antenna are chang('d

without touching the first antenna. Since, when the para-
nieters of the antenna are constant, the ratio of the current
induced in the antenna to the field intensity of the electro-
magnetic wave acting on the antenna retnains constant,
the left-hand side of (6-23) will rernain constant too no
matter how Wt! change the parameters of the second antenna.
Consequent1)1, the right-hand side of (6-23) wilJ also rernain
constant.
Thus, \\'"e shaJJ have the following relation for any
receiving antenna:
I (ZL -+- Zin)
EIIF(H, cp) e'r,~ = N =const, (6-24)
where E is the field intensity of the plane electromag·
netic \\~ave in the vicinity of the antenna:
I, the current in the antenna load;
ZL, the load resistance;
Z1n, the antenna input resistance when dealing with
a transmitting antenna:
h, the antenna effective length during transmission;
185
 F(9, cp) and ""' the antenna amplitude and phase
directional characteristics during transmission.
To determine the constant N use is made of the expression
(6-7) obtained for a symmetrical dipole. Since, as we saw
in Paragraph 4-9, for a symmetrical dipole
hF (O)e'•= !:._cos (kl cos 8)-cos kl
n sin 8 sin kl '
on comparing (6-24) with (6~ 7), we obtain:
N= 1.
Thus~ the current induced in the load of any receiving
antenna is expressed as:
zLe:z,n F(fJ, cp)e '~'.
1
I= (6-25)

The expression (6-25) can also be represented in a

different form. Indeed, by substituting (4-62) into (4-68)
and taking into account that Pr. == { 111• Rr., we have:
30k1h1F1 (8, cp)
D= R:: • (6-26)

Substituting the expressions (6-25) and (6-26), we obtain:

I ).,E , JDRz e-"~ (6-27)
== n (ZL +Ztn) V 120 •

The expression (6-27) (without the factor ei+) was obtained

by M. S. Neuman in the above-mentioned work. Subsequent-
ly. A. ~· Volpert (27] introduced into this expression the
factor e'~. which takes into account the antenna phase
characteristic.
The following conclusions can be draw.n from the expres·
sions (6-25) and (6-27):
1. The amplitude F ( 8, cp) and phase 'I' ( 0, q>) directional
characteristics of any receiving antenna, which give the.
dependence of the induced current on the angle of arrival
or the electromagnetic wave are found to be the same as
during transmission, provided the receiver and the trans-
mitter are connected to the same points of the antenna.
2. Any receiving antenna can be regarded as an oscilJa.tor
with an emf defined by the expression
1,== Eh F (0, tp)eli' , · :- (6-28)
186
or
G =-~ ,/DRr, e''t , (6-29)
o n V 120 '

with an internal resistance Z1n=R1n+iX 1n and a load ZL=

=RL +iXL.
3. The effective length of a receiving antenna, dtfined
as the ratio of the magnitude of the induced emf to the
magni ~ude of the fi~ld intensity when the electromagnetic
wave comes from the princapal direction,
I I eol (6-30)
t=IEI'
is found to be the same in the case of both reception and
transmission and is defined· by the expression
(6-31)
I

where D0 is the directive gain in the direction of the

maximum radiation. .
4. The maximum power dissipated in the load of any an-
tenna takes place when XL +X,n=O and RL=-=R,n and equals:
P = 'AE £• £_1'1_ (6-32)
max n• 960 ,
where T)=::t~ In
is the antenna efficiency (Rtn=R:to+R 1011c••
R losses. the resistance of the losses o the antenna).
The maximum power dissipated in the receiving antenna
load is the greater, the more directive is the gain and the
efficiency of. the antenna.
5. The current arising in any receiving antenna gives
rise to a reverse radiation of the antenna. Thus, part of
the power absorbed by the antenna is spent in the receiver
and part on reverse radiation.
6. The reciprocity principle leads to the reversibility
of the antenna. Any transmitting antenna may be used as a
rtceiving antenna and ice versa. . .

8-5. Effective Area of an Antenna

Let us dwell on one more concept of r~ceiving antenna
theory. lt is sometimes useful to introduce the notion of
the area of the front of an incoming electromagnetic wave,
161
from which the antenna absorbs energy. The power passing
through the area Srff perpendicular to the Poynting vector
equals:
EE•
P = 240n seff · (6-33)

On comparing (6 .. 33) with (6-32). which reprP.sents the

max irnun1 power dissipated in the load of the receiving Rn-
tenna, we obtain:
SeH D11
~-

A1 -
__.;
4n ' (6-34)

i.e., the effective area of the wave front, expressed in

fractions of the length of the wave from which the antenna
absorbs energy, equals the product of the directive gain by
the antenna efficiency, divided by 4n. This area is usually
referred to as the effective area of the receiving antenna.
Consequently, the larger the directive gain of the antenna,
the larger the space from which the antenna derives energy.
Equation (6-34) coincides with (4-71) so that we may say
that the effective area of an ideal plane antenna coincides
with its geometrical area. For example, the effective area of
a half-wave dipole equals (when T)= 1) Seff=0.522l2•

6-6. Mutual Resistance of Receiving

and Transmitting Antennas
The expressions mention~d in Paragraph 6-4 enable us to
write a fairly simple expression for the mutual resistance
of receiving and transmitting antennas. Designating the
transmitting antenna by the index 1 and the receiving
antenna by the inc.lex 2 (Pig. 6-4), \Ve shall, in accordance
with (6-27), write for the receiving antenna current:
_
'AEt, cosy
1.--n(ZL•+Zlnl)
v=D
R- .,,.Ta
I l':l
e'' (6-35)
120
where we have introduced y which is the difference between
the polarisation angles of the field of receiving and transmit-
ting antennas.
If we take into account (6-26). we shall, from (6-17) obtain
for the field intensity of the transmitting antenna:

(6-36)

188
From the substitution of (6-36) into (6-35), I
WE' obtain:

I z- . A ens V l,rD D R R-- e- i.kro AL (41" t- •h ) (6-37)

-- -I Jl "'I ~2 ~;- ,..., .,..., •
/ 1 2n (ZLz -t- Ztru) 1 1 - - r0

Making use of the Kirchhoff's equation

I. (Z L. 2 + Z, n z ) ·+ I,Z,. =-.:: 0,
where z,. is the mutual resistancf' of the antennas rPiated
to the currents at the feed points of the antenna, \\'e
obtain:
• A. cosy e- tkr,,
Z 1 1--' 21t VD 1 DI Rl: 1 R1:2 - - e' (,..'a+1!'~). (6-aS)
'u
This simple expression enables us to deterrnine the n1utual
resistance of t\\·o arbitrary antennas situated in relation
to one another in the zone of radiation in accordance with
their electric pararneters. In practice, this ex press ion n1ay
be used if the distance between the antennas satisfies the
ratio
(6-39)

where L is the longest din1ension of one of the antennas.

For exampJe, for two half-\\'ave parallel dipoles lying on
one plane, the expression (6-38) becotncs:
· 60i,. -
Z.. 11 =z--e tkr
., (6-40)
nr 0

and, as sho\vn by direct calculations, it is in good accord

with the data of V. V. 1"atarinov's tabJes for distances
exceeding one wave-length.
 PART TWO

Transmission Lines
of Radio Waves
---------------·----------------------------
CHAPTER SEVEN

Transmission Line Theory

7-I. General
Transmission lines of radio waves are devices for chanel-
ling high-frequency energy from transmitters to antennas
or from antennas to receivers. The transverse dimensions
of the lines are S!,Tlall in comparison with the wave-length
or are of the same order whereas their lengths are many
times greater than that of the wave-lengths.
The first main requirement expected of devices of this
kind is that, when transmitting energy from a transmitter
to an antenna and from an antenna to a receiver, they
should not radiate high-frequency energy and should
not, therefore, distort the transmission or reception
directional characteristics of antennas.
The second requirement is that they shouJd transmit high-
frequency energy with the maximum efficiency. Conse-
quently, the losses on heating the line conductors and isolators
should be kept down to a minimum.
The third requirement is that the high-frequency voltages
arising in the lines should be as low as possible to avoid
breakdowns in the line and thus, transmit the maximum of
power.
The final main requirement is that the line should not
aJiow any noticeable distortions of the signals transmitted
along the channel.
In accordance with the above-listed requirements, difter·
ent transmission lines are used for diff~rent wave ranges.
'

190
 Fig. 7-1, a and b show a two-wire and a four-wire syrn-
metricallines (feeders) in which a transverSe electromagnetic
wave is excited, accompanied by forward and reverse
currents, in the line. These lines are used on short and partly
on ultrashort waves.
Coaxial cables and hard coaxial lines (Fig. 7-1, c), in
which a transverse electromagnetic wave is excited, also
accompanied by forward and reverse currents, are mainly
used on metre and decinlt'tre waves.

(C)

~
{d) (e) tfJ
Fig. 7-1. Cross sections of transmission Jines of radio \\'aves.

Fig. 7-1, d, e ~hows a rectangular and a circular wavegui cfes

in which the lower types of transverse electric and partly
transverse magnetic waves are used. As a rule, the wave-
guid~s are utilised on waves shorter than 10 cn1.
Finally, on decimetre and centimetre waves, strip lines
are used (Fig.. 7-1, d, e) in which a wave of a mode close
to the TEM mode is excited and single-wire lines with a
thin layer of dielectric (Fig. 7-1, g) in which a so-called
surface wave is excited. Other lines are also in use.
The present chapter deals with the theory of these lines.
We presume that the reader is familiar with the theory of
long lines and the propagation of the main modes in wave-
guides~ These questions will therefore not be treated in any
d~tail ·here and the line theory will 6e considered chiefly
from the point of view of the excitation of radio waves.
The lines in the present chapter will be assumed free from
losses. The question regarding the evaluation of the energy
losses in the lines will be dealt with in the subsequent
chapters.
191
 The present course being mainly concerned with ultrashort
wave antennas, we shall begin our study of transmission
I i nes with wa vegu ides.
•
7-2. Rectangular Waveguide Theory
A rectangular waveguide, of which a longitudinal section
is shown in Fig. 7-2, is usually excited by means of a probe
introduced into the waveguide through an opening in its
broad wall. The end of the waveguide close to the probe
is shortened and the far end is connected to the antenna.
The direction of the electric current in the probe coin-
cides with the x-axis of rectangular coordinates. This current

2 ~To ontellllfl
z

Fig. 7-2. Longitudinal section of a rec-

.. tangular waveguide:
/-probe: 2-shortened wall; 3-coaxlal llne.

occupies inside the waveguide a certain limited volume

with a certain density and, relatively to the currents
induced on the waJJs of the waveguide, it is an external
(exciting) current.
In accordance with the above, let us examine the problen1
of the excitation of the wavegu1de by an ex ternaJ electric
current with a density J~ [28]. For simplicity, we shaJI
assume the \\'ails of the waveguide tcr be ideally conducting
and its length infinitely great. Let a be the internal
dimension of the narrow wall of the waveguide and b, that
of the broad wall.
The electron1agnctic field excited in the waveguide is
described by the equations (1-4) and (1-5). Using the
mirror method, it may be sho\\tn [29] that the direction of
the vector potential of the excited field coincides with
that of the current and has, therefore, only .th~ x-th com-
ponent Ae =A~. Consequently, in rectangular coordinates,
192
the equations (I -4) and (1-5) wil1 be written as:
•

1 k'Ae +~A~
Ex = i we' ;~
x iJx1

E ::-; I o1 A~ (7 .. 1)
1 iwe' oy ox'

(7-2)

Since the waveguide is an ideal conductor, the tangential

components of the electric field intensity at the walls
of the waveguide equal zero. In accordance with (7-1) this
occurs under the condition
A~=O when y=O, y=b;
oA~ (7-3)
ax =0 when x=O. x=a.

Thus, the problem is reduced to the solution of the

inhomogeneous waveguide equation (7-2) in the existence of
the boundary conditions (7-3).
Equation (7-2) is with separable variables and we shall
therefore seek its solution in the form of the product of
three functions, each of which depends on one varia ule only,
A; (x, y, z)= X (x) Y (y) Z (z) . (7-4)
•
The function X (x) is prescribed in the interval 0 ~ x ~ a1.
Let us represent this function in the form of a Fourier
series expansion:
GD

X (x)= ~ [a,. cos (7 x) + b,.sin (7 x)]

,

and, in order to satisfy the boundary condition (7-3), it

is evidently necessary here to assume the coefficient at
the sine to equal zero: b,.=O.
The function Y (y) is prescribed in the interval O~y~b,
so that we shall also represent this function in the form of
13-2122 193
a Fourler series:
CI'J

L fam cos ( n; u) + bmsin ( n; u)] .

v<u> = m=o '

and in order to satisfy the boundary condition (7-3), we

must here assume that am=O.
The function Z (z) is prescribed in the infinite inter-
val - oo < z < oo. We shall therefore represent it in the
for1n of a Fourier integral expansion:
CO

Z(z)= ~ g(x)e-txzdx.
x=-OD
Thus, the solution of the equation (7-2) which satisfies
the boundary conditions (7-3) may be represented in the
forn1 of an ex pans ion:
CC Cl) CD

A; (x, y, z)= L, L, ~ a,m (x)cos (~ x) X

n=o m x=-OD
=1

. . (nmb y.)e-'.xz dX.

X Sin (7-5)

To determine the unknown coefficients anm (x) in (7 -5),

let us substitute it into (7-2); we shall then obtain:

n=o tn= 1 x= -~

x a,m (x) cos(~ x) sin (n;t y) e-txz dx=- j~(x, y, z). (7-6)

Consequently, the distribution of the density of the ex-

citing current is represented in the form of an expansion,
using the sc~me systern uf eigenfunctions of the given bounda·
ry problem as for the expansion of the vector potential.
Applying to (7-6) the reverse Fourier transforn1ation, Jet
us multi pJy the left- and right-hand sides of (7-6) by
cos ( nn
,1 x)\ s1n
. (nm )+,x'z h , , , "b
- 6- y e , Y-' ere n , rn , x are prescr1 ed
values of n, m and x, and let us integrate the left- and right-
hand sides of the expression thus obtained over the whole
volume of the waveguide.
194
 Let us, further, make usr of the orthogonality conditions
of the eigenfunctions of the discrete · and continuous
spectrums:
2 a nrr nn' 0 when n =I= n',
tl = n';
b
2 \ . I Jl '11 ) • ( n nt · ) J 0 \\'hen m =f= m',
b .J
11=0
sm l b Y sm - b Y dy= \ 1 when m= m';
ea

.Jrr. 5 e-tc~e-x'lrdz=ll(x-x'), (7-7)

z=- GD

\Vhere en= I \Vhen n=O, en=2 when n= I, 2, 3 ...

Here 6(x-x') is the delfa-function, which equals zero
everywhere rxcept at the specific point x=x't where it
turns into infinity; moreover
' GC

5 6(x- x')dx= 1.
As a result of this transformation, we obtain:
Dnm ( K ) =nab
f r1 5 .e (
/x X ,
1
!I , Z
, ')
COS
(
a- X
Jtfl 1 )
Sin
• ( 1tftl
-by 1 )
X
V
e' ...z·
X ·····-- __,_,_ - dV
lx• +(:;p)' + (n:lr -k' J • (7-8)

\Vhere the integral is taken over the voJun1e which con-

tains the external currents
Now, let us substitute (7-8) rnto (7-5) and calculate
the integral
«t -lM(Z-2')

K=- CID Q ) ~ b .
Applying to his integral the residual theory, we obtain
for (z-z')<O: ~

(7·9)
and for (z-z')> 0:
e-V (z-z')
M=:t-- , (7-10)
'V
t3• 195
where
(7·11)

is the propagation constant.

Thus, the final expression of the vector potential is
r:tJ r:tJ

~ ~ n
A Xe ( X, y, z) = ~0 (Jtn
~laby cos a
8
) .
X sm
(nm \
b y) X

z' ==z

x Scos (~x') sin (n; y') e-vz S j~ (x', y', z')x

s z'=-r:tJ

z'=CID
xe+V%' dz' +e+vz ~ 1~ (x', y', z')e-vz• dz' ds, (7-12)
, __
Z
-·
where the index s of the sign of the integral indicates that
the integration is performed over the cross section of the
waveguide. The coordinates of the exciting current are
shown in bold type.
The calculations we have given are useful because a whole
series of conclusions of practical importance can be drawn
from the expression (7-12). Together with (7-1), this expres-
sion provides a cornplete solution to the problem of the
excitation of a rectangular waveguide by a transverse
(:l]ectric dipo]e lying in a direction paral1el to the narrow
wall of the waveguide.
Usually, in electron1agnetic field theory courses, only
the structure of the field in the waveguide is determined,
no determination being given of the absolute magnitude of
the field components. This is due to the fact that the ex-
citing (external) currents are not introduced into Maxwell 's
equations and the wave equations derived· from them. The
solution of the problem regarding the excitation of a rec-
tangular waveguide, i.e., the solution of the wave equation
with the right-hand side (7-2) enables the field components
in the waveguide to be determined also from the absolute
magnitude. This, in turn, enables to solve a series of
problems of practical importance, such as that of finding
the radiation resistance of a dipole exciting a waveguide,
which will be mentioned further.
Let us analyse the expressions obtained.
196
 l. An electric antenna excites an infinitely large number
of modes in a waveguide. To each pair of v.alues of n and m
there corresponds a definite mode which, in accordance with
(7-11), has a definite propagation constant y.
2. If the wave number k is a real quantity (as is usual-
ly assumed), the propagation constant y may be real or
imaginary, depending on the ~ransverse dirnensions a and b
of the waveguide and the values of n and m. A real value of
the quantity y indicates that the wave process is non-existent
and the field amplitude decreases as we move 8\\tay fro1n
the dipole tow.ards the axis of the waveguide in accordance
with the exponential law. In that case, the wave is said
to be damped (not propagated).
The imaginary value y=ia.. where

a= Yk·-u~~r-(TY I

indicates that a wave is propagated along the axis of the

waveguide with a definite phase velocity. The field ampli-
tude of this wave remains constant as we r~ove away frorn
the exciting dipole, since the walls of the waveguide are
supposed to be ideal conductors.
3. The wave-length for which the propagation constant
y equals zero at the given values of n and m, is known as
the critical wave-length.
The cri t ica J wave-length is ex pressed as
A __ 2n
crll nm- ¥(7)' +(n;y • (7-13)
'
It can be seen from this expression that the higher the mode,
i.e., the larger the values of n and m, the transverse dimen-
sions a and b of the waveguide being prescribed, the smaller
the critical wave-length.
4. When the wave-length of the oscillations becomes equal
to the critical wave-length A,= Acrlt nm• the amplitude of
the oscillations in the waveguide becomes infinite. In real
conditions, the walls of the waveguide have a finite conduc-
tivity, so that although at the critical wave-length the
amplitude does become large, it nevertheless remains finite.
5. To each propagated mode there correspond a definite
phase velocity and a definite wave-length in the waveguide,
which, in a boundJess medium, are related to the phase
197
velocity v and the wave-length "'· as follows:

Vpb= V V
1 I;
l- (1ert: nm) ,

1 ( ).,, )• (7-14)
- Acrlt nm

6. The propagated modes move away from the exciting

sources towards infinity in the forrr1 of travelling waves.
As can be seen fron1 the expressions (7-12), within the
volume of the exciting sources, there occurs a supetposition
of travelling waves moving towards one another in the
direction of the axis of the waveguide.
When a rectangular waveguide is used as a transmission
line of high-frequency energy frorn a transmitter to an
antenna or fron1 an antenna to a receiver, the dimensions
of the waveguide are chosen in such a way as to allow the
prop3gation of onJy the lowest rnodes, for which n=O and
m= 1. We shaJJ therefore dwell at length on this mode.
'• ~d d •I
------I '
a I tl'l
,.z I
r t~'
l
.._,--~--

:: z •J
Fig. 7-3. Explaininf;{ the calculation of the
field in a semi-tnfinite waveguide.

Let the exciting dipole be a linear dipole with a current

1:
moment land let the waveguide be semi!infinite, as shown
in Fig. 7-2. Neglecting the higher modes (n>O and m>l),
for whiC'h the field an1p1i tu des at a certain distance from the
exciting dipole are sn1all in comparison with the amplitude
of the propagated 1node, let us consider this dominant mode.
It may be shown that the field in a semi-infinite wave-
guide (Fig. 7-2) coincides with the field in an infinite wave-
guide (Fig. 7-3) in which the field of the mirror image of
this dipole relatively to the end wall of the waveguide is
added to the field of the dipole under investigation.
19/l
 In the region z>d, in accordance with (7··12), we obtain
for the dominant mode, excited by the antenna itself, and
for the field excited by the mirror image of the dipole:

• = - abta sin ( !!_

Axo1 b
W 01
y) sin ( !!:_b y') e- taoa <z-d).

Thus, the total field in the region z > d is:

'21~x l sin
Aex to0 1, = aba ( n y ) sin ( -n
·- [1'
) sin (a d)e-tao,z. (7-15)
01
b b .1 oJ

Similarly, in the region z < d, \VC have for the total

field:
2/e I
Ac. tot=_:_ sin ( ~ 11) sin ( ~ tJ' }\ sin (ex z) e- ian, d (7 -16)
x o1 abaol b ·' b .. ot

For this dominant mode, the expressions (7-1) are \Vritten

as:
E k2 Ae
xo1 = iroe' x o 1'
•

aA~.,l dA~Ol (7 17)

Hxos=O, H,o.=-a 1 , /fzos=- ay· -
In the above-mentioned expressions

a01= Vk•- (-i-'l· AcritQt=2b, Aoper<2b.

Thus, the lowest mode under consideration has only one
transverse component of the electric fieJd intensity. For
this reason, the wave of this mode is referred to as the trans-
verse electric TEot wave and, since the magnetic field has a
longitudinal cornponent, it is also referred to as the magnetic
H 01 wave. The field of this wave does not depend on the
coordinate x (n=O) and has one "ariation (m= I) in the
direction of the y-axis.
In the z >d region, the wave is a travel I ing wave and in the
z<d region, a standing one. Fig. 7-4 shows the distribution
of the amplitude of the electric field intensity in the trans·
verse and longitudinal planes of the waveguide.
The amplitude of the field in the waveguide depends on
the location of the antenna. It is seen from (7-15) and (7·16)
that if the dipole is placed near the lateral wall (JJ' =0, b),
the field of this dipole wiJI equal zero everywhere. In just
199
the same way, the field in the z>d region will equal zero
if we assume d=~. The field wiJJ be maximum for y'=-%-
and d=';. (

Having established the relation between the field intensi-

ty in a waveguide and the exciting current, it is relatively
z
d ~

Q E:~IJI E., I:
~"' ~ r~
~--IJ-~...
~
~ ~

•
Fig. 7-4. Distribution of the electric field in a waveguide.

easy to calculate the power radiated by a dipole in a wave-

guide. Indeed, if we ensure the conditions for the propa-
gation of only the H01 rnode, the whole of the energy will be
transferred along the waveguide only by that mode.
Applying the Poynting vector method and making use
of the expression {1-10). we obtain:
a b

P1:=-} s sEJmH;
.r=o u=o
01 dxdy, (7-18)

or, substituting here the expressions for the field intensity

in accordance with (7-17) and (7-1~), on integrating we
obtain:
Pz = I~' Rzd, i
where R1:d is the radiation resistance of th~ dipole in a semi-
infinite waveguide, defined by the expression
-fl
/~Id:
Y -
8 21 I . 1 2'& , • 1

1- ( Acrlt
"'· ot )
By changing the quantities g' and d, one can regulate
the radiation re4iistances of the dipole and, thus, match
its coupling with the coaxial feed Jine.
200
 Longitudinal or transverse slots of, as a rule, a length
of half a wave, are often milled in the walls of rectangular
waveguides and serve as energising and radiating devices.
In this connection, let us investigate the question of the
excitation of a waveguide by slots.
We know that slots can be regarded as magnetic dipoles.
Let us therefore assume the distribution of the density of
the external magnetic current to he prescribed in a certain
volume of the waveguide 128 J. Let this current have only
the y-th component j~. Then, the vector potential of this
current will also have only the y-th con1ponent A~. Iience,
in that case, in rectangular coordinates, the field equations
(1-4) and {1-5) will be written as:
oA~ oA~.
Ex= oz ; E,::::O; Ez=- nx '
a•AM 1 ~AM,
H = I -11 • H = kI A M -!- 11 : (7 20)
~ ; W!J. ox ay ' Y i C.O!J. " • au• -
I cJIAII •
Hz = . a!l
irof..L iJz -'
aaA~ a·A~
otA;' • M "l\\
ax• + ay• + oz1 + k A" =-I"· (7-21)
The tangential components of the electric field intensity
at the walls of the \\'avcguide will equal zero under
the conditions
A~=O when y=O, y=b.
(7-22)
oAM
a:
=0 when x=O, x=a.
The ·solutiori of the equation (7-21) in the boundary con-
ditions (7-22) proceeds in just the same way in the case
of the investigation of the excitation of the waveguide by
an electric dipole and is written as follows:

A~ (x, y, z)= '£ '£:bv cos(~ x) sin (~my))(

n=o m= I
z
X (a
nn X ') Slfl
. (nm y ') e-vz r J: (x', y'. z')x
b
S
I
COS
z'-=- aD
fiJ

xe+vr.' dz' +e+Yz S 1~ (x', y', z')e-va' dz' ds. (7-23)

•'••
201
 Of course, this expression coincides with (7-12) since the
boundary conditions for the vector potentials are the same.
Now, let us use the solution thus obtained to define the
radiation towards the inside of a waveguide of the narrow
transverse slot of a length equal to half a wave shown in
Fig. 7-5. a. Let a, y 0 , Z0 be the coordinates of the centre
of the slot. Let V o be the voltage between the edges of the

Q a
• IJ----- ~-IJ-....,.
(Q} (!J)
Fig 7-5 E xc-it at ion of a wavegui de by a si ot:
a- trn nsvt.·• St' slot h-longlt ud 1nc1l sJot

slot in the centre and et us assu1ne the voltage distribution

in the slot to be s1 nusoi da J:
U11 = U O(.'QS fl (!/ !/G).
1
- (7 -24)
Let us now identify the voltage in the slot \Vith the linear
ma~netic current /~1 = U v and let us \\'rite for the current
density in (7-23):
j~(x'. 11'· z')=U 0 cosk(t/'-tl,J6(x '-a)6(z'-Z0 ), (7-25)
where 6 is the delta-function.
Substituting (7-25) into (7-23) and considering from
now on onJy the lo\vest 1node (n = 0, nl= I), \\'e obtain:
AM = ~o sin(!!:_u)etlo.o,,z-zo)x
yo 1 abta0 , b
flo +A· 4

x S cosk(Y'-y.)sin( Tu')dy', (7-26)

u' = Yo- A.tt
where the negative sign (-) is taken for the z> z,
re~ion and the positive sign ( + ), for the z < Z0 region.

202
 At the same tin1e. it is assumed that the IO\\'est n1ode
is of the propagating kind (Aoper < 2b). -,
PE'rforming the integration indicated in (7-26). \\·e
obtain
AMuoJ = U 0 2k ( nA \ . ( :t ) . ( n )
iaba~, cos '4b) sm by sm by, X
X e+ lUol (Z-Zo). (7-27)
At the sarne tirnc, in accordance with the equation
(7-20), the con1ponents of the field intensity are:
-r U0 2k
Exo, = .- 2 cos ( ib"
rtA ) . ( 1l )
s1n b y x
uba 01
•

X sin ( TY.) e••o... (z-z,);

HVOI = -- I U
- -· 0 2k
COS ( --
C.t)lllJbfl
nA.)
4b
.
Sin ( -n 11 ) X
b · 1
, r 01

x sin ( ~ 1/ 0 ) elm,, cz- •al ;

•

H ZO I = ,..- J U 0 2lt :t
,... tWfl
-- --·-
ab2CJ. ~
( n A)
C'OS --
4b
COS
( -n·- L/ ) X
h ..
01

X sm (
\
TYo) e• I
<0,, (Z- z,) (7·28)
Thus, just as an electric dipole, a transverse slot generates
an infinitely large nu1nber of n1odes in a \Vavcguide. The
Jowe~t n1ode is the sarne as in the case of the electric dipole
investigated above, i.e., the H 01 mode.
Note that the expressions (7-28) apply to an infinite
\vaveguide. To define' the field in a scrni-infinite wave·
guide, one rYJust take account of the mirror unage of the slot
relatively to the end wall of the \\"aveguide. This presents
no dtfficulties and we shall therefore not dwell on it.
Let us now calculate the power radiated by a half-wave
transverse slot 1n an infinite waveguide. Making use of
(l-10) and taking the double value of the ·.ntegra1 over the
cross section of the \vaveguide when z :>zo (the power radi·
ated in the z>z 0 direction equals half the overall radiated
power). we have:
Cl b

Pz = ~ ~ Ex 01 H;01 dx dy. (7-29)

.C:O 11=8

203
Substituting here the expressions · £~& 01 and H 101 fro1n
(7-28) when z > Z0 , we obtain:

wherr G1:A· 1 is the radiation conductivity of a transverse

half-wave slot, defined by fhe exprt'ssion
2
41l In).\ . 2 n )
GIA 2 = 3
.. 2
cos ( 4b ) Sin
(
--;; Yo . (7-30)
w~tubu 0 1 \ 1

Let us now investigate the excitation of a waveguide by

a haJf-\\'aVE." slot rnilled in the broad waJJ of the waveguide
(Fig. 7-5, b). Let us represent the slot as a longitudinal
n1agnetic current of density j~. The vector potential of such
a current has only the longitudinal component A~. Thus.
in that case, the equations (1-4) and (1-5) are written as:
dAM
z .
EV = d:t '

H =-1- (7 -31)
z l (l)fJ

(7-32)

It is readily seen that the boundary conditions will

be written as:

dAM
0: ==0 when y=O. y=b;
oAM
(7-33)
a: = 0 when x==O, .x==a.
Solving the problem in the san1e way as above, the
solution which satisfies the boundary conditions \\'ill be:

A~(x,
~ ~
y, Z)=~ ~
e,. £ 111
2abycos
( nn
ax )
X
n=o na=o
t

cos (W::u )5cos(~x') cos(n,mu,) e-.vzSi~ (x: y~ z')e+vz'dz'+

s z'=-"'
Ql)

+ e+vz S i': (x', y', z') e-vz' dz' ds. (7-34)

-· Z' - ·

Let a, y 0 , Z0 be the coordinates of the centre of the slot

and V 0 , the voltage in the centre of the slot. Then, the
space distribution of the current in (7-34) \Vill be written
as:
'

j~ (x', y', z')= U 0 cos k (z' - Z 0 ) lJ (x' -a) 6 (y'- y 0 ). (7-35)

Substituting (7-35) into (7-34) and limiting ourself to the
in\'cstigation of the lo\vcst mode (n =0, m= 1), we obtain
for the ( Z0 - - } ) < z < (Z 0 -! {) region:

Uu
A .Mu 1 = abiaol ( :t ) ( n )
cos \ b y cos -b y o X
~

X e- ,a,,z s COS k (z'- Z0 ) eta.,z• .:lz' +

z' =zu -Att
.ro+A/.a
~ cos k( z'- Z0 )e - ta,,z'dz' . (7-36)
z' =z •

For the z > (2 0 1- A, 4) region, \\'e have:

1

AM
zo1
= ~· cos ( !!:_
abta01 b
u) cos ( ~b y ) o
e-iUoaZ X

Zo+~t ~

x ~ cos k (z' -z,) eta.,%' dz'. (7 -37)

:' =z 0 -J~o,'t

A similar expressiOn is obtained for the z < ( z,- !)

region. Performing in (7-37) the integration indicated there,
205
we obtain:
A~ 1 =!!~~:1 cos (Tu) cos (Tu.) cos ( !%.,1) x
X e- LaoJ (Z- Zol. (7-38)
Substituting (7-38) into (7-31), we obtain the following
expressions for the field components for the region
Z > (Z0 + 'A/4):

kllo sin(~
E-*'0 1 = anta 2
01
b
. u) cos(.!!.b y )cos (a, .!=.)
4
e-tao,cz-zo);
o o1

kU sin ( !!_ y) cos (!:... y ) cos (a !:. )

2
H yot = t(I)J.L
.I 0 e-laol (Z-Zo);
an b b 4 o ot

H m= - i~11 :~~cos ( T Y) cos ( f Yo) cos ( X !%01 {-)

X e- lUo, (%- Zo). (7 -39)

Thus, a longitudinal slot likewise excites an infinitely
Jarge number of modes and the lowest mode is the Hot mode.
Note that, when the longitudinal slot is situated in the
middle of the broad wall of the waveguide (y0 =~). all the
components of the field of the H01 wave are found to equal
zero everywhere. Slots of this kind are used in the instal-
lation of waveguide measuring lines.
Now, let us determine the power radiated by thE' slot
towards the inside of the waveguide. Making use of (7-29)
and substituting there (7-39). we obtain:

Pr. =} u: Gr,A/It
where Or.A 1 is the radiation conductivity of a longi-
tudinal half-\vave slot, which is defined by the expression
2
a ~A. 2=-
4t• b2
lJ)J&an u 01
CO~
2( n
-b lJo )
•
cos • (
\
aOI
A) .
-4 (7-40)

lt is seen fron1 (7-40) that the internal radiation conduc-

tivity of the longitudinal slot equals zero when the slot is
situated in the middle of the broad wall of the waveguide
(Yo=f> and reaches its maximum when the slot is situated
at the edge of the broad wall of the waveguide (y0 =0 or
y.,==b).
206
 The problem of the excitation of a waveguide by electric
dipoles and slots in the walls of waveguides placed in an
arbitrary way, can be solved in a manner analogous to the
above. However, we shall limit ourself to the cases of the
excitation of a rectangular waveguide mentioned above and
pass on to the radiation of a circular \\'aveguide .
•
7-3. Circular Waveguide Theory
Let us exatnine a circular waveguide of infinite length
having ideally conducting walls (Fig. 7-6). Let this wave-
guide be excited by a longitudinal electric dipole [30 ].
l,et a be the internal radius of the \\'aveguide and let us
introduce the cylindrical cqordina tes r, q>, z causing the
waveguide axis to coincide
with the z-ax is of the coordi-
-·~
•
--. -
nates. " iltI ~;
!-
z
Under the influence of the Za - r -
field of the dipole, electric lr

.. c
currents are induced on the •

internal walls of the wave- tube Fig. 7-6. Excitation of a circular

by a longitudinal electric
guide. These currents have dipole.
only longitudinal components,
owing to the fact that the primary n1agnetic field, i.e .•
the rnagnetic fjeJd of the electric dipole has only transverse
components and the corresponding secondary magnetic
field, i.e., the magnetic field of the 'nduced cur.rents wi 11
also have only transverse components.
As a consequence, the vector potential of the electromag-
netic field e·xcited in the waveguide has only the Ae=Ac2
longitudinal compont~nt and, in that f=ase, the equations
(1-4) and (1-5) will have the foJJo\\'ing expressions in
cy Jindrical coordinates:

(7-41)

(7 -42)

201
 As can be seen from the expressions (7-41), the boundary
conditions relatively to the tangential components of the
electric field intensity at the waJJs of the waveguide are
satisfied if the vector potential obeys the '!ondition
A:=o when r=a. (7-43)
Thus, the given problem is reduced to the solution of the
waveguide equation (7-42) in the existence of the boundary
condition (7-43).
"fhe equation (7-42) is of the separable variables kind and
we shaJI, therefore, seek its solution as a product of three
functions, each of which depends only on one independent
variable
A~(r, cp, z)=R(r)(J)(cp)Z(z ). (7-44)
Let us imagine the function R (r) to be dependent on the
radial coordinates and prescribed in the interval O~r~a
as an ex pans ion in a series of cy IindricaJ functions:
cm

R (r)= L (a,J,. (xmr) + b,Nn (xmr)],

m=1
•
(7-45)

where n is an arbitrary index.

The Neuman function N,(xmr) tends towards infinity
when r-+0 and since the field in the waveguide must have a
finite value on its axis, we assu1ne that b,.=O. It follows fron1
the boundary condition (7-43) that, when r=a, the function
R (r) should equal zero. We shall satisfy this condition if
we assume:
(7-46)
Consequently, the quantities x,a=rn, should be the
solutions of the equation (7-46). The -index m is the
numerator of the solution of this equation.
Let us represent the function tD(cp) depending on the azi-
tnuth coordinates and prescribed in the interval o~'P~2Jt
in the form of a Fourier series expansion:
ao
<n (q>) = L (an cos nq> + b
11 sin rup), (7-47)
n =o

where n should be a whole number, since the excited field

should be sirnple at all the points of the waveguide.
208
 Let us represent the function Z(z) depending on the longi-
tudinal coordinate and prescribed in the-· infinite interval
- oo<z<oo in the form of a Fourier integral:
.,
l (z) = Sg (x) e-txz dx. (7-48)
-ao

Thus,. the solution of the equation (7-42) will be sought

in the form of the following expansion, which satisfies
the boundary conditions on the walls of the waveguide:
GD CID Clt

A:(r, cp, Z)=L £ s ra:,.(x)cosncp+h~,.(x)sinncp)x

n=om=t x=-GD ·

X Jn (x,. r) e··IXZ dx. (7-49)

Substituting (7-49) into the equation (7-42) and taking

into account that the Bessel function satisfies the equa-
tion ·
(7-50)

we obtain:
Clt

S (k•-x•-x:_>[a~,. (x) cosncp +-b:,. sin ncp]x

n=o m =1 x=-«»

(7 -51)

We have thus represented the right-hand side of the equa·

tion (7-42), i.e., the distribution function of the exciting
current in the form of an expansion in the same system of
eigenfunctions of the given boundary probJen1 as for the
expansion of the vector potential.
To deterntine the coefficients of the expansion a:m(x)
and b:m (x), let us apply to (7-51) the reverse Fourier transfor-
mation. To this end, we shaiJ muJtipJy the left.. and
right-hand sides of this expression by the complexly conju-
}J
gate eigenfunctions, i.e .• by ~~ ~:: ,.'(x.' r)e+ ix"z, and shall
integrate over the whole volume of the waveguide. Taking
t4-2122 209
into account that
2
f0 when n =I= n',
5
"'

CQS n, cp cos n<p d<p = \ ~ when n = n,;

q>=O 8 n'
2n
r
J sin n',n sin n«> dm= I 0 when n =I= n',
T " T I n when n=n';
fiJ=O \
(7-52)
a
0 when m =I= m',
5J n•(x,.r) J n' (Xm• r) r dr= a' 'a
2 1 n'(Xm• a) when m=m'; I
·
r=o
Cl)

~ e-'" 1x-x'1 dz= 2n~ (x -x '),

(7 -53)

where the integrals are taken over the volume which con-
tains the exciting currents.
Now, Jet us substitute the expressions of the coefficients
V.'<: have found (7-53) into thr ini tj;tJ ~x press ion (7-49) and
nvestigate the integral:

which is calculated by n1eans of the deductions theory

and equals:
e ± V (z- .t. )
M=n----.
y (7·54)
where
(7-55)
210
 Thus, the vector . potential sought for is defined by
the expression -

A~(r, «p, z)=Sdv £ }~ 2;:~cosn(q>-«p')x

V n=o m=t
X J ,, (xmr) J n~..!!.'' ) e ± v<z--z'> (7 -56)
,2 •
J,., (xma)

In (7 .. 56) the positive sign (+)is taken for (z-z')<O

and the negative sign(-) for (z-z')>O.
·rhe value of xm is determined frorn the expression
r,m
)( m = a ' (7 -57)
\\'here r nm are the roots of the equation (7 -46) equa I, for a
number of valuE's of n and tn, to:
ro1 =2.405, ru =3.832, r 21 ~5.135, fJt =6.379,
ro2 =5.52, rt2=7.016, r 22 ~8.417, r3 2 =9. 760.
The components of the electric and n1agnetic fields are
determined fron1 the substitution of (7-56). into (7-41 ).
The ex press ion (7 -56) shows that a longi t ud i na I electric
dipole excites in a waveguide an infinite nu1nbcr of modes.
To each pair of values of n and nz there correspond a
definite mode and a definite propagation constant defined by
(7-55). Depending on the frequency of the excited osci J.
fations and the dian1eter of the waveguide, certain modrs are
propagated along the \Vaveguide: for them, the propagation
constant (when k is a real quantity) is an in1aginary quanti-
ty~ the other rr1odes are damped (they are not propagatt-d):
for them, the propagation constant is a rcaJ quantity.
Under suitable conditions, the propagated 1nodcs arc the
lowest ones since they have the largest cri tica I wave-
length, which is determined through identifying (7-55) with
zero. Since k=¥. the expression obtained for the critical
wave-length is:
A 2na (7 -58)
crit nm = r
nm
•

In the c&se of the lowest mode (n=O, m= I), the critical

wave-length equa Is Acrtt o1 =2.62 a. For the follo\ving n1ode
(n=l, nz=l), the critical wave-length equals i.,crit,. =
= 1.64 a. Const'quently. if the wave-length of the osci 1-
tations, corresponding to the propagation in an unbounded
t4* 'Jll
 .
medium. satisfies the condition 1.64a<A<2.62 a, only one
lowest mode with the indexes n=O, m= I is propagated in
the waveguide; all the other modes are damped; furthermore,
the damping will be more pronounced for higher modes.
Note also that the waves excited by a longitudinal
electric dipole in a circular waveguide are referred to as
transverse magnetic waves T Mnm) in accordance with the fact
that Hz=O. They are also known as electric waves (Enm)
in accordance with the fact thRt E,=~:o.
The wave-length and phase velocity of a propagating mode
in a waveguide are defined from the same expressions as
for a rectangular waveguide, i.e., from the expressions
(7·14).
Without dwelling on the higher modes, let us examine in
greater detail the lowest mode £ 01 in the case of the exci-
tation of a waveguide by a dipole, on condition that this
mode is of the propagating kind. Let r 0 , z0 be the coordinates
of the dipole (Fig. 7-6). Then, in accordance with (7·56),
the vector potential wiiJ be:
lel
Ae =-z1 _Jo(x,r)Jo(XIfo)e±l«ot(z-z~) (7-59)
zoJ 2nu taoa J'•< )
0 x 1a
'
where
_
ao,-
, /
V 1
k -
(2.405 )a .
' a
The expressions for the field cornponents (7-41) will be
written as:

(7-60)

The rest of the components of the field intensity equal zero.

The expressions (7-60) show that the wave of the Eat mode,
excited by the dipole, moves away from it towards infinity
to the right and left as travelling waves. At the same time,
the radial component of the electric field intensity and the
magnetic field intensity are in phase with one another and
212
the longitu dinal component of the electr,ic field intens ity
is in temporary quadrature with them.
Further, the magnetic lines of force of the field repre-
sent circles and Jie in the transverse planes, the electric
lines of force being confined to the walls and lying in the
longitudinal planes of the waveguide.
The picture presented by the electric and rnagnetic lines
of force in the longitudinal and tran~vcrse planes of the
waveguide at a fixed moment of ti1ne are shown in
Fig. 7-7.
Let us now compute the power conveyed by the Eoa wave.
In accordance with (1-10), the expression for the conveyed
•
power 1s:
" 2n ·

Pt = S S ErH; rdrdcp.
r =o rp-.:. 0

Substi tuting' here E, and H,t frorn (7-60) and perfor1ning

the stated integration, \\'C obtain:

(7 ·61)

If the £ 01 \vave is the only propagating mode, the power

defined by (7-6!) related to the square of the effective

~-~-....... - Electric lines of fore•

2 ---
hiiiiS of ""Zilf
~ Fig. 7-7. The field in a circular waveguide for the £ 01 wave.

value of the current in the dipole represents the radiation

resistance of the dipole in the waveguide. We shall have:
1 J!
(x,ro)
(7-62)
R Ed = 2na• (!)la_, - J~ (x1a) '

218
It is clear that the radiation resistance equals zero when
the dipole Jies at the wall of the waveguide (r 0 =a) and reaches
its rnaximum when it lies on the axis of the waveguide.
Now, let us pass on to the study of the excitation of a
circular waveguide by a longitudinal slot cut in the wall
of a waveguide. In this connection, let us investigate the
excitation of a waveguide by a space distribution of a
rnagnctic current having only a longitudinal component
jM :-= i':'· Similarly to the excitation of a waveguide by a
longitudinal electric dipole, the field in the waveguide is,
in the given case, defined by the vector potential, which has
ouly a longitudinal component AM=A~.
The equations (1-4) and (I -5) wi JJ, in the given case,
be \vritten as:

(7 -63)

J
(7-64)

As can be seen from the e~prcssions (7-63), the boundary

conditions are \\Titten as:
iJAM
0 / =0 when r=a. (7-65)

We shall \Vrite the solution of the equation (7-64) in the

san1e forrn a!; (7 -49): •
CID CID CID

A:'(r, «f, z)= ~ ~ ~ [a~m(x)cosncp+b~m(x)sinmp]~

n=o tn=l x=-e»

(7-66)
Furthertnore, the boundary condition (7 -65) wi 11 be sat is-
fied if we assume that
{7-67)
214
Conseque ntly, the quantities
0
xm are defined by the exprcs-
SIOn -
,
X
r nm
= -- .- (7 -68)
m a '

where r' nm are the roots of the derivative of the Bl'ssel

function.
Substitut ing (7-66) into ('7-64) and taking account of
(7 -50), we obia in:
Ql) QQ Cl)

~ ~ ~ (k
1
- x!- x~,) (a~n (x) cos ncp + b~n(x) sin mp] X
n=o m=t '<=_CID

x ,I , (x,r ) e- 1 x z dx
= - Jl·M 0 (7-69)
i\pplying to this expression the reverse Fourier transfor-
mation, just as \Vc did it earlier, and taking account of
the ratio (7-52), \Vith the exception that in the present
case •

a
~ J n' (xmr) J n' (X111 •r) rdr = ·

(7-70)

we obtain for the coefficients of expansion , the expressions

Further, taking into account (7-54), we obtain the required

expansion of the vector potential of the longitudinal rnag-
netic currents, satisfying the boundary conditions and the
216
principle of radiation at infinity, as:
QD CD ·M
A:' (r, cp, z) = 5dV L L ;;~;X
V n=o m=l

, J (x r) J (x r') ,
X COS n (cp - cp ) n ;. \ n m e ± 'V I z- .r), (7-71)
( I- xm 1
a1
) .1! (xma)
where the propagation constant y is defined by (7-55) and
the quantity xm by (7-68).
Let the•·e be an elementary longitudinal slot in the wall
of a waveguide. Regarding this slot as a tnagnetic dipole
with a mornent 1':'1 lying at point a, q> 0 , Z0 , we obtain:

(7-72)

..
Thus, here again, an infinite nun1ber of modes is excited.
The waves are of the transverse electric kind, since Ez = 0
and are designated as TEn,.; they are also known as
n1agnet ic waves, since H 1 =I= 0 and are designated as H n•·
The first few roots of the equation (7 -67) equal

,~1
I I

r o1 = 3.38 r 11 = 1. 84 = 3.05
, , ,
fnt=7.02 f 11 =5.83 r,, = 6.71.
•
The critical wave-length is defined from (7-58). For the
H 01 wave, it equals Acrlto• = 1.64 and for the H 11 wave,
Acr,ll11 = 3.42.
·rhus, the lowest mode excited by a longitudinal slo~ is
the Hu mode; it is the lowest relatively to all the modes of
not only the transverse electric, but also the transverse
magnetic waves. .
Let the diameter of the waveguide be such that the H01
and Htt waves are of the propagating kind. Let us calculate
the electromagnetic field components for these waves.
216
 For the Hot wave. we have:
Jl\\l x J · (x r)
·aOt (•15-•o
• )•
E• -- 2na
- Z
1ia
I 0
J 0 (x,a)
I ....L.
e~ I
'
01

I /Ml X J' (X r)
H= ± z 1 0 I e±/Goa (z-ro)· .(7 -73)
,. i<tlJL 2n:a 1 J 0 (x,a) •
J /Ml x'J (x r)
ff = Z - 1 0 I e±/Uol (Z-Zo)
z HOfJ. 2na 1 ia
01 J 0 (x,a) ·

As can be seen from the expressions (7-73), the H01 wave is

characterised by the fact that the electric lines of force lying
in the transverse plane are closed on themselves. The rnag-
netic lines of force are entirely situated in the Jongi tudinal
plane. For this mode, the electric currents induced on the
walls of the waveguide are transverse and, just as the field.
do not depend on the azimuth. The picture present<~d by the
electric and n1agnetic lines of force of this wave is shown in
Fig. 7-8.
For the H u wave, we have:
/Ml

"a tau r J- ~'a• J, (x,a)

1

(7-74)

211
·rhe H u wave has a more complex picture of the field. The
electric currents corresponding to this wave, induced on
the walls of the waveguide, have longitudinal as well as
transverse components. The transverse components of the

- Electric IIIHJS of {ore.

- - - Hognellc UIIIJS tJf f~
Fig. 7-8. Distribution of the field of the f/ 01 wave in a circular
waveguide.

currt'nts are at a n1aximum in the plane of the exciting

slot (cp=q>o), and the longitudinal components are at a
tnax imun1 in the plane perpendicular to the first one (cp-cp 0 =
= 90°). The picture presented by the electric and n1agnetic
lines of force of the H 11 wave is shown in Fig. 7-9.

A"" Eleclric lines of force

2 ---Magnetic lines offorce
Fig. 7-9. Dhtribution of the field of the H 11 wave in a circular
wavPguide. •

We have investigated the excitation of a waveguide by

a longitudinal electric dipole and a longitudinal slot.
We could have, in the same way, examined the excitation
of a waveguide by dipoles lying differently, for example,
as is often met with. by an electric dipole lying in a radial
direction.
liowever, the examples we have given are sufficient to
illustrate the general picture of wave excitation and study
the lowest rnodes, which are the ones applied in practice.
218
 7-4. Brief Information Regarding Cooxial Lines
Similarly to rectangular and circular \Vaveguides,
coaxial 1ines may be excited by electric dipoles as well as
by slots. arranged in various ways. The waves excited in
the line are of various rnodes. Ho\vever, the n1ain difference
bet\veen a coaxial line and a rectangular or circular waveguide
consists in that a transverse electrotnagnetic \vav~ (TEJ\\)
may be present in the coaxial line whereas no such wave
can be present in hollow tubes.
In a transverse eJectron1agnetic v:avr, the Jongi tudinal
components of the electric and n1agnetic fields equal zero
and the phase velocity of the \\'3Ve equals the velocity of
light in the given
rnedium. This type of f I,\~~
oscillationshasnoc riti- 2Aj ------.11-+~
~f- - - ._._ - - - .J.l--IJ
caJ wave-length and ; \\ I
n1ay be propagated no _._
rnatter ho\v smaJl the FiR. 7-10. Excitatiou of c1 corJxial line.
transverse di rnensions
of the line. Usually, the TEM \\'ave is excited in a coaxial line
by a transverse electric dipole in accordance \Vith the circuit
shown in Fig. 7-10, i.e., sirnply by connecting the tcrrninals
of the ernf generator to the internal and ex tcrnal conductors
or the line. Of coarse, a dipole of this kind excites also highl~r
n1odes, but the transverse din1ensions chl)Sen for the co,ixhll
line are 1nuch srr1aller than the
operating \\'avc-len~th so that these
higher modes are quicklydampcdancl
at a cer t ain sin a IJ d is ta nee fron1 the
point where the generator is <.~on­
--Electric lines titforce nected, there remains only the T'EM
-- -Hopnetic/t?esof(orce wave. The picture of the field oi
Fig. 7-11. The field in a this \Vave is shown in Fig. 7-11.
coaxial line. The electric lines of force begin on
the internal conductor and tertni·
nate on the external conductor and ha'!e only the radial
components. The n1agnetic Iines of force form circles sur-
rounding the internal conductor and have only the azilnuth
components. The currents corresponding to this n1ode have
only longitudinal components~ furthermore, the current on
the internal conductor equals that on the external
conductor and is of opposite direction.
219
 It is easy to show that the fieJd of the transverse electro-
magnetic wave in an infinite coaxial line is described by the
•
expressions

(7-75)
where the constant C depends on the power of the source.
Indeed, in cylindrical coordinates, when E=E,and li=-=H",
Maxwell 's equations (1-1) will be written as: · .,
an,,
-
iJz
~-== iweEr•·
- iJE r =.--: iwnH (7 -76)
az r ;p'

and the substitution of (7-75) into (7-76) shows immedi-

ately that the expressions (7-75) satisfy MaxweJJ 's equations.
These expressions also satisfy the boundary conditions on
the walls of the line.
The ratio of the electric field intensity to the magnetic
field intensity defines the wave impedance of the mediun·a
which fills the line.
~=~=
H,.. k
v~-
e .
(7-77)
•

The voltage in the z-section of the line will be defined

from
Ra
V~= S E',dr=Cln~e-'lu, (7-78)
r=R 1
where R, is the radius of the internal conductor of the
line;
R•• the internal radius of the erternal \Vire (tube)
of the line,
The current on the internal conductor is defined from
k ·~cz
1.=2nrH =C2n -e-' . (7-79)
• ' Wfl
The wave impedance of the coaxial line will be defined
as the ratio of the voltage to the current of the line and
will equal:
W = , ~~ _!_ In _& • (7-80)
J' e 2n R1
220
 If the transverse dimensions of the coaxial line are not
small in comparison with the length of the excited wave,
then, apart from the TEM wave, waves of the H,, and
En, modes can be propagated in the line and, in the first
place, a transverse electric wave similar to the H02 wave
in a rectangular wavt'guide. The critical wave-length of
this mode is defined from the approxin1ate expression
(7-81)
i.e., it equals the length of the mean perhneter of the
cross section of the coaxial line. And if the wave-length
of the oscillations corr~sponding to the propagation in
an unbounded space is srnaller than .n: (Rt ·t- R2), then the
wave of the mode under consideration will be propagated
along the line and superposed on the main wave. Usually,
a superposition of this kind is undesirable. so that the
transverse dimensions of the line are chosen on the basis of
the condition
Aopcr >n(R,+R 1 ) (7-82)
•
The condition (7-82) limits the application of the coaxial
line to the three-centimetre and shorter ranges of radio \\·aves.
Coaxial cables also belong to the class of coaxial lines;
they differ from rigid coaxial lines by their tlexibiJity. The
wave used in coaxial cables is the TEM wave, described by
the ex press ions (7•75)-(7 -80).

7-5. Single-Wire Line Theory

The utilisation of single-wire transmission lines with a
surface wave is of recent date, although the possibility
of the propagation of a symmetrical surface wave along a
cylindrical conductor of finite conductivity was established
by Sommerfeld as far back as 1899(31]. The reason is
that the field of the wave investigated by Sommerfeld
has a large extension in the space surrounding the conductor
and requires therefore that a considerabl~ part of the space
surrounding the line should be free from obstacles. Further-
more, a very cumbersome exciting device is necessary to
bring about an efficient excitation of the Sommerfeld wave.
In 1950, Goubau (32) showed that if the surface of the wire
was ribbed or covered with a layer of dielectric, then even in
the case of a conductor of ideal conductivity, the field of
221
the surface wave would be excited and wouJd, in the main,
be concentrated in the vicinity of the wire; its extension in
the transverse plane could be made insignificant and the
exciting device of the acceptable dimensions.
A schematic illustration of a single-wire transmission
line is shown in Fig. 7-12. It consists of a thin circular con-
I J I
. \
/ .. \
-------..J----------
.. .
\ ... \
2 2
Fig: 7-12. Circuit of a single-wire transmission line:
/-horn; 2-coaxlal line; 3-slngle-wlre line

duct or covered with a layer of dielectric, and of coaxial

lines at the transrnitting and receivtng ends with conical
horn transitions. Owing to the presence of the dielectric
Jay er on the internal conductor, the transverse electromag-
netic wave (TEM) which is propagated in the coaxial line
reaches the horn transition and is gradually transforn1ed into a
transversen1agnetic wave,
as shown in Fig. 7-13.
On leaving the horn, the
transverse magnetic wave
of the surface rnode E0
is propagated a long the
single-wire line. On
reach 1ng the horn tran-
Fi~. 7-13. Picture of the electric lines sitionat thereceivingend,
df force in a horn transition. Jhi.<\ 'illrface wave is
gradually transformed in
the horn into a transverse electromagnetic wave which is
further propagated in the coaxial line.· ·
Let us investigate the theory of the single-wire line. The
radius of the \\~ire without a dielectric covering is a and,
the radius of the wire with a dielectric covering a'.
The wave excited in the line has only a transverse com-
ponent of the rnagnetic field intensity H=H'P, so that in
cylindrical coordinates, Maxwell's equations (I-I) will be
writ ten as:
(7-83)

222
for the first equation and
•
(7 -84)

for the second one.

Substituting (7-83) into (7 -84), we obtain for the ntag-
netic field intensity, the folJo\\·ing \\'ave equation:

_!_ ~ r iJH, otH~ k2H - !.!.J_- 0

r or or + iJz 1 + 0 ~ r2 - t
(7 -85)

\\'here k 0 is the \vave numbcar (coefficit~nt of phase) for

free space.
It is readily seen by a direct subsUtu1 ion. ~hat the
solution of the equation (7-85) will be \Vritten as:

H ~ == A e - ;hrH ~ 2 , ( yr), (7-86)

where H~ 2 , (vr) is a Hankel function of tht' srcond orcler~

'\' = - i V/l·-k~, the propagation constant in a radtetl
cl i rect ion;
h, the propagation constant in the direction of thl!
•
z-ax JS.
Substituting (7-86) into (7-R3) and taking into account
that +f, !rli~' (·rr)] = '\'11~' (yr),
1 1
we obtain for the com·
ponents of the electric field intensity, the following cxpres-
•
SlOBS:

EZ =A .y e-•hz H<z) ("'f). (7 -87)

UH8o 0 I

In order that the field ex pressed by (7 -86) and (7 -87)

should be able to exist, the covering of the \\'ire should
possess a definite surface resistance, by which \\'e under·
stand the ratio of the tangential contpon(.~nt of the electric
field intensity to the tangential cornponent of th~ ruagnetic
field intensity on the surface of the layer. Let z. be this
surface resistance. Then
Ez when r = a ' .
Z1 =-= H (7 -88)
' 223
Substituting here (7-86) and (7-87), we obtain:
y H~1 , (va')
la== ifiJ8o H~s) (ya') • (7 -89)
We are dealing with a radius of the wire and a thickness
of the dielectric layt~r that are small in comparison with
the wave-length and we shall therefore assume I ya' I~ 1.
On that condition, the Hankel functions will approxi-
mately equal:
H~'> (va') ~-~In (o.89a'-{ ha -k: ;
ya' -+O n
2
H~ , (ya') ~-! 1
• (7 -90)
ya' -+o n a' V hs-kze
Then, the equation (7 -89) will be written as:
1 1
(h -k 0 ) a' { -{ \
ZI ~ l(J)!
In 0.89a' 1
h -k 0 J.1
(7 -91)
0

In the case of a surface reactance of the covering of the

wire, the equation (7-91) is satisfied on condition that this
reactance is of an inductive nature and if h>ko.
Note that the asyn1ptotic expression of Hankel function
is as follows:
H~•l (yr) ~ 2 e-I ( vr--!-n) y'
vr..., Cll nyr
and since, when h>ko, the propagation constant y is a
negative imaginary quantity, the field diminishes expo-
nentially in a radial direction, and the larger the quantity y,
i.e., the more lz differs from kn, th~ faster the field is damped
when the point of observation moves away fron1 the wire.
Such a wave is called a surface wave and since the phase
velocity v is smaller than that of light t, it is also called a
slow wave.
Let us also define the relative power conveyed by a
surface wave through a cylindrical surface of radius p,
surrounding the wire. In accordance with (1-10), the power
conveyed through a surface of radius oo>r >p will be
defined from the expression:

(7-92)

224
The full power P, conveyed outside the wire, will be
obtained if in the expression (7-92), the low~r limit be ex·
tended as far as the surface of the wirer== a'. 'fhen, the relative
power with which we are concerned will be defined by the
expression (1-P,IP). The substitution of the expressions
E,. and H. from (7-86) and (7-87) into (7·92) leads to the
lfXJO ..
_I
·- r-· r-

100 ~a'

-~
ZIJOR
11%
tfKJ
ID ~~ ~
~
I
4/J
~~

zo ~ ~
ID
1 !
-
""

•
2
.
--
ia'liit=!tl
'at 112 D4 0.1 ,
--

Fig. 7-14. Relative radii of a cylindri-

cal surface inside which 50, 75, 90
and 99% of the po\ver are propagated.

necessary calculation lormula which we do not give here.

The calculation based on this formula is represented in the
form of graphs in Fig. 7-14, which shows the relative radii
of the cylindrical surfaces pia' inside which 50, 75, 90 and
99% of the total power conveyed by the surface wave is
propagated, depending on the quantity a'Vh 2 -k:.
We notice that an increase of Vh'-k:, ·being prescribed,
leads to the decrease of the radius of the cylindrical surfa~
inside which a definite part of the power is conveyed. Let,
for example,
a'= lcm, h -k:=2X to-•,
1
y
tS-2122
Then, 50~o of the power is conveyed inside a cylinder of
radius p=6 ctn, 75·~~ of the power is conveyed inside a
cylinder of p= 16 cm, 90% of the power is conveyed
inside a cylinder of p=33 cm and 99~o of the power is
conveyed inside a cylinder of p=86 cm.
Apparently, the radius of the output opening of the
conical transition of a line should correspond to the radius
of a cy Jinder inside which the largest part of the power is
being convey~d. In that cnse, when the TEM wave in a
coaxial line is transformed into an Eu wave in a homo-
geneous line, the largest part of the power will be contained
in the surface wave and the sma11est part, in the higher
modes. The higher modes determine the radiation frorn the
open end of a conical horn and lead to the decrease of the
overall efficiency of the line.
Now, let us see how the surface resistance Zs depends
on the dielectric layer. The field in the layer obeys the
equations (7-83)-(7-85) when ~:: 0 is replaced in them by e
and k0 by k. In this case, the solution of the equation
(7-85) is sin1i Jar to the expression (7-86) except that instead
of the second Hankcl function, we n1ust take in it the linear
cornbination of the s(~cond and first liankeJ functions. Ho\v-
ever, owing to the fact that the radius of the wire is s:na 11
in con1parison with the wave-length and the thickness of
the layer is sn1all in comparison with the radius of the \Vire,
the ntagn~tic field intensity inside the dielectric 1ayer
(a';· -·r;_.:a) rnay be written in approxirnatcly the foiJo\ving
forn1:
H~-~e-•z
I 'h
(7-93)
' 2nr •
where I 0 is the current in the wire at the cross section z= 0.
This expression is accurate on the surface of the \Vire
when r=a. Wh~n r>a, the error wilr not be large.
Substituting the expression (7 -93) into the equation
(7 -83) and replacing e0 by c, we obtain:

E' ~
/ 0 -e
h -IhZ
(7 -94)
2nr we •

Suhstitu1ing further (7-93) and (7-94) into the equation

(7 -84), we have:
(i -95)
226
Integrating this expression from a to r, we obtain:
E lo h•-k• 1 r -ihz
t 2n icoe n a e • (7-96)
It can be seen fron1 (7-96) that, when r=a, Et=O. Thus
the expressions (7-93), (7-94) and (7-96) approxirnately
satisfy Maxwell 's equations and accurately satisfy the
boundary conditions on the surface of the conductor.
Dividing E, by Htf and assuming that r=a', we shall
find the surface resistancf of the dielectric layer
h 2 -k 2 , a'
Z5 == LW£
a In-·
a
(7 -97)

This resistance \\'ill have an inductive nature on condition

that h<k. Thus, the surface· wave is propagated with a
phase velocity deterrnined from the condition k 0 <h < k, i.e.,
the phase velocity of propagation lies within the interval
Vr w
4 .,_:_ __ < V <vr.w (7 -98)
V £/l!o
\\'hen:' t',. w is the vc]ocity of propagation of radio waves
in free space.
Regarding the expressions (7 -91) and (7 -97) as equal,
\\~e obtain:
(7 -99)

Let us
consider thin dielectric layers, for which
h 1 -k!~k~. Then, we may assun1e that h -k ~k:-k• 1 1

and \Vrite the. expression (7 -99) as:

(7 -100)
\\'here
G(a' ( 1z• -k:) =
'Jt/' 'J
= _ a' ;:-k! In ( 0.89a' }ih•- k":). (7-101)
The function G is represented in the form of a graph in
Fig. 7-15. . .
Here is an example of calculatton of a hne. Let the
radius of the line a'= 1 cm, the wave~ length Ao=50 ern and the
radius of the cylindrical surface, inside which 90~o of the
221
power is transmitted, (!=50 cm. What will be the thickness
of the dielectric layer if e/eo=4?
From Fig. 7·14 we find that for the values Q.la', which
correspond to 50 and 90% of the transmitted power, the
quantity a'Vh1-k: equals 1.21 x to-•. The quantity G
corresponding to this value is found from Fig. 7-15: it equals

f'" r" "

to_,
1
I

~
l», •
~
~~
,., I I
J Im-•
I

~ ·~
'rl- 1-
__..
-... f ltl~i.

ttr' to-z
Fig. 7-15. Curves for deternlining the 0
function in the expression (7-100).

1.68x JQ-fi. Substituting this value G into (7-100), we

find that the thickness of the dielectric layer should, in
this case, equal 5.6 x I0- 2 cm.

7-6. Other Transmission Lines

•
Apart from the transmission lines examined above, other
lines have appeared in recent years.
An1ong them we have the so-called H·like waveguide of
the open type with a surface electromagnetic wave [33).
The waveguide consists, as shown in Fig. 7-16, of the two
parallel ntetal plates of width h, separated by a dielectric
plate of thickness a. A surface electromagnetic wave, the
energy of which is concentrated in the vicinity of ,the
dielectric plate, may propagate along a waveguide of this
kind. The amplitude of the field nf this ·wave decreases '

118 -
exponentially in the direction of the x-axis of the waveguide,
and, in case the metal plates have a suffi~ient width h,
almost all the energy conveyed by the electromagnetic wave
is concentrated in the space betw~en the plates and onJy
an insignificant part of it is radiated
through the open lateral \Valls of the
waveguide.
The approximate theory of the H-like
waveguide is relatively simple to build.
Let the metal plates be ideal con- h
ductors and of infinite extension not only
in the direction of the ?-axis but also I
in the direction of the x (h= oo)-ax is.
Let the pararneters of the dielectric
plate be e, f.1 0 , a =0 and the parameters
of the surrounding medium, e0 , f.to, a=O.
Assume that the vector potential of the Fig. 7-16. I-f-shaped
electromagnetic wave has only the x-th '"aveguide.
component Ae=A~.
The \'ector potential inside the dielectric pJate obeys the
\\·ave equation ~
iJ2Ae a~Ae o=Ae
x+
---iJx~a iJy2
x+ x
cJzZ -t
kJAe 0
x= . (7-102)
\\~here
k= CJ.) ~~ EJ.1 0 ,
and the cornponents of the n1agnetic and electric field
\·ectors are defined fron1 the equations
tJAe oAe
Hx=O, f/v-:...-=~.!,
vz /fz=- iJx;
y
o' Ae' I cJ2 Ae
EX = 1
l(J)8
kJ Aex ·-t- d .x ' E =
X Y
a X
i(t)£ y iJX
, (7-103)
I a'A~
Ez = iwe. iJz iJx •
,
The electroniagnetic field outside the dielectric plate is
defined fr(Jm the sarne equations (7-102) and (7-103) by
replacing k and e by ka and so respectively.
The boundary conditions on the surface of the metal plates
consist in that the tangential components of the electric
field intensity equal zero, and on the surface of the
229
dielectric plates, in that the tangential components of the
electric and magnetic fields intensities are continuous.
Thus, the boundary conditions relatively to the field
vectors will be satisfied if the following co11ditions are
imposed on the vector potential

Acx;atr= Ae.t rllel = 0 when y= ± 2'

k . l
A~alr=A~dlel when X=± f; l
rlAt· 1 OAe (
_!_ '·
l'
JC
iJx
air =-
8
X
dx
dlel when =
X
±- ~2 . I
)
0

We arc interested in the lowest n1ode, propagated along

the axis of the waveguide and decreasing exponentially
outside the dielectric in the direction of the x-ax is. Inside
the dielectric, there will be a standing wave in the x-axis di-
rection. In they-axis direction, inside as \Veil as outside the
dielectric. there wi 11 also be a standing wave. Accordingly,
the solution of the wave equation (7-102) will be written as:

A~ dlel ==A: cos kxx cos ~ ye-'"z; I

a
cos k.-c 2 (7 -1 05)
Ae:re al r =
Ae0 (I e + t\r. cos
' b n ye -•u .
e-1~-z

These solutions will satisfy the boundary conditions (7-104)

if we in1pose on the wave nurnber kx and the attenuation
constant p the condition

(7-106)
•
On the othe! hand, frorn the substitution of (7-105) into
(7-102), we obtain the equations

(7-107)

Fron1 the equations (7-106) and (7-107) we may determine

the quantities a, p and k,., depending on the transverse
dimensions of the waveguide and the permittivity of the plate.
130
 Fig. 7-17 shows the curves of the ratio of the wave-length
In free space Ao to the wave-length A.x and the curve ol the
attenuation of P in a transverse direction. as a function of
the ratio ;- for e/e 0 =2.53.
0
~
Az

84

0 IJI 0.2 113 0.4 0 0.1 ll2 03 04 p.

Fig. 7-17. The relative wave-length Ax and attenucttion p
i n an 1-1-Ii k e wa veg ui d t'. .

Fig. 7-18 sho\\·s the relative wave-length in the wavcguide

A.w depending on the transverse dimensions of the waveguide
a and b. In the expressions cited, we u~----
have taken the following designations: Ap
2n tO ~
k 0 =~
.. '
1"1--
'-'--- fJ •

Ao "-w
Thus, in the systern under consider-
ation, there arises a surface wave, Q6
the phase velocity and wave-len~th Ill
of \\'hich rnay be detcrn1ined frotn the fl4
curves in Fig. 7-18 and the decrease in 02
a transverse d i rcct ion, frorn the curve tt
in Fig. 7-17. We see that at a certain, A;,
con1parativeJy short distance from the O fJt 0.1 0..1
dielectric plate, the field decreases to Fig. 7·18. Dependence
such an extent that one rr1ay consider- of the w'1ve-lenJ,tth in
a \\'a\' eg uid c on its
ably reduce the dirncnsions o f t he meta J transverse dimensions.
plates in the direction of the x-axis,
\\'ithout substantially changing the picture of the field in the
\\'aveguide and, consequently. without any substantial radi·
at ion of such a \\'aveguide.
231
 Fig. 7-19 shows the graphs of the dependence of the ratio
of the power propagated along a waveguide in the vicinity of
a dielectric plate within the limits of the width h to the
full power propagated along a waveguide with infinite
surfaces when e/r.0 =2.53. From these

llZ
I curves, one may determine the trans-
verse dimensions of the waveguide

(Jf
for a ratio of the powers ;ex
1
full
=0.001;
0.01 and 0.1.
Note that only transverse compo-
0 1 z J 4 nents of the electric currents arise on
Fig. 7-IY. Dependence of the m~ta] plates, since Hx=O. This
the transrnitted power on is quite a favourable circumstance
the transverse dimensions since it enables to dispense \Vith
of a waveguide. waveguide flanges when joining the
separate sections of the waveguide.
Among the relatively new transmission lines used in the
microwave range, we also find the strip waveguide, the
cross section of which is shown in Fig. 7-20.
The line consists of a main plate, which plays the role
of a screen, of a thin layer of dielectric and of a metal strip
printed onto the dielectric.
The dielectric is usuaJiy
polysterene and the strip 5
is made from copper, silver
or some other n1eta 1.
The dimension of the strip
b and the thickness of the

3
Fig. 7-20. Cross section Fig. 7-21. Dependence of the
of a strip line: wave-t~r1gth on the width
1- rnt•lal strip; 2-dlelectric; of the plate (h=0.15 cm;
.9-maln plate f=4, 700 Mc/s; polysterene
dielectric).

dielectric h are chosen much smaller than the width of the

main plate, so that the effect this plate has on the electro·
magnetic wave can be taken into account by constructing
the mirror image of the strip relativ.eiJ to the main
RII
plate. Thus, one may approximately regard the line as a
twin-plate symtnetrical feeder along wh1ch a transverse
electromagnetic wave is propagated.
The main part of the electron1agnetic energy propagated
along the line is concentrated in the dielectric b~tween the
strip and the main plate and this is all the more true as the
din1ension b of the strip is larger. That is why the wave
f.O
D.l
:k_ I

.......
.... ..
0.8 / •

Q7 ~

/
0.6 A
0.5 ~ --
I.
Ql-
! J 5 10 10 JO .f(J 70 IOIJ
Fig. 7-22. Dependence of the wave
impedance of a strip line on its transverse
dimensions.

phase velocity in the fine is close to that which we would

have if the dielectric which surrounds the line were honlo-
geneous. Fig. 7-21 shows the curve which characterises
the wave-length in the line, depending on the width of the
plate. We see that the wave-length in the line lies in the
interval A., <A,., <Ao where A.,= y"}..,/ , A.o is the wave·
e e,,
length of th~ osci I lations propagated in air.
The wave irnpedance of a strip Jine may be detern1ined
from the theoretical curve shown in Fig. 7-22 (W' -~ Jf-';
is the wa\'e impedance without account of the edge effects
in the strip line).
 CHAPTER EIGHT
Methods for Matching the Line to the Load

8-1. Twin-Wire Line Equivalent to a Waveguide

In n1any respects, a waveguide with one unique propa-
gating wave n1ay be reduced to an equivalent twin·wire line.
The question arises as to what we are to understand by the
voltage, current and wave in1pedance in this twin-wire
line.
To be explicit, let us investigate the Hot \Vave in a
rectangular waveguide. As we know, the field of this wave
in the wavcguide is defined by the expressions (7-15)-
(7-17). Passing from the waveguide with an H01 wave to
an equivalent t\\'in-wire line, the voltage in the latter may
be understood to be the integra t of the electric field intensity
Exo 1 along the x-axis. In this case, we obtain:
a P a
(n-b y ) sin In-b y ,\) x
~x=o
• 2/"xiak
V = E ~ 01 dx = . b sin t
z >d uoeu u 01 \

•
•
X sin CX. 01 d e-'"o,z.
We see that the voltage between the walls of the waveguide
depends on the transverse coordinate y and, by replacing
the waveguide by an equivaiE'nt twin-wire line, the defi-
nition of the voltage of the latter is found to be non-
unt• que.
The current in the equivalent twin-wire line should
naturally be understood to be the longitudinal cornpuuenl
of the total current flowing on one of the broad \\'ails of the
234
wavrguide. We obtain:
b ,
41~/b ( n \
I
z>d
=
5
11=0
HY01 dy = -:-- sin - y') sin a
tabn b 01
de-iarJt'
•

The voltage to current ratio \vill detern1ine the wave

impedance of the equivalenl 1\\'in-wire line, which will be
expressed as:
U 120n na . ( n ) (8-1)
W= T = Y I - ('J..t2bji 2b sm 7i y '
where we assun1rd ... ~~= 120n.
Jl E
Thus, the wave intpedance of the equivalent line is
likewise non-unique. To be rnore certain, the vol1agr of
the equivalent line may be understood as its maxin1urn
value in the wavf."'guide. Then, the \\·ave inlp('dance is defined
by the expression
J 20n ne,
w= V' I= o:i-:zbft 2b · (8-2)

In a sirni Jar way, one n1ay corn pare any n1ode in a rec-
tangular. circular or other waveguide \Vith th<' corresponding
wave in an equivalent twin-wire line. In so doing, one has
to ascribe to the equivalent twin-wire line the phase
velocity characteristic of the mode under investigation in the
\\·aveguide. As we shaJJ see below, the uncertainty in the
choice of the wave impedance of the equivalent line ruay be
removed by introducing the concept of the relative resistance
of the line. .

8-2. Transmission Lines of Finite Length

We ha\ e been considering lines of inftnitc length and
assuming that they guide travelling waves. In practice,
any Iine is finite and has some sort of a load at its end
(antenna, receiver input, absorbing resistance, etc.).
In the general case, part of the energy is absorbed in the
load and part is reflected from it. That is why there occurs
in the line a superposition of the reOected wave on the
incident one.
Making use of the concept of the equivalent twin-wire
line, one may represent any sort of load as the con1plex
236
resistance Zt, connected to the end of the line (8-1). Then,
the voltage and current at the cross section z of the line may
be represented by the expressions (we are dealing with
lossless lines):

- ulnc ,jcaz- ~ e-ira

I ,- (8-3)
w C' w '
where U1nc is the complex amplitude of the voltage of the
incidrnt wave at the end of the line;
Uren, the complex amplitude of the voltage of the
reflected wave at the end of the line;
a.=~ , the phase coefficient;
W, the wave impedance of the line.
One usually introduces the concept of the coefficient of
reflection of the wave relative to the voltage, by which
one understands the ratio of the voltage of the reflected
wave to the voltage of the incident \\'ave at the cross
section z of the line:
(8-4)

This expression shows that the coefficient modulus of

reflect ion remain s consta nt wherea s its phase change s along
the line in accordance with the linear law.
I a l •I z~·t I
I I I
I I
I I

Fig. 8-1. Twin-wire line of finite length .

•
The ratio of the voltage to the current at the cross section
of the line determ ines the equiva lent resista nce of the line,
expressed as:
Z.-wl+
-
p,
I -p. . (8-5)

Thus, the equivalent resistance of the line depends only

on the wave impedance of the line and on the coefficient
of re flection. Frorn (8-5) we obtain the following useful
expressions enabling us to determine· the. coefficient of
• ·,

23b
reflection from the wave impedance and equivalent resistance
of the line or load resistance: ,
Zz-W ZL-W 86
P.=Z1 +W' Po=zL+W' (-)
By analogy with (8-5), the Pquivalent resistance of the
line at the cross section l-f- z is defined t y the expression
zz+l
= W 1 +
Pz+t = W + Pze
1- Pz+t
1 -iau
I - Pze-u•l
•

Substituting here the value Pz fro1n (8-6), \'le have:

~ + i tan a.l
ZZ+l=W z . (8-7)
1 +i Wtan al •

This i~portant expression relates the equivalent

resistance at the cross section z of the line to the equivalent
resistance at the cross section z+l. This expression leads
to the follo\ving particular expressions:
a) when al= 180°, Zz+'A.t,=Z,, i.e., the equivalent
resistances at cross sections of the line spaced half a wave
apart are equal to one another;
b) when al=90°, Zz+A 1, ·Z~=-= W't, i.e., the product of the
equivalent resistances at cross sections of the line spaced
a quarter of a wave-length a part is equal to the square of
the wave impedance of the line;
c) when Z = W, Zz+t= W, i.e., when the load of the line
is equal to the wave impedance, the equivalent resistance
of the line at· any cross section equals the wave impedance,
d) when Z_.=O, Zz+ 1=iW tan I, i.e., when the line is short-
circuited, its equivalent resistance becomes reactive and
changes along the line in accordance with the tangential Jaw;
e) when Zz= oo, Zz+ 1=-iWcot al, i.e., in the case of an
open-end line, its equivalent resistance likewise becomes
reactive and changes along the line in accordance with the
cotangential law. ~
Apart from the coefficient of reflection, the concept of the
travelling-wave ratio is also introduced, under which the
ratio of the voltage at the node to the voltage at the antinode
of the line is understood
(8-8)

287
 The voltages at the node and antinode are defined by the
expressions
umln =I V aei'J.Z 1-1 V oe-laz f;
V ma.x =I uaeia.z I + I uoe-iaz t,
and the travelling-wave ratio is therefore expressed through
the coefficient of reflection by means of the expression
K_ I -I Po I (8-9)
- 1 +I Po I '
where p 0 is defined from (8-6).
If the load resistance is purely active (XL = 0), the
expression (8-9) becomes:
RL-W
l- -·----:':':':
RL t· w
K= RL -W ,
l + RL +w
and the travelJing-wave ratio is then defined by the sitnple
•
expressions
K=; L
when RL>W; K=RL when RL<W . (8-10)
W
Owing to the fact that, in the first case, a voltage antinode
is establi shed at the end of the line, RL represe nts the
resistance of the line at the anti-
I node Ra . In the second casf,
I a voltage node is at the end of
'I I
-,..----!,~---'
~
the line and, then, RL represe
the resista nce at the node Rn.
nts
I • In • 1 The expression (8-8) is used
Fig. 8-2. Explaining the cal- to determ ine the travelling·\\'ave
culation of the line equiva- ratio experimentally, when the
lent resistance. voltage at the node and antinode
of the Iine is rneasured with the
help of a measuring line. The expressions (8-9) and (8-10)
are used when the load resistance is known.
Note that instead of the concept of the traveJJing-\vave
ratio, the concept of the standing-wave ratio is often used
by which we understand a value which is the reverse of the
travelling-wave ratio.
It is clear that when in (8-7), the cross sect1on z coincides
with th~ voltage antinode {Fig. 8-2), thr resistance Z, is
288 ..
purely active. Then K= !a and (8-7) is written as:
I +iK tan altt
Zl .J = W K + i tan- a/ a-, (8-11)

where la is the distance between the voltage ant inode

towards the generator and the cross section in which the
equivalent resistance is defined.
Note that the uncertainty in the expressions (8-5)-
(8-11) with regard to the choice uf a wave hnpcdance for
the twin-\vire line equivalent to the \vaveguide 1nay be
removed if we n1ake u-;e of the concept of the relative
equivalent resistance of the line Z~ ==Zz/W.. 'fhcn, the
expressions (8-5), (8-7) and (8-11) becon1c:
,
,_ 1 + pz • , Z. ·-+· t fan ul
Z Zz+l = · ----;- ;
z- 1 - pz ' J ·+ i Z 1
t a 11 ul
(8-12)
, l+iK tan ala
Zla = K-t- ita n u/--
a
"

Consequently, in order to deterrnine the relative equivalent

resistances, the phase coefficient of the Jine, the traveJI in~­
wave ratio and the position of the voltage antinode
(or node) on the Jine have to be known. l"he equations
(8-12) are founrl to be valid for any transtnissiQn Iine with
any tnode.
When ca]culating the input resistance of a line, when
re-evaluating the eqtuvalent resistance frorn one cross
section of the hne to another, when defining the resistance of
the load of the line at prescribed travelling-wave ratio
and position of the voltage antinode (node) in the line, as
well as in a number of other cases, it is quite convenient to
make use of the so-caiJed circular diagrams of impedances
(34 ].
The use of these diagranls reduces the calculations to a
minimum and saves n1uch time, whi 1st the accuracy is quite
satisfactory for engineering purposes. .

8-3. Narrow-Band Matching of the Line to the Load

The transmission lines connecting a transmitter with
an antenna or an antenna with a receiver are usually tuned
on a travelling wave.
239
 The traveJJing-wave type of operation makes it possible
to decrease the relative power oi the losses and, consequent-
ly, to raise the efficiency of the line. This type of operation
reduces the voltages in the line, which eliminates the
possi bi li ty of a breakdown at high transmitted powers. This
type of operation is further characterised by the absence of
oscillating power in the line, which leads to a widening of
the pass-band of the line. Finally, the load on the generator
rernains constant regardless of the length of the line.
As a rule, the antenna input resistance differs fr<1m· the
wave impedance of the line, so· that there is always an
electromagnetic wave reflected from the load. l'o eliminate
this reflected wave, some sort of reactive element absorbing
no energy is introduced
into the line. Then, in ad-
- dition to the wave reflected
from the antenna, there
K=f l(<f appears yet another wave,
_a_ _ _ _ ____. ~ reflected from the rPactive
element. The proper choice
of the size and position
of the reactive element
n1akes it possible to level
Fig. 8-3. Tuning the line on a
these two \vaves in ampli-
travelling wave by means of an tude and make them to be
inductive stub. opposed in phase. They
will then cancel out and
there will only be the travelling wave propagated from
the generator towards the load in tne region between the
connection point of the reactive element and the generator.
In the region of the line between the load and the con-
nection point of the reactive element, the type of operation
of the line will remain the same all along i.ts length.
V. Tatarinov has shown (2] that to tune the line on
a travelling wave, a reactive shunt has to be inserted into
it at the cross section where the active part of the equiva-
lent conductivity of the line equals its wave conductivity.
At the same time, the conductivity of the inserted
shunt should equal the reactive part of tht. equivalent con-
ductivity of the line at this section and be of opposite sign.
Let the reactive shunt be a close-ended twin-wire stub
connected to the line in parallel (Fig. 8-3). Let the wave
impedance of the stub equal the wave imp~dance of the line.
240 -
Its resistance will then be: I

Zstub= iWtan alstub· (8 .. 13)

This resistance is connected in parallel with the resistance
of the line at the cross sect ion aa defined by (8-11 ). Let
the total conductivity at the cross section cuz or the line
equal its wave conductivity
+
1 K i tan 111 8 1 J
+
W 1 iK tan ala + iW tan a/stub == W ·
The real and imaginary parts of the left-hand side being
equal to the corresponding ones of the right-hand side. we
obtain the following calculation forntulas:
- JlJ?
cot a.la =V K, tan a./•tub= l-K . (8-14)
Thus, to determine the length of the stub and the place
of tts connection into the line, the travelling-wave ratio
and the position of the antinode in the line has to be known.
These values are usually detern1ined experimentally.
lt should be noted that when K-• o, the length of the
stub /stub and la ~'Aw/4, i.e ... the stub is very short and is
connected close to the voltage node betwe~n this node and
the generator. When K~l. the length of thestub/stub-+Aw/4
and la--+Aw/8, i.e., in that limit case, the length of the
stub is close to a quarter of a wave-length and it is con-
nected close to the points situated at a distance 'Aw/8 fron1
the antinode, towards the generator. Thus, the inductive
stub is always connected in the region situated at a distance
of an eighth of a \Vave-Iength frorr1 the voltage node towards
the load.
Note also that, apart iron1 an inductive stub, a capacitive
stub is also used although ntore seldorn, it is connected
into the line in the region comprised between the voltage
anti node and the voltage node, towards the load, i.e., where
the equivalent resistance of the line is of an inductive
nature.
It is evident that. instead of one, two or three reactive
elements, lying at a certain distance from one another can
be inserted into the line. The proper selection of the
impedances of these elements will make it possible to obtain
a travelling-wave type of operation; furthermore, these
reactive elements may be connected into the line not only
in paralleJ but also in series.
16-2122 241
 l~he reactive elements may also consist of lumped in-
ductances or capacitances or cotnbinations of both, connected
into the line in series as well as in parallel. Elements of this
sort are utilised in the medium and long wave ranges and
are investigatrd in Chapter T\velve.
In wave~uides, reactive shunts are made in the shape of
stubs and diaphragms; in a coaxial line, the reactive shunt
is rnade in the shape of a segment of coaxial line. We shall
dwell on this in greatt!r detail in Chapter Nine. At this stage
\ve shall only poiut out that the method for obtaining the
travelling-wave type of
____ .....,./ operation cited above re-
ntains valid for any trans-
rnission lines.
__w___IJ_. ~M o W 2l Note also that it is quite
convenient to calculate the
Fig. 8-4. Tuning a line on a
1ravelling wave by means of a parameters of matching
quarter-wave-length insert. stubs by means of the
diagram of impedances
,. already mentioned.
Another method for tuning on a travelling wave
consists in the insertion into the line of a segment of
quarter-wave-length line with another wave impedance, as
shown in Fig. 8-4. The quarter-wave-length line is inserted
between the voltage node and antinode.
1..-et the wave ilnpedance of the line which is being tuned
equal W and the wave irnpedance of the quarter-wave-length
insert equal Wtnsert· Then, if the insrrt is connecten in the
rt~gion cornprised between the voltage antinode and the
voltage node, towards the generator, the equivalent resistance
of the line at the J,JOiut~ uu \\lilJ IJe:

and the equi\'a lent resistance at the points bb of the line

wi 11 be:
__ W'2in,~rt _
lf'2insert t."
Rbb- R - w l\o
aa

And if the wave in1pedance of the insert is chosen so that

the line equivalent resistance at the points bb be equal to
the \Vave in1pedance of the principal line Rbb= W, a travel-
ling \vave in the line will thus be set up. Consequently,
'

242
 the necessary wave impedance of the quarter-\vave . Jength
insert is defined from the expression
Wtnsert= WfJII(. (8-15)
When the quarter-wave-length insert is situated so that there
is a voltage node at the points aa, the equivalent resistance
of the line at these points equa Is:
Raa=WK.
The equivalent resistance at the points bb \Vi 11 then he
_ u·'2n c:,r r t _ w2in
1 ~C" r t
Rbb- Rac'
--
WK
-.

Since the resistance is equai to the \vave irnp('danct• nf

the principal line Rbb= \t:'. the condition of the tuning of
the line on the travel1 ing wave is obtained here too
'

W Insert= W lfK.
•
(8-16)
In the first case, the necessary wave .tmpedance of the
quarter-wave-Jength insert is found to be larger than the
wave impedance of the line; in the second case, it is sn1al Ier
than the \Yave in1pedance of the line. The choice of this
or that method of tuning is govcrued by considerations of
assembly and of eonvcniency of tuning.
Note that a purely travelling wave in the line occurs.
only in the region of the line comprised between the gener-
ator and the points bb of the insert. In the region of the
line from the Joad to the points aa of the insert, the traveJ-
Iing-wave ratic1 will remain unchanged. In the region of
the insert itself, the travelling-wave ratio will equal JIK.
The two methods for tuning the line on a traveJiing wave
cited above ensure the matching of the line to the load in a
narrow pass-band, since the tuning of the line is effected ou
a fixed frequency. A change of frequency wi IJ lead to a
change of the load resistance on the transmission line. a
change of the magnitude of the travelling-wave ratio and
of the position of the voltage node and antinode in the line.
Consequently, the necessary impedances of the tuning
elements and the points at which they are inserted into the
line will also change.
The magnitude of detuning of the line 1n ~he eve~t of a
frequency change is all the larger the greater ts the d1stance
16• 241
in wave-lengths between the load resistances and the tuning
element, because when the difference between these distances
is very large, the magnitude of the displacement of the
voltage node and antinode along the line is found to be
relatively large.
Such a narrow-band matching (with a band of the order
of 2~~) is quite sufficient for many radio engineering devices.
However, it is often necessarv that the transmission line
should be matched with the load in a wider frequency band
(of the order of 5-10%). We shall deal with this point in the
next paragraph of the present chapter.

8-4. Broad-Band Matching of the Line to the Load

As we saw above, to match the line to an arbitrary load
by means of one reactive shunt or a quarter-\vave insert is
possible only on one fixed frequency. On that frequency,
the travelling-wave ratio equals unity and when the frequency
changes, it rapidly decreases.
From a strict theoretical standpoint (35 ), it is impos-
sible to bring about a complete matching of an arbitrary
resistance with the wave impedance of a transmission line,
i.e., to tune the line on a purely travelling wave on all the
frequencies within a certain frequency range. Complete
matching can be obtained only on certain fixed frequencies
within a prescribed range, for example on the medium
frequency and the two extreme frequencies of a range. To this
effect, one should make use of several tuning elements.
However, such a n1atching is not the optimum one from
the point of view of the pass-band width. A broad-band match-
ing is obtained if no such special conditions are i1nposed,
provided we ensure that within the prescribed frequency
band, the· travelling-wave ratio does . not fall below a
definite value. Thus, in certain ~adio-engineering devices,
the travel Iing.. wave ratio in the transmission frequency
band should not be lower than 0.6; in other cases it should
not ue lower than 0.95.
In order to get a better understanding of the idea of
broad-band matching, use is n1ade of the RL .. xL diagram
which shows how the load resistance of the line RL +iXL
changes when the oscillation frequency changes.
The characteristic aspect of the RL-XL diagram is shown
in Fig. 8-5, where the values· of RL app~ar on the real axis
I

244
and the values of XL on the imaginary axis. The values of
the frequencies corresponding to the given values of the
resistance RL + iX L are marked by dots on the curve. On
this diagram, the curves of the constant travelling-wave
~atios form circles; t~eir centres .lie on the real axis and they
Intersect the real axts at the po1nts WK and W!K.
If the curve of the load resistance intersects the circumfer-
ence of the minimum permissible travelling-wave ratio

K=fl85 ' K=ll65

Fig. 8-5. R-X-diagram of the load resistance.

Kpcrm• the portion of the resistance curve which is inside

this circle represents the matching band. Let, for example,
the permissible travelling-wave ratio K~rm=0.65. Then,
the frequencies on the curve of Fig. 8-5 lying between
f. and fto wiJJ, evidently, determine the n1atching band. If,
on the other hand, we take K~rm=0.85, then, as can be seen
from Fig. 8-5, the load resistance curve lies outside this
circle and the necessary matching does not take place. Thus,
the matching band is determined, on the one hand, by the
dependence of the resistance Rt +iX L on the frequency and,
on the other hand, by the circumference of the permissible
travelling-wave ratio.
By inserting matching elements into the transmission line,
one can distort the load resistance curve and transfer
it on the R-X plane from one region to another. The
problem of broad-band matching consists precisely in the
245
transfer of the most part of the resistance curve into the circle
of the pern1issible travelling-wave ratio.
As shown in Fano's work mentioned earlier, there is a
limit maximum frequency band. determined, for a pre-
scribed change of the load resistance, by the circle of the
pernlissible travelling-wave ratio attainable when there is
an infinitely large number of matching elements. for a load
which can be reduced to the series resonance circuit from
RL, LL, CL (36 J, this limit relative frequency band is
defiued as:
( L\/c)
fo
Q= n
In I + Kt,erm
, (8-17)

I - Kpcrm
where fo is the resonance frequency of the series circuit
of the substitution of th~ Ioad;
Q , the q ua I i ty of the c ire u it of the subs t it ut ion
of the load.
A satisfactory approxirnation to the relative matching
band. defined by (8-17) is obtained by the use of one parallel
reactive elen1ent connected at the points of the series
resonance (i.e., in the voltage nodes) of the system that is
being tuned. In n1any cases, the use of a large number of
clerncnts is not indicated, because of assembly complications
and difficulties of tuning.
When a parallel matching element is being used, the load
susceptance curve G+iB is more convenient to be used than
the resistance curve. Let the load susceptance on the plane
G-B (Fig. 8-6) be represented by the curve 1 and let the
circumference of the rninirnum perntissibJe travelling-wave
ratio K,erm be aiven.
As shO\\'n by the broad-baut.l 1uatcJJing tiJcory, an approxi·
n1ation to the maximum frequency band, when one element
of matching is used, is obtained when the curve I is
rolled up into the loop 2 inside the circle of the permissible
traveiJing·wave ratio. as shown in Fig. 8-6. The apex of
the loop rests on the point G0 = w.~ R and the intersecting
ends of the loop, on the point G1 =wKtm of the real axis.
At the san1e time, W1m represents the wave impedance of the
imaginary line chosen such as to cause the maximum value of
the active admittance in the curve I to coincide \\'ith the
maxirnum value of the active admittance in the loop 2.
246
 8 •

,._K
u-,- ~117
3
'
c

Kperm

F1g. '8-ti. G-8--diagraru of t~Jl• loJd cHJHllltanc«.-.

To determine the paran1eters of the rnatching cle1nent, let

us plot the active and reactive adrnittances corresponding
to the curve I, against the frequency (Fig. 8-7). Let us 111ark
out the points Gt=G0 K2 , corresponding to the frcqucncirs
ft and {2 , on the curve of fhc active adrnittance. Evidently,
the difference bet\veen these frequencies ~fr:.=-:f:.-ft will
approximately define the matching band.
On the frequencies ftt fo and !2, the conductivity of the
compensating device shouJd have a magnitude equal to
that of the susceptance of the n1atching d~vice and be of
opposite sign:_,B., -Bo, -82 • Let us draw through the
point -80 a straight line approaching as close as possible
to the points -8 1 and -8 2 • It is readily seen that this
straight line approximates the admittance curv~ of a parallel
short-circuited stub, defined as:
1
' 2:tlstub
Bstub=-- \l'l -CO ~ - - •
" stu~ "stub
On the wave-length Af stub corresponding ·to the frequency
fo stub' the admittance o the stub should equal zero. Frorn
here we define the length of the stub as:
I = Aostt!!! = I (8-18)
stub 4 4/ 0 •

where v is the phase velocity of the wave in the stub.

247
 The incJinat ion of the straight line of the admittance of
the stub relatively to the /-axis determines the magnitude
of the wave impedance of the stub. If, for { 2 , the straight
Jine passes through the point -81 , the \Vavc im~edance of

&,8

~----~~rc ---

f,
Fig. 8·7. Curves of the active admittance and susceptance
of the load.

the stub wi 11 be defined as:

(8-19)

In case a stub \Vith a wave impedance, defined by (8-19),

is di fficuJt to realise because the magnitude of W,tub required
is too sn1all, one can use two short-circuited stubs of lengths
ls and l2, connected in para lie I to the same points of the
matching device. The sum of the lengths of the stubs should
248
obey the condit ion
1 + 1 = )..os2tub = 2/ f' (8-20)
I :1
o~:~ tub •
It \vill 1hen be possible to define the r~lation bet\\·een the
lengths I 1 and l 1 frotn the expression
B., -_ I ( eo t 2nf
- wstub 2 l
1
+ eo f ---
2n/ 1/
1
)
, (8-21)
' V tJ
where f. and 8~ are he same values as in (8-19). The
advantages of a stub of this kind consist in f ha 1 \Ve n1ay vary
at wiiJ the magni tude of the wave irnpedance of the stub
wstub•
After the wave impedance and the length of th~ coni pen-
satin~stub have been determined, one can specify tht~ rolling
up of the curve I into the loop 2 caused by the stub (F:ig.
R-6).
Since the ~wave in1pedance of the real line W di fTers fron1
that of the imaginary line wlm• we nlust also include into
the 1ine a quarter-wave insert \Vith a \Vave irnp~dancc de·
fined as
W.,== VWW 1m· (8-22)
Furthern1ore, the length of the insert shou Jd be taken equa I
to Ao/4.
As can be seen in Fig. 8-6. a stub ot this kind transfrrs
the adn1ittance loop 2 into the admittance loop 3 and, at
the san1e time, somewhat nar-
rows the n1atching band. This ~..&--i
decrease may be determined ~ U 4 /
when re·eva luating the curve 2 ~ #"

into the curve 3 with the he I p

of the impedances diagrarn. W ~
Two quarter·\vave inserts
or an exponential line can be
used to transfo rm the admit-
tance loop 2 into the loop 3.
Note that a double quarte r- Fig. 8·8. Diagram of a broad·
wave transfo rmer has a band matching.
considerably broader band
than one quarter-wave insert.
Since any nJode in the waveg_uide_ may always be_ repre·
sented by an equival~nt twin_-~t~e hne and the equtvalent
resistance of this liue 1n the vtctntty of the voltage node can
149
be repres~nted as the resistance of the series resonance
circuit, the broad-band n1atching rnethod described above
turns out to be applicable to any transmission line \Vith an
arhi trary load. The circuit of a matching devic£: of this kind
is shown in Fig. 8-8. ·
The broad-band matching can also be effected through the
series insertion of a reactive stub in the vicinity of the
\'oltage node. In this case, it is convenient to calculate the
parameters of the n1atching dcvic~ by means of the resistance
diagram R-X.
Note in addition that the broad-band matching device
should be placed as c1ose as possible to the end of the line,
i.e.. in the first voltage node (or antinode) frorn the load,
because at these places, the degree of change of the equivalent
r~sistance of the line with the frequency is at a mininli.trri and
the rnatching will therefore have a broader band.
r:inaJiy, note that in the case of a broad-band matching,
the dependency curve of the travel Iing-wave ratio on the
frequency has two tnaxirntuns \vhereas in the case of a
narrow-band rnatching, this dependency curve, as \Vas
pointed out l'arlier, has one maxirnuJn.

8-5. Matching the Line to the Load

in a Broad Frequency Range
We have investigated ru~thods of narrow- and broad-hand
n1atching or a lint"' !~>a load, which consist in the insertion
into the line of reactive ciPrnents to cornpensate the retlect·ions
fron1 the load. l'hesc n1ethods are used wht!n the load
represents a narrow-hand resonance systen1.
In so1ne cases, the load lt!~i~lanc~ d~pend~ but little
on the frequency. Thus, for example, the input resistance
of a short-wave travell ing-\vave antenna in .a broad frequency
range ren1ain~ alrnost constant. In cases of this kind, the
matching of a constant resistance (or a resistance which
changes only slightly with the frequency) to the \vave
irnpedance of the transn1ission line is obtained by means of
so-caJJed diverging lines. There are several types of di-
verging lines. We shall investigate the exponential line (37)
represented in Fig. 8-9.
The equations for a long line are written as
dU dl
dx =-IZ,, ii=- UY,, (8-23)

250
\\'here Z1 and }' 1 are the resist anee and conducti\'ity ot
the 1ine per unit length.
Excluding the current frorn the first equation (8-2:i),
we obtain the follo\ving equa1ions for the voltage in the line:
dzU-1
d In~ -~'l_- l }" U = 0 (8-24)
d.x dx dx 1 ' ·

Let the pararneters of the line change along the lt.~ngthas

z
Z1 == 0 eb" }'" 1 = }'" e-bx
t C) t
(8-25)
where b is a certain positi vc quantity.
Let Co and L0 be the linear capacity and inductance
at the beginning of the line and let us assume that the
line is free from losses. Then ...... ,.,_____ '
Z o = i wL o, Y o = i wC o, ,. z • f
and \\'C obtain for the \\'ave im-
pedance of the exponential line
the expression
w
w= t~= wi"', 1
(8-26) Fig. 8 . 9. Exponential line.

and for the \\'ave nunllJer, the expression

k~ =-=- zly 1 = - ZOY o=WZLOCO.
At the san1e tinu~. the differential equation for the volt~tge
(8-24) will be written as
dzU
1
rE!_ +·k 2U=0.
-b dx (8-27)
dx o

An hnmediate substitution will readily show that for a

line of infinite length. the solution of the equation (8-27)
is written as:
(t -ta )~
U=Ae , (8-28)
where
(8-29)

We see that, as the wave moves, the amplitude of the voltage

in the exponential line increases in accordance with the
e~ .. exponential law. Further, from (8-29) we see that there
261
is a critical wave-length which equals:
" 4n (8-30)
"'cru=-;; ·
For the oscillations to be propagated In the Hne. the
condition A,<~ should be conformed with, i.e., the wave-
length in free space should be smaller than critical \\'ave-
length. .
The phase velocity of the wave is greater than that of
light and is defined as:
VI

The wave of the current in the line is determined by substi-

tuting (8·28) into the first ex press ion (8-23). Taking (8-25)
into account, we obtain:
1=- z. -2 -ia.
1 ( b )
Ae
- ( !!_
1 + ia) 1t
• (8-32)

Thus, as the wave moves in the opposite direction to that

of the voltage, the an1plitude of the current decreases in
accordance with the exponential law.
The equivalent resistance of the line in the x-section is
defined by the ratio of (8 .. 28) to (8-32)
loeb~ U7
l -- - -- ---r==::::::;:==:::; (8-33)
~- ({-1a)-.~.+Ji1-( 2!J•
and in accordance with the transforrnation of the current,
the voltage is also transfuruaeu along the line in accordance
\Vith the exponential law. It is complex and for this reason,
in order to eliminate the reflection in a line of finite length,
the load resfstance should contain a certain reactive
cotn ponen t.
The quantity b Is usually chosen so small that 2 ~ 1.
0
!
Then. tn accordance with (8·33), the equivalent resistance
of the line is close in magnitude to the wave impedance
z"'~ Wand the line may serve as a transformer of active
resistances.
Let the load resistance of the exponential line equal
the wave impedance ZL = W. Then, after substituting this
expression and (8-33) into (8·6), when ! ~I. \\'e obtain
2 0
approximately:
IPI= I
Z-ZL
Z+ZL
I~4k0 •
b (8-34)

lt follows from (8-34) that the reflection coefficient p will

be all the smaller as the quantity b is smaller. This is at·
tained by way of a continuous change of the distance between
the wir~s of the line.
In accordance with (8-26), the length of the exponential
line necessary to match the active load resistance Zt = W
to the wave impedance of a homogeneous line Wo is defined
as:
I W
l=b'lnw . (8 35)
•
Thus, having prescribed the permissible rnagnitude of
the reflection coefficient, \\'e determine from (8-34) the
magnitude b of the longest desirable wave. Then, by
substituting the value b thus obtained into.(B-35), we find
the necessary length of the exponential Jine l.
Let, for example, W=600 ohms, Wo=3Cl0 ohms, p=0.05
and A=50 m. Then we shall obtain b~0.0252 and 1~27.5 rn.
It is clear that the length of the exponentia' line should
be of the order of half a wave-length on the longest
desirable wave. On all shorter waves, the reflection coefficient
will be smaller than p=0.05.
Thus, the exponential line turns out to be a convenient
device for matching active resistances in a broad frequency
range.
 CHAPTE~ NINE
Parameters and Elements of Transmission Lines

9-1. Two-Wire and Four-Wire Lines

We shall begin our study of transrnission line devices
by the simplest ones, viz., the aerial two-wire and four-
wire lines, usually called feeders. They are open lines and,
as indicated earlier, are used on short and partly on ultra-
short (metre) waves, in the main for antennas uti Iised on
ntajor radio communication lines.
An aerial twin feeder (Fig. 7-1, a) is usually n1ade of
copper or bi-n1etal wires (3 to 6 mm in dia), spaced from
5 to 40 cm apart (D). The wave impedance of a feeder of
this kind is calculated frorn the expression

D+
W = 276 log [ d Yr (~.)' + 1J, (9-1)

which, \V hen D ~ d is reduced to the simpler expression

W = 276 log ~. (9-2)

On short waves, twin feeders are, as a rule, used with a

wave impedance W -=600 ohms. The feeder is fixed on wooden
supports 3 to 5 m high by rneans of ceramic insulators of the
bar or stick type as shown in Fig. 9-1.
As a rule, aerial four-wire feeders are used for feeding
antennas from a powerful transmitter. The wave impedance
254
of a feeder of this kind rnav• be defined as:
__ ,(

2D D'
W= 138 log d
• I --i ,
+ D"l. (9-3)

provided the .wires lyin~ in one vertical plane (Fig. 9-2)

have a potential of one stgn and the \\'ires lying in another
vertical plane have a
putrnt ia I of the other
•
s1gn.

~-o, -----~
Fi(Z. 9-1. Fix()tion of twin feeder. FtJ.!. 9-2. four-wire
feeder.

In the case of a four·wire crossed ft~ccter (wires lying

crosswise have a potential of the sarne sign) the wave
in1pedance is defined as
'l.V
W==- 1381og---~· --. (9-4)
dl 1
+!:!_!
ot 2

lJsuaJiy, the fo!.lr-wire crossed feeder utilised for transmit-

ting antennas .has a wave impedance w~aoo ohn1s. lt is
n1ade of a 5 rnm wires spaced 30 to 40 ern apart.
Four-\\·ire crossed feeders used for receiving antl~nnas of
the professional type are made of 1.5 mm copper or bronze
wires spaced at a distance Ds=D 2 =3.3 ern apart. The
\\'ave impedance of a feeder of this kind equals 200 ohrns.
The feeders are tuned on a traveJ J ing wave by rncans of
inductive stubs inserted in parallel.
All the lines investigated earlier were without losses.
Under real conditions, part of the energy propagated along
the line is lost on heating the wires and insulators. As a
consequence, the field amplitude oi the propagated wave is
attenuated as the exponent e-~z. where ~ is the attenuation
constant.
25S
 We are now confronted with the problem of defining the
attenuation constant of air feeders. Let P0 be the power
conveyed by the eJectromagnetic wave at a certain cross
section of the line. Then. at the distance z from this section
in the direction of propagation of the wave. the transmitted
power wi 11 be:
p ,-- p 0 e-•"'
and the power of the losses per unit length of the feeder
u· i I I be dt.' fined as:
dP, 2AP
dz = - t' ,.

The attenuation constant will therefore be defined as:

A I
dPz (g S)
t'=-2P, dz · -

In real transmission lines, the energy losses are small,

so that the expression (9-5) is used for the approximate
calculation of the attenuation in all cases. This approxi-
rnation consists in that when defining the conveyed power
and the power of the losses, we shall assume that the field
is unperturbed, i.e .. that it does not differ from the field
in loss less lines.
To begin with. let us find the attenuation of a twin air
feeder. The energy losses on heating the insulators in an air
feeder are sn1all in cotnparison with the losses on heating
the wires and will therefore be neglected. Let R, be the
resistance of one of the wires of the feeder per J tnctre of
length. Then, the power or the losses equals d~~ =-21! R..
and the power conveyed along the feeder will be P.=I!W.
Frorn this we obtain:
•
(9-6)

We know that the resistance of a copper wire to high-

frequency currents is defined as:
I
R1 =, ·V'"t.83llohms/m], (9·7)

where r is the radius of the wire in mmi

A, the wave-length in m.
256
 If we again designate by R, the r~sistance <'f one wire
per I m of length, the attenuation constant of' a four-wire
feeder (neglecting the losses in the insula tors) is defined as:

~ == 2~ . (9-8)

Owing to the fact that the wavf! in1pedance of a four-wire

feeder is about half that of the wave impedance of a twin
feeder, the damping coeffic&ent of these feeders is fouud to
be of the same order.
As for the efficiency of the feeder. as was pointed out
earlier, it is found to be at a maxin1un1 when the travelling-
wave ratio equals unity. Indeed, in the case of a travelling
wave, the power at the beginning of the feeder equals
2
P 1 =I!W. and the power at the end of it equals P 2 =-= l!e- ''l W
where I o is the magnitude of the curren-t at the beginning of
the feeder. Hence, for the traveiling-wave type of operation,
the efficiency of the feeder wi 11 be:
11=e-•~ 1 • (9-9)
.
Now, let us suppose that the travelling-wave ratio 1n
the feeder does not equal unity. Then, representing the
type of operation in the line as the superposition of the
reflected \vave on the incident one, \VC shall \Vrite the powl'r
for the incident \VC.tVe as P 1 i = 1:w
at the beginning of the
line and as P, 1= I:e-·~'W at the end of the line~ for the
reflected wave, we shall \\-rite the power at the beginning
1 1
of the line as P, 0 =l:e-·~ Wp and the power at the end
of the linP. as P" = l,e-'f.lW p•.
Hence, the expression for the efficiency will be:
(9-10)

where p is the reflection coefficient..

Fig. 9-3 i 11 ustra tes the dependence of the feeder efficiency
on the travelling-wave ratiowhenPl=0.06. When the trave.f-
Jing-wave ratios are relatively high (K>0.3), the efficiency
depends but little on the magnitude of the travelling·
wave ratio. At smaJJ values of the travelling-wave
ratios (K <0.2), this dependence becoanes pronounced. the
'efficiency falls rapidly as the travelling-wave ratio decreases.
'The transmission line load should be such as to make
t7-2122 261
the travelling-wave ratio higher than 10% even prior to
tuning the feeder. Otherwise, a large part of the energy
might be wasted on heating the wires in the region of the
feeder between the load and the tuning elem~nts, where
the reflection from the load is not compensated for.

tor
Qd

0.6
Q4
02
K
o o.t a2 oJ 04 a5 0.6 0.1 o.a a1 to
Fig. 9~3.Dependence of the efficiency on the
value of the travclling~\vave ratio.

Furthermore, at Jow travelling-wave ratios, there is an

increase of the voltages in the line. The effective voltage
iu the antinode is defined
-.
as:
,/PW (9-11)
Ua= V K •
-
and at low travelling-wave ratios, this voltage may exceed
the permissible magnitude.
We know that the existence of high voltages in feeders of
short-wave antennas may· give rise to overvoJtages, attended
by the appearance of .zlow discharges. As was shown by
M. S. Neuman [38 ], in the case of thin wires (d~2 rnm),
the cri t ica I voltage in the short-wave range, i.e., the voltage
for which we still have a glow discharge·ntay be approxi-
mately defined as:
Uc:ru= Vl9 +0.029A.',
where A is the wave-length, in m;
U cr\t• kW.
For thicker wires, the critical voltage exceeds this value.
Energy losses in air feeders are also caused by radiation
(the antenna efiect of feeders). This radiation is all the
greater as the distance between the wires is greater in com-
parison with the wave-length. The radiation of feeders,
RSB
 at a prescribed spacing of the wires, also increases should
there occur any dissymmetry in the feeder, leading to the
appearance, apart from the usual antiphase wave. of an
additional co-phasaJ wave (at every given moment, the
currents in the wires are directed in the same direction).
This dissymmetry arises because of the differences between
the ea paci tances of the wires relatively to the surrounding
objects· and the earth, the differences bet,veen the lengths
of the .wires of the feeder and the differences between the
loads on the wires at the end of the
feeder, etc.
On short waves, the distance between
the wires of the feeder is sufficiently
small in comparison with. the wave-
length so that, provided the symmetry in
the feeders is observed, the losses on a
radiation may be neglected. On uJtrashort Fig. 9_4_ Twin
waves {metre and decimetre waves), screened feeder.
the spacing between the wires is not
so small in comparison with the wave-
length. In this wave range, the electric dissytnmetry of the
wires exerts a greater influence, so that the radiation of the
feeder increases considerably.
To reduce the radiation of feeders, as well as to make
them stand weathering, twin feeders with a relatively sntall
spacing between the wires are sometimes covered with a
dielectric. In other cases, the twin feeders are screened
(Fig. 9-4). Screened feeders are free from radiation and pro-
tected against weathering. However, due to the introduction
of the screell' and especially of the dielectric, there is an
increase of the damping in feeders of this kind.
The wave impedance of a screened twin feeder with a
solid dielectric filling may be defined as:

The phase velocity is lower than the velocity of light In

free space in lines of this kind and is defined as:

169
 9-2. Coaxial Lines
In the metre, decimetre and partly centimetre wave
ranges, extensive use is made of coaxial lines, which take the
form of rigid lines or of flexible coaxial cables. These lines
are also utilised in the short- and medium-wave ranges.
Rigid coaxial lines are usually made of brass or copper
tubes (often silver-plated) and the jnner wire is fixed by
1neans of either dielectric washers or metal insulators.
'fhe washers are ~paced sufficiently wide apart and ·are
made of a dielectric with small losses. The only drawback
of such a line is that the
4
conductivity of the washers
depends on the moisture
I C
IJ To ltx!tl content of air, which may
~
'~
11

lead to extensive damping.

I :I
Lines with metaJ insu-
~ lators (Fig. 9-5) are free
from dielectric losses but
are utilised in a compara-
Fig. 9-5. Rigid coaxial line with tively narrow frequency
metal insulators. band, due to the fact
· that metal insulators are
resonant. Lines of this kind are generally used in the centi·
metre wave range, where constructionally the line is compact.
To broaden the pass-band, the metal insulators are usually
of the broad-band kind. This is achieved by the broad-band
n1atching method described above. To begin \Vith, an in-
sert ab of length Ao/4(Al is the mean wave-length of the range)
is connected into the ine to lower the wave impedance of
the line and set up a voltag~ nude at the poiut b of the line.
Then, a stub (a metal insulator, in fact) of length Ao/4 is
connected to the point b of the line ta compensate the
reactance in the frequency band at that point of the line.
Finally, yet another quarter-wave insert be is connected
into the line to transform the resistance compensated at
the point b of the line into a resistance at the point c of the
line close to the wave impedance of the line in the frequency
band.
The attenuation in the coaxial line (neglecting the losses
in the insulation) may be defined as:
A_R.+R. I

(9-12)
.. - 21f t
260
where R, is the resistance of the inner wire per unit length;
R., the resistance of the outer wire per unit length.
For copper wires, the resistances R, and R. rnay he
calculated f~om the expression (9·7).
The wave impedance. of a coaxial line is calculated
from the expression
138 D
W= y log 7 , (9-13)
.. 8/80

where D is the diameter of the inner. ..surface or the outer

conductor;
d, the diameter of the outer surface of the inner
conductor.
The choice of the transverse
.. dimensions of the coaxial
line is based on two requirements: minimum attenuation
and maximum admis-
sion of power by the , J
line, on condition that
higher modes are D
absent. For an air
dielectric, the first re-
quirement leads to a 2
wave impedance of the Fig. 9-6. Flexible coaxial cable:
line equal to92.6ohms 1-Jnternal wire: 2-external wire: 8-poly..
ethylene; 4-braldlng.
and the second re-
quirement, to a wave
impedance equal to 44.5 ohms. Usually, the wave impedance of
the coaxial lines equals 75 or 50 ohms. The effective value
of the voltage set up in the coaxial line may be determined
from (9-11 ). As for the b~eakdown voJ tage in the Jine, it
may be defined as: ·
D/1
d. d D
(9-14)
Ub,= S
d/1
Ebrijidr= Ebr'7fln d'

where Eb, is the electric field breakdown voltage of the inner

conductor of the line.
For air, the amplitude of the break11own voltage of the
electric field is usually taken to be Eb .m x==30,000 V/cm.
Flexible coaxial cables usually consist of a flexible inner
conductor, an entire dielectric (of polyethylene), a metal
braiding which plays the role of the outer conductor and of
a protective·c:over (Fig. 9·6).
261
 In a coaxial cable, losses in the insulator are added to
those in the conductors: rnoreover, the former predominate
and, on the whole, determine the attenuation. The attenu-
ation in the cable increases rapidly as the frequency
increases, so that flexible cables are used as transmission
lines only on waves longer than 10 cm. .
When high powers are being transmitted at high frequen-
cies the dielectric gets very warm and there is a risk of break-
downs caused by overheating of the cable. That is why the
operating voltagein coaxial cablesseldomexce eds3to5.kV.
The wave impedance of coaxial flexible cables may be
determined tram (9-13). For standard cables, it usually equals
50 or 75 ohms. The phase velocity of propagation of the wave
in the cable is inferior to the velocity of light and is defined
as:
V- Vt
- Vs/80 •

Table 9-1 gives the data of a number of coaxial cables

manufactured in this country.
Table 9-1
Wave
Diameter Per· I m- Attenuation. db/m at frequency Oper·
- 1. Mc{s
Type
rnm
D, I d,
mm
m it- pedance
tlvlty w.
e/eo ohms 10 100
I I
aoo 1. ooof3. ooo I
atlng
volt-
age, kV

PK-3 9 0 1. 37 2.27 75 0.017 0.070 0.13 0.26 0.52 5.5

Pl\-6 9.2 2.55 2 19 52 0.0165 0.052 0.12 0.27 0.59 4.5
PR-19 2.4 0.68 2. 11 52 0.0565 0.20 0.32 0.60 I . 1 I
•

PI\ -20 7.2 l . 11 2.23 76 10. O!l!.l 0.087 0.15 0.33 0.68 1

Rigid coaxial lines can be tuned on a travelling wave by

means of either a quarter-wave insert or an inductive stub.
A convenient shape for the quarter-wave insert is that of
a tu be enclosing the inner conductor of the Iine. It is
inconvenient to move the inductive stub along the line, so
that one uses a two-stub transformer (see Fig. 9-7). By
changing the length lstubt of the first stub, one can change
the position of the voltage node on the line and the length
of the stub is chosen in such a way that at the connection
points of the second stub, the active part. of the line
262
equival~nt cond~ctivity should equal the re\'erse magnitude
of the ltne wave Impedance. Then, through, the proper choice
of the length lstubJ of the second stub of the transformer
the reactive part of '
the conductivity of ,. f-t •1
the line is compen- - liJiood
sated at these points
and the Ii ne is there- ,1'11161
by tuned on a travel- J ~~
V///~
1ing \\,.ave. The stubs _t_
V/~I/
3
are usuaiJy spaced 8 A, ?

apart. To tune the Fig. 9-7. Two-stub inuuct•ve tr.:Jnsforrncr.

Jine on a travelling
\V a ve, one a Iso occasio-
, . l - •

na Jly uses a three-stub

transformer. Kt
Wi
When tuping a co- ~
axial cab1c on a tra- ~
velling wave, a single \

inductive stub can be

used along with two- Fig. 9-8. Conical transition.
stub and three-stub
transforrr1ers, which are connected to the line by n1eans of
special tees.
For matching two coaxial lines with different wave
impedances or with the same wave impedances but di ITerent
cross sections, conical
transitions are used,
which imitate an ex-
ponentialline. A coni-
cal transition of this
kind is shown in Fig.
9-8. 'fhe length l of
• I "4 the conical transition
is taken equal to
several wave-lengths .
Fig. 9-9. Rotating joint of a coaxial line: . In the transmission
l-frlctJonal contact: 2-section A. 1ines of radar an-
tennas, use is often
made of the rotation of one part of a coaxial line relatively
to the other part. At the same time. a special choke is used
in which the frictional contacts take place 1n the current
263
nodes; this reduces the energy losses in the transitions and
ensures a continuous contact. A rotating device of this kind.
is shown in Fig. 9-9. As can be seen from this figure, close·
ended half-wave lines are connected in series with the inner
and outer conductors at the cross section A of the line; the
half-wave lines enable to achieve a contactless short, at a
high frequency, of the two lines rotating relatively to one
another. The mechanical contact is effected in the current
nodes of these half-wave Jines.
•
9-3. Wavegulde Lines
In the 10 cm and shorter wave range, waveguides and,
in the first place, rectangNiar waveguides with an Hot mode
are extensively used as transmission lines of the antenna
systems. Circular waveguides are also occasionally used
as waveguide channels in antenna installations.
Let us first d\\·eiJ on the characteristics of the H01 mode in
a rectangular waveguide. A fairly detailed analysis of the
rectangular waveguide theory was given in Chapter Seven.
Under real conditions, the energy lost on heating the walls
of the waveguide has to be taken into consideration. As a
consequence, the amplitude of the field of the wave decreases
along the waveguide as the exponent and the attenuating
constant can be determined from (9-5).
The problem of wave propagation in a waveguide with
non-idea11y conducting wa11s has no exact solution,
therefore, usually the attenuatin~ constant is determined
approx imateJy, on the basis of the unperturbed field.
As we know, this field is defined by the expressions
1
E =A .k sin (' -n y )- e-ta,,,.
ou.oe \ b '
H y U1 =-A 0 ia. 01 .sin(~b u)e-IG~bZ• I (9-15)
11 101 = A 0 ~cos
b (~b u) e-tao.z J I

\vhere A. is a constant quantity;

a,o1 == k Vl-(A/Acrlt osl 1

·

The power conveyed along the wavegufde will be de-

termined from the substitution of the expre~Sions (9-15) Into
'
(7-18). We shall have:
1
p = A a k «cuab •
z 0 4(1)!

Owing to the fact that the waveguide conductivity is

finite, the tangential component of the electric field intensity
at the wall~ of the waveguide is not actually equaJ to zero.
Since the conductivity of the walls is high and the skin
effect very pronounced, this component of the field intensity
can be .obtained from the Leontovich boundary conditions
-·-
,, - - roJ& •
e I -t --
roe
Substituting here, instead of. H1, the values of the unper-
turbed magnetic fieJd from (9-15), we obtain the approxi-
mate expressions for the tangential components of the
electric field intensity at the walls pf the waveguide E,.
Thus, the approxirnation consists in that the tangential
component of the n1agnetic field intensity is taken unper-
turbed, i.e., determined from the expressions (9-15), the
tangential component of the electric field intensity at the
waJJs of the \\'aveguide being determined approximately
from the Leontovich boundary conditions.
The power absorbed by the \Valls of the waveguide can be
calculated by the Poynting vector method. Thus, the power
absorbed by the lateral waJJs per unit length of the waveguide
will be

The power absorbed by the broad walls per unit length

of the waveguide will be:
b

(-d::) broad
= Y: S(H.,.,H;o.+H,
II=O
01 H; •• ) I dg=
o~=o
,I (OJ.I. I bk1 ..
a:: V 2a Ao 2 •
Substituting the expressions obtained for the power con.
veyed along the waveguide and the power lost on heating
the walls of the \\'Sveguide into (9-5), following a nun1ber
265
of transformations, we obtain:
~= jl§.!.~ I+~ A' (Np'm], (9-16)
2a a A, b "-~rtt

where all the dimensions are given in metres.

The expression just cited shows that the damping
increases when the operating wave-length 'A approaches the
critical wave-length A,crlt· Then, the wave-length in the
waveguide A,w tends towards infinity and the attenuating
constant P also tt'nds towards infinity. In addition,· the
attenuating constant depends on the dimensions a and b.
~rhus, in the case of a constant value of the dimension b and
a decrease of the dimension a, the attenuating constant
increases. However, the dimensions of the \vaveguide cannot
he chosen frotn the tninimum attenuating condition due to
the fact that to eliminate the higher rnodes in the waveguide,
the ditnension a should be smaller than 'A/2 and the dimen-
sion b stnaJJer than 'A. Vll1en the ratio alb=0.5, as generally
adopted in practice, the attenuating constant in the centi-
lnetre wave range, in the case of copper waveguides, is
higher than 0.005 db/n1.
Usually, rectangular waveguides have an air insulation,
so that the energy losses on heating the medium inside
the waveguide can be neglected. In the case of a dielectric
with losses, the propagation constant y is defined as:

~) =
1
'\' 01 = r
-. /(.::_'
bI
2
-W f.l8 (} - · i (I)P,
---,------:----

Hence, the attenuating constant \Vill be approximately

defined as: •
k tan li
~.==2 , (9-17)
A,'£
·1 -
A~rlt 01
where
2.n 0
k=~-e, tan~=-.
(1)8

The general attenuating constant is determined by the sum

of the expressions (9-16) and (9-17).
266
 •

As for the maximum permissible power,.which a rectangu-

lar waveguide with an air filling can admit. it is defined as:
Pmax=6.63X IO-"E~rmab~ K (W], (9-18)
where Eperm is the maximum permissible field intensity,
which is usually taken equal to 30 000 Vjcn1;
K, the travelling-wave ratio in the waveguide.
In t~e expressions (9-18), all the dimensions are given
in metres. It can be seen from this expression that when
the wave-length approaches its critical value, the permis-
sible power in the waveguide goes down. The maximum
power in the waveguide is obtained when~ ~0.5. However,
this ratio is critica1 with regard to the appearance of the
A
H02 mode, so that one usually chooses 2b ~0.7.
Circular ,waveguides are utilised with the lowest modes:
H1t. EoJ and H 01 • Of these, as was indicated above, the Hu
mode has the highest critical wave-length (A.cru=3.42 a), so
that, in that case, the transverse dimensions of the tube
are the smallest. The Eot (l\cru=2.62 a) and Hot (A,·rlt= 1.64 a)
modes are excited in the case of tubes oi larger dimensions
and require therefore the use of filtering devices to elirr1i·
nate the excitation of the lower modes.
·, As for the attenuation caused by the energy losses in
the walls of the waveguides, in the case of the 11 t, and E01
1modes, it increases with the frequency much in the same
:way as in a rectangular waveguide and in coaxial lines. The
IH01 mode is remarkable for the fact that the attenuating
'constant decreases when the frequency increases.
: The attenuating constant can be calculated in the same
lway as for the Hot mode in a rectangular waveguide. Fig.
'9-10 shows the curves of the attenuating constant plotted
:against the frequency, in a copper tube of radius a=5 cm,
I
~for the three modes under consideration (39 ].
It can be seen that for the Hot mode, a very small
attenuating constant can be obtained. However, the
' utilisation of this mode in practice is rather difficult because
I a special device is required to filter out the Hu and Eot
modes, which are unavoidably excited when any tnhomo·
~geneities are introduced into the waveguide. One of the ways
'of obtaining a pure Hos n1ode is to utilise a circular wave-
guide ma-de of copper rings insulated from one another.
261
Rings of this kind favour the excitation of transverse cur-
rents with which the H02 mode is associated and prevent the
excitation of longitudinal currents with which the H11 and
8 01 modes are associated. Thus, in the milli!lletre wa ve
range, a waveguide with an Hot modl: can be uti lise d as the
most efficient transmission line in antenna instai'Jations.
(Jf
~
••
1t
If,
~~
:H,;.
~~
'

~
.,j.

fltrJ f

o 2 ~ 6 a m
~N m
Q
~

Fr81/UIIncv, tllousolltl8 Nt;b

<

Fig . 9-10. Atte nua tion constant of the modes ID

a circular waveguide.

l·
Q

Fig. 9·11. Quarter-wave-length insert in a rectangu·

lar waveguide. •

To tune waveguides on a travelling \Vave, quarter-wave-

length inserts, reactive stubs, reactive diaphragms and
waveguide stub transformers are utilised.
A quarter-wave-length insert in a rectangular waveguide
wit h an H0 , mode is shown in Fig. 9·11 . It is con nec ted
betweeb the node and the ant ino de of the ele ctr ic field in-
tensity (from the node towards the generator). The insert, is
a metal plate soldered to the broad wall of the wavegutde.
The insert reduces the narrow dimensioll of;. tile cross section
,
268 ...
of the waveguide, thereby reducing the wave impedance of
the twin line equivalent to the wavegu1de. The required
value of the insert wave impedance can be calculated with
the help of (8-16) and since the wave impedance of the equiva-
lent twin line is determined from (8-2), the transverse
size of the insert \Vi 11 be defined as:
a'=aVK. (9-19)
The reactive stub is inserted into the waveguide in a di-
rection parallel to its narrow wall and, for a smaJI length
(l <iv/4), it acts as a capacitance connected in parallel to
..... d---..
•

11 d a

.._,___ "
(/}J
....
__..

Fig. 9-12. Diaphrag1ns:

a-Inductive : b-capacltlve .
'
•
the equivalent twin line. l'he stub is connected at a distance
l 1 from the antinode towards the load, calculated frotn
(8-14). The calculation of the necessary length of the stub is
fairly difficult. so that its length l is determined experi-
mentally.
The waveguide is fairly often tuned by means of an
inductive or a capacitive diaphragm (Fig. 9-12). The dia-
phragms are ,thin metal plates lying in the transverse plane
of the waveguide. In the case of an inductive diaphragm, the
dimension b of the waveguide decreases and the dirnension
a of· the waveguide decreases in the case of a capacitive
diaphragm. A decrease of the dimension b of the wavegulde
in a twin line is equivalent to the insertion of a shunting.
inductivity and a decrease of the dimension a of the
waveguide is equivalent to the insertion of a shunting
capacitance. .
Let us investigate the theory of the inductive diaphragm.
Let the diaphragm be connected at the section z=O of an
infinite wavegulde and let a wave of the Ho• mode conte
towards it from the left, in the direction of the rising
values of z.
269
 Under the influence of this wave, the diaphragm becomes
excited and, as a result, a wave is formed after passing
through the diaphragm and a wave of the same mode Hot is
reflected from it. In addition, we get the excitation of
higher modes. It is evident that the electric field set up at
the opening of the diaphragm has a direction which coin-
cides with the x-axis; furthermore, the amplitude of this
field does not depend on the x-coordinate; in the direction
of the y-axis, it changes in accordance with a definite law
and on the rnetal part of the diaphragm, it equals zero ..
The electromagnetic field excited by the diaphragm can be
regarded as the field of the surface magnetic currents
induced from the left and right on an entire, infinitely thin
and ideally conducting metal plate \vhich cornpletely screens
the waveguide. Thus, we assume that the wave which
reaches the diaphragm is completely reflected from it and, in
accordance with (7-15) and (7-17), the primary field in the
region z~O is defined as:
E~= sin (;... y) {e-ta.,z_e+ta.,,z);
(9-20)
H~= :~ sin ( -F y) [e-ta.,z + e+ta.,,z],
.
where
an= Vk•-( TY.
In these expressions, the electric field amplitude of the
incoming wave is assumed to be equal to unity; the longi-
tudinal component of the magnetic field intensity will no
longer be needed and is not given here.
A secondary field is ~npPrpo"Prl nn the primary field in
the region z~O: it is set up by the surface magnetic cur-
rent, which coincides in direction with the y-axis and is
defined as •
JM= [n, E)= liz, lxJ Ex (y)= 1, E~ (y),
where i~, 11 , I, are the unit vectors;
Ex (y), the electric field intensity at the opening
of the diaphragm.
This surface current density can be associated with the
space current density in the region z ~ 0 by means of the
delta-function
j: = E~ (y) 6 (z-0). (9-21)
2'10
And, consequently, th~ secondary field in, the region z ae=; 0
can be determined frorn the expressions '(7-20) and (7-23).
Substituting (9-21) into (7-23) and integrating over x and z,
we obtain: .
(I)

~ 2
A11M (y, z)= ~bVom •
sm
(nrnb y ) X
JEs(y')sin (T y') dy',
xev....z (9-22)

where v... = y ( r- k and the integration

:n:bm
1
is per-
fortned within the limits of the opening of the diaphragm
b-d ~ b+d
2 ~Y~ 2 ·
We have taken into consideration the fact that, due to
the mirror itnage of the magnetic current in the metal
partition of the waveguide (diaphragm), the field is doubled
in comparison with the field of the same current in an
infinite waveguide. .
The surface n1agnetic current superimposed on the right
side of the partition is defined as:
JM=[n, E]=[-iz, lxlEx(y)=-i 1 Ex(Y)
and is associated \vith the space density of the n1agnetic
current in the region z ;;;;:= 0 by the ex press ion
j~=
-Ex {y) fl (z-0). (9-23)
Substituting (9-23) into (7-23), perfor1ning the inte-
gration over ,x and z and doubling the field on account of
the mirror image of the current in the partition. we obtain
for the region z ;a, 0:
GO

A;' (y, z)= ~

1
b!,.(:n:; y) x
sin

x e- v•.,, Jr E s (JI') sin (nmb y')\ dy'. (9-24)

The intensity of the secondary· electric field is

determi r1ed from the substitution of (9-22) and (9-24) into
(7-20). as a result of which, for z ~ 0, we obtain:
QO

E~= L j. sin(;~ y) eYv•' SE~(JI') sin (:n:;' y') dy' (9-25)

mz"l

271
 •
and for z~O:

JEx (y') sin ("'" y' )dy'.

Cll

, ~ b2 sin
E.¥==~ (nm y ) b e-Yom' b
nt=l

The superposition of surface magnetic currents on the

metal partition forms an opening and the continuity of the
electric field in the diaphragm thus obtained is satisfied
by the expressions (9-25) (when z=O, the primary field
I

Ex also equals zero). ·

The transverse component of the secondary magnetic
field intensity is determined from the substitution of the
expressions (9.. 22) and (9-24) into (7-20), as a result of
which, for z ~ 0, \Ve obtain:
«<

1 11 =
1'-1" 00
I ~ 2iYom •
~ b ·Sin
(nmb y ) X
f.l.,n=1 ·
xev...., SElf (y') sin (n;; y') dy'
and for z ;;;;:: 0 (9-26)
CID

1 · ~ 2ly0111 •
H ,11 =:t - Ca> ~ b Sin
( nm )
b Y X
J.l.m=t
X e-YomZ sElf (y') sin (n:: y') dy'.
For z=O, the transverse component of the intensity of
the total magnetic field should be continuous within the
Iimi ts of the opening:
Hu + H 11
; , .
l,.;;r, _ == H I,=
0 11 T0 a-t the opening.
Substituting here the expressions (9-20) and (9-26), we
obtain: •
sin ( T y) =
• 2
= -~I !!::" sin ( ~; y) sElf (y') sin e'; y') dy' (9-27)
•
at the opening.
Here, the unknown function E.¥ (y') is the integrand, sq that
the equation (9-27) is the integral equation relatively to
this unknown function. ~ ·
212
 The higher modes generated by the. diaphragms are
attenuated in its vicinity and at a certain, rather high,
distance, there is only the propagated H 0 , wave. The total
field of this wave in the region z ~ 0 is defined as:
, , ( n \ .
E"=E1(+E%= sin ,by) [e-ta.,z+ p 0e+lt~a,z]; (9-28)
, , a sin ("'
Hy=Hu+Hu=(l); by ) [e-ta.,~z-p,e+lausZ],
where · •

Po= -I+~ SEx (y') sin ( 1- y') dy' (9-29)

is, evidently, the coefficient of reflection of the H 01 wave

coming from the diaphragm;
The relative conductivity of the waveguide for the H0 ,
wave at the connection points of the diaphragm is determined
by the ratio of the quantity ~ H, to the quantity Ex in
ant "
the expressions (9-28) and is equal to the sum of the relative
conductivity of the infinite waveguide, eqt1al to unity, and
the conductivity of the diaphragn1 iB':
1+ t.8,_1-~
- I +Po ·

1-lence, the conductivity of the diaphragtn ;s defined as

(9-30)

Thus, having defined from the integral equation (9-27) the

intensity of tJie electric field at the opening of the diaphragm
E"(y'), we can calculate from (9-29), the reflection coef-
ficient p0 and then, from (9-30) the conductivity of the
diaphragm B'.
The solution of the integral equation (9-27) is difficult
to obtain due to the fact that the functions sin (~"!. y) are
not orthogonal within the limits of~ the opening of the
diaphragm. i.e., on the sect ion
b-d~ ~b+d
2 -== y ;;:::::: 2 •

Thdf~solutioo
of the integral equation relatively to Es(Y)
Is given in L. Levine 's book [40 J to which the reader is
tB-2122 218
 referred. The conductivity of the diaphragm in a quasi-static
approximation is found to be defined as:
B' = -~ cot (~) .
1
(9-31)
The quantity 8' is negative, so that the diaphragm under
investigation is inductive. If we take account of (8-13)
and (8-14), we obtain for the calculation of the magnitude of
the window d of the inductive diaphrag1n. the expression

~ cot ~) (9~32)
1 2
;;; = ( •

In a sirni lar way, we can obtain the expression for the

calculation of the capacitive diaphragm. The calculation
of the window d of this diaphragrn is done from the expres-
•
SIOfl

1-K
VK 4a
=~In csc
(nd) , (9-33)
20
where K is the travelling-wave ratio in the waveguide;
Aw, the wave-length in the waveguide.
These diaphragms arc connected at a distance la from the
antinode of the electric field intensity towards the generator
in the case of an inductive diaphragrn and towards the load
in the case of a capacitive diaphragm, determined from (8-14).
The inductive diaphragm is preferable because it practi ..
cally does not lower the power admitted by the waveguide,
since the distance between the broad walls of the waveguide
retnains unchanged. In the capacitive diaphragm, the distance
between the broad walls of the waveguide decreases and,
consequently, so does the magnitude of the admitted po\ver.
For the insertion into the waveguide of stub transformers,
so-called tees are used, installed in the plaqe of the magnetic
or electric vector of the field of the waveguide (Fig. 9-13).
In the case of an £-plane tee, a transverse slot is miiJed
in the broad wa11 of the waveguide and the branched wave-
guide joined to the slot. Since the transverse slot is excited
by the longitudinal component of the electric current,
in a twin line equivalent to the waveguide. such a tee
consists of a stub connected in series (Fig. 9-13, a). In
the case of an H-p1ane tee, the slot is milled in the narrow
wall of the waveeuide and the branched waveguide ,tf>ined
to it. This slot is excited by the transver~e
' .
components
214
 (tl) (/))
Fig. 9-13. Tees:
a-tee in the £-plane; b-tee In the H-plane.

of the current in the waveguide, so that in a twin line

equivalent to the waveguide, such a tee consists of a stuh
connected in parallel (Fig. 9-13, b). A two-stub tuner
consists of two £-plane tees shorted out at the end by
means of short-circuiting plungers. The distance bet\\·een
the axes of the tees is made equal to an .odd nutnber of
quarter·wave-lengths in the waveguide. Tees (stubs) of the
proper length always make it possible to compensate the
reflection from the load of the waveguide and obtain in it a
travelling-wave type of operation (narrow-hand matching).
The travelling-wave tuning elements described above
concern a rectangular waveguide with an H 01 wave, Similar
tuning elements can also be utilised in a circular waveguide.
Now. let us discuss the other elements utilised in wave-
guide channels. On the whole, we
shall have in m,ind elements utilised
in a rectangular waveguide.
The waveguide channel must
necessarily be assembled of separate ,zz:;!2Zz!t2=~:ss:s:s~~
sections and poor contacts at the
junctions may cause considerable tz::::;!:ZZ~~ =s~;:ss:::s:a:t
energy losses. Moreover, due to ox~­
dat.ion of the metal, the contact ts
not permanent. Considerable energy
reflections may occur from the place 2
of junction of two sections of t~e Fig. 9-14. Choke flange:
waveg~Jde channel, thereby d•s- 1-wavqulde; 2-nan1es :;3-
turbinf'the matching of the wave- rlnt perforation:
waaber ensurln1
4- rlnr
her met le
guide to the load. To avoid these aealln1
2'/S
 undesirable occurrences, the waveguide sections are gener-
ally joined by rrteans of choke flanges, as s~own in Fig. 9-14.
A coaxial perforation is made in the left flange of the junction;
its length AB is equal to a quarter of a wave-length.
In addition, the left flange has a circular groove forming
together with the right flange a radial line BC of length also
equal to a quarter of a wave-length. Thus, at the point C of
the flange, two sections of the waveguide are electrically
short-circuited. At the point B, where the flanges are in
actual contact, there arises a current node. Owing to the fact
that the resistance of this contact is connected in series to
the input resistance of the coaxial quarter-wave-length line
c 8
AB shorted out at the end, the
rnagnitude of the resistance of this
tz==ZZZZ: ~a: ~~zz:z:::~c::Q contact and its lack of permanency
A do not affect the operation of the
wave in the waveguide. The energy
lost on heating this contact js also
"lg. 9-15. Plunger with small.
a throttle. In the case of a circular waveguide
with an E 01 wave, transverse electro-
magnetic waves are set up in the coaxial line AB and radial
Jine BC of the clioke-ftange junction, the lengths AB and
BC of the groove being chosen appropriate to this mode.
In the case of a rectangular waveguide, an H11 mode is set
up in the coaxial line AB, the critical wave-length of which
equals AcrltH =1tdmean. \Vhere dmean is the mean diameter
11
of the coaxial line. An H11 mode is likewise set up in a radial
line, with one variation on the periphery. Consequently, the
lengths AB and BC of the grooves are chosen so as to suit
the H11 modes in these lines.
In the case of the H01 mode in a circular waveguide,
flanges can be dispensed with, since this mode is accompanied
by transverse currents on the walls of the waveguide. The
existence of a poor contact or even a clearance between two
sections of the waveguide does not affect the propagation of
the wave in the waveguide.
Special mobile plungers provided with the above-described
throttle are utilised for producing a short-circuit in wave-
guides. Such a plunger is shown .in Fig. 9-15. The length
chosen for the groove AB is equal to a quarter of a wave-
length and direct contact of the pl with the wavtgtiide
takes place at the points 8, where .oQIUrs a current
''

Z16
node. The length of the groove BC is also equal to a quarter of
a wave-length and, for this··reason, a current antinode is
produced at the points C and, consequently, an electric
short-circuit of the waveguide.
Let us now pass on to the description of waveguide-to-
coaxial-line and waveguide-to-waveguide transitions. Tran-

(a) (IJ)
Fig. 9-16. Rectangular wavegulde-to-coaxfal-li ne transitions:
1-waveJuide; 2-stub: 3-coaxlal line .
•

sitions of this kind find frequent application in practice;

the main considerations in their design are that they should
be matched and should not narrow the pass·band and
diminish the magnitude of the admitted power.
One type of rectangular waveguide-to-coaxial-line tran·
sition is iJiustrated in Fig. 9-16, a. In a transition of this kind,
the proper choice of the length of
the coaxial stub lstub and of the
length of the section of rectangu-
lar waveguide d makes it possible
to obtain in the coax ia I Iine, by
means of which the energy is fed to Fig. 9·17. Rectangular-to·
the waveguide, a travelling-wave circular waveguide tran·
sition.
type of operation. Another
type of waveguide-to-coaxiai-Jine
transition is illustrated in Fig. 9-16, b. In this transition
the internal conductor of the coaxial line, introduced into
the waveguide, thickens out, assuming a..·drop-like shape. A
transition of this kind enables to widen the pass-band a~d
Increase the power passing through the transition.
Fig. 9-17 illustrates an Ho~ mode r~c!angular wave~uide
to H1tmode circular wavegutde transttton. The trans1t!on
is accomplished by a gradual deformation of the cross sect1on
Z71
of the tube into a circular cross section. If the length of this
transition is of the order of one wave-length or more, in an
equivalent twin line, it represents a dispersive line with
gradually changing parameters. Such a transition is of the
broad-band type.
H01 mode rectangular waveguide to Eo 1 mode circular
waveguidc transitions are very frequently used. This is
because the Eo 1 mode is symmetrical relatively to the
waveguide axis and is therefore convenient for use in rotary
joints.
Fig. 9-18, a illustrates a transition of this kind, obtained
by means of a stub placed on the axis of the circular wave·

WtM1 Hme
Et,
£"' WDM1H,
Aall
4
WoveN"'
(a) (/J)
Fig. 9-18. Rectangular-to-circular waveguide tran-
sitions:
a-transition by me::1ns of a coaxial stub: 1-stub; b-tran-
aiUon with a resonance ring: 1-lnsulator: 2- metal rlnfl.

guide and introduced into the rectangu1ar waveguide through

its broad wall. Although in a circular waveguide whose dimen-
sions have been chosen so as to allow the propagation of an
Eot mode, we n1ay a Iso get the propagation of an Hss mode,
the latter is not excited because of the symmetrical position
of the stub. Thus, in this device, we get the excitation of a
pure Eot mode.
Another transition of this kind is illustroted in Fig. 9-18, b,
in which the Junction of the rectangular waveguide with the
circuJar one is accomplished by means of a rectanguJar aper-
ture in the lateral wall of the circular waveguide. In that
case, the excitation is favoured of both the Eot and H u modes
in the circular waveguide. In order to eliminate the Hst mode,
the short-circuited section of circular waveguide is given a
length of half a wave-length of the Eo 1 mode (AwE 0 ,/2) and a
metal resonance ring is placed in its transverse plane at a
distance of a quarter of a wave-length of the H1• mod~
(AwH 1114). This metal ring is excited by .the.'
Hss mode and
•
278
 short-circuits it. At the same time, it allows the free passage
of the E 01 mode, since the electric I ines of force of the E0 ,
mode are directed perpendicularly to the ring. In a twin line
equivalent to the waveguides, the short-circuited stub foraned
by a section of round waveguide is connected in series into
the line and allows· the free passage of the £ 01 1nodr, whi 1st
preventing the excitation of the H 11 mode. The nl('tal ring
is usually fixed on a dielectric support with sn1all loss.es.
A rotary joint in a waveguide channel is forn1cd h1 means
of an inductive section in a circular \Vaveguidt~. The latter
I r- T I

•
11
f

11,, I

(11) (/J)
Fig. 9·19~ Rotary joint:
a-transition with a coaxial stub: /-Inductive ~~'"Uon: 2-t(ears2
3-motor: 4-stub; 5-matchinR tunt!r:
b--transltlor; with a resonance ring: /-absorbing r!ng: 2-fal1erlng
metal ring

ts similar to the inductive section shown in Fig. 9-9. In

a rotary joint, rectangular-to-circular waveguide transitions
are used along with circular-to-rectangular waveguide tran·
sitions sintilar to the ones described above. Fig. 9-19, a illus-
trates the arrangement of a rotary joint of this kind with an
£ 01 wave in a circular waveguide in which the upper part
rotates relatively to the lower one around the axis of the
circular waveguide by means of gears. Fig. 9-19, b illustrates
another rotary device in which a transition from a rectangu-
lar to a circular waveguide with a resonance ring is used.
In this case, special absorbers are utilised to reinforce the
279
 ..
suppression of the H11 mode in the circular waveguide.
FQur longitudinal slots are cu,t in the circular waveguide
disposed on the periphery at 90° angles to one another. The
slots are excited by the Hu mode and oscillations of the Ho,t
mode arise in the tuned coaxial resonator. The coaxial reso-
nator is formed by the external surface of a circular wave-
guide and a metal tube enclosing the waveguide. An absorbing
ring is placed in the anti node of the electric field intensity of
the resonator and the Hu mode in the circular waveguide is
thereby suppressed.

9-4. Switching DevIces and Duplexers

Various switching devices are fairly often used in trans-
nlission lines. A brief description of some of then1 is given
bE' low.
The necessity may arise, for example, when two trans-
mitters of different frequencies operate on the same antenna.
Fig. 9-20shows the circuit
Antenna of a device of this kind
in which two feeders,
one frotn transmitter 1
.. and one from transmitter
2 are connected to the
colTtmon feeder of an
antenna at the point 0.
As suggested by S.l. Na ..
denenko, a combined stub
is connected in parallel
to each of the feeders at
~ distance of a quarter of
liJ tron.rmillerl To lrtm.tmillar 2 a wave-length from this
A, Az point. The combined stub
Fig. 9-2u. Two transmitters employ- consists of a short·
ing one antenna. circuited line of a length
equal to an even number
of half-waves of one of the transmitters. The length
of one of the arms of the stub between the connection
points to the feeder and .the short-circuit points is equal to
half the wave-length of the other transmitter.
Thus, when the transmitter 1 operates on the A1 wave, the
stub abc of length nA1/2 represents an infinitely large re-
sistance and allows the free passage of t)le. A.• wave, and· the
'

180
 stub klm. which lies at a distanc;e A1/4 frotn the point 0
and one arm of which has a length equal to ~1 /2, short .. circuits
the feeder 2 at the point land sets up an infinitely large
resistance of the feeder 2 at the point 0. When the transmitter
2 operates on the A2 wave. the stub klm of length n1 2/2 repre·
sents an infinitely large resistance and allo\vs the free passage
of the A2 wave, and the stub abc, which lies at a distance
A.2/4 from the point 0 and one arm of which has a length
equal to A2/2 short-circuits the feeder 1 at the point b and
sets up on this wave an infinitely large resistance of the
feeder I at the point 0. Consequently, both transmitters
operate on the same antenna, A
independently of one another.
If we take into account the
finite resistance of the wirt's of
1he combined stub, we see B
that there is satisfactory degree
of independence between
the transn1itters when the A1
and iw 2 wave-lengths are
shifted off the frequency by
5 to IO,iJ.
Another type of decoupling
device is the hybrid rin~. Q
which consists of a twin feeder, Fig. 9-21. Circuit of hybrid ring.
a coaxial lint~ or a wave-
guide (Fig. 9-21). The overall length of the ring is chosen
equal to 1.5 A.0 and the distance between the terntinals A
and 8, A and D, 8 and Care chosen equal to A.o/4 . If a gener-
ator is connected to the terminals A, a standing wave is
set up in the·annular line. provided all the other tern1inaJs
are unloaded; furthermore. a vcltage antinode is produced
in the section a, a voltage node at the terminals C and once
again a voltage anti node at the terminals B and D. Now, if
loads are connected to the terminals 8, D and C. no power
whatsoever is delivered to the load C and, on condition
that the loads are equal, the power is divided equally be·
tween the terminals Band D. Converselyrwhen the generator
is connected to the tern1inals C, no power whatever is de-
livered to the load connected to the termina I, A, but is divided
between the loads at the terminals 8 and D. Similarly, the
power fed to the terminals 8 is not delivered to the terminals
D but divided between the loads at the terminals A and C.
281
 It is readily seen that in the case of equal resistances
of the loads connected to the terminals 8, D (Z 8 =Zo). the
input impedance at the terminals A. connected to the
genera tor, is defined as:
W'
lA =2z o ,
B
where Wo is the wave impedance of the annular line.
Let W be the wave irnpedances of the n1atched lines connected
to the terminals of the r111~~ the condition of full rnatching
of the annular systcn1 is then reduced to the expr~ssion:
W 0 =W V2
Thus, on the principal frequency / 0 , the decoupling (ratio
of the powers) brtwE;\en the terrn1nals A and C equals infinity
and the trave11ing-wave ratio in the feeder A equals unity.
When the frequency is shifted ofT the pri nci pa I frequency
duE' to the fact that the hybrid ring is thrown out of balance,
th~ ciecoupling and mat,~hing deteriorate. Calculations show
that when the fr(_\quency ts sh 1fted off by 10 °Al (f=O. 9 f0 ).
the dccoupling is lo\vered down to 24.6 db. When /=0.9 / 0 or
f= 1.1 / 0 , the travelling-wave ratio in the ferder is lowered
down to K == 0.88, and when f =0.81 0 or f= 1.2{0 it is low-
ered do\vn to K = 0. 72. When the junctions of the hyhrid
ring are not parallel, as sho\vn in Fig. 9-21, but series
junc:t ions. the resistanc~ loading the generator at the termina Is
A is defined on the principal frequency as:
·>w'0
1-

ZA = ~z­
n

and the condition for the opli1uuu' lilatchiug of the annular

systen1 is defined as:
w .
W o= Jl"2.

The ser•es junction in thE' hybrid ring occurs, for example,

in the rase of a rectangular waveguide. when. E-p1ane
tees are used.
The hybrid rtng ctrcuit can be used for the simultanenus
operation of a tran~mitter and a receiver on a contmon
antenna on the same frequency. In that case, the transmitter
ts connected to the terminals A, the receiver, to the termi·
.,
282 •
nals C and the antenna, to the tetminal~ B. A matched
absorbing resistance is connected to t.fte terminals D.
Under these conditions, half the po\\·er of the transn1itter
is radiated and the other half is spent in the absorbing
resistance. Likewise, only part of the pO\\'er received by
the antenna is spent in the receiver.
The decoupling device shown in Fig. 9-22 finds fairly fre·
quent application in practice. This is the so-calJed magic tee
in a rectangular waveguide with an f/ 01 \\'ave. As can be seen

D 8
0
•
0

"
Fig. 9-22. Magic tee.

from the figure it is a cornbination of an £-plane tee and

H-plane tee. The device operates in a way similar to the
hybrid ring described above. When the power enters arn1 A,
it is not delivered to arm C but divided equally between arrns
B and D On the other hand, when the power enters arn1 C,
it is not delivered to arm A but divided equally between
arms 8 and D. Magic tees are matched by means of stubs,
as shown in Fig. 9-22, the inductive stub being connected to
arm A and the capacitive stub, to arm C. In the case of loads
matched in this way, no reflections of tbe powfr take place
when it enters any of the arrns, Furtherrnore, when the power
enters arn1 8, it is not delivered to arm D but divided equally
between arms A and C, Similarly, when the power enters
arm D, it is not d(\livered to arm B but divided equally
between arms A and C.
283
 •
The balanced bridge junction, shown in Fig. 9·23, a, finds
application in the centimetre wave range. It consists of two
sections of a rectangular waveguide joined along the narrow
sides through rectangular windows of appropriate dimensions.
In the case of matched loads, the high .. frequency power
entering arm A of the bridge is divided equally between arms B
and D, none being delivered to arm C; furthermore, the field
intensities in arms B and D have a 90° phase shift relatively
to one another. In the ab~nce of a window, the resultant
field intensity in arm B has a 45° phase shift relatively to .the

,_~
6 .D
8
ll "1lt
I I

it
{0)

Fig. 9-23. Balanced bridge junction:

a-circuit of the junction: /-principal waverulde:
2-auxlllary wavegulde: a-coupling section; b-vector
... diagram of the fielii.

field intensity in that arm. This is shown in the vector

diagram in Fig. 9-23, b.
From the physical standpoint, the wave incident from arm
A excites the window in the common wall of the waveguide
in such a way that the magnetic current of this window, which
changes in accordance with the travelling-wave law, radiates
co-phasaJ fields of equaJ amplitudt' in the direction of arms
B and D and produces no radiation in the direction of arms
A and C. The total field of the incident wave and of the wave
excited by the window in arm B corresponds to the vector.
diagram in Fig. 9·23,b. These properties of a bridge conne~ion
are maintained in approximately 12% of the
band.
r. fr!M~Jlenc.y~·\"'· J •
- ~~~.~ ·~ · .-· _ ,
Now, let us give a brief explanation regardfng·antenna
duplexers utilised in pulse radars, when transmitter and re-
ceiver are connected to the same antenna. During the transmis-
sion period (transmission of energy pulse), aiJ of the power
should reach the antenna, without branching off to the input
of the receiver. During reception of the signaJ.. . reftected from
the target, all of the received power should reach the input of
the receiver, without branching off towardS the transmitter.
The alternating connection of the antenna to either trans-
mitter or receiver is accomplished by means of a duplexer
the speed of response of \vhich should be 2 to 3 M/sec and the
switching frequency, 50 to 3.000 times per second. Switchings
of this kind are effected automatically, by means of spark
gaps and resonance line~
Tb ontenno

2
Tronsmiffer
J 6

10

Haceiwr
Fig. 9-24. Circuit of a duplexer with
quarter-wave and half-wave short-
circuited lines.
,
Fig. 9-24 shows the antenna duplexer circuit of the metre
wave range employing twin or coaxial lines. Quarter-wave
and ltalf·wave resonance lines are used in this circuit, where
they play the role of voltage set-up transformers. The reso-
nance line L1 is connected to the stub L2 at the points I, 4
near its short circuit, and the r~sonance line L5 is connected
to theJin~ t, and Le at the points 7, 8 and 11, 12, also near
their short circuits. ·~
·During the transmission period, the spark gaps connected
at the intensity antinodes of the resonance lines break down
and short-circuit the resonance lines; the points 1, 4 and 7, 8
of the lines L2 and L,. are short-circuited and the resistances
of these lines
'.
at the input at the points 2, 3 and 5, 6 are of
286
infinitely large magnitude. The energy pulse from the trans-
mitter passes freely into the antenna without branching off
towards the receiver input.
During reception, the spark gaps are deionised and the
input resistance of the resonance line L1 at the points I, 4
becomes infinitely large while the input resistance of the
stub L2 at the points 2, 3 becomes ir~finitcly small and the
branch of the transmitter at the points 5, 6 is disconnected
from the antenna. On the other hand, the input resistance of
the line L 6 at the points 1/., 12
I is transformed by the resonance
line L3 into the points 7, 8
To without change and the line
A L, becomes loaded on a matched
1" 1b antenna load. The signals picked
2 11 up by the antenna pass freely
to the input of the receiver.
Duplcxers used in the cen-
timetrewave rangearesimi Jar to
Fig. 9-25. Twin hybrid-ring those with the above-described
duplexer: circuit, which makes use of
1-load; 2-cavlty resonator waveguide lines and cavity reso-
' nators. There are other types
of transmit-receive duplexers. Fig. 9-25 shows a twin balance
duplexer which makes use of two hybrid rings. The hybrid
rings are coupled with the waveguides I and I I, into which
spark gap cavity resonators are inserted at a distance A-/4
in branch 8 and at a distance 'A/2 in branch D.
During transmission, the resonators break down and branch
8 sets up an open circuit in the ring, while branch D sets
up a short circuit. Since they are -E-pJane branches, the
energy from the transmitter reaches the antenna via ADC, no
energy whatever passing via A BC. During transmission,
that part of the power which leaks through· the spark cavity
is directed towards the second hybrid ring along the connect-
ing waveguides I and //and becomes absorbed in the load re·
si stance without reaching the receiver. This constitutes the
advantage of the circuit under consideration over the circuit of
the duplexer represented in Fig. 9-24, where the power which
leaks through the spark gap does reach the inputofthereceiver.
During reception, the power of the signals from the antenna
branches off into branch C without reaching the. output
of the transmitter, passes through branrlles ~ and D along
286
the waveguides I an~ If to the second ring and is deliv{\red
in full to the receiver input. Since the path difference of the
waves from the antenna to the load in the second ring along
the waveguides I and // is equal to half a wave. no power is
being dissipated in the load during reception.
The abo'(re-described balanced bridge junctions are also
utilised in antenna duplexer~ in the centintetrc wave range.
Tbonter1no

•
From~
mognf'lt;on
..._..._.,._ Energy ol' tronsmitler
--. -- ---teolrg.,,.,e through discharge tviJes
{0)

Hognetron.....,._

• ... Signal received

(/J)
Fig. 9-26. Ex:1iaining the princ1ple of the operation of a sym-
.· n1etrlcal duplexer with balanced bridge junctions:
a-during t ransmlsslon; b-durlng reception.
Fig. 9-26 shows the circuit of such a duplexer consisting
of two balanced bridges connected in series through duplex
spark gap cav1ty resonators.
During transmission, the transmit-receive boxes (T-R
boxes) break down and short-circuit the first bridge at the
outputs B and D. As a result of the 90° phase shift of the
voltages, the reflected signals cancel out in the transmitter
arm and add up in the antenna arm. Thus, the power of the
transmitter reaches the antenna. Here again, due to the 90°
phase shift of the voltages, that part of the power which leaks
through the T-R boxes •s absorbed in the matched load and
does not reach the input of the· receiver.
During reception, the power from the antenna foJiows two
paths and appears in full at the input of the receiver, no
power reaching the output of the transmitter or the absorbing
resistance.
281
 9-5. Devices Employing Perrltes
Lately, extensive application has been made of ferrite
waveguide devices. The ferrites, which are in a permanent
rnagnetic field, enable to obtain waveguide systems that do
not satisfy the reciprocity principle. The systems can be
divided into three groups: gyrators, isolators and circulators.
The gyrator, the symbol of \Vhich is shown in Fig. 9-27, a,
causes a 1800 change of phase of the wave when the wave
moves in the direction of the arrow and no change of phase

z
I 2

I ' 3

(a) (IJ) (fJ)

Fig. 9-27. Conventional signs of the waveguide ferrite devices.

when the \vave moves in the reverse direction. Gyrators are

used as cornponents of more complex systems, in particular
of circulators. ,
The isolator, the symbol of which is shown in Fig. 9-27, b,
is a device for the admission ahnost with out attenuation of
the waves moving in one direction and theintense absorbtion
of the waves ntoving in the opposite direction. Ferrite
isolators are used for decoupling generators from loads and
n1atching antennas to feeders. This is the most common
device utilising ferrites.
A circulator is a device which, as a rule, has four inputs
in which the energy is delivered as follows: signals fed to
input 1 are delivered onJy to input 2, from input 2 the signals
are delivered only to input 3, etc. In the conventional circuit
of the circulator (Fig. 9-27, b), the arrow shows the direction
of energy transmission. Circulators with a different number of
inputs are also possible. The direction of energy delivery can
be regulated by changing the direction of the permanent
magnetic field, in which case the circulator operates as
a non-contacting high-frequency commutator. In addition,
circulators can be used for decoupling a receiver from a trans,.
mitter continuously employing only one. ant~na, ~ ·for
other purposes. ·~ ··
'·'
2BIJ
 Apart from the devices described aboy.e, ferrite phase
S\\'itchers are also used, which utilise the property of ferrites
to change the value of their permittivity under the influence
of an external permanent magnetic field. Ferrite phase
switchers can be of two types: those that satisfy the
reciprocity principle and those that do not.
Here is a brief account of the physical and chernical proper·
ties of ferri tes.
Ferrite is a very hard and brittle ceramic material usually
of a dark colour. It does not readily yield to mechanical
treatment and can only be polished.
By virtue of its electric and magnetic properties, ferrite
is a unique material. On the one hand, it is a ferromagnetic,
i.e., a material akin in properties to iron (the relative permea-
bility of ferrites varies within a wide range depending on the
wave range. the composition of the ferrite, the temperature,
etc. It can pe less than unity on centimetre waves and can
reach as high as 2,000 at low frequencies). On the other hand,
it is a dielectric of high speci fie resistance amounting to
1Q6-:-l08ohms/cm. As a result, electromagnetic waves can be
propagated in ferrites just as in ordinary dielectrics. Note that
the relative dielectric constant of ferrite lies within 10-:-20.
Ferrite is prepared from magnetite FeO· Fe 20 3• The bi-
valent iron in magnetite is replaced by some bivalent rnetal
the dimensions of the ion of which are close to those of the
ion of the bivalent iron. Suitable metals are Mn, Mg, Ni,
Zn, Cu. Be, Li, Co, etc.
Combined ferrites, in particular lithium-zinc, manganese·
zinc and nickel-zinc ferrites find wide application.
The ·productlon process of ferrite is approximately the same
as that of ceramics. Powdered magnetite and oxides of the
corresponding metals are very thoroughly mixed, finely
ground and moulded after the addition of some binding ma·
terial,such as paraffin. This is followed by preliminary kilning
at a temperature of 300-400°C during which the binder is
burnt out. Then comes firing at a temperature of 1100·
14000C. The production process of ferri tes- is con1p lica ted by
the fact that the properties of ferrites are influenced by a
whole series of factors.
The gyromagnetic properties of ferrites require an ex pia·
nation.
The motion of an electron around the nucleus of an atom
gives rise. to an orbital magnetic moment, due to the motion
t9-2122 28J
 of the electron in an orbit, and to a spin, due to the motion
of the electron around its axis. A large amount of theoretical
research and a series of classical experiments have shown
that the dominant role in the nature of ferromagnetism is
played by the spin and not the orbital moments, which can be
neglected since they are always directed in a chaotic rnanner.
In n1ost elements, the spins of all the electrons are mutually
compensated. Ferromagnetics, which have no such compen-
sation on one of the orbits, are an exception. It is these
.electrons with the non-compensating spins that are responsible
for ·the pheuomenon of ferromagnetism.
We know that aJJ ferromagnetic materials consist of a
large number of regions or domains of spontaneous magneti-
sation. Each domain contains quite a large nutnber of atoms.
Within the limits of a dornain, the magnetic moments of the
atoms are parallel to one another, so that. even in the absenc~
of an external magnetic field, the domains are magnetised
to saturation. When a ferromagnetic is introduced into a
magnetic field, the separate domains become orientated in a
direction close to that of the external field.
All these phenomena are very complex to analyse, but we
can get a quaJitatiye picture of the phenomena in a ferrite by
considering the behaviour of one electron. This simplified
theory was put forward by Polder (41 ]. Without dwelling
on the n1athematical calculations of this theory, we shall
attempt to clarify its physical aspect.
As rnentioned earlier, ferromagnetic pheno1nena are due
to the possession by the electron of a spin, which n1ay be
roughly pictured as the rotation of the electron around its
axis. As a result of this rotation, the electron has its own
ntechanica I n1ornent, like any rotaltng body, as weiJ as its
own n1agnetic moment, the appearance of "'hich can be ex-
plained by regarding the rotating electroR as an elementary
winding with a current. Thus, the electron possesses on the
one hand, the properties of a gyroscope and, on the other
'hand, those of an elementary magnet.
Let us imagine that the electron is located in a permanent
magnetic field. Since the electron possesses the properties
of an elementary magnet, it is subjected in the magnetic
field to a pair of forces, which tend to turn the axis of the
electron in such a way as to cause it to be parallel to the lines
of force of the magnetic field (Fig. 9-28, a). However, due
to the gyroscopic effect, instead of becomi~g orientated along
I

290 -
the lines of the field, the electron begins to wobble; moreo\'er,
the end of the rnagnetic moment vector of the electron de-
scribes a circumference (Fig. 9-28,b). The circular frequency of
wobbling is proportional to the ntagnetic field app,lied:
roo= fy~ H, where y is a certain negative constant. Such a
wobbling n1otion would continue indefinitely if it was not for
the losses, due to which, the t:nergy of the wobbling is rapidly
dampttod and the electron axis bt'co1nes orientated in the
direction of the lines of force of the magnetic field. The
duration of the damping of the wobbling has been found
from experitnents to equal approximately 0.01 M/sec.

H z
N

V
(a) (DJ
Fig. 9-28. The electron in Fig. 9-29. Cornbined efTect
the permanent maqnetic of the magnetising and
fie Id: high-frequency fields on the
o-lorce~ at"t I01! on I he el{·l·t ron: electron.
b- wobbling of the t'h·ct ron

Now, in addition to the rnagnetising field, directed along

the z.. axis (Fig. 9-29) let the electron be subjected to the
action of a high-frequency 1nagnetic field h of frequency m,
directed along the x-axis. As a result of the action of the
high-frequency field, the electron starts wobbling anew, but
this time the wobbling is not dan1ped and proceeds at the
established pace with a frequency CJ> and not roo. The phe·
nomena described are close to those which take place in a reso·
nant circuit: the wobbling occurring uqder the effect of the
magnetising field resembles the natural oscillations of the
circuit, the combined action of the magnetising field and the
high .. frequency field resembling the forced oscillations.
Thus. at the established pace of wobbling, the end of the
vector of the magnetic moment m moves on a certain closed
curve. Polder has shown that this curve is an ellipse lying in
291
a plane perpendicular to the direction of the magnetising
field. Let us represent the vector of the magnetic moment m
as the sum of the constant vector m, directed along the lines
of the magnetising fieJd plus the vector m2 rottting with a
frequency w and lying in the xy-plane (Fig. 9-29).
It is readily seen that when the amplitudes are not too
high, the magnitude of the vector m2 is proportional to the
magnitude of the vector h. The rotating vector m2 can, in
turn, be represented as the sum of two vectors oscillating
with a frequency ro, having a goo phase shift relativeJ·y to
one another:
(9-34)

where lx, 11 are unit vectors;

"'· k, constants, corresponding to the magnetic sus-
ceptibility in the case of an isotropic medium.
They are defined by the expressions given
below.
Thus, under the action of high-frequency oscillations, there
arises in a magnetising ferrite medium, an internal high-
frequency magnetic field, characterised by the magneti-
sation vector m2 • The magnetic induction vector of the
high-frequency field b is the sum of the vector J.toh and the
magnetisation vector m2 , i.e.,
b x= (f.L 0 + x) h"; \
• (9·35)
b1 = ikhx. I
These expressions reveal the existence of a difference be-
tween a ferrite and an orcfin~try isotropir medium, in which the
rnagnetic induction vector is parallel to the magnetic field
intensity vector. This difference consists in that when a ferrite
is subjected to a magnetic field with a ·unique component
hJ(, the vector b has two components b and b1 . When the
ferrite is subjected to a high-frequency fiefd with a component
h1 , the same reasoning yields:
bx= -ikh1 ;
b1 =<Jt 0 +x)h1 •
· In the general case, when t.he electron is subjected to
the effect of an arbitrarily oriPntated. fieJd;.h, the relation
. ~

292
between the vectors b and h is as follows: I

bJC =ph"- ikh,;

by= ik/J X ·t- flhJ': (9-36)
b, = J.1 0 h,.
This same relation can be represented in an abbreviated
form:
b=/IJ.LHh,
where 11 J.L U is the tensor of the pernteability with the
following coefficients:
f.' -ik 0
ik . f.' 0 .
0 0 JL 0
The qu~ntities t.t and k have been calculated by Polder.
Thf'y are expressed as:
~' = J.' -J_ I VI Mro«L. • '
0 I (I) I - (1)1 '
0
(9-37)
k = _ I YI Mro ,
Ct> 0• - w'
•
where M is the magnetisation of the ferrite (by the magnet-
ising field).
The expressions (9-37) are valid for the electron. In the
case of a ferrite, they give a good qualitative coincidence
with experimental data.
The existence of a complex relation between the vectors b
and h greatly complicates the solution of the field equations
even in such a simple case as the propagation of a linearly·
polarised wave in an infinite ferrite medium. The analysis is
considerably simpler in the case of waves with a circuJar
polarisation. This is not surprising, since the magnetisation
vector in a ferrite is a rotating one.
Owing to the fact that the direction4 of wobbling of the
electron depends only on the direction of the magnetic field
lines, it is· likewise convenient to relate the direction of
rotation of the wave incident on the ferrite to the direction of
the magnetic field lines and not to the direction of wave
propagation, .as is usually done. Let the wave, with vector
b rotating clockwise when looking towards the magnetic
298
field lines, be right-handed rotary. In the case of the left-
handed rotary wave, the vector h rotates anticlockwise.
Let the ferrite be subjected to a right-handed rotary wave.
In the case of this wave, the relation for the cotnponents hy
and hx is h1 =-ihx, i.e., the component h1 has a 90° phase
Jag relatively to hx. Let us calculate the components of the
vector b account taken of (9-36):
bx= (f.t- k) hx;
b1 = - i (~ -k) hx, or by= -ibx.
Con~equently, the vector b is polarised exactly as the
vector h and the relation between these vectors will be
the sa1ne as in an isotropic mediun1. The perrneability, in
the case of the right-handed rotary wave, is expressed as:
= -k= lVI M
J1 + ~ flo + Ctl -m ·
0
(9-38)

In the case of the left-handed rotary wave,

h1 =ihx
and we get a similar picture, but the permeability is now
expressed as:
(9-39)

Thus, in the case of a rotary \\rave, the permeability is

a scalar quantity but it is not the same for waves with a
left-handed polarisation as for waves with a right-handed
one. The relative pertneabllity JL, is plotted as a function of
the 1nagnitude of the magnetising field for waves with a
left- and right-handed polarisation in Fjg. 9-30.
It is seen iron1 Fig. 9-30, a that at a definite value of the
magnetising field, the quantity fl+ undergoes a pronounced
change. This vaJue corresponds to the coincidence of the natu-
ral frequency of wobbling with the frequency of the forced
oscillations Ctl. This phenomenon is known as ferroresonance.
The same can be seen from the expressions, but the latter do
not take account of losses whereas the graphs do.
The dotted curve in Fig. 9-30, a, characterises the losses
in the ferrite and the sol id curve, the quantity f'+. In the
resonant region, the losses in the ferrit~ are ·very high. From
•

294 -
 the physical standpoint. this is quite understandable, ~ince,
\V hen the natural frequency of wobbling ·coincides \\'i th the
frequency of the external field. the electron wobbles \Vith the
maximum possible arnplitude and abstracts the maxin1un1 of
energy fron1 the field. Furthermore, a considerable part of
this energy is spent on heating the ferrite. In the vicinity of
the resonance, it is the perJne-
ability that undergoes the 1nost II
6
pronounced chang~. since, as
5 I I
a resu J t of the pronounced 4
wobbling of the electron, the 3
I \
vector m2 has its max in1u1n 2 I \
\'alue. f I \ H
/
No such thing occurs in the 0 •
case of a left-handed polarised -f
\vave (Fig. 9-:30, b) due to the -2
fact that, in the case of a right- -a H,.
handed polarised \Vave, the -4
(tl)
direction of rotation of the wave
and of the vector m2 coincide
and the delivery of energy by
the wave to the electron is con- Jlr-
tinuous, whereas in the case of '/ ,................._ __
a left-handed polarised wave, , , N
the rotations are opposed and 7~----------
the electron receives energy at (IJ)
one part of the period and gives Fig. 9 _30 The relative mag-
it up at the other part of the netic permeability fLr+ and flr-
period. plotted as a function of the
Let us new return to the magnetising field.
propagation 'in the ferrite of a
linearly-polarised wave, which we shall represent in the form
of a superposition of two waves of circular polarisation
with opposed rotations.
Since the perm ea bi Iit ies of the left-handed and right-
handed waves are different, so are the phase coefficients:
(X+= (I) VBf.'+' a,_= (J) J{ lf1_.
As seen from Fig. 9-30,a+<a,_;consequently, these waves
have different phase lag over one and the same distance.
If, at a certain point, the vectors are directed as shown in
Fig. 9-31, a, then, at a distance corresponding to a wave·
length, the picture for a wave with a left-handed polarisation
291
is as represented in Fig. 9-31, b. The polarisation plane of the
resultant vector will turn by an angle
1
8= 2 (a_ -a,+)l. (9-40)

where l is the length of the wave path in the ferr·ite medium.

lt should be carefully noted that for an observer looking in
the direction of the wave motion, the rotation of the polari-
sation plane occurs clockwise if the wave is propagated in the
direction of the lin~s of force of the magnetising field,. and
anticlockwise, if the wave is propagated in the opposite
direction to the lines of force of the magnetising field. In other
words, the direction of rotation of the polarisation plane does
h 11 not depend on the wave propa-
,' r gation direction if the observer
h-' t
'h-~ looks all the time in the
direction of the magnetising
(Q) (DJ field {or against it).
fig. 9-:":Jl. Illustrating the Fara- The polarisation plane
day cft'ect: rotation effect is known as the
a-reprcsentotion of a linearly· Faraday effect; it is frequently
polarised wav~ In the form of two made use of in various micro-
waves with a rotary polarisation;
b- turn of the polaris a tfon plane of
the resultant vector of the magnetic
wave ferrite devices.
f1cld 1ntenslt y. All the above-given expla-
nationsconcern the propagation
of a plane Jinearly-polarised wave in an infinite ferrite medi-
um. In devices used in practice, the ferrites are placed in
waveguides. In particular, to utilise the Faraday effect, a
ferrite rod is placed on the axis of a circular waveguide with
an Hu wave. At the san1e time, the longitudinal magnetising
field is set up by a solenoid \vhose .axis coincides with that
of the waveguide.
It is always desirable to get the maximum possible angle
of rotation of the polarisation plane for mfnimum reflections
frorn the ferrite rod, minimum losses, the smallest possible
size and weight of the device. Let us consider the main
points to be taken into consideration \\'hen choosing the
size of the ferrite rod. The most important is the choice of
the rod's diameter. When the diameter of the rod is small,
only a very small part of the energy is propagated in the
ferrite rod, so that the magnitude of the angle of rotation .or
the polarisation plane by a distance equal to the unit length
of the rod e. is small. Bigger diameter .of· t~ rod means sn
' '•

216
increase of the amount of energy propagated in the ferrite
and with it, of the angle 81• ,·
Due to the high permittivity constant of ferrite, even in
the case of a relatively small diameter of the rod, almost
all of the energy is propagated inside the ferrite and a further
increase in diameter does not lead to an increase of angle
61 (Fig. 9-32). There is, therefore, no point in using ferrite
rods of a large diameter.
Fua·thermore, bigger rod diameter means an increase of the
losses and, due to the high permittivity of ferrite, an increase
of the equivalent diameter of the
waveguide and this favours the ap-
pearance of the higher modes. Thus,
in practice, only the initial region
of the curve is utilised (Fig. 9-32).
The influence of the length of the
ferrite rod on the magnitude of the
angle of rntation of the polari·
sation plane is seen from (9-40);
it follows that the angle e is
directly proportional to the length • I
Fig. 9-32. Angle of ro-
of the ferrite. Ho\vever, the length tation of the polarisation
of the ferrite rod should not be too pi ane by et distance equal
great since, at the same time, one to a unit length of the
has to increase also the length of ferrite rod plotted as a
the solenoid and the whole device function of its diameter:
dt-dlameter of the h·rrilc
becomes too bulky. Moreover, an rod: d-dlameter of the wave-
excessive length of the ferrite leads guide.
to a narrowing of the bandwidth
due to the fact that the two sources of reflections, i.e.,
the ends of the ferrite rod, are spaced too far apart. To
reduce the reflection from the ends of the ferrite rod, they
are given a fairly elongated conical shape, but, even so,
the reflection cannot be avoided altogether.
Let us no\\' consider a number of devices based on the Para·
day effect and, to begin with. the isolator, which is illus·
(trated in Fig. 9-33,a. As was mentioned earlier, in an isolator,
the attenuation is different for waves propagated in differ·
lent directions. Let a wave be propagated from left to right.
From the segment of rectangular waveguide I, the wave is
•delivered to the continuous transition 2 from the rectangular
to the circular waveguide. The wave propagated in the circu-
lar waveguide is an Hsa wave. Owing to the ferrite rod 3,
which is magnetised by the solenoid 4, the polarisation plane
of the 1111 wave rotates to the left by 45° (at the same time
the wave moves against the direction of the lines of force of
the magnetising field). The 45° turn is achieved through the
appropriate choice of the diameter and length of the ferrite
rod and the magnitude of the magnetising current. Subse-

, z 6

I t 7 I

(Q)

I
(b) I I
•

I (C) I I
I

t
I
(d) I
I

(e)

Fig. 9-33. Isolator adapted on a circular \\'aveguide.

quentJy, the wave again reache~ the continuous transition 5

from the circular tC' the rectangular waveguide. But the
rectangular waveguide 6 is turned 45° to the left relatively
to the waveguide ~I and, as a result, the ·wave passes freely
through the isolator. The two absorbing plates 7 and 8 are
perpendicular to vector E and have practically no influence
on the propagation of the wave. The structure of the field in
various regions of the isolator, when the wave is propagated
from left to right, is shown in Fig. 9-33, b.
Now, let the wave be propagated from right to left (Fig.
9-33, c). The propagation takes place approximately as in the
preceding case. However, owing to the fact that the angle of
rotation of the polarisation plane does not depend on the
••
direction of wave propagation, the polarisa.tion plane etT~cts
a furtht:'r 45'.) turn to the left in addition 'to the rotation of
the waveguide and the E vector is found to be paraiiE:'I to
the broad wall of waveguide I. This n1eans that the wave
cannot get into waveguide /. At the same tirne. as seen in
Fig. 9-33, c, theE vector will be parallel to the absorbing
plate 8 and the wave is intensely absorbed by this plate.
If the .wave has not been con1pleteJy absorbed by plate 8,
owing to the turn of the polarisation plane to the left, the
wave reflected from the plate reaches plate 7, with the E
vector para lie I to that pia le (Fig. 9-33, d) and is absorbed
once again. Finally, that part of the energy which is not ab-
sorbed by plate 7 is delivered to the wavegui'de I (Fig. 9-33, e).
Thus, in the device under consideration, the wave propagated
from right to left will be considerably attenuated.
A cjrculator based on the sarne principle can also be
constructed (Fig. 9-34, a); it differs frotn an isolator in
having two additional outputs.
Let us investigate the way waves are propagated in a
circulator when it is fed from different inpu.ts. If the signal is
delivered to input /, the picture of the fields in the various
sections of the circulator is as shown in Fig. 9-34, b.
Here, as in the case of the isolator just exan1ined, the
wave passes freely to input I I; the segments of waveguides
Ill and IV are not excited. As can be seen from Fig. 9-34, b,
from the point of view of the structure of the field of the
circular waveguide, the inputs of these segn1ents are analo-
gous to longitudinal slots in the centre of the broad waJ I of a
rectangular waveguide with an Hot wave, wh1ch, as we know,
are not excited.
When the signal is fed to input I I, we get the picture
of the fields shown in Fig. 9-34, c, where the window coupling
the segment of rectangular waveguide Ill with the circular
waveguide is similar to a longitudinal slot cut in the centre
of the narrow wall of a rectangular waveguide and is,
therefore, quite intensely excited. Thus, the signal from
inpui I I reaches only input Ill. The processes which take
p.lace when the circulator is fed from input Ill (Fig. 9-34, d)
and frorr. input IV (Fig. 9-34, e) can be investigated in a
similar way. As mentioned earlier, the circulator should
deliver energy from input I only to input I I, from input I I
only to input Ill and so on. The circulator described accom-
plishes such a delivery when all the inputs are well matched
299
to the circular waveguide. Inputs I and // are well matched
by means of continuous transitions, but the same cannot be
said of inputs Ill and IV. Thus, if a signal is fed to input
I/, the main part of the energy teaches input Ill, the energy
reflected from input Ill reaches input /~'and that which was

J/l
.....
. ~
•

- •. I •

I - E -;;.. ' • ll
I
I
•

(Q)

(/J)

I
(d)

{8)

Fig. 9-34. Circulator adopted on a circular wavcguide. The

arrows sho\\· the direction of wave motion.

reflected from input IV reaches input /. This is the disad-

vantage of the circulator described.
The circulator shown in Fig. 9-34 can also serve as a
non-contacting commutator. If we change the direction of the
current in the solenoid, the rotation of. the polarisation
plane occurs in the other direction and the energy transfer
takes place in a different order: fr-om input I to input IV,
from IV to Ill, from Ill to /I and from I/ to I. . ·
A ferrite rod placed In a circular wave~uide mayA also be
utilised as an electrically regttlated p JtWerter. in wbicb lf
•
800
the Faraday effect is undesirable and measures should be
taken to eliminate it. ;.
One of such measures is the utilisation in the waveguide
of a field with a rotating polarisation. A phase inverter of
this kind is shown in Fig. 9-35. In outward appearance it
also resembles the isoiator described above. But contrary to
the isolator, the phase inverter has two polarisers (/ and 2).
Polariser I transforms the H11 wave with a linear polari-
sation into a wave with a rotary polarisation. The operating
princi pie of a polariser is the fa) Iowing: a dielectric plate is

• f I •

•
Fig. 9-35. Phase shift er.

placed in a circular waveguide at an angle of 45° relatively

to the dir:ection of the vector E of the H1s wave. The Hu wave
in the waveguide can bt' regarded as the sum of two waves,
the vector E of one of which is parallel to the dielectric plate
and that of the other one, perpendicular to it. The wave, the
vector E of which is parallel to the dielectric plate, is re-
tarded by this plate to a greater extent than the second wave,
so that a pl}ese shift takes place between the two waves.
The length of the plate can be chosen such as to obtain a
phase shift of 90°. Then, after the plate, there will be waves
polarised at a 9CY' angle to one another and having a 90°
relative phase shift. Two such waves taken together produce a
wave with a circular polarisation. The wave utilised in a
phase shifter is usually a left·hand rotary wave, because, in
this case, the losses in the ferrite are lower than in the
case of a right-hand rotary wave. ~· ·
· The rotary wave passes through the section with the
ferrite. A change in the magnetising field leads to a change of
the magnetisation of the ferrite and, consequently, also of its
permeability (expression (9-39)). A change of permeability
leads to a .chaQge of the phase velocity of propagation of the
301
wave In the section with the ferrite and, consequently. to a
change of the phase of the osciiJations at the output of the
phase shifter. In depoJariser 2, the rotating field is again
transformed into a linearly-polarised one. The arrangement
of depolariser 2 is similar to that of polariser 1.
The value of the maximum phase shift accomplished by
such a phase sh i fter depends, on the whole, on its length, as
well as on otber factors. It Jies
within a range of several hundred
degrees.
Devices with ierrites used in circu-
lar waveguides have a number of
shortcomings, the need for a rectangu-
lar-to-circular waveguide transi-
tion, the presence of a long solenoid,
the bulkiness of these devices atnong
0 them. The devices have a number
of other shortcomings, in particular
the difficulty of obtaining small
reflections from them in the frequen-
cy band.
The best results can be obtained
in the case of devices in which the
Fig. 9-36. Explaining the ferrite plate is placed in a rectangu-
effect of the rotation of
a magnetic field in a rec- lar waveguide. The operating princi-
tangular waveguide. ple of devices of this sort can be
roughly explained as follows. Let us
consider the picture of the magnetic lines of force in a rec-
tangular waveguide with an H 01 wave. This picture is shown in
Fig. 9-36 for four instants differing from one another by an
eighth of a period.
As can be seen from these illustrations, at a certain
distance from the middle of the waveguide. the magnetic field
forms a rotattng vector of constant magnitude. The direction
of rotation of the h vector depends on the direction of wave
propagation.
Let us place a ferrite plate in the waveguide, so that
the region with the rotating 1nagnetic field is situated in
the ferrite (Fig. 9-37). The magnetising field should be made
parallel to the narrow wall of the waveguide. The phenomena
which occur in this case are similar to those considered
earlier in the case of an unbounded ferrite medium and in a
circular waveguide. Since the dirPction of rotetion of the field
802
 intensity vector depends on the directiQn of wave propa..
gat ion, the permeabiJity will not be the satne for waves propa-
gated in different directions; furthermore, it is determined
in accordance with fJ.+ and f.l-· The phase velocities of the
waves moving from left to right and right to left are different
too and, consequently, also the phase shifts. If the dimensions
of the ferrite plate and rnagnctising field are chosen in
such a way that the phase shifts in the forward and reverse
directions di tier by 180°, then such a plate
in the rectangular waveguide acts as a H
gyrator.
A gyrator of the type described can be
utilised as an element of -more complex
devices, such as circulator~ (Fig. 9-38). Fig. 9-37. Ferrite
A circulator has two directed couplers 1 plate in a rectangu-
and 2 with a transitional attenuation of 3 lar waveguide.
db (an attenuation of this kind ensures
the transmission of half the power into the second channel),
a ferrite plate causing a 180°+'\1' phase shift in the forward
direction (indicated by an arrow) and a phase shift
equal to 'P in the rev~rse direction and, in the symn1etrical
channel, it has a dielectric plate causing a 1P phase shift in
both directions.

I __ AI I __ ./I
.1/[
--8 -- lY
DieltJCtric plots
Fig. 9-38. Circulator adapted on a rectangular
waveguide.

Let us investigate the behaviour of the ctrculator when it

is fed at input /. The energy cannot reach input I I I, since
both couplers 1 and 2 are directed and transmit energy into
waveguide 8 in such a way that it moves in th~ same dir~ction
as in waveguide A. The energy can reach anput /I tn two
ways: directly through waveguide A or through coupler I,
303
waveguide B and coupler 2. The wave propagated in the
first way acquires a phase shift corresponding to the distance
between input I and input // plus l80° plus 11' on account of
the ferrite.
The wave which followed the second path acquires the
same phase shift, it is obtained due to the fact that when
the wave passes through the coupler with a transitional
attenuation of 3 db. it acquires a goo pbase shift and when
the wave passes through two couplers, the phase shift equals
180° plus a 1p phase shift on account of the dielectric p'Iate.
The signal reaches input Ill likewise in two ways: through
coupler I and waveguide B as well as through waveguide A

, •
2 3

~(a) (IJ)
Fig. 9-39. Resonant isolators: Fig. 9-40. Structure of
a-resonant Isolator with one plate; /-ferrite the field in a waveguide
plate; b-resonant Isolator with two plates: with a ferrite:
2-ferrlte plates.
/-forward wave; 2-reverse
wave; a-partially filled
wavegulde.

and coupler 2. Moreover, because of the ferrite, the phases of

the waves propagated in these two ways differ by 180° and
cancel out.
In a similar way, we can investigate the behaviour of the
circulator when it is fed at the other inputs and ascertain
that signals pass from input I I to input I 1I, from input Ill
to input IV and from input IV to infut /. The described
circulator gives quite a high transitions attenuation between
the inputs, which should not be coupled (e.g., I and I I I;
I I and IV, etc.). It amounts to approximately 30 db. It is
impossible to obtain such a high transitional attenuation
in a circular waveguide circulator.
With the help of a ferrite plate placed in a rectangular
waveguide, one can obtain devices possessing the properties
of isolators. In particular, resonance isolators have found
fairly wide application. The magnitude of the magnetising
field is chosen large enough to brit:~& •bout a ferroresonance-
and, consequently, a high attenuatiQnt fQf ~tbe wave which
'
'

804
 causes a right-hand rotation in the ferrite. There is no reso-
nance for· the wave moving in the opposite direction and it
progresses without noticeable losses. The circuit of an isolator
of this kind is shown in Fig. 9-39, a. In such an isolator,
it is possible to obtain a ratio of the losses when the wave
moves in the forward direction to the losses \vhen the \\·ave
moves in the reverse direction of about 18 db. The- best
results are produced by the isolator illustrated in Fig. 9-39, b,
in which the ratio of the forward losses to the reverse ones
reach approximately 22 db. However,
in the latter case, large fields have to be H
utilised and the system must besufficiently
long because the attenuation per unit ,
length for the forward wave is lower here
than in the isolator shown in· Fig. 9-39, a.
Apart from resonant isolators, we should
mention the application of isolators
making use ofthenon-reciprocal distortion
of the field in a waveguide with a ferrite
plate. Fig. 9-40 shows the structure
of the field for the forward and reverse Fig. 9-41. Isolator
in a coaxial line:
waves obta-ined experimentally. It is seen t-ferrlte olate:
that the reverse wave is concentrated 2-dlelectrlc plate.
in the vicinity of the ferrite plate. If the
ferrite plate is coated with an absorbing film, the reverse wave
is absorbed to a considerably greater extent than the forward
one. This feature particularly is made use of in isolators.
Isolators of this kind are more convenient than resonant iso-
lators, because they require considerably smaller magnetising
fields. The ratio of the losses for waves moving in different
directions obtained in th-is case is of the order of 12-13 db.
It can be seen from the brief account given here that the
technique of the application of ferrites in waveguides has
reached a fairly good level of development. The same cannot
be said o.f the utilisation of ferrites in coaxial lines. Placing a
magnetised ferrite plate in a coax ia I 1ine does not lead to the
appearance of irreversible properties. However, if a dielectric
is placed near the plate (Fig. 9-41) ana the configuration of
the plate is matched to the dielectric, it is possible to obtain
an isolator with a ratio of the losses of the order of 20 db in
the 10 cm range and around 10 db in the 30 cm range. Apart
from isolators~ no other non-reciprocal devices are used in
coaxial l.ines at the time of writing.
20-2122 ··•.:
 PART THREE

Types of Antenna .Devices

CHAPTE~ TEN - •

U ltrashort-W ave Antennas

10-1. Antenna Types

The term ultrashort-wave antennas is generally applied to
antennas used on waves shorter than 10 m. It covers antennas
of the 1netre, decimetre, centimetre and millimetre wave
ranges. Ultrashort-wave antennas are subdivided into several
types: simple wire and slot antennas, multiple wire and slot
antennas, horn-type antennas, lens antennas, reflector an-
tennas, helical antennas, dielectric antennas and impedance
antennas. These types are, in turn, subdivided into several
varieties which wil1 be dealt with below.
Ultrashort-wave antennas are used in radio cornmunication
and broadcasting, television, radar, radionavigation, radio-
astronomy, radio control. They can be simple or multiple,
of low directivity or highly directional. These antennas are
utilised in ground and ship installations and on flying
vehicles of various kinds.
Depending on their purpose, ultrashort-wave antennas are
expected to satisfy the most varied requirements. The follow·
ing requirements are cotnmon to all antenna types: maximum
possible simplicity of construction and q~aintenance, high
efficiency, relatively wide pass-band, high breakdown electric
voltages and steady performance over a prolonged period.
We shall begin by an investigation of simple wire and
slot antennas, before passing on to more complex ones.

10-2. Simple Antennas of the Wire Type

Simple antennas of the wire type are used as self-
contained radiating and receiving devices and as components
of many multiple antennas. They find frequent
,.,.
application
'
as feeds of multiple antennas, for example parabolic
reflectors. ,·
The simplest of antennas is the haJf.. wave dipole fed by a
twin feeder. When the twin line is connected to the dipole at
its gap (Fig. 10-1, a), the travelling-wave ratio in the line
is about 0.15 so that the line has to be tuned on a travelling
wave. A suitable circuit is shown in Fig. 10-1. b. The proper
choice of the din1ensions Land l of the shunt ensures operation
close to the travelling-wave type \\·ithout supplen1entary

-
1.· . I
-DJr
\
0
7t •
a
/= Of!if. ~

l= 0.15A
' ' ' '
.(Q) (C)

Induct/vs
.,.__l I· sfu/J

Fig. 10-1. Half-wave dipole fed by a twin feeder:

- symmetric::tl half-wave dipole: b-dlpole with shunt feed: c-loop-llke
Pistolkors dipole; c-range shunt dipole.

tuning eletnents. An additional advantage of such a feed is

that a voJ tage node is set up at the point a of the dipole
securing the dipole to its support without insulators.
To increase its input resistance, the half-wave dipole can
be given a loop-like shape (Pistolkors's circuit). As shown
in Fig. 10-1, b, two elements of the dipole are shorted out at
the endcJ a twin line being connected to the gap of one of
them. The currents in these two elements are in eo-phase and
since the distance between them is smaJJ in comparison with
the wave-length, the directional diagram of the loop dipole
coincides with that of the usual half-wave dipole.
20* 801
 The input resistance of the loop dipole can be calculated
from the power balance. Let the current in the antinode of
each of the elements of the dipole equal I. The· power radi-
ated by the dipole will then equal:
Pr, = (21) 1 X 73.1.
On the other hand, the radiated power equals:
• J•
Pr. = Rab· -
Hence, the input resistance of a loop-like dipple is:
Rab=4 X 73.1=292.4 ohms
and the travelling-wave ratio in the feeder wiJJ be around
0.5, Sb that no additional adjustment of the feeder is
necessary.
The loop-like dipole offers the additional advantage that
it can be secured to its support without insula tors, since a
voltage node is set up in the centre of one of its elements.
The loop-like dipole is generally used as a component of tefe ..
vision receiving antennas and antennas of the wave channel
type.
The dipoles considered above are not of the broadband
type. The input resistance of any of these dipoles undergoes
a sharp change in the event of a change of wave-length, giving
rise to mismatch between dipole and feed line and a fall
of the travelling-wave ratio in the feeder.
Fron1 this point of view, the so-called multiple-tuned
shunt dipole proposed by G. Z. Eisenberg (16] is a consider-
able improvement. The circuit of this dipole is shown in
Fig. 10-1, d. The action of an inductive stub (shunt) connected
to the dipole at the poiut~ ub is .similar to that of the reactive
stub utilised in the case of broadband matching (see Para-
graph 8-4) and which, within a certain waye range, backs off
the input reactance of the dipole.
The proper choice of the points of connection of the stub,
its geometrical dimensions, the angle a and the wave im- .
pedance of the feeder enables to obtain a satisfactory matching
of the dipole to the feeder in about a quadruple wave range
(from lw~6.5l to "-~ 1.51). Within that range, the travelling-
wave ratio in the feeder does not fall below 0.3. Matching
is likewise favoured by the fact that the dipole consists of
broad plates (or of thick wires) which have a fairly l.o\tl wave
impedance. .: ·~~~
•,,
 The ordinary as well as loop-like half-wave dipole can also
be fed by means of a flexible coaxial cab~ or a rigid coaxial
line. In t~at case, th~ transition from the asymntetrica I line,
the coaxtal cable be1ng such an asyrnmetrical line to the
symmetrical dipole is effected by means of various ba'lancers,
the circuits of which are shown in Fig. 10-2 and 10·3.
Let us examine the U-elbow type balancer (Fig. 10-2, a).
If t.he coa~ial c~ble is ~onnected direct~ to the symn letri·
cal d1 pole, t.e., tf the tnternal conductor of the ea hie is
connected to ane of the arms of the dipole and its external
conductor to the second ctrm of the dipole, the current flo\ving

. 'I
b
4

~ ·I1
- Q~ r-

I 11

I ~
I I •
,~
~

l,+,...., (C)
{DJ
Fig. 10-2. Feeding a half .. wave dipole by a coaxial cable:
a-feed of a symme trlcul dipole using a U-t'lbow: b-feed of Joop·ll ke dipole using
the U-elbo w: c-f~ed by mcan.~t of a metal Insulator.

along the internal surface of the external conductor of the

cable flows out of the cable, partly into the dipole and
partly branching off towards the outer surface of the externa I
conductor of. the cable. As a result, the currents in the
arms of the dipole are unequal. In addition, there arises
a radiation of the cable , on account of the currents which
branch off outside of it. This leads to a distortion of the
directional diagram. The role of the balancing U-elbow is
precisely to prevent branching of the currents. To this end,
the difference between the lengths of the segments of cable in
the U-elbow is taken equal to As /2 and the external coatings
of the segnaents of cable are shorted out at the ends. The
length of the segment of cable lt in the balancer is chosen
such as to enable the obtention of a travelling wave in the
main coaxial cable.
In the circuit of Fig. 10-2, a, the input resistance of the
antennas equals R.b=73 ohms, and the resistances loading
309
the segments of cable equal Ran=Rbo=36.5 ohms. If the wave
impedance of the cable forming the U-elbow is given a value
W =70 ohms, and the length of the segment of cable lt is
assumed equal to a quarter of a wave-length in the cable, the
resistance at the points c loading the main cable will be:

and the travelling-wave ratio in the main feeder will be close

to unity. ·
In the circuit of Fig. 10-2, b, the loop-like dipole has an
input resistance Rab~292.4 ohms and the resistances loading
the cable segments are equal to Rao=Rt,o~; 146.2 ohms. If
the length of the cable segment /1 is taken equal to zero,
the resistance loading the tnain cable is:

Re= ~I!!! :loo 73.1 ohms

and a travelling wave is set up again in the main cable.
Another balancer circuit is shown in Fig. 10-2, b. Here, two
tubes with a short-circuiting bridge forming a quarter-wave
metal insulator are joined to the sym•netrical dipole. A co-
axial cable is inserted through one of the tubes the external
conductor of which is connected to the tube at point a and
the internal conductor to the other tube at point h. This
eliminates the excitation of the external surface of the cable
and ensures that both halves of the dipole are symmetrically
fed. The resistance loading the cable is equal to Rab=73.1
ohms, and if we take a coaxial cable with a wave impedance
equaJ to 70 ohms, a traveJJiug wave wiJl be set up in it.
Note that here, the metal insulator plays also the role of a
stub, which backs off the reactive component of the input
resistance of the dipole and widens the antenna pass-band.
In the centimetre wave range, the half-wave symmetrical
dipole is fed by means of rigid coaxial lines. Here too,
special balancers are used, which eliminate the excitation of
the external surface of the coaxial line external conductor.
Fig. 10·3, a shows a feeding device of a half-wave dipole
in which the coaxial line is enclosed within a quarter-wave-
length tube short-connected with the coaxial line at one end~
The new coaxial line thus formed, which is often referred to
as a "quarter-wave" sleeve, represen~s a metal insulator,
810
which prevents the current from branching otf to the external
surface of the outer conductor of the coaxjalline and ensures
the symmetrical feeding of both halves of the dipole.
I~ another device (Fig. 10-3, b), the symmetrical dipole is
exctted by means of two longitudinal slots rnilled in the
outer conductor of the coaxial line. As we kno\v. a TEM '\\'ave
is propagated in the coaxial line, accompanied by lon~itudinal
currents. If longitudinal slots are mi Jled in the coax ia I line,
they will not be excited. However, if, after that, at the
place Where the slots ate situated, We ('Onncct the internal
~ ¥z-- . . .
r;.z-~., I
I I I I ~~
I ~~
I ..,.
\:==tilt LJ :=::==:~
I I r1
I 1I I -
I I I I
l --r L
•

(Q) .... ~

J
t!J) (C)
Fig. 10·3. Feeding a syrntnetrica l dipole by a coaxicAf cable:
a-feeding by means of a quarter-wave ~teeve: b- feeding by means of ll)ngJ.
t udlnal slots: c- feeding by means of a double coaxial stub.

conductor to the external conductor by means of a crosspiece,

i.e., if we short out the TEM wave, the current which
flows along this crosspiece will generate higher-order modes
and, in the first place an Hu wave in the coaxial line.
Although this mode will not be propagated along the line
(Aoper>A<- rtt 11,,=2nRmean), it is accompanied by the local
transverse components of the currents which intersect the
slots in the external conductor and, thereby, excite the ex-
ternal space. Consequently. on the external surface of the outer
conductor of the coaxial line, there arise transverse electric
currents which excite the symmetrical dipole connected at
that spot; this is seen from Fig. 10-3, b, which shows the cross
section of the line at the place where the dipole is connected to
it. The slots can be of any length, but the most intense exci·
tat ion of the dipoles occurs when the slot length is resonant,
311
f.t., when l ~ A./2. The coaxial line is shorted out at the end
and the distance from the end of the line to the short·
circuiting crosspiece is taken equal to approximately a quarter
of a wave-length.
Feeding a symmetrical dipole by means of a broadband
ba Ian cer is shown in Fig. 10-3, b. Here, the coaxial· line is
enclosed within a tube of length lt+l 2 shorted out at the
ends; the tu be has orifices for the passage of the arms of
the half-wave dipole. One of the Rrms of the dipole is connected
to the end of the outer conductor of the coaxial line and •

the other, to the internal conductor of the coaxial line,

\Vhich, becoming thicker at the length /2 is joined to the tube.
The overall length of the
balancing and con1pensating
~ ~ double coaxial stub is taken
~ ~ equal to half a wave-length
SIR'
and the ratio of the lengths
of the stubs 1, and 12 is
I• ~ chosen so as to ensure the
~
broadest possible pass-band.
A symmetrical (vertical)
dipole may be fortned by the
- continuation of the internal
(o) (IJ) conductor of a coaxial
Fig. 10-4. Feeding a vertical dipole line or cable and a quarter-
by a coaxial cable: wave-length sleeve, as
G-excltation by means of a quarter·
wave-length sleeve; b-excltatlon by
shown in Fig •
10 4 a In
• ' •
means of horizontal screening stubs. that case, the external
surface of the quarter-~ave·
length sleeve is excited and together with the quarter-wave ..
length projecting internaJ conduc.tor it forms a half-wave
dipole. In another case, instead of a quarter-wave-length
sleeve, four horizontal wires of A/4 length each are connected
to the external conductor of a coaxial line (Fig. 10-4,b). The
wires intercept the displacement currents branching off from
the vertical dipole and thus screen the external surface of
the coaxial ·Jine. In the horizontal wires, the currents flow·
towards one another and their radiation cancels out. In the
main, the radiation is determined by the vertical quarter-
wave stub. Of course, instead of wires, we can connect the disc
to the external wire of the cable, but this will not lead to
any signi-ficant change of the radiati.on and, at the 5ame ti.ine,
wiJJ complicate the construction of th~ antepa.
''•
812
 We have investigated a number of feed circuits of a shnple
dipole. There are many other feed circuits, but the ones
n1entioned above are the most frequently used in practice.

10-3. Simple Antennas of the Slot Type

In the first part of the book, we discussed the theory of
slot antennas lying on infinite plane conductors. Making use
of the duality principle and of the mirror method. we deter-
mined 'the directional diagrams of elementary slots. We
established in the first approximation the voltage distri·
bution law in a bilateral slot of finite length excited by a
concentrated mm f. We also showed that in the case of resonant
(half-wave) slots, this distribution did not depend on the
distribution of the exciting rnmf.
Thus, we defined the directional diagrams and radiation
conductivities of the half-wave slot lying on an infinite plane
screen, not l{)nly for a bilateral slot but also for a unilateral
slot excited by means of a waveguide or a hollow reso-
nator.
Knowing the voltage distribution in a resonant slot, \Ve
also defined by means of the eigenfunctions rnethod the di-
rectional diagrams and the radiation conductivity of a half-
wave longjtudinal slot lying on an infinite circular cylinder.
In literature, one can also find the strict solution of the
problem of the radiation of slots lying on bodies of regular
shape, such as an infinite eiJipticaJ cylinder, a sphere, prola-
te or oblate spheroids.
We found, in particular, the strict solution of the problern
of the radiation of a circular slot with a uniform voltage
distribution lying on a prolate perfectly conducting spheroid.
In real conditions, surfaces on which slot antennas are
located are finite and of irregular shape, so that no strict
solutions of the problem can be obtained and we have to seek
approximate solutions adequate for engineering purposes.
Usually, resonant half-wave slots are uti Used, so that the
voltage distribution in the slot is assumed sinusoidal. If,
at the same time, we have in mind sttrfaces whose linear
d·in1ensions are of the order of one wave-length or more and
whose curvature radii are of the same order, we can, as may
be deducted from the theory of the half-wave slot lying on
a circular cylinder, assume the radiation conductivity of
the slot 1o be equal to the radiation conductivity of a slot

'''
lying on an infinite plane screen, i.e., equal to Gstot =
=1,028 milliohms.
Moreover, if we take into consideration the minimum
dimensions of the resonator by means of which the half-wave
slot is being excited, then one such resonator may be forn1ed
by a segment of short-circuited rectangular waveguide with
an Hot mode of length equal to a quarter of a wave-length in
the waveguide. Such a resonator plays the role of a metal
insulator and its conductivity equals zero. Consequently, the

21
~-------21------~~

a
a
(IJ)

(a)
Fig. 10-5. Half -wave slot antenna with a quarter-wave resonator:
a-first method ol Inserting th~ coaxial cable Into the resonator: b-second method
oi Inserting the coaxial cable Into the resonator (longitudinal section).
-
input resistance of the half-wave slot antenna with a quarter-
wave resonator will, in the voltage antinode (in the centre
I
of the slot), equal Rslot = 0 slot =970 ohms.
In order to tnatch the input resistance of the antenna to
the wave impedance of the coaxial line \\~hich feeds the
antenna, the line should be connected closer to the edge of
the slot, at a distance l defined frem the expression
• t kl w
stn = Rslot '

where W is the wave impedance of the feed line;

2n .
k =A,;
l is the distance from the edge of the slot to the
point of connection of the line.
Thus, we arrive at the circuit of the simple slot antenna
shown in Fig. 10·5, a.
Designing considerations make it convenient to abut. the
resonator with one broad side agaill$t thE; screen and to
814
insert the feed line into the resonator through its other broad
side. The longitudinal section of such a device is shown in
Fig. 10-5, b.
As regards the directional diagrams, they di iTer, of course,
from those that are obtained when the slot is situated
in an infinite screen. It can be asserted that, owing to the
fact that the radiation along the slot is zero, the amount

t/1.==05; Hh_=0.5 t!A =fl.f; H/A. =Q2b

{0) (b)

l/A=I; H/A=I l/,t=t5; %=t5

(C) (d)
Fig. 10-6. Directional diagram of a slot
ant-enna in the E vector plane .

of diffraction in the plane of the slot (in the plane of the

magnetic vector) is onJy slight and the directional diagram
depends but to an insignificant extent on the dimensions of
the screen. In the front half-space, it approximately coincides
with the directional diagram of a slot milied in an infinite
screen; in the rear haJ f-space, there is a small lobe. In a plane
perpendicular to the slot (in the plane of the electric vector),
the directional diagram depends to a considerable extent on
the dimension of the screen 2L and depends only slightly
on the dinrension of the screen 2H. In that plane, the wave •
11$
diffraction effect at the edge of the screen is pronounced.
In that plane, the slot radiation is superimposed by the radi-
ation which appears when the waves are reflected from the edge
of the screen. i.e., due to the di tTraction of the radio wave
which leads, on the one hand, to the radiation oi energy
towards the shadow sidepl the screen and to the distartion of
the directional diagram on the irradiated side of the screen,

H/A=0.25 Hh=QJ/8
(a.l (IJ)

NA=O.J7S
(C)

Fig. 10· 7. Directional diagram of a slot an·

tenna in the H vector plane .
•
on the other. Due to the interference of the waves radiated
directly by the slot and, as a result of diffraction at the edges
of the screen, there appear radiation minimums and maxi-·
mums, whose number is all the greater as the dimension 2L
of the screen is larger in comparison with the wave-length.
Experimental directional diagrams of a slot antenna in
the electric vector plane depending on the dimensions· of the
screen are shown in dotted-line curves in Fig. 10-6. A& can
be seen, these curves do confirm what has been said ahove.
The theoretical curves obtained by mean~ o~an approximate
"i
lA

816
 theory are shown in the same figure. These curves have been
calculate~ from the distribution of the surface electric
current on the irradiated side of the screen, this distribution
being assumed to coincide with the distribution of the
surface electric current in an infinite screen (42).
Fig. 10-~ shows the theoretical· directional diagrams in
the magnetic vector plane, calculated on the assumption that
the dimension 2H is finite. and the dirnension 2L infinite.
In that case, the screen represents an infinite strip which
is a degenerated infinite ell i,ptical
cylinder. The theoretical calcu-
lations were based on the eigen-
functions method [43).
A half-wave slot antenna with
a resonator in the form of a quarter-
wave metal insulator, can be
placed on a segment of a circular
cylinder as shown in Fig. 10-8. A
cylinder of this kind may be afforded Fig. I0 ·8. Cross section
by, for example, the structure with of a circular cylinder
a quarter-wave reso-
of a tower or the fuselage of an nator and longitudinal
aircraft. The theory of this antenna slot:
has, on the whole. been expounded /-slot; 2- resonator.
in Chapter Five and its directional
diagrams and input conductivity related to the voltage
in the antinode shown in Figs. 5-14 and 5-16.
In practice, not only unilateral but also bilateral slots
lying on planes of limited dimensions are utilised. Thus, in
Fig. 10-5, a bilateral slot antenna is obtained when the
resonator is, removed. The directional diagrams are now
bilateral and approximately the sarne as in Figs. JQ.. 6 and 10-7
for the illuminated region of the screen.
As regards the input resistance of a bilateral slot antenna,
in the case of a sufficiently large screen in comparison
with the wave-length, it can be defined from the expressions
(2-54b) and (2-55b). For a resonant half-wave slot, it equaJs
R1n =485 ohms.
Fig. 10-9 shows the experimental t1.1rves of the input
resistance in the vicinity of the resonance of a bilateral slot
antenna with a sufficiently large screen. The length of the
slot 1=10 cm, and the width of the slot, equal to 2. 4, 6 and
8 cm. The measurements were taken in the 170-240 Mc/s
frequency range. It is seen from the curves that the resonant
•

811
 Q40 042 044 fl41 048 1150 Q52 054 1158
Dhm H
,.
600 q !
.
'J~ i
J• J ~
500
j 'I J
f '
~
••
I
I

400 fl
4-..... 1
J..,J_ 7ft~ •

I', '
JUO
I , &I
r/j T
I ~ ~
200
/ i "~ ~
~"
~
100
' I '-....._ .
T
0 1

-zoof70 f6(J

Fig. 10·9. Experimental curves of the input

resistance of a bilateral slot antenna. Width of
slot: 1-2 cm, 2·4 cm, 3-6 cm, 4·8 cm.

Ill
length oi the slot is so mew hat less than ha If a wave-length
and decreases as the \vidth of the slot increases.
The plane short-circuited dipole shown in Fig. 10-10
can also be regarded as a slot antenna; it is utilised as a
component of turnstile antennas of trarisrnitting television
stations (44 ].
This dipole represents essentially one half of a bilateral
slot antenna. It is evidE'nt that, in the case of sufficiently
large linear dimensions of the screen relatively to the wave·
length, the input resistances of such a dipole are approxi-
rnately twice as large as those shown in Fig. 10-9. Wht!n

~
~

-.
•t/
f - ' - ' c::s
•
2 - '

J I I I f
Fig. 10-10. Plane short· Fig. 10·1 I. Ring slot
circuited dipole: antenna:
/-slot; 2-- metal plane: 1- slut: 2-coaxlal rc,o·
-feeder. nator; 3-cable; 4-cylln·
der.

the scieen decreases, the volume and character of the

input resistance of the antenna change. B. V. Braude has
shown that the antenna with the \videst pas..'i-baud is
obtatned when the ratio of the dimensions of the screen a
and l is equal to a,'l = 1.5. For these dimensions and for
d/'A=0.09, lz=2 ern within the range 0.25< < 0.4, the i-
antenna input resistance changes insignificantly and
the antenna pass-band is quite wide in that region.
Ring slot antennae; arranged on a circular cylinder find
an application as non-directive antennas (Fig. I0- JI). Here,
the circular slot is executed by means of a resonator in the
shape of a quarter-wave metal sleeve, similar to the one
shown in Fig. 10-4, a. The resonator is fed by means of a
cable at the point a.
When the average length of the periphery of the resonator
n (r1 +r 2) is less than the operating wave-length, a TEM
819
mode is excited in it. The voltage in the slot is distributed
uniformly and we get an antenna si~ilar to a symmetrical
wire dipole, whose radiation depends on the length of tHe
exciting cylinder.
If the average radius of the resonator is greater than the
\Vave-Jength, i.e., n{rt +r 2) >A-, then, apart from the 1'EM
mode, an Hu mode is excited in the resonator the critical
wave-length of which is equal to. the average periphery
of the resonator. If, at the same time, the length l of tbe
resonator is taken equal to } , there is no excitatiOn of the
TEM mode, while the excitation does take place in the case
of the H u mode whose wave-length is defined by the expres-

l - (
A.
Acrlt
)2
to be distributed in accordance with the cosine law, transverse
components of electric currents wiJ I arise on the excited body
and the radiation of the antenna
will be substantially modified. Due
o to the transverse current com-
ponents, there will occur a radiation
along the axis of the cylinder. As
shown by theoretical analysis and
practice, such an antenna radiates
uniformly in all directions .. More-
over, the directional diagrams
Fig. 10-12. Excitation of depend but little on the length of
an antenna by a wave the excited cylinder.
H11 in a ring slot. Fig. 10-12 represents the feed
circuit- of an antenna at two dia-
metrically opposed points of the resonator. The voltage at
point a differs in phase from the volt~ge at point b by 180°,
so that, for any length of the resonator l, the voltage in the
slot will be distributed in accordance with the cosine law, as
shown in the figure. Here, the length of the resonator is
taken equal to A\\/4.

10-4. Construction of Simple Wldeband An·tennas

In the two preceding paragraphs we gave an account of
various simple wire and slot antennas. The present paragraph
deals with a number of questions. regarding the construction
\ I :-.;.

120
of wideband as \Veil as multiple-tuned antennas of a simple
type. ,
The magnitude of the frequency band in which each
antenna can be utilised is one of its important paran1eters;
it is determined by the degree of permanency of such charac-
teristics as the travelling-wave ratio in the feeder and the
directional diagrams in the major planes of the antenna.
Depending on the relative operating frequency band, all
modern antennas can be roughly divided into the followin~:
groups: ·a) narrowband antennas with a relative bandwidth
of less than 10%; b) wide band antennas with a relative be~nd­
width of from 10 to 50~6; rnultiple-tuned antennas, for \Vhich
the ratio of the upper and lower lin1its of the operatinf! band
exceeds 1.5 : 1.0.
The problem most frequently rnet with in practice is that
of designing wideband and multiple-tuned dipole antennas.
The operating band of these antennas is lin1ited n1ainly by
the frequen~y dependence of their input resistance. That is
why we shall, to begin with, dwell on the problem of designing
wideband antennas by matching them to the feeder. There
are two possible ways of increasing the ntatching band of an-
tennas: designing a wideband antenna with a barely changing
input resistance close to the wave impedance of the feeder, or
the application of wideband matching ntethods to rxisting
antennas. It follows from the theory of the wideband matching
of a line to a load dealt with in Paragraph 8-4, that such
a matching cannot in itself solve the probJen1 of designing
wideband antennas and represents but a certain correction
of the behaviour of the antenna input resistance, possible
in a limited fr~quency band. Hence, wideband matching plays
an auxiliary role, the main task in solving the problem of
designing wideband antennas being the design of radiators
\\·ith a barely changing input resistance in the frequency
band.
Let us elucidate the reasons of the frequency dependence of
the antenna input resistance. Assutne the antenna in the form
of some transition device accomplishing the tran~forn1atian of
the coupled electron1agnetic waves of tire feeder into freely
propagated space waves. It is evident that if this transit ion
device constitutes a sufficiently continuous and electrically
extended transition between the segments in which there are
only coupled and space waves, then the reflections which arise
upon the transformation of the modes are not high and can
21-2122 821
partly cancel out, on account of the phase difference between
theJn. The total reflection in the feeder which supplies such
an antenna is small and is not critical relatively to the oscil-,
lation frequency and the length of the transition device.
The input resistance of such an antenna depends but very
little on the frequency, whereas the existence in the antenna
of seg1nents with an abrupt change of dimensions gives rise
to pronounced reflections. If there ,jre several sources of
such pronounced reflection~ :1nd the distance between them
is commensurable with the wave-length, then even an insig-
Z
......,....,,r:( 'A~

'
I
/S
/
' '\
I \
I \
I \ I

.,..~~~~rw 'PII~ftiiiiPIII~IIflftllll!!~ __, ~n

(/J)

Fig. 10-13. Asymmetrical coni cat dipole:

a-circuit of dipole: b-equlvalent circuit.

nificant change of frequency will lead to a considerable

mutual phase difference between the individual reflected
waves and a pronounced frequency dependence of the input
resistance. Let us illustrate this on the example of the
asymmetrical conical dipole shown in Fig. 10-13, a.
As we know, a cone of infinite extension over a metal
surface constitutes a homogeneous transtnission line with TEM
mode and wave itnpedance Wc=l38 log cot 'i'/2l45]. In
the case of an infinitely long dipole, iJs input resistance is
not frequency dependent and equals the wave impedance.
In the case of a conical dipole of finite length, there appears
in it a reflected TEM wave; moreo\'er, one can consider the
source of the reflections to be the surface s, on which part of
the incident TEM wave is transformed into a sum of azimuth-
synlmetrical TM waves formed on the outside as well as
inside space. In accordance with the above, the equivalent
circuit of the dipole can be represented, as shown in Fig.
10-13, b, as a segment of homogeneous line with a wave
impedance W, and length L. loaded with .a certain complex
'•
822
resistanceZA which itself depends on Land the wave irnpedance
Wt.. Thus, the total re flection in the leeder is due tu the
presence of two partial re nections: frorn the surfaces (end of
the dipole) and from the apex of the cone (\\•hen Wl:;6W 1).
In the case of an asynunctrtcal d1 pole of d1 fferent forrn,
e.g., cylindrical, the equivalent circuit will cornprise a
scgrnent of, in this cast~, in:unuogeneous transn11sston line
\Vtth variablt! wave irnpedanccs for the ·rEl\\ \\'avl~:> along its
ll'ngth;. d istri bu ted re tlt'C-
tions fron1 each segrnent
\vill be added to the retl<.~c­
ttons fron1 the end and
beginning of this segrnent.
The ruost significant retlec- -·-
tion fron1 the end of the
dipole can be checked -
ei tht~r by con1 pen sating it
through another reflection I
or by reducing it directly 0
through the appropriate 11+11 ~ 0.25A
cho1ee of the form of the Fi~. 10-14. Asy•nn1ctrical sle<~ve­
dtpole. The idea regarding stub dipole and its current distri·
the cornpcnsation of the bu lion.
end re llect ion can be i 11 us-
tratcd by rneans of the exan1ple of the asytnluctrJcal
sleeve-stub dipole with a raised feed point shown in FiJ,!. 10-14.
In a dipole of this k1nd, the current wave reflected fron1
the upper end is cornpensated at the feed point by another
current wave of opposite pha~e arising fron1 the reflection
oiT the metal,plane at the low end of the dipole. In other
words. tht:» capacitive reactance of the upper half of the dipole
is cornpensated by the inductive reststance of the lower half
of the dipole, connected in series with it. A satisfactory
cornpen~ation can be obtained only through the correct choice
of the geornetr1cal dirnensions and in a li1nited frequency
band. not exceeding J0-15 '}'o .
To continue, returning to the conical dipole in Fig. 10·13
and constdering that its length L is not·· inferior to 0.2-0.25
of the \\,.ave-length, let us investigate the inftuence of the
angle at the apex of the cone ..p on the total reflection coef·
ficient in the feeder. As the angle 11' increases from zero, to
begin with. there is a decrease of the re fleet ion coefficient
fro•n the base of the dipole due to the rapid fall of the wave
21· 823
impedance and its approximation to the quantity ZA. As for
the quantity ZA, it remains almost constant owing to the
fact that the main role in the radiation is played by the
equatorial regions of the spheres (it is through them that the
largest part of the radiated power comes out into the
surrounding space). In a certain region of the mean values of
the angle 1J>, the coefficient of reflection from the end of the
dipole falls down to its stnallest valye; moreover, the wave

aB~--~~~--~----+---~

02t---+ ---+--... .;

0 at a2 o.J a4 as
Fig. 10-15. Coefficient of reflection in a
conical dipole with an angle'\l'=l5 ° plotted
as a function of the ratio -~ .

impedance of the cone approaches the wave hnpedances of the

lines generally used (50-75 ohms). A further increase of the
angle ..p leads to the increase of the coefficient of reflection
from the end, due to the fact that the radiation occurs through
a narro\ving ring-like slot and the resistance lA acquires
a large reactive part . Thus, there is a certain average opti-
mum n1agnitude of the wave impedance of the dipole for
which the reflection frorn the base and that from the apex of
the dipole are at a minirnum. .
Let us now examine the dependence of the coefficient
of reflection at the base of the dipole on the frequency or,
which is absolutely the san1e, on the length of the dipole. The
typical curve of the moduJus of the coefficient of reflection
IPI plotted as a function of the ratio of the dipole length to
the wave-length LIA obtained experimentally l46) is shown
in Fig. 10-15. It follows from this graph that IpI becomes
'

324
sufficiently small when the length of tht' dipole exceeds
0.2-0.3A. Such a dependence of the --·nJodulus of the
coefficient of reHection is explained hy the fact that as
LIA increases, the ratio of the transverse corr1ponents
of the E and H vectors in the space T~\ waves t~nds
towards thE:' value of the wave impedance of the TEM
\\'aves in air equal to l'''"f-1 0 /11 0 = 120 n. ohn1s. It should
be notPd that in thin (for exantplc '1'<:-l ') or too
••

50

40 ' f/J=Jo· I
3.0
\
20
- ..._
"'\ -
I -r '
.....~
• /
tO l/:4.=0.875
L/1
i

0.2 04 060Bt0 !4 t8
(a)
{b)
Fig. 10·16. Standing \\'(1\"C ratio and rlirecUonal die~grdlltS or a conical
d i pule ('I~=: 30°):
a-experimental curve of th(• SWC Jllotted as a iunct Ion nf the rnt \o !r, b-expc:rl·
mentat dln•ctlonal diagrams

thi\:k \for example ~'>80") conical dipoles, there occurs a

pronounced reflection frorn the end e\'en for significant
lengths of the' arms, due to the abarp di fl"erence between the
\\'ave impedance of the dipole and the equivalent loading
resistance ZA. Thus, a conical dipole with an optimurn wave
impedance may fJe regarded as a 1nulti pie-tuned radiator
from the point of view of the input resistance on condition
that L>0.2 Amax• where Amaa is the maximun1 wave-length of
the operating range of the antenna. Fig. 10-16, a shows the
experimental characteristic or the standing wave ratio in a
coaxial f~der with a wave impedance W,=50 ohms for one
of the designs of the conical dipole. lt can be seen from the
figure that such an antenna is satisfactorily matched in a
sevenfold frequency range. However, the directional dia·
grams of this coni ea! dipole change substantially already when
8ZS
L >0 . 75A (Fig. 10-16, b), which is due to the change of the
electrical dimension of the dipole when the frequency
changes.
Another example of an antenna with wideband properties
with respect to the input resistance is the disc-conical radiator
shown in Fig. 10-17, a, which consists of a cone and a disc
8.0 ....
swc -
7.0
6.0
5.0
4.0
9.0
2.0
f.O "= O/A
0 at5 t130 0.45 010 am
(IJJ

ajA=Qf/5 O/!l=Q280 o/Jt=Q450

(C)
Ffg. 10-17. Disc-conical radiator:
a-c ircu it of radi ator : I-eone~ 2-d lsc; h-ex perl men tal curv e of the SWC plot ted
as a func tIon of the ratio CJ{A; r-PY f"Pr i mP.ut al cllrectiona! diag ram s.
~·

between which an exciting voltage is appli~d with a coaxial

feeder. The con~ is connected to the outer cda t i ng of the feeder,
the disc, to the inner conductor. For a wave impedance of the
feeder of 50 to 75 ohms, tht disc-conical radiator is matched
up to K =0 .6 at all frequencies exceeding a certain limit
frequency determined by its geometrical dimensions (46].
Figs. 10-17. b and 10-17, c show the characteristics obtained
experimentally of the matching of a disc-conical antenna
with dimensions a/'Amax =0 .17 5, b/'Amax =0 . 25, c/'Arnax =0 . 275 to
a coaxial feeder whose wave impedance equals 50 ohms, as
well as the directional diagrams on a number of frequencies.
'126
A study of these figures reveals that the operating rarige of
this antenna is also limited by the frequency· dependence of
the directional diagrams.
Analysing aJI that has been said regarding antt"nnas of
wideband properties with respect to the input resistance, the
following conclusions can be drawn: 1) such antennas can
only be constructed with dimensions comparable to the maxi-
mum length of the operating range wave; 2) their directional
diagrams have a pronounced frequency dependence due to the
fact that frequency changes lead to changes of the electric
lengths of the radiator elements.
Owing to the fact that dipole antennas of low directivity
with good matching characteristics and constant directional
diagrams in a wide frequency range are required for many
practical radio-engineering purposes, we shaJJ investigate a
number of fundamental considerations which can serve as a
basis for designing such antennas.
A starting point can be afforded by the electric model-
ling principle, which establishes the identity of the input
resistance and directional diagrams of two separate antennas
in two different frequencies if the shape of the a'ntennas is the
same and the dimensions change proportionately to the ratio
of the frequencies. In the case of antennas of infinite ex-
tension, whose shape is defined only by the anguJar dirnensions
(e.g., for conical dipoles), a change of scale proportional to
the change of frequency does not lead to changes in the an·
tenna. That is why antennas which are defined only by angular
dimensions have the parameters of frequency-independent
antennas. The construction of finite antennas defined by angu-
lar dimensions is impossible. However, one may attempt to
find such forms of infinite stru~tures as are defined only by
angular dimensions, the behaviour of the end part of which
will, above a certain limit frequency, tend asymptotically
towards the behayiour of the corresponding infinite structure
from the point of view of the constancy of the input
resistance as weiJ as from the point of view of the constancy
of the directional diagrams.
With regard to plane antennas, the cla~ of curves out-
lining infinite structures defined only by angular dimensions
can be found in the following way ( 47 ]. Let us introduce
the polar coordinates r, cp and cause the curve r(q>)
which limits the form of the plane antenna to pass into
itself in the event of changes of scale along r·and to di tier from
8Z7
 •

the original curve only by a turn of a certain angle ~elative1y

to the axis of the coordinates. Mathematically, tbis can be
written as follows:
Kr (<p)=r (q>+ y),
where K is the coefficient of chang~ of sea le; y, the angle
of rotation of the whole or1ginal curve, corresponding to
such a change of scale. It is obvious_ that the angle y
should depend on K, but neither K nor y should depend
on r and cp. This condition leads to the following relations:
( ) dK- iJr ('P+Y). (10-1)
r cp dY- or '
K dr (q>) = ar (q> + Y) ( 10-2)
dcp d<p ,

moreover, the following equa 1ity shou Id a Iso be satisfied:

dr (q> + y) or (cp + y) dr (q> +V) ( 10-3)
ay = (f[cp +v> = aq> • .

Taking account of (10-3), we shall obtain fro1n (10-1)

and (10-2) the following diflerential equation for the
function r (cp):
r (()') dt!!S. = K d~ (r.p) or ~d' = ar (I 0-4)
V q> ~

where a=f ~ is a certain new constant. The solution

of the ~quation (10-4) is represented by· a logarithmic
(equiangu Jar) helix
, (q>) = Aea '<P +'~>a, •
In this ex press ion, A anJ~ cpo represent certain constant
·~quantities defining the scale in the r-direc(ion and the origi-
na1 angle of rotation of the winding. The parameter a is
equal to the contangent of the angle 'l't between the radius r
and the tangent to the curve.
Thus, we may conclude that the shape of plane infinite
frequency-independent antenna~ should be limited by
logarithmic helixes having the same rnagnitude of the
parameter a. The latter is indispensable to avoid. the
overlapping of separate parts of the antenna at the origin of
the coordinates where the excitation should take place. An
example of such a structure is shown in· Fig.,, 10·18.
a2B
 A similar analysis for finding out the class of frequency-
independent infinite structures can be performed in a three-
dimensional space in the spherical coordinates r, 0, q>
taking into consideration the condition
Kt (8, <p)= I (0', cp'), (10-5)
where f (8, q>) is th~ original shape of the antenna; I (8', rp')
corresponds to a rotation of the antenna in space, equiva-
lent to a change of the dimensions by K times. It is

Fig. I0-18. Infinite structure defined by

angular dirnensions.

readily seen that the condition (10-5) is satisfied by a

sol uti on in the following form:
r= ea tcp + Q>u, F (8),
where F (8) represents any function, and the direction of
the change of scale should coincide with the axis of the
spherical coordinates.
Of great importance for the proper understanding of th~
operation of an antenna formed by the finite part of the
structure shown in Fig. 10-18 is the following experimentally
discovered fact: the intensity of the current in the arm of the
antenna, when it is excited in the centre b~ an emf generator,
sharply decreases (lOO times and more) after it has passed
that turn of the structure whose periphery is approximately
equal to the wave-length. It is precisely due to this peculiar
automatic "cut-off" of the current at the spot, defined by the
frequency of the exciting generator, that the electric dimension
of the radiating part of the antenna is maintained constant
829
on different frequencies. As a result, such an antenna of finite
dimensions can have an almost constant input resistance
and the same directional diagram in a twentyfold and
even larger frequency range. The lower limit of the operating
range is defin~d from the condition of the equality of the
wave-length to the periphery of the last turn of the antenna.
The upper Jimit of the operating range is determined by
the commensurability with the wave-length of the central
region of excitation of the antenna where the geometry of the
helical lines is disturbed due to the existence of a connection
with the feed Jine. It is a Iso import ant to observe the following
conditions with regard to plane antennas defined by angu-
lar dimensions: the form of the metal part of the antenna
should coincide with that of its slot (additional) part.
It is clear that a plane antenna defined by the angular
dimensions can be simultaneously regarded as an electric as
well as a magnetic (i.e., "slot") antenna. Let Zt be the
input resistance of the plane antenna and note that in ac-
cordance with the duality principle, an interchange of the
metal and "slot" parts of the antenna leads to a change of
its input resistance, which becomes equal to Z2 , this being
expressed as (see Paragraph 2-9):
- (601t)l
Z.- z .
I

On condition that the forms of the electric and magnetic

parts of the antenna coincide, we get the equality Z1 =Z 2
and the input resistance of such an antenna is equal to Z1 =
Z2 =Zin=60n: ohms on any frequency. In antennas of finite
dimensions, this condition regarding the coincidence of forms
is, in a sense, equivalent to the concept of the optimum wave
impedance of the transmission line formed by the arms of a
plane antenna and constitutes an additionat guarantee that
the latter is multiple-tuned. An example of the fulfilment of
this condition is shown in Fig. 10-19.
Fig. 10-20 shows the experimentally investigated con-
struction of a multiple-tuned plane antenna l48 ], whose arms
are bounded by four logarithmic helixes defined by the
equations:
r,=Aeacp, r 1 =Mr,, r 1 =Aet~<cp-n>, r.,=Mr,.
This antenna, which is cut out of a metal sheet of large
dimensions, is fed by means of a flexible coaxial cable placed
••

830
along one of the arms of the antenna. When the number of
windings equals 1.5, the directional diagram· of the antenna
consists of two wide lobes whose maximums are perpendicular
to the plane of the antenna. The radiated field in the n1ain
directions has, in the operating frequency range 20 : l,
an elliptical polarisation with a coefficient of eJlipticity
exceeding I : 2.
Owing to the fact that for an antenna of this kind, a change
of frequency is equivalent to its rotation in space relatively

Fig. 10-19. Plane antenna of

infinite di men si ons.

to the axis perpend icu Jar to

the pI a ne of the sheet, the Fig. 10-20. Equiangular helical
width of the lobes at half- antenna in a rnetal sheet of
power oscillates periodi- large dirnensions.
cally from 40 to 50° in
the wholE' of the operating range. The travelling-
wave ratio in the feeder with a wave impedance W 1=50
ohms constitutes no less than 50% on all the frequencies of
the operating range. A change of the parameters a and M
has no significant influence on the multi ple-tunin~ properties
of the antenna and leads mainly to a change of the average
magnitude of tts input resistance. However. the best results
are obtained when a=0.30 and M =0.62.
The idea of an automatic cut-off of the radiating current
at the spot determined by the frequency of the generator has
also been put to use in plane antennas with a "logar•thm1c
peri<;>dicity" of the parameters as a function of the frequency
(49-50 ). The form of such an antenna 1s shown in Fig. 10-21.
The antenna arms consist of peculiar plane structures with
alternating depressions and projections. The geometry of the
structures ts characterised by relative period t - RR: 1
,

881
coefficient of form a = ~: and angles a. and p. In the
structures genera11y used, the condition a= "VT'is fulfilled.
At the origin of the coordinates, the arms of th~ antenna
are divided by a gap to which the exciting voltage is applied.
As a feeder, one can utilise a twin-wire symmetrical line,
as well as a coaxial cable. In the latter case, as shown in
Fig. 10-21, the cab1e is coated only along one arm of the
antenna and is electrically con-
V nccted with it along its entire
length. The inner conductor of the
coaxial ea ble is connected to the
beginning of the other arm of the
antenna.
It has been experimentally
established that the magnitude
~-----.~~:&~ of the current in the arms of
such an antenna increases away
from the centre, reaching its
maximum at the place where the
resonant project ions and depres-
sions of a pprox ima tely quarter-
wave-length are located.
Fig. 10-21. Plane antenna B d th · t f t
with logarithmic periodicity eyon ese potn s o resonan
of paratneters as a function inhomogeneities, the rnagni tu de
of the frequency of the oscil- of the current sharply faJJs by
lations. 100 and more timt's. A decrease
of frequency of the radiated
oscillations lt~ads to the successive appearance of the
resonances of the inhomogeneities that are more distant from
the centre. It is readily seen that the ratio of the two frequen-
cies ft </2 for which the resonances of the neighbouring
inhomogeneities occur constitutes l~xactly the quantity 't=
= ftlf 2• Accordingly, the directional diagrams and the antenna
input resistance are found to be the periodical function of the
logarithm of the oscillation frequency. The change of the an·
tenna characteristics vi thin the limits of one period is not large
granted the appropriate choice of the parameters of the struc-
ture. The exact repetition of these changes fron1 period to period
ensures the satisfactory behaviour of the antenna character-
istics within a considerable frequency range. The limits of the
operating range are determined by the resonant frequencies
of the extreme inhomogeneitie~ (i.e .• those that are closest to
882
and farthest from the centre of the antenna)~ A magnitude of
the operating range of 10 : 1 and more can easily be achieved.
The low dependence of the antenna input resistance on the
frequency within the limits of one period can be explained
as follows. The resonant projections and depressions on the
antenna arms are in fact the electric and magnetic dipoles,
the coefficients of the reflection from which are approxi-
mately equal in n1agnitude and opposite in phase. As a result,
the reflected waves cancel out at the feed point. The sarne
phenomenon was observed during the investigation of a
sleeve-stub dipole (Fig. 10-14) with a raised feed point.
The average value of the antenna input resistance with
a logarithmic periodicity of th~ parameters represented in
Fig.l0-21 usually exceeds 100 o~ms and depends on the choice
of the angles a and ~- In the event of the equality a=-P=
=45'\ the form of the metal part of the antenna coincides
\vith that of its slot part and, in accordance with the expres-
sion Z1 ·Z 2 ='(60n) 2 • its input resistance should amount to
188 ohms. Measurements made with an experimental model
of such an antenna gives a somewhat smaller value ("-150
ohms). A further decrease of the average value of the input
resistance, desirable from the point of view of a better match-
ing of the antenna to the coaxial cables generally used. can be
obtained by decreasing the angle ~. but this leads to a greater
unevenness of thP antenna input resistance within the limits
of one period.
The directional diagrams of the antenna represent two
wide lubes orientated perpendicularly to the plane of the
arn1s. The radiation in the antenna plane ts low for all
directions. A cbaracteristic feature is that, in the maxin1um
radiation direction, the vector E is polarised in a direction
parallel to the x-axis, the radiation with a polarasation
parallel to the y-axis being small. The width of the lobes
of the directional diagram depends to a great extent on the
relative period of the structLre 1. For example, when the
parameter 't changes from 0.81 to 0.25 in an antenna in
which angles a=P=45°, the width of !he lobes at half-
power from 73 to 38° graduaiJy decreases. When the arms
of the antenna shown in Fig. 10-21 are rotated about the
x-axis by an angle less than 180°, a unidirectional radiation
can be obtained in the direction of the acuteness thus
formed. At the present time antennas with a logarithmic
periodicity of the parameters are being subjected to intensive
388
research and improvements. However, they are already
used quite extensively as individual antennas, multi pie-
tuned feeds of parabolic and lens antennas, antennas for
radio communications on short waves, etc. [51 and 52).

10-5. Multi-Unit Antennas

In the preceding two paragraphs, we investigated simple
antennas of the wire and slot types. We shall now investi·
gate multi-unit antennas of the highly directional as well
as non-dir_ectional types.
To begin with, let us examine a eo-phased antenna con-
sisting of an array of half-wave wire dipoles. The theory of
A the eo-phased array of
~21 dipoles was investigated in
~: ~~ / .. ..
4, ' - - Chapter Four. Here we shall
~'2 ~ investigate the feed circuit
of such an antenna, its con-
... .. ... ...
- -J..
- - - struction and application.
>< )C Fig. 10-22 shows the most
... ..
~ - _ ..
- -
. comrnonly used circuit of
eo-phased feed of an array
)< of dipoles. The half-wave
.. dipoles are arranged in one
A
I""' -
vertica I plane with their
6 centres spaced half a wave-
t, t, length apart and are con-
nected in pairs to a twin
Fig. 10-22. eo-phased antenna. vertical feeder, forming a
so-callerl section. For the
dipoles to be fed in eo-phase, the wires of the feeder should
intersect between the stages. The two segn1ents shown in
Fig. 10-22 are connected by the feeder ·in parallel. The
arrows show the direction of the currents in the dipoles
at a fixed instant. The dipoles are taken somewhat shorter
than half a wave·length, so that they should be tuned in
resonance.
The input resistance of each pair of dipoles can be defined
by the approximate expression (2-50):

R =- -~!mean (2-50)
a Rl:
J34
where W w is the wave impedance of each pair of dipoles,
which can be defined by (2-43Y,-
Rl: mean , the radiation resistance of each dipole,
defined by (4-18) and averaged for all the dipoles
of the antenna so as to simplify the calculations.
Because each pair of dipoles is connected to the feeder
at a distan~ of half a wave-length from one another and
the length of the feeder 11, which connects the segments,
is usually taken a multiple of half a wave-length, the antenna
input resistance is expressed as:

R -- w~
e- 2nRr. mean '

where n is the nun1ber of stages in the antenna.

The wave impedance Ww is usuaiiy of the order of a few
hundreds of ohms and the radiation resistance RE. mean of the
order of 70 to 80 ohms. Hence, the antenna input resistance
Re, loading the antenna feeder is close to the wave
impedance of the feeder and the type of operation obtained in
the latter is cJose to the travelling-wave type, without the
use of any tuning elements.
The antenna radiates the maximum of energy In a
direction perpendicular to its plan~. The directional diagrams
of the antenna can be calculated from (4-8) and (4-9).
In order to get rid of reverse radiation and raise the
directive gain, a reflector is generally used, placed at a
distance of approximately a quarter of a wave-length from
the antenna. It can be a tunable system, i.e., exactly like
the antenna system shown in Fig. 10-22. In that case, it is
made passive and tuned by means of a short-circuited stub
connected to the points e of the reflector. In the metre
wave range, e.g., for radar purposes, it is convenient to
use a reflector consisting of a wire-net, the dimensions of
which are not inferior to those of the active antenna system.
Such a reflector is aperiodical and requires no tuning. In
the decimetre wave range, the reflector can be in the form
of a r•gid metal plane. This reflector is convenient from the
constructional point of view because the active antenna
system can be fastened to it.
Fig. 10-23 represents the photograph of a decimetre wave
antenna with a solid reflector. The antenna is mounted in
an aluminium box, the back wall of which plays the role of
us
 the re Hector. the active
antenna system being
fixed to it by means of
quarter-wave metal in-
sulators. The insulators
are ·rods connected to
the dipoles at the
voltage ttodes. In order to
widen the antenna
pass-band, the dipoles are
made with a low wave
irnpedance in the shape
of laths 17-17.5 cm long
and 5-6 cm wide. The an-
tenna has five stages and
two sections. The sections
are formed by means
of a twin feeder and con-
nected with one another
by a single wire tubular
the
line placed above the
Fig. 10-23. A photograph of
eo-phased decimetre wa-ve antenna. screen (the second wire
of the line is its mirror
image in the screen). The transition from the single
to the twin feeder is effected by means of a single-wire
U-clbow, connected to the sections in their middle. The
single-wire line is fed by a coaxial cable the coating of
which is connected to the screen. The type of operation
established in the cable is close to the travelling-wave type
in the whole band of the operating frequencies (K>O. 7).
The coaxial cable has a wave impedance of 70 ohms and
its attenuatio n equals 1.9 db per 1 km on the 50 cm wave.
The antenna is placed on special supports and is protected
from weathering by a thin radiotransparent lid.
Atnong the non-directional multi-unit antennas, mention
should be made of the turnstile antenna utilised in tele-
casting stations. Television antennas rnust satisfy specific
requirements.
First of all. the antennas should radiate uniformly in
all directions in the horizontal plane since television trans-
mitters are usually s1tuated in the centre of the region
which they serve. In the vertical plane, the directivity
should be sufficiently large in order that the radiation
836
 should be concentrated in the horizontal plane. Moreover,
the antenna should radiate electromagnetic "energy with a
horizontal £-field polarisation, due to the fact that the
sources of interference on ultrashort waves are industrial
installations of all kinds whieh radiate an electromagnetic
field mainly with a vertical polarisation.
An important requirement is that concerning the pass-band
of the antenna system. The antenna together with the feeder
should possess a sufficiently wide pass-band in order to
avoid distortions of the transmitted image. Furtherrnore,
the antenna-feeder system should be free from feeder echo.
which gives rise to a ghost image. The echo wil I be insigni fi-
cant if the wave reflection from the
antenna is negligibly sma~l in the
entire pass-band.
The antenna should also be reliable
from the mechanical point of view 7
because it has Jto be placed on very
high supports where it is exposed to
high winds and quite often to icing. •

It should also be Jightning-proof, this

being achieved through the direct ;>. /
grounding of the antenna elements.
In order to obtain a rAdiation with Fig. 10-24. Arrange-
a horizontal polarisation, untform in ment of t·ross-shaped
all directions in a horizontal plane, elements in television
antenna.
use is made of the cross-shaped
elements described in Paragraph 1-7,
which radiates a rotating field. Such elements. which, in
the siJnplest case. consist of half-wave dipoles, are placed
horizontally in several stages spaced 'A/2 apart and fed
in eo-phase (Fig. 10-24). In the case of an even number
of elements, the radiation in the direction of the axis
of the system (up and down) is zero and the maximum
of radiation occurs in the horizontal plane. The dipoles
situated in one vertical plane are fed by one feeder and
those situated in the other vertical plane are fed by
another feeder. The feeders are tuned on atravelling wave
and connected in parallel; moreover, one feeder is a quarter
of a wave-length longer than the other, thus ensuring a
90° phase shift of the currents in the dipoles of the cross.
The antenna circuit, usually called a turnstile antenna,
is shown in Fig. 10-25.
22-2122 881
 In order to reduce the reflection from the antenna in the
transmission band to a negligibly low level, the dipoles
should have a wide pass-band, i.e., their input resistance
In the frequency band should chanRe but insignificantly .
...
- 334 - 1•286; I

' , n
" /

)( ><
8
, ~

_)
,' ~
~ '"
~ '114- .....
'

' ~ 11

tt,t lffuiJe:S
~,
-tO I I
=:

fig. 10-25. Feed circuit of Fig. 10-26. Plane Braude antenna.

turnstile antenna.
The plane dipole described in Paragraph 10-3 is of this
type. In order to reduce the resistance to wind and
strengthen the dipole, it is made of several tubes. A dipole
of this kind is shown in Fig. 10-26. The antenna system
consists of 11 tubes 10 mm in diametrr; its dimensions for
180 R,(K
160 d.. ~
140 \;:
f20
fOO
" ~
~ -.....
80
60
40
20 ~ Nc/..f
-zo
0
f50~ 200 -~ aa'IJ
-40 ,I X
•
-60
-80 11LA -
-tOO L-- .. L c - - ...J --

0225 QJOO Q975 Q450

Fig. 10-27. Input resistance of d plane antenna.

operating in a 155 to 180Mc/s frequency band are: a=460 mm

and l=286 mm. The dipole is short-circuited to the support
at the points BB and insulated from it at the points AA.
Fig. 10-27 shows the experimental curves of the antenna
input resistance. It is seen that, in th~ 155,-180Mc/s frequency
188
band, it is almost constant and equal to 140 ohn1s. Hence,
it is found convenient to feed this antenna by means of a
coaxial cable with a wave resistance of the order of 70 ohms.
Two such antennas arranged crosswise and fed with a 90~
phase displacement form a turnstile
element. The antenna is light, so that
it can be of the stacked type. In
the case of three turnstile elements
spaced 'A/2 apart, the directive gain
of the antenna equals 3.2.
A plane antenna of this kind has
been devised by the American firm
RCA (53]. An elen1ent of this antenna
is shown in I:ig. 10-28. It is short-
circuited to the pole at the points BB
as well as AA and fed in the middle
at the points CC. Its input resistance Fig. 10-28. RCA plane
in a wide frequency band is of the television antenna.
order of 150 ohms.
Now, let us examine the multislot an-
tenna disposed on a circular vertica I tube
and sho\vn in Fig. 10-29. It consists of four
longitudinal slots milled in a tube iv/8 in
diatneter spaced at a distance A apart.
The slots are 3/4A. long each and are fed
by a coaxial cable.
In this case, the length of the slot, equal
to 3/4A, is found to be resonant due to the
fact that the slot is shunted by the inner
cavity of the tube, which, owing to the
relatively small diameter of the tube with
respect to the wave-leng th, represents an
Jnductive resistance. In order that the
resultant input resistance of the slot should
be purely active, the natural (external)
resistance of the slot should be of a capaci·
Fig. 10-29.
Cylindrical tive nature. This, as we kn~w, occurs
$lot antenna: for a length of the slo-t higher than 'A/2.
/-slot; 2- .. cable The increase of the resonant length of the
slot can also be explained in a somewhat
different way. Let there be a twin line in free space. Then,
its wave has a phase velocity equal to that of light. If
this line is shunted with inductances, the phase velocity
339
of the wave in the Hne increases, and all the more so as the
shunting inductances are smaller. An increase of the number
~f shunts eventuaJJy leads to the formation of a tube with
a longitudinal slot.
The antenna radiates a horizontal polarised field and its
directional diagram in the horizontal plane has a low
directivity (Fig. 5-15). In the vertical plane, the antenna
is highly directive. Such an antenna is use.ful for broadcasting
on ultrashort waves.

10-6. Slot Waveguide Antennas

In the centimetre wave range, wide use is made of slot

waveguide antennas. They are mainly utilised in radar
systerns. The slots are usually cut in the broad wall of a
rectangular waveguide and an H 01 tnode used for their ex-
citation. The slots are made resonant, i.e., of a length
equal to approximately half a wave-length and disposed
in such a way that their excitation should be in eo-phase.
In this way, they form a eo-phased array and radiate
maxin1um energy in a direction perpendicular to the plane
of the waveguide.

(a) (D)
Fig. 10·30. Resonant slot waveguide antenna:
a-antenna with transvciSe slots: b-antenna with longlt udlnal slots.

Fig. 10-30, a represents a slot waveguide antenna con-

sisting of transverse slots in a waveguide spaced one wave·
length apart. Fig. 10-30, b represents a slot waveguide
antenna consisting of longitudinal slots arranged in stag·
gered rows on both sides of the waveguide axis and spaced
half a wave-length apart in the waveguide. The end of the
waveguide is short-circuited by a plunger, th~ last transverse
•

840
and longitudinal slots lying at distances of Aw/2 and Aw/4
respectively from the plunger. ,
Due to the presence of the short-circui ting plungers, the
waveguide has a standing wave type of operation and the
antennas are therefore known as resonant slot antennas.
Let us dwell in greater detail on the question of the
excitation of the slots. As we know from expressions (7-16)
and (7-17), in the absence of slots; the field of the Hof wave
in a waveguide short-circuited at one end is expressed as:
Ex 01 =C i~·e sin ( f y) sin (a. z);
1

HJOI =Ca01 sin ( 7J-.y) cos (a01 Z); I ( 10-6)

H zo 1 = - C ~-cos ( ~ y) sin (a 01 Z), J

where C is a constant quantity, independent of the co-
ordinates
•

It is seen frotn the expressions ( 10-6) that the tra ns-

verse slots are situated at the points Z0 Aw, I ~· "-w•
=}
2 { Aw, ••. , i.e .. in the anti nodes of the transverse component
of the magnetic field intensity, the longitudinal slots
being disposed at the points Z0 ={Aw, t{Aw, 2{Aw'"'
when y < : and at the points Z0 = ! A..,, I { Aw, 2 }"-w• •••
when y> : , i.e., at the antinodes of the longitudinal
component of the magnetic field intensity. At the points
indicated. the fields are of equal magnitude and phase, so
that the slots are excited in eo-phase and with equal
amp I i tu des.
The intensity of excitation of the slot. depends on their
location relatively to the waveguide axis. Thus, if the
longitudinal slots are disposed in the middle of the broad
wall of the waveguide (y0 = t>.
Hz 01 =0 and the slots are
not excited. The maximum excitation of the longitudinal
slots occurs when y0 =0 and Yo=b, in particular when they
.941
are situated on the Jateral walls of the waveguide, and the
maximum excitation of the transverse slots occurs when
y0 =-%-· In other words, the most intense excitation of the
slots occurs at those points of the waveguide at which the
con1ponents of the surface electric currents, which intersect
the slots. have their rnaximum value.
Let us now turn our attention tu thejact that the slots
in the waveguide can be regarded as receiving elements.
They receive energy propagated along the waveguide. of
the H01 wave. Part of the received energy is radiated by
thern into the surrounding space and the other part (in
the absence of losses) is radiated into the inner space.
Considering a slot waveguide antenna fron1 this point of
view, let us, for simplicity, imagine an infinitely long
waveguide with one single half-wave transverse or
longitudinal slot. Let this slot be excited by an Hot wave
propagated in the z-direction. The electromagnetic field of
the wave is written as:

(10-7)

Re~arding the slot as a magnetic receiving dipole, we can

find the magnitude of the magnetic current induced in it.
in a \\tay simi Jar to that which was adop1ed when investi-
gating the receiving t'lPrtric dipole in Chapter Six. i.e.,
fron1 the expression
• (10-8)

where M is the equivalent mmf induced in the slot;

a~. the external conductivity of the slot radiation;
G. the internal conductivity of the slot radiation.
For a transverse slot (Fig. 10-31), the mmf is expressed as:
11=11o +AI•
M= ~ H!~c1 cosk(y,-y)dy.
ll•llo-At•

142
 Substituting here the expression Hyat from (10-7),
performing the integration and substituting into (10-8),
we obtain the expression of the voltage (of the n1agnetic
current) a pp Iied in the centre of the slot in the
following form:
/ 110 = G 1
M 2k B cos ( -n -A \l ~In n y ) e-tanlfo
• ( ---
.z + G mJ.t.U01 b 4 ) ~ b 0 •
(10-9)

The reverse radiation of the slot into the wavegu ide

(field reflected from the slot) can be defined fron1 (7 -28).
The electri c field in the region z < Z0 is expressed as:

E 0
XOl == 2k I ~0 n -
cos ( - A. ) Sin
• ( --
n y ) Sin
. ( -n y) e+l«ol (Z-Zu).
0
aba'01 b 4 1 b. • b

Substituting here the expression /~ from (10-9) and di-

viding by th~ expression E~~l~ for the incident field, \Ve
obtain the expression for the reflection coefficient of the
wave from the slot:

p= E!o
. = 1 1 4k - cos
2
1 (
-n - -A) X
E nc
1
XOl
G£ + G oo•aaba r
3
01
b 4

)'!Sjfl 1 ( ~- !fo) p+ill.,,l (Z-Zo),

If we assume here that Z=Z0 and take into account that

the internal radiation conductivity of the half-wave trans-
verse slot is expressed as (see Paragr aph 7 -2):
4 1
G=. k 8 cos•
.:Oflabao 1
r,T{) sin (TYo)' 1
(7-30)

the expression for the reflection coefficient wi 11 be:

(j
Po=G z+G' (10-10)
If we now reduce H01 waveguide to the equivalent twin
line and take account of the fact that the transverse
resonant slot in this line will be represented by a series
resistance, the coefficient of the reflection of the wave
from this resistance will be ex pressed as:
(10-11)

343
where r' is the relative (normalised) resistance
( r' = ~) .
\ I

From a comparison of (10-10) with (10-11), we obtain

for the rcJ~tive resistnnce the expression
r' = ~~ . _ (10-12)

The external conductivity of t~e slot radiation in the wave-

guide G t can be approximate1y defined frorn the expres-
sion G= w;
2R
where R:t=73.1 ohms and W=377 ohms.
Substituting this expression as wetJ as the expression G

----u
Fig. 10-31. Calculation of the reaction of
the tr.:1nsverse slot.

from (7-30) into (10-12), we obtain the well-kno\vn expres ..

si on (54]
' -
' - '0 523 (' Aw)
~
a A, a • ( nA,' . I
ab cos \ 4b.) Sin ( n
b- Yo )
• ( 10
- 13)

Thus, the slot antenna represented in Fig. 10-30, a, can be

reduced to an equivalent loaded t\\'in line (Fig. 10-32).
In order that a travelling-wave type of operation should
be established in that 1ine (at its input), it is evidently
necessary that the condition r'n= 1, where n is the number
of slots in the anienna, should be fulfilled.
For a longitudinal slot in an infinite waveguide (Fig.
10-33), the mmf is expressed as:
A
r:z;:z 0 +-
4

M= ~ lf!~i cos k (z- z,) dz.

A
•= ...
-'o--

144
 SJ,Jbstituting here the expression H~:~ frorn (I 0-7). inte·
grating and substituting the expression thus· obtained into
( 10-8), we obtain:
b
2k --
1:: = Gz ~a iw: B cos ( (101 ~) cos ( TYo )e- la.,z.. ( 10-14)

In accordance with the expressions (7 -39), the reverse

radiation .of the slot into th~ wctvegu ide (reflected field) is:
0
ExoJ 2k/~
= 01U0
. cos ( a - A) cos (--
n y ) sin { --
n y ) e-i«oa (Zo-Z)
01 °
1 4 b 0 \ b •
•

Substituting here (10-14) and dividing by the field of the

incident wave £!~~·. we obtain for the coefficient of

r' r'
----------v
Fig. J(J .. 32. Equivalent loaded Fig. 10-33. To the calculation of
I in e. the reaction of a longitudinal slot.

rellection the expression

2
I 4k b 1 ( A)
P=- 0 l +G a~ , WtJ.Uo• cos a 01 4 X
- -

't( cos2 ( TYo) e+ ;a., I (Z-Zo). ( 10-15)

A~suming here that Z=Z0 and taking into account that

the internal conductivity of a longitudinal haJf-wave slot
in an infinite \vaveguide is expressed a3 (see Paragraph 7-2):
Gl = -4k
2
b
an
2

WJ.UI01
cos 1 (
t.X 01 - A)
4 cos
1 ( n
--b Yo ) (7-40)
we obtain: ,
(j
p,=- Gz +a. (10-16)
If we now reduce the waveguide to the equivalent twin
line and take into account that the longitudinal slots in
that line will be represented by shunting resistances, the
845
coefficient of reflection in the line wil J be expressed as:
g'
Po=- 2+g' ' (10-17)
where g' is the relative conductivity (g'=gW). From a
comparisoiJ of (10-16) with (10-17), we obtain:

g' = ~~ . - (10-18)

The radiation conductivity of the s1ot Or, can be approxi-

2
mately expressed as Gl: = ;; • Substituting this value,
as well as the value G from (7-40), into (10-18), we obtain:

g'=2.09 ~ ~wcos• ( !X01 ~) cos• ( Tuo). (10-19)

Thus, the slot antenna represented in Fig. 10-30, b, can

be reduced to the equivalent loaded twin line (Fig. 10-34).
In order that a travellingMwave type of operation should
be established in that line (at its input), the condition
g'n= 1, where n is the number of slots in the antenna,
should evidently be satisfied.
Let us now calculate the directional diagrams of resonant
slot waveguide antennas. The problem of the calculation
of the directional diagrams of such a system has not been
solved. However, for engineering purposes, one can make
use of the duality principle. In particular, this concerns the
longitudinal plane of the antenna, in which the dimensions
of the antenna are large in comparison with the wave·
length and the approximate expressions for this plane are
well in agreement with experiments. The antenna di-
rectional diagram consisting of transverse slots can, in the
longitudinal plane, be calculated from th~ expression (4-8)
concerning a eo-phased array of electric dipoles,
sin ( n~d1 sin a)
1
(4-8)
1
sin s1n a
2

where, in this case, d1 ="'w·

Of course, this expression characterises the radiation
of the antenna in the half-space containing the slot:) and
does not define the radiation in the shadow region.
846
 To calculate the directional diagram in the longitudi-
nal plane of the antenna consisting of longitudinal slots,
the expression (4-9) is used

1 cos (-]-sin~) sin (m:d sin ~) 1

F <P>=-
m CllS
p •
su1 l
I kd
2
1
si n P. )
t'
' • (4-9)

where, in· th\s case, d1 = 'kw/2.

This expression is likewise not valid for the field in the
shadow region of the antenna.
Note that the level of the side lobes in the case of an
antenna with transverse slots is considerably larger than
in the case of an antenna with longt tudinal slots. Th 1s is
due to the large distance (equal
to Aw) between the transverse
slots.

-
2 4 •

-- • ~

.=~I' :~g'
~
:. :g'
Ill
-
Fig. 10-34. Equivalent loaded line. Fig. 10-35. Directional
diagram of slot antenna in
transverse pi ane.

As regards the directional diagrams in the antenna trans-

verse plane, in the case of an antenna with transverse slots,
the expressions, obtained in accordance with the duality
principle produce results which are sti 11 satisfactory But
the results are unsatisfactory \VhE'n the duality principle as
used for antennas with longi tudinaf slots. The diagrams
calculated from the expressions obtained for elliptical
cylinders agree more satisfactorily. Thus, F1g. 10-35 giv~s
the directional diagram of the antenna sho\vn in Fig. 10-30. b,
in the transverse plane. In this calculation. thE' rectangu'lar
waveguide with the dimensions b=47.8 nim and a= 16 mm
was replaced by an elliptical cylinder of eccentricity e=0.9
155 ].
Apart from resonant antennas, non-resonant slot wavegu1de
antennas are also used; they differ from the former in that
the waveguide is loaded at its end on an absorbing resistance
841
in such a way that a travelling wave is set up when
there are no slots in the waveguide. The slots can be spaced
at a distance d= 'Aw/2 apart in the case of longitudinal slots
and d= Aw in the case of transverse ones, where 1Jpon they
are excited in phase with the same amplitudes. The di-
rectional diagrams are then calculated by means of the
expressions n1entioned above. in particular the expressions
(4-8) and (4-9). -
UsuaJly, the slots are spaced at distances which differ
somewhat from 'A /2 and 'Aw. They are then excited with a
w certain phase shift and
the maxi1num radiation is
inclined at a certain angle
to the axis of the wave-
guide in the direction of
propagation of the wave.
Moreover, the angle of
inclination of the maxi-
rnum radiation can be con-
trolled by changing the
Fig. 10-36. Excitation of slots by frequency of the osci J. .
reactive stubs.
Jations or of the critical
wave-length in the wave-
guide. The necessary distribution of the amplitude and phase
of the field in the slots can be specially chosen, for example,
through an appropriate displacement of the slots away from
the rniddle line of the broad wall of the waveguide.
Slot waveguide antennas can also be obtaineri by dis-
posing the slots on the narrow wall of the waveguide. Of
course. transverse slots on the narrow wall of the waveguide
will not radiate, since they are sttliated along the line~ of
electric current.
Fig. 10-36 represents a slot waveguide antenna in which
the slots are dtsposed in the middle of the broad wall of the
waveguicie and excited by means of reactive stubs. The
stubs are inserted into the waveguide on both sides of the
slots and arranged along the lines of the electric field of
the ffoa wave. rfhe electric current excited by the H01 wave
in these stubs spreads along the wall of the waveguide,
crosses the slots and thereby excites them. The neighbouring
stubs are excited. by the Hot wave in antiphase and since
they are situated on both sides of the slots, the latter are
all excited in phase. The advantage of the··l e.xcitation by
)

848
means of reactive stubs consists in the fact that the
intensity of excitation of the individual, slots can be
regulated by changing the depth of insertion of the stubs.

10-7. Director Antennas

Direttor antennas or antennas of the waveguide channel
type, often called after the name of their inventor Vagi
(56J, consist of wire half-wave dipoles arranged in parallel
in one plane (Fig. 10-37). One of then1 is an active dipole
and the rest are passive; one of the passive dipoles, situated
behind the active one, plays the role of a retlt\ctor and the
other dipoles, situated in front of the active dipole, play
the role of directors.
We know from the theory of coupJed dipol~s that for
a passive dipole to act as a reflector, its reactive resistance
c:?,J
3 f I q~ t?,-i:t?~ _;7·~
I { ~
I ' ~~
Direction of
• 1'110:CI~m
''llfll~hon
f
Fig. 10-37. Director antenna: Fig. 10-38. Design of di-
J-act ive dipole; 2-dir,•clors: a--reflector. rector antenna:
/-feeder.

should be inductive. That is why the length of the reflector

is taken somewhat greater than "A/2. A passive dipole
becomes a direc~or when its resistance is capacitive. Hence,
the length of the director is taken somewhat sn1al Jer than
A/2. Only one dipole is utilised as a reflector because the
subsequent reflectors would be only ~·eakly excited and
would not exert any noticeable influence on the radiation
of the antenna. As for the number of directors, it can be
large, because the radiation of the antenna is directed
towards the directors and, consequently, they are fairly
intensively excited in series, forming a waveguide channel.
The number of directors varies from 2 to 10 and more.
Director antennas are utilised on metre and decimetre
waves; the dipoles are rigid and fastened in their middle
to a metal rod (Fig. 10-38). This rod does not become
excited due to the fact that the electric field of the antenna
349
intersects it at right angles. The active dipole is usually
made in the shape of a loop-like dipole and fastened to the
rod in the middle of its uncut part, so that the antenna is
mounted without insulators, and is quite compact and rigid.
The distance between the active dipole and the reflector
is from 0.15 to 0.25A, while between the active dipole and
the first director as well as between the neighbouring
directors, it is from 0.10 to 0.35 A. In GFder to obtain the
maximum radiation in the main direction, the distances
between the directors as well as their length are speciaJJy
chosen on the prescribed frequency of excitation so as to
get definite ratios between the currents in the dipoles. It
is evident that the magnitudes of the currents should be
approximately equal and close to the magnitude of the
current in the active dipole. As for the phases of the cur-
rents, they should Jag behind by a definite magnitude from
dipole to dipole towards the end of the antenna. Under
this condition, the fiPids of the individual dipoles reinforce
one another in the main direction whereas in the other
directions the radiations cancel out. As a ruJe, the radia-
tion maximum of the antenna towards the directors coin-
cides with the radiation minimum towards the reflector.
In this way, the antenna sets up a unidirectional radiation.
The input impedance of a director antenna is quite
sensitive to frequency changes of the oscillations, so that it
has a narrow bandwidth (the width of the pass-band of
the antenna amounts to a few per cent). In the case of a
large number of directors, the tuning-up process, which
consists in the proper choice .of the lengths of the directors
and their spacing, is most tedious because a change of
length or position of one of the dipoles leads to a change
of the amplitudes and phases of the currents in all the
dipoles. For ibis reason and also because the inten-
sity of excitation of the
directors towards the end
of the antenna somewhat
decreases, the number of
a directors usually taken is
H-A 4-o;---- small and the width of the
directional diagrams of
.-o~d.+.tt.l director antennas adopted
Fig. 10·39. Calculating the radi- in practice is never less than
ation of a director antenna. 15-20% at half-power.
•

850
 If the ratio of the amplitudes ~nd phases of the currents
in the dipoles is known, the directional di-agrams can be
calculated from the expressions of the antenna array (see
Chapter Four). For the equatorial plan~ of the dipoles
(Fig. 10-39), the antenna field intensity in the radiation
zone can be calculated from the expression
E:u=EA [~e+tkdRcosa + 1 + '/:1 e-•kdcosa+
.+ Ii Dt e-ilkd c.osa -1- . . . +I IDN e-iNkd cos a] ' (10-20)
A A
where
/R, I A, I o m are the complex arnplitudes of the
currents in the reflector, the actively fed dipole
and the m-th director respectively;
601
E A=i A e -ikr A is the field of the active dipole
'A
in the zone of radiation;
a is the angle between the direction towards the
point of observation and the antenna axis. The
expression ( 10-20) is based on the assumption
that the distances between neighbouring directors
as well as between the first director and the
actively fed dipole are the satne.
For the meridional plane of the dipoles, the field is
calculated from the expression (10-20) tnultiplied by the
ratio
cos ( -i-sln a)
cos a •

To calculate the complex current amplitudes excited in

the dipoles, use can be rnade of the coupled dipoles theory
(see Chapter Three). For the director antenna, the KirchhoJT's
equations are written as follows:
0=1 RZRR +I AZAR+ /o1ZDtR + ... -t-loNZDNR;
UA=/RZRA f-/ AlAA + /o1Zo1A + · ·. + foNlDNA;
o:;:: 1RZRo1 + 1AZAo1 + 1o1Z o1o, + ... +
+loNlDNDt; (10-21)
• • • I I e I I • I I I • • • • • • • •

0=/RZRoN +I AZAoN + /otZo,oN + ... +

+foNlDNDN•
851
The difficulty in calculating the complex current an1pli·
tudes consists, in th~ first place, in that the mutual re-
sistances of dipoles of arbitrary length are not tabulated and
their calculation is complicated because one has to know
the N (N2-f:"D of these values. Anothf>r difficulty in calcu-
lating the currents is due to complexity of computing
the detcrtninants of the equations Q.0-21). This explains
why the calculations effected so far have not gone beyond
antennas with a number of directors not exceeding. four.
Once the currents in the antenna dipoles have been
calculated, the determination of the input resistance of
the actively fed dipole from the foilowing expression
presents no difficulty:
IR lot IoN
Z A= 1 ZR.I\ + ZAft. + 1A
· Zn1A -1- ••• -f- 1 l DNA· (10-22)
A A
After that, the directive gain of the antenna can be
defi nP.d from the ex pression (4-68).
The above-mentioned n1ethod was applied to the calctt·
Jation of a director antenna consisting of one active
dipole and from one to four directors (57}. The distances
betwE'en the dipoles \Vere assutned to be the same and the
n1utuat resistances were taken frotn the tables for half-\vave
dipoles. The directional diagrams, input resistances of the
active dipole and the amplification coefficients of the
antennas were calculated. We quote the calculation data
regarding an antenna with four directors, for three different
spacings between them: d=O.l lw, d=0.2J, and d=0.3A-.
The data \vere obtained <t,C\ R function of the magnitude of
the natural reactance of the directors. The natural reactance
of the actively fed dipole was taken equal to zero.

3
11/A •.(If 3 d~=Q2
I 3 --~~-----

2 .,.____...11o.....l~::;:t---::; 2 .,_,_:.. .... ~--t-----1 21-~

f r-----::;llli----"~1-N-___, I ~~::t--~ "'- ~_... f
1
0 0 ' ,... ~ n~_..,........r:;.~~...,
tao" o· 60 · tzo" tao'o" 60. at/' mo•
Fig. 10-40. Directional diagraans ol a diE~ctor antenna.

852
 Fig. 10-40 represents the directional diagrams of the
antenna in a plane pE'rpendicuJar to the~ dipoles (solid
lines) and in the plane of the antenna (broken lines) in the
case of a reactance of
the directors X =-40 8
ohms. It can be seen fo.'t e
that an increase of
'
-~ ~
the length of the
antenna leads to a
• ~ '
decrease of the width -
of the major lobe of the 2 . .
directional diagram. -tBI'
Xohm
-t20 -ao -40
L
Shown above the curves u
are the vector dia-
grams of the currents -
in the dipoles, figure .11..-u
1.- T £
0 designating the - -, ,...
currept in the active
dipole and figures 1, 2, V
.....

3 and 4, the currents

in the directors in the
"f •
\
order of their position, l'tnw '2
starting from the -160 -120 -80 -40
active dipole.
Fig. 10-41 repre-
sents the amplification d
coefficients of the f=ll3 ~ e
D 6
antenna (e= 1. 54 )
~
plotted as a function - "
of the magnitude of the 2.. ~
~ 2
reactance of the direc-
tors. The curves indi- xlhn
cate the magnitudes-160 -120' -40
of the amplification .
coefficients in the di- F1g. 10-41. Antenna amplification coef-
rection of the directors ficients curves:
/-forward~ 2-back
(forward) and in the
reverse direction (back). It is seen that the maximum of the
antenna radiation (e==5 -7) occurs in the vicinity of the
reactance of the directors which is approximately 40 ohms;
moreover, the forward radiation maximum corresponds to
the backward radiation minimum.
23-2122 853
 Fig. 10-42 represents the curves of the active part of the
input resistance of an actively fed dipo.le. On comparing
these curves with those in Fig. 10-41, it is seen that when
the antenna is tuned on maximum radiation !orward, the
input resistance of the acti',tely fed dipole decreases and has
a value of R10 =20- 30
ov ohms instead of 73 ohms
,._....., -....10 for tme single dipole.
There are other
40 methods for computing
1----+---t----lf---f:.=--4-~ .......~ 20 director antennas, such
as the approxirnatemeth-
.....tJ~i..-..-~-L,..;;~....J,.-"--_.0 od based on reducing
_..__-~
-
Fig. 10·4!l. Curves of the active part
the system of equations
of the input resistance of an active ( 10-21) to a difference
dipole. equation with linear
coefficients [58].
We quote the experimental data of a director antenna
consisting of a series of directors spaced 0.34 A, apart and
of the same length. Table 10-1 gives the width of the
directional diagram at the 0.25 power level and the power
gain of the antenna. Table 10-2 gives the lengths of the
directors corresponding to different number of them, for
which the Jevel of the first minor lobe is of the order of 30%
of that of the directional diagrant major lobe.
It is seen from tables 10-1 and 10-2 that, as the number
of directors increases, the directional diagrams become
narrower, and the power gain increases. The necessary
length of the directors decreases as their nutnber goes up .
..
Table 10-1 Table 10-2

Number of Width of direction- Power • of

Numl:ier Length of director
directors al diagram In aaln directors In wave ·lengths
degrees •

30
20
22
26
-
21
42
30
0.385
0.40
13 31 15 20 0.407
9 37 13 13 0.414
4 46 8 10 0.42
7 0.423
6 0.434
I ~~

184
 It should be pointed out that a director antenna can be
regarded as an axial radiation antenna of retarded phase
velocity. Indeed, dipoles shorter than half' a \vave-length
can be regarded as elements with capacitive resistances
connected to the transmission line in a way similar to that
in which capacitive resistances are connected in parallel
to a twin line. The phase velocity of the wave in such a
system is less than that of light. The actively fed dipole
can be regarded as an element which excites a retarded
wave in such a line.
Thus, we can apply to the director antenna the axial
radiation antenna of retarded phase velocity theory and
use the corresponding expressions to calculate its parameters.
Thus, the width of the directional diagram of the antenna
can be evaluated from the expressions (4-26) and (4-27)
and the directive gain, from (4-72).

10-8. Helical Antennas

Helical antennas, which radiate a field of circular
polarisation in the direction of their axis, find extensive
application in the centim~tre and decirnetre wave range,
more seldom in the metre \Vave range.
3
f

..,
Fig. 10 ..43. Helical antenna: Fig. 10-44. Equivalent
l-hellx; 2-dlsc; a-coaxial line. circuit of the helix.

The antenna consists of a wire helix several wave-lengths

long and of winding diameter of the order of a third of a
wave-length (Fig. 10-43). One end of the helix rernains free
and the other is connected to the inner wire of a coaxial
line: The external conductor of the feed coaxial line is con·
nected to a metal disc with an aperture for the inner con·
ductor of the line. The disc plays the role of a counterbalance
and prevents the penetration of electric currents to the outer
surface· of the external conductor of the coaxial line.
 A travel Iing wave of electric current arises in the he Iix
and the antenna radiates maximum energy along its axis
in the direction of motion of the current wave. In order to
get a clearer understanding of the operating pr!nciple of a
helical antenna, Jet us represent it in the form of plane
circuJar loops of diarneter D = 'A/n, spaced at a distance S
a part on the axis and series fed by a single-wire feeder
(Fig. 10-44). Let us investigate the radi-ation of one of the
windings and assu.tne the current in the winding to be
distributed as I 0e-'k 1 , where I o is the magnitude of the current
z :r at the beginning of the
· d" 2n
wtn tng; = - ; l ts
k ·
I,cosk/ ifoslnkl lw
the length along the
winding.
~~ ~v 'fhis expression can be
written as I ocoskl-
-il o sinkl and thus repre·
sent the current as the su-
Fig. 10~45. Current distribution perposition of two stand ..
helix winding. ing waves with a 90°
phase shift, the amplitude
of one of which changes along the winding in accordance with
the cosine law and the other, with the sine law (Fig. 10-45).
We obtain four bent half-wave dipoles oscillating in phase
pair by pair: one pair of dipoles is orientated in the di-
rection of the y-axis and radiates maximum energy in the
direction of the z-axis with the vector of the electric field
intensity coinciding with the y .. ax.is; the other pair of
dipoles is orientated in the direction of the x-axis and radiates
n1axin1um energy also in the direction ot the z-axis, but this
time, the vector of the electric field intensity coincides
\vith the x-axis. Since the pairs of difJoles oscillate with
a 90° phase shift, the field radiated in the z-axis direction
has a circular polarisation. At a certain angle to the z-axis
the polarisation of the field is elliptical and in the xy-planet
it is linear.
Note that, as a result of the eo-phased oscillation of the
dipoles and mutual coupling, the radiation resistance in
each pair is rather high. The input resistance of each
\vinding is close to its wave impedance and if, in addition,
account is taken of the fact that neighbouring windings In
the helix oscillate almost in phase due~ to the small value
'

856
of S/A, it will become clear that a travelling-wave type
of operation is set up in the helix. The phase velocity of
the wave in the antenna (in a single-wire feeder Fig. 10-44)
is somewhat inferior to the velocity of light, and we get an
antenna with a retarded phase velocity radiating along its
axis. From this point of view, a helical antenna is akin
to a director antenna.
If the diameter of the helical antenna is small in coln-
parison wilh the wave-length (D~A/n). the currents at
diametrically opposed points of a winding are opposed in
direction and the radiation resistance of the winding is very
small (loop effect). As a result, a standing wave type of
operation is established in the antenna, the radiation along
the axis of the helix being zero and the radiation maxirnum
of each winding and of the whole antenna occurring in the
transverse plane of the antenna.
When the diameter of the antenna is large (D~'A/n),
the currents at diametrically opposed points of a winding
and in neighbouring windings are again out of phase and, ow-
ing to mutual influence, the radiation resistan~e of the wind-
ing decreases, the travel-
ling-wave type of operation
is upset and the radiation
from individual elements
in the direction of the axis
of the winding cancels out;
the radiation maximum of ..-
the antenna is at a certain
angle to the antenna axis. -
~

The picture of the radi-

(b) (C)
ated field corresponding (a)
to the three types of an- Fig. 10-46. Three types of helical
tennas described above is antenna:
shown in Fig.I0-46. a-non-direct ional radiation; b-axlal
radla tlon; c-conlcal radla tlon.
The qualitative picture of
the electric current distri-
bution in a helical antenna is confirmed by theoretical ·re·
sear.ch on wave distribution along an infinite helix conducted
by S. Kogan [59 and 60 ]. In accordance wit~ hi~ in~esti·
gations, in the general cas~,. we get the excttat.ton an an
infinite helix of a superpos1t1on of three travelling waves
of current of different phase velocities; the first type of
current wave of increased phase velocity predominates In
851
a helix of small diameter (kD~l), and the third type of
curr~nt wave of slow phase velocity predominates in a
helix whose diameter approaches one·third of a wave-
len gth ; moreover, for frequencies hig her than the cri tic al
one, which is approximately expressed as:
CDcrlt COS a
~ D'
v1 (I+ sin a) ..
2
there remains only the third type of current wave, corre-
sponding to the axial radiation type of operation.
f'ig. 10-47 shows the experitnental curves of the current
amplitude distribution in the wire of an antenna with the
I - Open end
4
(0) 3
2
f

Op8n end

f 2 3 5m
Fig . 10-47. Cur ren t dist ribu tion in a heli x:
a-no n-di rect iona l radiation; b-a xla l radi atio n.

following parameters: number of windings n==7; angle of

inclination of the windings a.= 12°; antenna diameter
D= 22 .5 cm; win din g pitch of the hel ix S= 15 cm [61 ].
The upper curve corresponds to the current distribution on
a frequency f=2 50 ~\c/s (D~O.l87 A), aod the lower curve,
on a frequency {= 450 Mc/s (D~0.338 A.). It is seen that
in the first case, there is considerable reflection from the
end of the helix and a standing wav-e is set up in the.
antenna. In the second case, the reflection from the end of
the antenna occurs too but the amplitude of the travelling
wave in the middle region of the helix is more or less
constant. Fig. 10-48 represents the schematic picture. of the
current distribution in the antenna in the axial radiation
type of operation. It can be said that the reflected wave 4
is superposed on the incident wave 2;- Rioreover, bi&ber ' •

818
 •

modes arise at the end and beginning of the antenna to

v1hich the attenuated currents 3 and I co,rrespond. The
current amplitude of the reflected wave 4 constitutes about
20% of the current amplitude of the incident wave 2 and
in the first approximation, when investigating the antpnna
radiation, the currents 1, 3 and 4 can
be neglected.
As regards the phase velocity, Open
end
measurements show that in an axial 2
radiation 'helical antenna the deceler- ~:::::=~ !11======:~4-b::::=~
ation coefficient of the phase velocity /Jisfonce .,.,.,n,.., heliz
of the current wave in the wire £= Fig. 10-48. Se hem at ic
= v1 is found to be frequency de- picture of the current
u distribu tion.
penden t. Thus, for a helical ant~nna of
seven windings with an angle of inclination a=l2° , a
diameter D=23 cm and relative length of winding LI'A=
=0. 72-:-1.2, the deceleration coefficient changes from a
value &= 1.67 on the frequency /=300 Mc/s to a value
= 1.1 on a frequency {=500 Mc/s.
'=
If the deceleration of the phase velocity and the fact
that a helix winding has a certain winding pitch are taken
into account the following condition should be observed
to obtain a circularly polarised field in the direction of the
bel ix axis:
k~L-kS=2n
or
S+A.
I~= t • (10-23)
"
However, if ,we regard a helical antenna as a slow phase
velocity antenna then, as we know, to obtain the maximum
directive gain, the phase shift of the field radiated by the
first and last elements of the antenna should equal a. Hence,
instead of the condition (10-23), we obtain:
k;L-k S = 2n + ntl
or ,

(10-24)

Thus, if the condition observed in the direction of

the principal radiation is (10·23), we obtain a circular
859
polarisation; no such polarisation is obtained if it is
(10-24) but the directive gain is then at its highest.
Empirical expressions based on experimental data have
been obtained for a helical antenna with an angle of incli-
nation a= 12°-:-16° for a number of windings larger than
three. The expressions are given below.
The width of the directional diagram at half-power is
52 -
-
L
-nS
(10-25)
-A -A
The width of the directional diagram at the radiation
•
zero ts:
115
2Po = r- (degrees]. (10-26)
_!::_ 11 ~
A Y A
The directive gain is
D= 15 ( {- r n {-. (10-27)
The antenna input resistance is
· R1n ~ 140 { [ohms). (10-28)

Due to the fact that the deceleration coefficient of the

phase velocity of the wave ' increases to some extent as
the wave-length increases, the relations
( 10-23) and (10-24) are but little
affected by changes of wave-length
and the antenna maintains its di-
aective properties in a wide wave range.
This wave range lies between approxi-
mately 0.7 Ao anp 1.2 Ao, whereA0
Fig. 10·49. Conical is the wave-length for which the
helical antennas. optimum dimensions of the antenna
have been chosen.
Apart from the cylindrical helical antennas described
above, conical helical antennas also find application. Ex-
amples of two helical antennas facing a screen with their broad
part are shown in Fig. 10-49. In one case, the helix is fed at
the base, in the other, the .bel ix is fed from the apex.
Measurements effected on conical helixes with the following
parameters: n=IO, «=6°, Do=20 cm, D•ax==60 cm show
'
'
160
that the axial radiation of antennas fed from underneath
is maintained in approx intately a twofold' wave range and
that of antennas fed fron1 above, in appro xima tely a three·
fold wave range.

10-9. Dielectric Rod Antennas

Dielectric rod antennas en1ployed in the centimetre
wave range represent dielectric rods of circular or rectangu-
lar cross se~tion, several waye-lengths long, excited by a
segment of circular or rectangular metal waveguide (Fig.
10-50). In the case of a circu lar cross section, the excitation
of the rod is caused by an Hu wave in a circular waveguide
and in the case of a 2 . 1
'
rectangular cross sec-
tion, the excitation of d
the rod is caused by an
Hot wave in a rec- ~-----l------~
tangular waveguide.
The wave excited Fig. 10-50. Dielec tric rod antenna:
ctric rod, 1-dleJ ect rlc rod; ~-metal wavegulde;,
in the diele 3-coa x In I line
which is the prolon-
gation of the metal waveguide, is an electromagnetic
wave, akin to the H 11 and Hot \Vaves in metal waveguides.
This electromagnetic \Vave is propagated along the rod with
a phase velocity inferior to the velocity of light in free space.
It has the structure of the field shown in Fig. 10-51, from
which one can see that the predominant direction of the
electric field in the rod coincides with the x-axis. The
cotnponents of,th e electric field inten sity in the rod, coin-
ciding with the y- and z-axes, have an insignificant value;
furthermore, they are of opposite direct ions in different
parts of the cross section. Hence, fJOiarisation currents
j= o.l( e-eo )E are excited in the rod which, on the whole,
are transverse and coincide with the x-ax is of the rod.
Thus, a dielectric rod antenna can be regarded as a
continuous system of transverse radia tors disposed in the
direction of the z-axis of the rod. The distribution of the
amplitudes of the polarisation currents in the direction of
the system axis can be considered as constant and the distr i·
but ion of the phases of the polarisation currents as changing
linearly. Hence, a dielectric rod antenna can be reduced to
a travelling-wave antenna with a slow phase velocity. As
861
we know, an antenna of this kind radiates energy along
its axis in the direction of motion of the v1ave in the antenna.
The phase velocity of the wave in the rod is defined by
the ratio of the amount of electromagnetic, energy conveyed
:r by the wave inside
the rod to the tot a I
amount of eJectro-
. rnagnetic energy
conveyed by the
11 wave along tbe·rod.
If the rod is thin
relatively to the
wave-length, most
(DJ of the energy is con-
•
(a) veyed by the wave
- Electric field into the space sur-
---Mognetic field rounding the rod,
Fig. 10-51. Field structure In a dielectric the phase velocity
waveguide: is mainly deter-
a-In cross section: b-In longitudinal section.
mined by the sur-
rounding medium and becomes close to the velocity of
light. If the rod is thick relatively to the wave-length, most
of the energy moves inside the rod, the phase velocity is
mainly determined by the internal medium and becomes
close to the velocity of light in the given medium. 1'hus,
the phase velocity of the wave lies in the interval v1>v>;• ,
where L/Bo is the relative permittivity of the rod. etea
The above-cited arguments with .regard to the structure
and phase velocity of the eJectrom_~gnetic wave in a dielectric
rod are based on the theory of waves in an infinite circular
dielectric cylinder. In accordance with this theory, an
infinitely large n~ber of 1nodes, symmetrical as well
as asymmetrical) can be present in a cylinder of this kind.
The wave we have described is an asymmetrical wave of
n1ixed mode (superitnposition of electric and magnetic
waves) the field of which is determined by all six corn·
ponents of the E and H vectors. The peculiarity of this type
of oscillations lies in the fact that a critical wave-length
is equal to infinity, i.e., a wave of this mode can theo-
retically be propagated in a rod of any thinness. The phase
velocity of this wave along the rod is represented by the
curves shown in Fig. 10.. 52 as a functl'Oil, Of the ratio of the
 rod diameter to the wave-length at various dielectric
permittivities of the rod. ~-
lt can be seen from the curves that in the case of thin
rods (d/~1), the wave phase velocity irl the rod is close to
that of light, and when the diameter of the rod approaches
half a wave-length, it begins to fall rapidly. This
diameter (d/A~0.5) is precisely chosen as the operating
diameter for the rod dielectric antennas.

Q6t--t-.......+--'

a4 ~~~~~~~~~~l==t=~~~~::~
0.2.,___...

tJ o.z a4 a1 tl8 to t2 t4 t6 t8 2.0

Fig, 10-52. Relative phase velocity in an
infinite dielectric cylinder. ·

The approximate calculation of the directional diagram

of the antenna can be approximately calculated with the
help of the expression (4-22) which we shall write as:

sin [~L (¥-cos a)] (10-29)

-1.~-cos a.
A V

where ~ is the angle between the antenna axis and the

· direction towards the point of observation;
L/A, the length of the antenna in wave-lengths;
vfv,.w, the relative phase velocif'J of.,. the wave in the
antenna. ·.
The expression (10-29) concerns a travelling-wave an·
tenna·with point sources and does not take into account the
radiation directivity of the element of the rod length.
Indeed, a~ can be seen from the picture of the electric field
in the rod (Fig. 10-51), the polarisation currents in the
··rod have transverse (in the direction of the x- and g-axes)
as well as longitudinal components and the superimposition
of the fie14~ set up by these current components forms
a rather complex radiation characteristic of the element of
the rod length, in particular in the transverse section.
However, the combination factor of the antenna (10-29)
"cuts out, from this complex radiation comprised
within relatively narrow angles a (up to 20-30"). But the
radiation within these angles is on the whole defined by
the components of the polarisation currents, parallel to
the x-axis, and for the element of the rQ_d length, it has a
low directivity in the planes of the H as well as E vectors.
Hence, the expressiou (10-29) must fairly accurately define
th~ major lobe of the direction diagram and is not valid
for the minor Jobes. . r

Another reason why the expression (10-29) is only ap-

proximate is that it takes no account of the radiation set
up by the polarisation current wave, reflected from the
rod end. However, the wave phase velocity in the antenna
is close to that of light (d/'A ~ 0.5) and the coefficient of
reflection of the current wave from the end of the rod is
not high (of the order of 0.15). Hence the radiation of the
reflected wave, directed in the opposite side, can be
neglected.
Since it is a travelling-wave antenna of sJow phase
veJoci ty, the die1ectric rod antenna has an optimum
length which is ex pressed as:
L 1
(4-25)
-r-= 2(~-t)"
At the same tilne the width of the antenna directional
diagram can be evaluated from the expressions (4-26) and
(4-27) and the directive gain can _be expressed as:
L
0=1.2-r. (4-75)
Fig. 10-53 shows the#' experimental directional diagrams of
a polyrod antenna of uniform rectangular section with
approximate dimensions 'A/2x A/3 for three different lengths:
LI'A=3; 6; 9. The same figure gives the measured values
of the directive gain in decibels. An examination of the
curves reveals that the dielectric antennas have large minor
lobes. Note, in addition, the absence of radiation zeros
between the lobes. · •
This can be explained by the attenuation of the wave
during its propagation in the dielectric. ~:,od, due to the
'

164
energy lost on heating the dielectric. As regards the width
of the major lobe and the directive gain, they are in fair
agreement with the expressions given above (to within
20%).
The radiation of dielectric rod antennas can be regarded
from a point of view somewhat different from that mentioned
above when we were considering the radiation of polari-
sation currents excited in the dielectric rod. Indeed, when
the elec~romagnetic energy conveyed by a wave in a metal
waveguide reaches its extremity, it is partly radiated
f r-r--r--~....._,.--.--

0.8 t+-----1~-+----+~
t
~ Q6 /lmp/1/lctd/Qn .....,_+-! j.o-:.--t---+-
~ dJJ•f26
tS Q4 .....,__..,_.....,_...........~""'1--·--t-- ~--~--.............
.t
ja2
~
0
-sll -40 -30 -20 0 4050

Fig. 10-53. Experimental directional diagrarns.

directly into the surrounding space and partly passes into

energy of a retarded (surface) wave in the dielectric rod.
The latter part of the energy is conveyed by the retarded
wave to the extren1ity of the rod where it is radiated, due
to a gap in the continuity of the medium.
Thus, the qirectional diagram of a dielectric rod antenna
of uniform cross section can be said to be formed by the
radiation of energy in two sections: in the section of the
transition from the metaJ waveguide to the dielectric wave·
guide (rod) and at the end of the dielectric waveguide.
Since the field amplitude of the retarded wave diminishes
exponential1y away from the rod in a radial direction, the
field amplitude distribution of the radiating surface at
the end of the rod is pointed towards the axis, the radi·
ation from the end of the rod being highly directional and
without lobes [62 ]. On the other hand, the radiation at
the junction of the metal and dielectric waveguides origi·
nates (in the main) from an area equal to that of the wave-
guide cross section and has a low directivity. Interfering
86S
with one another, the two radiations form a directional
diagram of the antenna depending on the length of the rod
as well as on its diameter.
If the diameter of the rod is large in comparison with
the wave-length, the larger part of the energy passes into
energy of a slow wave and the smaller part is radiated
directly at the junction of the waveguides. But due to the
great deceleration of the wave, the am~itude of the field
decreases more sharply as we move away fr~m. t~e rod and
the radiating area at the end of the rod dtmtntshes, that
leads to the widening
•• of the major lobe of
d, the resu It ing direct ion-
al diagram and at
1-4----1. _ __.,... the same time to the
Fig. 10·54. Conical dielectric antenna. relative decrease of the
size of the minor lobes.
If the diameter of
the rod is small in comparison with the wave-length, the
larger part of the energy is radiated directly at the junction
of the two waveguides and the smaller part passes into
energy of a slow W@Ve and is radiated at the end of the die-
lectric rod. However, since the phase velocity of the wave
in the rod approaches that of light, the decrease of field
amplitude away from the rod becomes low, the radiating
area at the end of the rod increases, which leads to a more
directional radiation from the end of the waveguide. As
a result, the width of the major lobe of the directional
diagram narrows down, out the relative magnitude of the
minor lobes increases.
Thus, it is desirable that the dlatneter of the rod at the
junction of the two waveguides should be large relatively
to the wave-length in order that the phase velocity of the
wave should be small and the larger part of the delivered
energy should pass into energy of a slow wave in the rod.
At the same time the diameter of the rod at its end should
be small in comparison with the wave-length, in order
that the phase velocity of the wave should approach the
velocity of light and the radiating area at the end of the
rod should increase. ·_ .
This leads us to the idea of a dielectric antenna with a
variable (by jumps or continuous) cross section. Fig. ·10.54
shows the structure of a conical rod &t\,teltna of circular
'

866
section. It is readily seen that the conical part of the antenna
can span either the whole length of the antenna or only
part of it, the other part remaining uniform .
Fig. 10-55 shows a dielectric rod antenna (63 1 of rec-
tangular cross section linearly tapering along a little over
half the length of the rod (the length of the rod being 6A).
At the beginning, the rod has a square section of side 'A/2
and at the end, a rectangular section of sides 1/2 and A/4.
The same figure shows the curve of the wave phase velocity
measurea at various sections of the rod as well as the
experimental directional diagram. We see that the phase
. '4--t--
11
=
~lid
I

tO ..
-
• V
0 L.,S). ""'
""'
O.D
0 """
aao 05 to t5 20 25 3.o .15 4.0 45 5.0 s5 60
) f Distance from ere/fer I, in wove ·length
-,
~aa
I' Ampli/icot1011•ta5ti1J
·~
'

~
' -
l.jA!.·B
~Q6
_J
Q4
, '
',
b.
Q2 A

vi7 V v \t ...ft..

0
-50-NJ -.10 -20 -to o to <X"
20 30 40 50
Fig. 10-55. Directional diagra1n of a tapenng
antenna of rectangular section.

velocity in the tapering part of the rod changes from sec·

tion to section in accordance with the linear law, remaining
constant in the uniform portion of the rod. The minor
lobes are insignificant as compared with those in Fig. 10-53.
A conical antenna of circular cross section the tapering
part of which extends approximately as far as the middle
of the length of the rod has nearly the same directional
di'agram; the overall length of the rod equals 6A. The
diamet er of the rod decreases from d, =0.5 A. to d 2 =0.3 A..
A few more points should be mentioned regarding the
properties of dielectric rod antennas. To begin with, it
should be pointed out that several dielectric antennas can
867
be connected in parallel, thus forming a eo-phased array
giving a high-directional radiation. Owing to the fact that
surface waves are propagated along the rnds and the field
of the waves diminishes exponentially as we move away
from the rods in radial directions, the interaction between
the rods in the array is insignificant. They operate almost
independently one from the other and the directive gain
in the main dirPction is proportional 19 tpe number of
elements in the system, i.e., to the number of rods.
Secondly, since rod dielectric antennas have no resonant
elements, they themselves have a wide pass-band. The
width of the pass-band of a dielectric antenna is conditioned
by the resonant properties of the exciting element (of the
dipole in a metal waveguide). ThirdJy, fhe dielectric from
which the antenna is made should have low losses, otherwise
the efficiency of the antenna will be poor and, in addition,
the exciting dipole in a metal waveguide should be arranged,
as shown in Figs. 10-SO and 10-54, outside the dielectric.
This leads to an increase of the efficiency of the antenna,
due to the fact that the higher modes generated by the
dipole are attenuated in the vicinity of the dipole and do
not penetrate into the dielectric medium .
...

10-10. Horn Antennas

Horn antennas, and in particular antennas forrned hy the
open end of a rectangular or circular waveguide find appli-
cation in the centimetre wave range. Antennas of this iype
are utilised independently or as radiators in more complex
antennas, such as lens antennas and parabolic antennas.
Let us first of all consider the radiation from the open
end of a rectangular \Vaveguide, excited by an HoJ wave.
Let a segment of rectangular waveguide t>e shorted out at
one end and open at the other and let it be excited by an
elementary dipole (Fig. 10-56). On reaching the open end
of the waveguide, the electromagnetic energy conveyed
by the H01 wave is partly reflected and partly passes into
space and is radiated in all directions. The radiation from
the open end of the waveguide can be calculated by means
of the equivalency theorem (principle). Enclosing the
waveguide with a surface s adjoining its outer surface, one
can neglect the radiation of the currents flowing on the
shadow sides of the waveguide in view o,f their low intensity
~

868
and take account only of the radiation of the equivalent
currents on the illuminated portion of thf'..,·surface s, i.e.,
at the open end of the waveguide.
The equivalent currents on the surface s can be calcu-
lated from the expressions:
J e = [ H, n] . J M == [ n, E'),
where n is the perpendicular to the surface s, external
relatively to the point of observation.

11
~

Fig. 10-56. Radiati on fro1n the end of a rectangular

waveguide:
/-antenn a; 2-coaxla l cable •

To simplify the problem, the field at the output

aperture of the waveguide is assumed unexcited and
expressed as:
b' =B. k" sin I
I n- Y)' e-:UoaZ•
XOl lWE \ b '
H yot = -- Bia o1 sin ( nb y) e-tao•'·'
H ZOI ----
-
B ~cos(
b ~-
b
u) e-tao,z.
Then the equivalent surface currents an. L'xpressed as:
e M E~
J x--- H yol' JU = - XOI

and their ratio equals:

..
J~ J20n
~- = VI- ();{2b)a.

Taking this into consideration and inte~rating tht' radiat~d

field over the open end of the wavegutde as was done 1n
24-2122 369
Paragraph 4-7, we obtain the expression of the radiated
field in the plane of the electric vector
ouR·atr
E,~,,)=-A, n•ro [I+ V l-(A/2b) 1
cosct]X

sin(~ sin a)
X ka e-lkr, (10-30)
slna -
2
and in the plane of the magnetic vector
E,y, ,,=-A. ~:[Vl-(A/2b) 1
+cosp]x
cos ~sin P)
\1
X ('A/2b) e-lkro (10 31)
- sin 1 ~ • -

In these expressions, ro is the distance from the centre of

the aperture to the point of observation of the field; ex and P
are the angles between the direction ro and the z-axis in
the xz- and yz-planes respectively.
The width of the directional diagram at half-power in
the electric plane is defined by the same expression as for
an ideaJ pJane antenna, i.e.,
(4-45)

Thus, for standard waveguides, we have a) A0 per ~ 3.4 ctn,

a= 1.28 cm, b= 2.86 cm and 2a..,,= 136°; b) Aoper= 10.5 cm,
a=3.4 cm, b=7.2 cm and 2a..1.=157°.
The width of the directional diagran1 at the radiation
zeros in the electric plane is expressed as:
sin a.0 = ~ . · (4·42)

However, for standard waveguides aA > 1, therefore the·

directional diagram in the plane of the electric vector
has no zeros.
In the plane of the magnetic vector, the width of the
directional diagram at half-power is expressed as.:
.,._ (4-52)
810
and the width of·the directional diagram at the radiation
zeros as: sin P. = l.S ~ , ,
derived from (4-51).
For standard waveguides Alb>l, hence the directional
diagrams in that plane likewise have no zeros.
As seen from (10-30) and (10-31), when the operating
wave-length approaches the critical one (A .... 2b), the radi·
ation in the opposite &
direction(cx=P= 180°) t ti/.Z
increases due to the (u
fact that when 'A-+2b, a8
the ratio J';' /J~ tends a7
towards infinity, Le., a6
in that case, we get 05
a predominance of the Q4 N..,
electric field at the a~ ~
open end andJor small Q2 )
ft

a/A,, the directional af - C%

diagram in the E-
vector plane becomes 0 20•11?60 rw'toout?ff't40'liirt80~o·'
•

non-directional. As ~
regards the H-vector, tO ~.................--,.....__,..----.~.,.--~~..,
in that case the di- 0.9 ~r+-----+--+---t----+--~-r--t--r--1
rectional diagram Q8 t--~-+--+---t---t----r---t---r--r----1
acquires the shape of Q71---+-\--+--+---+--+--1r---+-;--r--t
a "figure of eight": QB~---+~ -t---+--+-+--+--+----t-----t-r
The expressions a5 .-....- ---~~-~---+--+---t---+-t---t
( 10-30) and ( 10-31) Q4 1--1---1"'-~·---+-+--+·---+---t--r---t
are approximate since 03 .__~-.+--:w--+--t--t--t---r-"""1----w
they do not take ac- Q21--+---+-.r-1~~-+-lf--+---+--t--t
count of the distortion af ...__.._....__~
of the field during ·
reflection of the elec- o
tric waves at the open
end of the waveguide Fig. 10-57. Directional diagranJs of
rectangular waveguide:
and their branching a b
1==3.2 cm; 1:"~0.32: r==0.71.
off to the outer surface
of· the waveguide. .
However. in practice these expressions are found to g1ve
sufficiently accurate results. .
Fig. 10-57 gives the power directional diagrams tn the
E- and H-planes. The solid curves are the calculated ones
871
and the small circles correspond to the measured values.
The broken line curves have been constructed with an
account of the reflection from the open end of the waveguide.
The directive gain of the open end of a rectangular wave-
guide can be calculated frorn the expression
D ~ ~ [I+ JI'I-(A/2b) 1 1
} ab (l0-3 2)
n YI- ('}..f2b) 1 A1
.... •

As a radiator, the open end of a waveguide has a fairly

wide pass-band. The pass-band is, on the \vhole, determined
by the device which excites the waveguide, i.e., the radiator
connected to the coaxial cable. In order to widen the pass-
band, this radiator usuaJiy is made thicl< and has a specia1
shape, as shown in Chapter Nine (Fig. 9-16, b).
To decrease the reactive part of the input resistance of
the radiator, its l~gth is taken close to 'A/4.
The aciive part of the radiator input resistance is
adjusted by changing d (the distance from the radiator to the
short-circuiting plunger) or the appropriate choice of the
distance between the radiator Yo and the lateral wa 11 of the
waveiuide. Indeed, if a travelling wave is set up from the
radiator towards t~e open end of the waveguide, the active
part of the radiator input resistancr. can be determined
from the ex press ion
Rin= Y l V- JJ./E
2
2l Sin
-b- • 1 (2n
'1
d) Sin. • \' -bn Yo ) '
(A/2b) 2 a "'w
where l is the effective length of the dipole. Thus, the
equality bet\veen the radiator input resistance and thP
wave impedance of the feed coaxial cabJe, i.e., complete
matching, can be achieved by the proper ciJuice of d and Yu·
Another circuit of the exciting radiator is shown in
Fig. 10-58, where use is made of a transverse stub, which
makes the radiator rigid, electrically more reliable, and
widens its pass-band. Other circuits for connecting coaxial
cables to waveguides are also applied.
To lower the reflection from the open end of the wave-
guide and increase the directivity, horn radiators are
employed. Thus, Fig. 10-59 shows a so-called sectoral horn
flared in the H-plane, in which the dimension b gradually
•
Increases.
The wave arising in a sectoral horn is similar to the
wave in a rectangular waveguide and the picture of its
r.

.81J
field is shown in Fig. 10-60. A sectoral horn differs from a
waveguide in that the wave front in it form~ a cylindritaJ
surface, the phase velocity of the wave is a variable quantity
which is a function of the ratio bi"A and the field at a great
2 d A I Secflon through A-A

A
Fig. 10-58. Circuit of excitation by means
of a transverse stub:
1- mntchi ng dhphragm; 2--pl unger.
:c

Fig. 10-59. Sectoral horn, widening in the f/-plane.

r"\ , .....

- "'Yf\f~d I
b
~ ~---::;...' \,"'!",.,J \~I .._1'I '"I·
'J,'

'-'

t-ig. 10-60. The Held in a sectoral horn.

distance from the threat of the horn assumes the character

of a purely transverse wave. The wave . phase velocity is
1
approximately expressed as v= y v 1
and in the
1 - ('A/2b)
vicinity of the horn opening, it approaches the velocity of
light, which leads to a decrease of the reflection of the
wave from the radiating surface of the horn.
873
 If the flare angle of the horn G>o is sma11, .the wave front
at the opening is approximately plane and the radiation
characteristic in the yz-plane can be calculated by means
of the expressions (10-31) as well as (4-52) and (4-50).
Thus, the major lobe of the directional diagram narrows
down by about as many times as the dimension b of the horn
opening increases relatively to the dimension b of the
opening of a rectangular waveguide. ....
However if the flare angle of the horn <I>o is large, the
wave front at its opening differs considerably from a plane
NN one and the radiation instead of being
concentrated in a narrow beam will,
on the contrary, be dissipated in all
directions (the directional diagram of
•
the horn will be \vide) .
The difference of phase of the field
in the middle and at the edge of the
Fig. 10-61. Determin- opening can be deterrnined from the
ing the distortion of
the field phase at the approximate expression obtained from
horn opening. the geometrical din1ensions in Fig.
10-61
2~ nb2 n b «Do
11'=-r MN ~ 4'AR = T A. tan 2 , (10-33)

where R is the length of the horn.

Thus, the field phase at the horn opening changes in
accordance with the square law. As revealed by detailed
calculations, maximum radiation in the principal direction,
at a prescribed length of the horn, occurs when the differ-
ence of the field phase 'f'= 135°. This is due to the fact
that an increase of the flare angle uf aperture «Do leads, on
the one hand, to an increase of the relative dimension of
the horn bliv and a narrowing of the dh:ectional diagram,
and, on the other hand, to an increase of the difference of
the field phase 1J> and a widening of the directional
diagram. As a result of the action of these opposed factors,
the directive gain is found to be at a maximum precisely
when the angle of the difference of phase ,=
135°. The
horn corresponding to the condition "i'= 135° is referred
to as optimum horn.
The directional diagram of an optimum horn and of horns
In which w< 135°. can be calculated from the ~presslons
(10-30) and (10·31). -~.

874
 Apart from sectoral horns Oared in the H-vector plane,
se~toral horns flared in the E-vector plane are, also empJoyed
(Ftg. 10-62). In that case, the width of lhe directional
diagram in the H-plane is :r:
the same as at the o)'len
end of a waveguide; and in Q
the E-plane, the width of •

the directional diagram de-

creases as the dimensioJJ a
of the horn increases, pro- Fig. 10·62. Sectoral horn flared in
vided the flare angle <D\J is the £-plane.
taken small enough. In
that case, the difference of phase of the field corresponding
to the optimum is 1j)-=90° and the directional diagram
in the plane electric vector can be calculated from (10-30).
Pyramidal horns of rectangular cross section (Fig. 10-63)
are widely used. These horns enable to narrow down the
directional djagrams in the H-plane as well as in the
E·plane.
In a pyramidal horn we get the formation of a spherical
wave, the phase velocity of the wave is var-iable and as it
moves towards the open end of the horn, the wave front

Fig. 10-63. Pyramidal horn.

turns into a plane one with a phase velocity approaching

that of light. As a result, the reflection of the wave from the
opening of the horn is insignificant: the horn matches the
waveguide to open space.
The phase distortions of the field at the opening of a
pyramidal horn can be determined from the ex press ion ( 10-33)
in the H-plane and from an analogous ex,pression (replacing
b by a) in the £-plane. If the distortion of the field phase
at the opening is insignificant (11' small), the directional
diagrams of the horn differ but little from the directional
diagrams of the eo-phased plane and can be calculated from
(10-30) and (10-31).
818
 However, smaiJ phase distortions necessitate the use of
Jong horns. To reduce the length of the horn, the distortion
tolerated for the field phase amounts to 1J'= 135° in the
H-plane and 1J'==90° in the E-plane. As indicated above,
a horn of this kind is referred to as an optimum horn and
a sufficiently accurate definition of its directional diagram
is given by the expressions (10-30) and (10-31).
As regards the directive gain of sectoral and pyramidal
optimum horns, it can he approximately calculated frona
(10-32) which, due to the fact that A/2b~I. canbewritten
as:
- 4nabk
D - 'A• •' ( 10-34)
where kt=O.Bl.
The coefficient k 1 =0.81 applies to a eo-phased surface.
If we take into account that the field phase at the horn
opening is not constant but changes in accordance with
the square law, it will be found that this coefficient k1
equals 0.64.
Fig. 10-64 shows the directional diagram of a pyramidal
horn with dimensions a/A=9.4, bllv= 11.6 and RIA-=41.
I
?~
D.8 ·~~11" • 1 ,
~
! QB D-Af:
1/J "Tiia
~~ -V
~

~ Q4 ·' ~
\
:.b ' I
~02 ,
Q
020 f6
-r~-r1 Jq
12
8 4 0 4 11 f?. IIJ ,•

Fig. 10-64. Directional diagram of a pyramidal horn:

l-in the E-plane; 2-in the H-plane .
•
The directive gain of this horn is of the order of D~900
(28 db). The directional diagram in the H-plane is charac-
terised by relatively small side lobes, which is due to the
eo-sinusoidal distribution of the field amplitude in the
H-plane of the horn.
Apart from horns of a rectangular cross section, horns of
a circular cross section, i. e., conical horns, also find appli-
cation. Horns of this kind are formed by flaring the open
end of a circular waveguide excited by an . Ht 1 wave. The
816
radiation of a conical horn is similar to that of a pyrarnidal
one and it likewise has optimum dimensions, which can be
regarded as intermediate between those of optimum £-plane
and H-pJane sectora l horns. The same directive gain corre-
sponds approximately to the same dimensions of optimum
pyramidal and conical horns.

10-11.
•
Stow Phase Velocity Impedance Antrnnas

Apart fron1 director, helical and poly rod antennas, which

belong to the class of slow phase velocity axial radiation
antennas, slow phase velocity impedance antennas (antennas
of surface waves)
have also found
application in
recent years in the Q
centin1etre · wave
range. An antenna
of this kind is
shown in Fig.l0- 65.
The antenna con-
sists of a ridged sur-
face of length Land
width b, excited by
a pyramidal horn.
The surface wave
excited by the horn Fig. 10-65. lntpedance antenna.
is propagated along .
the impedance,structure without change of the field amplitude
(taking no account of energy losses in the structure). On
reaching the end of the structu re, the energy conveyed by the
slow wave is radiated. Moreover, part of the energy is radiated
on leaving the horn. The proper choice of the length L of
the ~mpedance structure, depth of the grooves l formed
by the ridges of the structu re and dimension a of the
opening of the horn produces a rather intensive radiation in
the direction coinciding with the direction of motion of
the wave in the antenna. As shown by theory and experi-
mental investigations impedance antennas radiate in a
way simila r to polyrod antennas, so that the corresponding
expressions of travelling-wave antennas may be applied
to them.
871
 Let us investigate the formation ot the surface wave. Let
the magnetic intensity of the surface wave be expressed as:
H,-H
- 0
e-P%-IIu I
(10 .. 35)
i. e., assume that the H vector coincides in dia·ection with
the y-axis and does not depend on that coordinate. The
dependence of the field on the x-coordinate is supposed to
be exponential (p is a real positive quantity). The field
is assumed to depend on the z-coordlhate in accordance
with the travelling-wave law (h is a real positive quant.ity).
Since the magnetic intensity n1ust satisfy the wave equation
o1 H iJ1 H
ox•~ + oz•' + k:H1 = 0, we get the following relation
between p and h:
(10-36)
Since p is a real quantity, h > k 0 , ~nd the phase velocity
of the wave v = -F
is lower than the velocity of light
m -
v, ==r.
0
The electric field intensity is defined from the equation
rot H = iroe 0 E and.. has only two components
E = h H e-px-ihz.
" WEo o ' (10-37)
E = _ .P H e-px-ihz.
z uoeo o

The transverse component of the electric field intensity

E" is in phase with the magnetic · intensity HJl and the
energy flow density in the direction of the z~axis
S11 = ~Efi; is a real quantity. The longitudinal component
•
of the electric field Ez has a phase displacement of 90°
relatively to the magnetic field intensity H 1 , and the energy
flow density in the direction of the x-axis S"= -iEaH;
is an imaginary quantity.
For the slow wave mentioned above to be maintained,
the surface impedance of the structure should be defined
by the quantity
(10-38)'
'
818
and since p. is a real and positive quantity, it must be
purely reacttve and of an inductive natur~.
A sur!ac~ !mpe~ance of this kind can be obta ined by means
of an tnftntte rsdged stru ctur e (Fig. 10-66). Inde ed, the
intensity of the mag- ~
netic field arising insid e ~·
the ridg ed structure (in
the grooves) has the
same component as in
surrounding space, viz.,
Hrtdee=ff,}dae. Assume Fig. 10-66. Ridged surface.
further that the fie) d
inside each cell of the ridged structure does not depend on
the z-coordinate. This approximation occurs on condition
that ·
h(s+-r)~l,
where s is Jhe width, of the groove;
-r, the width of the rib.
Then the wave equation for the magnetic field intensity
will be
i:JIHrldge
11 + k'Hr ldge _ 0
ox• 0 11 -

and its solution will be written as:

H~ldge=H~Idge coskox. (10-39)
Hence the components of the electric field intensity wi 11 be
ridg e_ Erldg e -0
BX - 11 -
and
Erldg e = - .ko Hrldg e sink X. (10-40)
z I(I)Bo 0 o
The expressions (10-39) and (10-40) represent a TEM wave
in the grooves •nd satisfy the boundary conditions on the
surface of the grooves.
The surface impedance averaged over the period of the
ridged structure will be:
Bridg e k0
zrtdg e == s 2
=- s . tank l. (10-41)
'r + s H rldre 'r + s 'CIJSe o
11
Identifying (10-38) with (10-41), we obtain:
p= f;,
k, tan k,l. (10-42)
81,
It follows that in order to maintain a surface wave, the
depth of the grooves should be less than a quarter of a wave·
length (kol <90°). When s-• 0 or k0 l--. 0, p--. 0, the phase
velocity of the wave approaches that of light and the attenu·
ation of the field intensity in the direction of the x-axis
decreases.
Let us now investigate the radiation of the ridged
antenna. The directional diagrarn in the yz-plane is de·
termined by the dirnension of the antenna and can be calcu-
lated from the corrf:'sponding expression for the radiation
of a pyramidal horn. In the present case we are interested
in the directional diagram of the ridged structure in the
xz-plane. Several methods are available for calculating the
radiation hut the most convenient is to use the equivalent
surface currents princi pie [64 ). Turning to Fig. 1Q.. 67 and
2

Fig. 10-67.
'
~------t----~~
Diagra1n of the antenna longitudina l section:
i -ribs; :--conductin g surface.

assuming, for simplicity , that the surface behind the

ridged structure is an ideal conductor of an infinite length,
we can integrate the radiation of the electric and magnetic
equivalent surface currents over the contour outlining of
the antenna (broken line in Fig. 10-67). At the same time
we can, as is usually done, neglect the integration over the
external surface of the horn and limit ourselves to the
integration over the open end of the waveguide and over
the surface x= l. However, due to the mirror image of the
electric and magnetic surface currents distributed over the
surface x= lt the radiation of the electric currents is
cancelled and that of the magnetic currents is doubled. Thus,
the radiated field will be defined by the double value of
the magnetic surface currents distributed over the surface
of the ridged structure and the radiation of the electric
and magnetic surface currents distributed over the open
end of the waveguide and over its mirror image relatively
to the x=l plane.
In case of adequate matching of the ridged structure to
the horn, the larger part of the energy d~,livered to the
'
l,

880
open end of the waveguide (horn) passes into energy of the
surface wave and the smaller part is ,radiated directly.
Hence, the radiation is, on the whole, determined by the
surface magnetic currents J.~ of the ridged structure and
the directional diagrarn is expressed as:
sin ~ ( :.t- cos a\
A V J
f(a)~· - - (10-29)
n.L v1
T v--cosu )
n1oreover, the maximun1 directivity of the radiation occurs
for the condition
(4-56)

tO~--~----~----~-----------------
aa~~-+----~----~------~----~----~
~~f
as~~~~---+-----+----~----~----~

a?~--~~------~-----~-------~------+----~
~
la6~--~r------;-----;-------:-------~--~
a
-l Q5
~ Q4 ...,___ ~,....._--l-----~-----+----4~--,______.
~
aa~--~~~~·--------:·------4·-----4·----~

Q2t-----+~ r+--+----.......~--~-----+----1

Qf~---~--~+4~~~~~~-----~-----~

tO 20 30 40 so 60
Angle, degrees
Fig. 10-68. Directionttl diagrarus ot irnpedance antenna:
/-exciter and surface are in phase: 2-exciter and surface are in antlphase.
,I.

Fig. 10-613 shows the theoretical and experimental direction-

al diagrams of an antenna of length L=7.33 A, in which
the condition of the optimum (4-25) is observed. The solid
curve is the experimental one and the broken curve, the
calculated one, no account being taken of the radiation
381
from the open end of the waveguide. The same figure shows
two calculated curves, plotted with account of the radiation
of the open end of the waveguide. for various phase
relations of the field of the surface wave at the beginning of
the impedance structure and of the field at the waveguide
aperture. It is seen that the theoretical and experimental
directional diagrams coincide fairly well and that it is
possible to obtain a matching of the horn to the impedance
structure for which the direct radiation of the horn affects
only on the side lobes, the major lobe changing but little.
Note that, in reality, the surface behind the ridged
structure (Fig. 10-65) has finite dimensions (b and d) so that
the mirror method used above is not valid. When calcu-
lating the radiation of the antenna, the radiation of the
rnagnetic as well as e1ectric surface currents have to be
integrated over the impedance structure. At the same time,
neglecting the reflection of the surface wave from the end
of the impedance structure, which is particularly justified
in case the depth of the grooves towards the end of the
antenna gradually decreases down to zero (Fig. 10-67),
one can regard the surface currents distribution to be the
same as in the case of an infinite plane. Then the directionaJ
diagram, with account of the radiation of the open end
of the waveguide and when d=O, is expressed as:
-:-----~--. sin nL ( ~- cos a
(10-43)
v nL(t'a
--,:- -; -cosa ) '

(0 <(X< :t).
Here, in contradistinction to the __ expression (10-29), the
first factor appeared as a result of the superimposition on
the radiation of the magnetic current J~ of the radiation
of the electric current I:; these currents are distributed
in the antenna in accordance with the expressions (10-35)
and (10-37) and the laws
JM _ P H e-Pt-th•
11 - - . •
Ui>80
and
e_ H e-Pl-t"h•
J. - 0 •

1t is seen from (10-43) that when the ratio ~ ls close

to unity, the radiation in the direction Qf the antenna
>
~

882
axis (a=O) is close to zero and the maximum radiation
of the antenna is directed at a certain angl,e to the plane.
Fig. 10-69 shows the experimental directional diagrams
of an impedance antenna of length L=7.33 A. The solid
~~~~~~~~~~~ n
~~~~~~~~~

~~~4~~~~~~1~~~~~~~~~~~
~-~~~~~++~~~~~~~~-~~~~+~~~
~~~~~~r+~~~~~~~~~4-~~~~~
l~'~
~

moJ~tla20~ 2102801'do 32fJ311Jro o 20 10 6J 100 120 t4o tlo tao

Ang/e,degrear .
Fig. 10-69. Directional diagrarr1s of an impedance antenna with a finite
~ plane... tt

. l
curve applies to an antenna in which the ratio of the width
of the ridges to the width of the grooves -.:Is equa Is 3, and the
dotted curve applies to an antenna in which this ratio equals
1/3. In both cases, the dimension d of the antenna is taken
equal to 'A/2. It can be seen from this figure that the radi-
ation maximum is approxin1ateJy at an angle a= 15° to the
plane, the radi~tion m
in the direction [ 1
0 r the p1a ne cons t i. ~~,.,.,.,"""""""',.,.,.,.~~"'9'PP9..............
tutes approxi· f •Z
mately 20% of the
power relatively to Fig. 10·70. Layer of dielectric (I} on a
the radiation in metal plane (2).
the main direction
and the radiation in the shadow regien (l80°<a<360°),
which is not taken into account in (10-43)t is found to be
insignificant.
Note that the impedance structure of the antenna repre-
sented in Fig. 10-65 can be formed not of a metal ridged
structure but of a dielectric layer (Fig. 10-70). In that
883
case, the electromagnetic field in the dielectric, for the
surface wave described by (10-35) and (10-37), is repre-
sented by the foJJowing ex press ions: ,.
H~= H~ cosgxe-•hz;
EdX = Q)£h Hd0 cos g"te-ihz.
• t
I

· Ed = - . g Hd sin tJxe-ihz (10-44)

z t(J)£1 0 b - t ~

\V here g= V k-~---h-1 :
k, =:k, ve:-teo;
e,/e 0 is the relative permittivity of the dielectric
layer.
It is readily sf'en that the expressions (10-44) satisfy
MaxweiJ's equations and the lioundary conditions on the
metal plane coated wtth a. dielectric layer.
From the expression (10-44) one defines the surface
irnpedance of the dielectric layer ·
"

E~ i
Zd =/id= - iroe, tan gl. (10-45)
1/ •

ldentifying (10-38) wTth (t0-45), we obtain the condition

for which the surfa~\ wave will be n1aintained: ·
e
--L p cos gl =g sin gl. (10-46)
80

It is readiJy seen that the equation (10-46) is observed

when ko<h<k 1 and gl<90°. This equation serves to define
the magnitude of the propagation constant h and, conse-
quently, of the phase velocity v.
In the existence of suitable matchi,ng at the horn output,
a sufficiently accurate definition of the directional diagrarns
of an impedance antenna with a dielectric.Jayer is given by
the expressions (10-29) and (10-43). Experience shows that
in1pedance antennas with a dielectric layer have higher
losses than ridged antennas. The efficiency of these antennas
is adequate only on decimetre waves.

10·12. Lens Antennas

Lens antennas of various designs find fairly frequent
application in the centimetre wave range. In these antennas
use is made of the optica 1 propert-ies of el.ectro1uagnetic
884
waves, since the dimensions and curvature radii of the
surface of the antennas greatly exceed tl)e wave-length
of the excited oscillations. The purpose of lens antennas
is the formation of hfghly directive radiation diagrams
and the. transformation of a cylindrical or spherical wave
front into a plane one.
Fig. 10-71 represents a di~lectric lens on which a spqeri-
cal wave of .length Ao is incident from a source F. Due to
the convex form of the irradiated surface of the lens, the
d

.. t>

n ,_,......

Fig. 10·71. Dielec,ttlc Flg.t0-72. Explaining the c~lcu­

lens. latlon of the lens profile!~
I-hyperbollc sflape of surface.

"'
spherical wave front is transformed ia it. into a plane one
.
with a wave-length A.= ,VAo • The output surface of the
£/Eo
lens is flat, so that, on leaving the lens, the wave front
remains flat. Because the dimensions of the output surface
of the lens are large in comparison with the wave-length,
its radiation is highly directional.
Let us now turn to Fig. 10-72 and define the lens profile.
Let a ray fall from pojnt F on point P on the lens surface
and let " be the angle of incidence of the ray and IJ, the
angle of refraction. We have:
•
sin~ ==n sin p,

where n= Vs/so is the coefficient of refraction.

The equation of the lens profile will be determined from
the condition that the length of the optical path (electric
length) from the equal phase surface undergoing transfor-
mation to the transformed one should be constant in any
direction, i.e., FP=FQ+nQQ'. Hence we obtain:
r==f +n(r cos8-/)
25-2122
or
r- (n-1) I
_;,...~.;....;_ (10-47)
-n cos 9-1'

where f is the focal distance, i.e., the distanc-e from the

focus F to the apex of the lens Q;
8, the angle between the axis of the lens and the
ray:
r, the distance between the focus- and the reference
point of the lens.
The equation (10-47) is that of an hyperbola. Thus, ·the
convex surface of the lens has an hyperbolic shape.
The thickness of the lens d is determined along the
optical path and is related as foJio\vs to the diameter of
the lens D:
d-1.. ( D
- n 2sin 9max
-f) •
(10-48)

The expressions (10-47) and (10-48) enable to calculate the

profile and dimensions of the lens.
Since the output surface of the lens is a eo-phased surface,
it can be regarded as an ideal plane antenna. In the case
of a point source (Hertz
. dipole), placed in the focus
of the lens F, the lens
should be circular and its
output surface will have a
circular form. In the case of
a I inear source, placed on
the focal line F, the lens
should be cylindrical and
its output surface will have
a rectangular form.
Fig. 10-73. Metal-plate lens. Dielectric lenses have a
wide pass-band, due to the
fact that the coefficient of refraction n is frequency independent
In a wide frequency range, including e\"en the highest ones.
However, they are relatively seldom used on account of
their weight, considerable energy losses and relatively
high cost.
High phase velocity metal-plate lenses find application
In antenna technique. Lenses of this kind consist of parallel
metal plates spaced at a distance b apart and formf.ng a
concave surface (Fig. 10-73). The electric field vector E
886 ~
excited by a source situated at the focus of the lens, should
be parallel to the plates. Then, the space between two
neighbouring plates represents a waveguide in which we
get the excitation of an H 01 mode with a phase velocity
01
v= y 1
• Thus, the medium formed has a coefficient
J - (A.J2b)
of refraction lower than unity,
(10-49)
It is clear that the distance between the plates b should
be taken larger than half a wave-length but smaller than a
wave-length, i.e., {<b<A, in order that only one
mode, viz., H01 , should be propagated betwPf•n the plates.
Consequently, the coefficient of a
refraction of a lens of high phase fJ
velocity lies within the interval
0<n<V0.75.
The equation of the lens D
profile wi 11 be defined from -----. f _ __.,
Fig . 10-74. The ray incident from
focus F on point P of the input n I
surface of the lens is refracted In Fig. 10 •74 . Explaining the
accordance with the law of geo- calculation of the lens profile:
metrical optics sin a,= n sin p. /-elliptical form of surface.
The optical length of the path
F Q should be equal to the optical length of the path FP'.
i.e., FQ=FP+nPP'.
Whence we. obtain the equation of the lens profile
f=-r + n (f-r cos 8)
or
(1-n) f
r-~-~
-l-ncos8 (10-50)

which is the equation of an ellipse.

The thickness of the lens d is related to its diameter as
~

follows:
d=.!_
n
(f- 2 sinDOmax ) • (10-51)

A lens consisting of identical plates of elliptical profile

focusses electromagnetic energy in the electric vector plane.
25• 887
If the lens consists of rectangular plates of different thick-
ness (d), which a Iso form an elliptical profile, it focusses
electromagnetic energy in the magnetic vector plane. In
order that the energy should be focussed both in the
electric and magnetic vector planes, the lens should consist
of plates of elliptical profile of different thicknesses.
Thus, a high phase velocity metal-plate lens transforms
a spherical wave front (in the case of a g.oint radiator) or a
cylindrical wave front (in the case of a linear radiator)
into a plane front and enables to obtain a co-phasal surface
at the output of the lens. The construction of these lenses
is very simple. The use of a metal-plate lens in combi-
d nation with a horn, for example, con-
I siderably reduces its length. Thus, in
order to obtain an opening size b=
=40 A. in an optimum horn, the latter
should be given a length R~BOO i.,
214--- F whereas the use of a lens at the opening
of the horn enables to reduce the
length of the horn down to the size of
the opening, i.e., R~40 A.
Fig. 10-75. Explaining Metal-plate lenses have a relatively
the calculation of the narrow pass-band, due to the fact that
distortion of the field the coefficient of refraction, as can
phase in a lens.
be seen from the expression (10-49),
is a function of the frequency. Let
us calculate the distortion of the field phase at the
opening of the lens in case of a change of frequency and
estimate the pass-band of the antenna. Let "i' 1 be the field
phase at point 1 of the lens and 1f'1 , that at point 2 (Fig.
10·75). Then, the di fferencc of phase of the field at these
two points wi 11 be:
~=11'.-\1>, =kf-k (R + nd).
Let n 0 be the coefficient of refraction for the wave-length
1 0 in the middle of the pass-band. On that wave-length,
the difference of phase 1J> will be zero. On the wave-length
A,+ ~A. the coefficient of refraction will be:
n==n,+;l UA tl = llo 6A
and the difference of phase will be expressed as%
·~-k.d~ln•n.41. r~r>
But since
~
an
aA
I n=no == Y
- "'• 1
(2b)
1- (A,0 J2b)1 =
-

no ==
Ao
(2b) 1 I-
- noAo '
n:
the difference of phase of the field will be written as:
1-n~ A/ d
1J' ~ 2n --. (10-52)
no '" ).,
The permissible value of distortion of the phase is taken
equal to 'lj> = T. In order to avoid considerable reflections
when the wave is incident from the radiator on the lens,
as well as to avoid the use of a lens of excessive
thickness the coeffic.ient of refract ion is usually taken equal
to n,=0.5. Then, the antenna pass-band equals:
2
t' = 33 ~ [
0
/ .1 . (J0-53)
As can be seen from this expression, the antenna pass-band
is inversely proportional to the thickness of the lens d,
and if d~ Ao, the pass-band is very narrow.
In order to reduce the thickness of the lens d and widen
its pass-band, as well as make it easier to manufacture,
the lens is usually zoned (Fig. 10-76). The irradiated part
of the lens is stepped, and the steps are taken of a depth (t)
such that the rays refracted in neighbouring zones of the
lens arrive at the output of the lens with a phase shift equal
to 2n.
Eat.-h of the zones of the lens has its focal distance and
the equation of the profile of the m·th zone will be:
(I-n) f"'
'• = ~1--n-c......
os--=--8 • (10-54)

where f.==f+(m-1) t, m= 1, 2, 3...

The size t is precisely chosen so that on the calculated
wave-length ).,, the field at the output. of the lens should
be eo-phased:
k,t-k,nt-=2a.
from which we derive:
(10-55)
 As a consequence of the zoning, there are non-irradiated
areas at the lens input, which cause shadow regions at its
output (Fig. 10-76). These shadow regions somewhat lower
the directive gain of the antenna. The lowering of the
directive gain is evidently due to the dispersion of part of
the energy on the steps.

'1'/

2 .......-

\
\
Fig. 10-76. Zoned lens: Fig. 10-77. Explaining
1-shadow areas. the calculation of the
frequency band.

Let us examine the width of the pass-band of a zoned

lens. The field phase at point I of the lens (Fig. 10-77)
equals 'I\ ==k (R+nd,). and that at point 2 equals ~' 1 =kf.
On the calculated wave-length A0 , the phase difference will
be:
w=k 0 {-k 0 (R + n 0d,)=- k 0 (M -I) A0 ,
where M is the number of zones of the lens.
On the wave-length "-o + L\A., the phase difference will be:
11'= (k 0 + L\k) f- (k 0 + llk) [R + (n 0 + ~n) d1 ] =
= -(k +Ilk) (~1-1) A0 -(k 0 -t ~k) ~nd 1 =
0
== -2n (M -1)-(M -1) 'A 0 ~k-k 0 d 1 lln.
Discarding 2~ (M -I) and taking into a'ccount that
8nl
lln = :s:r L\A. = -
1-n:
A. 11A
UA ll=llo fto
and
akl
Ilk == en; ,_=J.o A'A= -
4A
k 0 I; ,
we obtain for the phase difference
~·~ 2n(M-l) ~+2n l-n: ~lit. (10-56)
l, •• le .·A1
890
 lfweassumethat 'lj)'=T and n =0.5. ~he expression for
0

the pass-band wi 11 be:

1
2
~1 = 50
['/,]. (10-57a)
0
(M-1)+1 .5 ~
0

A comparison of (10-57a) with (10-53) shows that a

zoned lens has a wider pass-band than an unzoned one.
Indeed,. the relation of the quantity d1 in· (10-57a) to
the quantity d in (10-53) is dt=d-(M-l)t and the
expression (10-57a), account taken ~
that 1=2A0 , becomes: ~......,.__-
2Af: 50 [ 1 /.]. (10-57b)
fo 1.5~-2(M-l)
Ao Az(~~==~~~
Thus. at a prescribed value of d,
an increase of M leads to an
increase of the pass-band.
As regards the distribution of the
Fi~ . . J0-78. Explaining
field amplitudes at the output the calcu Iati on of the
opening of a metal-plate ]ens of high field ~nnplitude in the
phase veJoci ty, in the case of a lens.
uniform irradiation of the lens, the
field an1pJitude increases towards the edges of the lens due
to the fact that when the angle 8 increases, the va Iue of
the surface at the lens output, corresponding to the same
increment of the angle 9, decreases, which can be seen from
Fig. 10-78. The distribution of the field amplitudes at the
lens output when it is ilJuminated by a spherical wave
(spherical let1s) ts expressed as
1
E- , / (1-n cos8) F(O)
V t• (l-n)1 (cos 0-n) '
where F (8) is a factor which takes into account the direc·
tivity of the radiator.
For a cylindrical wave front (cylindrical lens), the
distribution of the amplitudes at the output of the lens
is expressed as:
E- yl (t <_!.:;;>~~:: ::..n) F (8).
The proper choice of the directional characteristic of the
radiator F(9) enables to obtain the desired distribution of the
891
field amplitudes af the lens output. Moreover, the directional
diagrams can be calculated from the corresponding ex·
pressions of the eo-phased plane antennas. The directive gain of
a lens antenna, in particular a horn lens, can be expressed as:
4ns
D= A,l k,, .
where s is the area of the output surface of the lens:
k,, the coefficient taking account ef the distribution
Jaw. of the field aruplitudes at the output of the
lens which is usually equal to 0.5-0.7. ·
Let us now examine the requirements concerning the
accuracy with which metal-plate lenses should be made.
4tl The accuracy is determined by
~t:. the permissible phase distortions
which occur when the dimensions
of the lenses do not correspond to
the calculated ones. Let fld repre-
F sent the inaccuracy to size of the
thickness of the lens (Fig. 10-79).
Then the phase distortion will be:
1J> 1 = k 0 dd ( 1- n0 ),
Fig. 10-79. Explaining
the calculation of the and if we take the permissible phase
accuracy to size of the lens.
distortion to be '1', = the per- T,
missible tnaccuracy to size of the thickness of the lens will
be expressed as:
lld- 16 (~no) (10-58)
lt follows that the closer i~ no to unity, the higher the
inaccuracy to size of the lens that can be tolerated.
Let now ~b represent the inaccuracy to size of the
distance bet\\t·een the plates within the limits of a certain
thickness d' of the lens. Then, the phase distortion will be:
~.-k,d':: !lb.
on
But iJb• -
J -n:
bnn , and
.If we assume that V. = n
, the per-
8
missible inaccuracy to size of the distance between the
plates will be:
(10-69)
' '

'
 Apart from metal-plate lenses, use is also made in antenna
technique of artificial dielectric lenses, of slow phase
velocity. These lenses consist of metal balls, circular discs,
lates of rectangular or other form, as well as metal tapes.
fn fact they are artificial dielectric lenses. Fig. 10-80
represents a lens of this kind, consisting of metal balls of
a diameter small in comparison with the wave-length,
spaced apart at distances small in
comparison with the wave-length.
The prafi le of this lens has an
hyperbolic form and is described
by the equation (10-47).
Let us briefly investigate the
artificial dielectric theory. Let
small meta 1particles be suspended
in a certain volume. Under
the influence of the field of the Fig. 10-lll. Lens antenna
electromagn~tic wave incident on consisting of metal balls.
the medium thus formed, the
electric charges of the particles are displaced and form elec-
tric dipoles, the moment of each of which will be designated
by Pe· If there are N such dipoles in a unit of volume, the
polarisation of the medium will be P=Np,. The electric
induction will then be expressed as:
D=s 0 E+P.
On the other hand, if a is the effective permittivity of the
medium,
D=eE
whence we obtain:
I

and if we take into account that the moment of the dipole

is proportional to the acting electric field P.=a,E where
a, is the electric polarisa bi Iity or suscept iveness of the
dipole, the expression for the effective relative permittivity
will be:
(10-60)

Assume that the particles are Ideally conducting balls.

Let the electric field acting on a ball be E.; its meridio-
nal component will then equal Ee-= E. sin e. The field
191
excited by one ball (the influence of the neighbouring ones
is not taken into account) will be:
E" _ pe stn e.
8- 4ne0 r 1
From the boundary conditions on the ball surface
E;+ E;=O. we obtain:
-
where a is the ball radius.
Hence, the electric susceptiveness of the cl ipole is:
ex,= 4ne 0a 1 • (I 0-61)
Substituting this expression into ( 10-60), we obtain:
~=I+ N4na•.
Bo

At the same time, 1he coefficient of refraction of the

medium will be:
n= ..V/ e.!::" =
0
Vl + 4nNa'. (10-62)

The expression for the electric susceptiveness and

effective coefficients of refraction of the media formed of
n1etal particles of a different shape is:
a) for an ellipsoid of major axis 2a and minor axis 2b
(10-63)

b) for a thin circular disc of radius a, the electric

vector being parallel to the plane of the disc,
.
16
a,= 3 s 0 a,
•
n= y' 1+-af:la,16 ••
(10-64)

c) for a thin rectangular tape of width a and relatively

large length b (b ~A), the electric vector being parallel to
the narrow side,
cz.= ~ s 0a!, n= V 1+ ~ Na•; (10-65)

here ex, is the polarisability of one unit length of tape;

N, the number of tapes per unit surface of longi-
tudinal aection of the medium.
'

894
 On defining the coefficient of refraction of the arti-
ficial dielectric, we proceeded from the'· existence of the
electric dipole moment of the particle.
How~ver~ the excited metal particle possesses also a
magnettc dtpole moment. Indeed, the electromagnetic wave
incident on the metal particle excites a hirrher current on
the irradiated side of ~-....
the particl e than on fi Direction .,~/
~'
'\o
its s~adow side. -.:he Hz Df ff:t"$, 11\ a
electrtc current tn- J 1
duced in the particle ''"' ,./ ...,IJI.,;,.J....
can be represented, as (a) = '~
shown in Fig. 10-81, Fig. 10-~1. Influence of an el.ectromag-
as the superposition of netac wave on. metal partaclea:
. l curren t a~ ball. b-dlsc.
the d tpo e
(solid curves) and loop current (dotted curves). An ele-
mentary electric loop can be regarded as a magnetic dipole
whose axis is perpendicular to the plane of the loop.
J

If p,. is the magnetic moment of the dipole (of the

electric loop) and N, the number of magnetic dipoles in
a unit volu1ne, the magnetic induction will be:
8=Jj 0 ll + p,.N.
Since the magnetic moment of the particle is proportional
to the intensity of the acting magnetic field p, =a.,H ,
where am is the magnetic susceptiveness of the particle,
the following expression will be obtained for the effective
relative permittivity of the medium
~ == 1+ Na, · (10-66)
Jl.o fJ.o
Let us examine the magnetic field of the loop electric
current induced on the ball (Fig. 10-81). I. et H, be the
magnetic field intensity of the electromagnetic wave inci-
dent on the ball The incident wave component of the
magnetic field intensity ,radial to the ball will be:
H,=H ,cos9. .
The radial component of the intensity of the magnetic
field set up by one magnetic dipole (no account taken of
the influence of the neighbouring ones) is expressed as:
,- p,. cos 8. .
H r- 2nJ.L.r'
From the boundary, condition
, on the surf.ace of an ideally
conducting ball li,.+H,.=O, we obtain
1
p,= -H,2nJl 0a •
Since p,. ==arH~' \ve shall have for the magnetic sus-
cepti veness o the particle:
a,. == - 21tJL oa•. ...
"'
Substituting this ESxpression into (10-66), we find:
~= l-N2na•. {10-67)
Jlo
Thus, the effective permittivity of the medium decreases
and the coefficient of ·refraction of the medium, instead of
being defined by (I 0-62), is defined as:

n= , /
V
eJ&
2ollr
== Vo + 4nNa')(l-2nNa') . (10-68)

In the . case of, for example, a disc lying relatively

to the incoming wave as shown in Fig. 10-81, b, the loop
effect occurs too, hJ.Jt the area of the loop is small, so that
the magnetic moment is insignificant I see (1-16) ]. In that
case the exprt'ssion (10-64) is sufficient. If, on the other hand,
the discs should be pJaced in the plane of the incoming
wave, the loop effect of the induced current cannot be
neglected and, as a result of the decrease of the permit-
tivity of the medium, the coefficient of refraction will go
down. The same reasoning goes for particles of a different
shape.
1'"hus, to avoid lowering the coefficient of refraction of
the medium (at a prescribed number of particles in a unit
volume), the particles should be disposed in such a way
that their ndnimum dimension should coincide with the
direction of the wave progress in the medium.
Fig. 10-82 represents sketches of metal-dielectric lenses
consisting of discs and tapes.
Note that to reduce their weight and dimensions, arti-
ficial dielectric lenses, just like metal-plate ones, are zoned
(Fig. 10-83). The equation of the profile of each zone of
the lens is expressed as:
r == {!l-;-l) 1. , (10-69)
• 11 cus8-l ~ · ,.
~
where
/,.=/+(m-1)1, m=l, 2, ~•.••
The dimension t of the lens is defined as:
t= ~0 • n-1
(10-70)
In the operating wave range, the unzoned artificial
dielectric lens constitutes an aperiodic system. Due to the
fact that the equality (I 0· 70) is •
upset, ~oned lenses are frequency
dependent. The phase shift of
the field at the output of the

Fig. 10-82. Metal-dielectric lenses ~onsfsting of discs and tapes.

lens between the first (central) and last zones is e1·

pressed as:
,P=k(M -l)t-kn(iW -l)t= -k(M -I) A•.
On the calculated wave-length A0 , this
phase shift ~equals -2n(M-I), i.e.,
the fields are in phase. On the wave-
length A.=i"o + 4A., we have k=k, + &k== ,
==k. -k0 f•
and the phase shift equals:
L\1
11'= 2n (M -l)~-2n (M -I).
Discarding 2n (M -1), we obtain th~ Fig. 10•83. zoned
djstortion of the phase at the output artificial dielectrlo
~t~lms •~.
'i''==2a(M-l)~ ·
If we regard 11?' =.!} as the permissible phase distortion,
it will be found that the pass-band of the artificial
dielectric lens Is expressed as:
2t!f - 50 [0/ ] (10-71)
fo -M-1 o·

The accuracy to size of artificial dielectric lenses is

defined by the same expressions as for metal-plate ones.
The observatk>ns made earlier regardin« the formation of
shadow regions in zoned metal-plate lenses fully apply
also to artificial dielectric ones. ·
Note further that part of the energy radiated by the source
p· and incident on the lens is reflected from it. That
part of the energy which is reflected from the illuminated
(input) surface of the lens is, on the whole, dissipated and
distorts the directional diagram of the antenna, whereas
that part of the energy which is reflected from the shadow
(output) surface is, on the whole, focussed in the radiator
and leads to a change of the coefficient of reflection in the
line feeding the radiator. The coefficient of reflection from
the output surface of the lens is expressed as:
n-1
P=n+l · •

This same quantity also approximately defines the coef-

ficient of reflection in the line feeding the radiator. If we
regard p=0.3 as permissible, it will be found that the
coefficient of refraction of the lens should equal n= 1.86
(n=0.54 for a metal-plate lens). To reduce the coefficient
of reflection, it would be desirable to take n close to unity,
but this would necessitate a lens of greater thickness. For
this reason, other means are used tu teduct! the reaction of
the lens on the radiator. Thus, Fig. 10-84, a represents a
lens of slow phase velocity one part of the surface of which
is displaced by a quarter of a wave-length relatively to the
other part. As a consequence, the waves reflected from
neighbouring elements of the lens reach the radiator with a
180° phase shift and, therefore, cancel out. One can also make
use of the inclination of the output surface of the lens, as
shown schematically in Fig. 10-84, b, which prevents the
rays reftected by the lens from being focussed. Use is also
made of quarter-wave surface layers of dielectric, the coef·
ficient of refraction of which is taken equal to the square
root of the coefficient of refraction of the ·lens. Sometimes
898
the layer is of the same material as the lens but has quarter·
wave depressions, and that leads to a lowering of the effective
value of the coefficient of refraction. The quarter-wave layers

(a) (b)
Fig. 10-84. Methods for reducing the eft'ect of the lens
on the radiator:
a-displacement of one part 9f the surface relatively to the
other; b-lncllnatlon of the lens output surface.

on the input and output surface of the lens lead to a consider-

able decreaJse (theoretically down to zero) of the coefficient
of reflection of the wave from the lens.
Let us now consider the accuracy of coJncidence of the
radiator with the focus of the lens. There can be two kinds
p

..,___,_...,.
Q' Q

(Q) {b)

Fig. 10-85. Removal of the radiator out of the focus of

the lens:
a-displacement of the radiator Jn the dtrectlon of the lens axla~
b-dlsplacement of the radiator tn a dlrec.tlon perpendicular to the
axis; 1-wave front.

of displacement of the radiator: a) in the direction of the

axis of the lens and b) in a direction perpendicular to the
axis. In the first case, when the radiator is removed from
the focus to the point 0 (Fig. 10-85, a), the difference of
phase of the field at points P and Q' at the output of the lens
will be:
'=k [OP-(OQ + nQQ')],
891
and since
· OP ~ FP-FP'==f+ nd -a cos&;
OQ=f-a;
QQ' ==d,
we get
~= ka (1- cos 0).

If we regard 11'= -j-, as the permissible phase distortion,

the permissible displacement of the radiator along .the
axis of the lens equals:
•

a= 16 (1-cos 0) · (10-72)

The angle em••

is usually of the order of 30°, so that
displacing the radiator along the axis by rather large
distances causes an insignificant distortion of the
directional diagram.
It should be noted that when the radiator is displaced
in a longitudinal direction, the phase distortions, as can
be seen from the expression obtained, are of a symmetrical
nature relatively to the axis of the lens and are a para-
bolic function of ~he distance from the axis of the lens
(for small angles O'~i'~ka:). A phase distortion of this sort
does not change the maximum radiation direction of the lens
antenna but does change the form of the directional dia-
gram (widens it).
A slight displacement of the radiator out of the focus
in a direction perpendicular to the lens axis (Fig. 10-85, b)
leads, on the whole. to a Iinear distortion of the field phase
at the output of the lens, i.e., to a rotatton of the wave
front. This, in turn, rotates the directional diagra•n by an
angle a., appr\lximately equal to the angle by which the
radiator is removed out of the focus (sin a.~7> without
distorting the diagram. Large displacement angles of the
radiator give rise, apart from a linear distortion. to a
noticeable cubic distortion of the phase, leading to a
distortion of the form of the directional diagram, the relative
increase of the side lobes on one side of the major lobe
and their relative decrease on the other side. .
The displacement of the radiator out of the focus In a
direction perpendicular to the axis is m~d,"use of in radar
t~

400
for scanning directional diagrams in the horizontal or
vertical planes or for conical scanning. ,However, due to
distortion of the directional diagram, it can be done only
within the limits of a double or triple width of the major
lobe of the directional diagram.
There are methods for scanning the directional diagrams of
lens antennas in a very wide range of angles. Let us
describe one of such rather widespread lenses, named after
Luneberg, its inventor, The lens is a sphere (Fig. 10-86) with

2 4
3

Fig. I0-86. Tra- Fig. 10-87. Longitudinal section of a

jectory of the rays Luneberg lens:
in a Luneberg lens. 1-dlelectrlc; 2-metal plates; .9-area of
radiation: 4-axi~ of symmetry.

an inhomogeneous dielectric, the refractive index of which

changes according to the law

n= r 2-\a .
, / ' r )' (10-73)

where a is the radius of the sphere;

r, the distance from the centre of the sphere.
A lens of this kind focusses a para lie I beam incident on it at
a point of the surface of the sphere. If, in accordance with
the law (10-73), the lens has the shape of a cylinder, it will
focus a parallel beam in a line lying on the forming element
of the cylinder. In practice, the lens can be made of two
circular metal plates, one plane and the other curved, the
space in between being filled with a homogeneous dielectric
medium (Fig. 10-87). If the vector of the electric field
intensity is parallel to the plates, the coefficient of
refraction is expressed as:

(10-74)

26-tlll 401
Identifying (10-73) with (10-74), we obtain for the variable
distance b between the plates, the expression
b= Ao , (10-75)
2Yio--2+(f r
where a Is the radius of the plates;
e/e 0 , the relative permittivity bf ..the medium;
A0 , the wave-length in vacuum.
To irradiate the lens shown in Fig. 10-87, use is made
of the open end of a rectangular waveguide, the broad walls
of which narrow down to a dimension equal to the distance
between the plates at the input of the lens.

10-13. Parabolic Antennas

Parabolic antennas find very wide application in many
fields of radio engineering, especially in radar. They are
utilised in the centimetre, decimetre and, sometimes, metre
wave range.
Parabolic antennas are given the shape of a paraboloid,
a parabolic cylind~er or a parabolic cylinder bounded by

((I) {/}) {C)

Fig. 10·88. Typ~s of parabolic antennas=

a-paraboloid; b-parabollc cylinder; c-parabollc cylinder bounded
by parallel planes. •

parallel conducting walls (Fig. 10-88). The paraboloid Is ex-

cited by a dipole antenna placed in the focus and transforms
the spherical wave front into a plane one. The parabolic
cylinder is excited by a linear antenna lying on the focal
line and transforms the cylindrical wave front into a plane
one.
In these antennas, just as in lenses, use is made of the
optical properties of radio waves. The geomet.rical properties
I >,

•
40J
of the parabola are such that the rays directed from the
focus and reflected from the parabola become parallel to the
axis of the parabola, and the length of the path from the
focus to the parabola and then to the line parallel to the
x-axis passing through the edges of the parabola is the same
for any angle 8 (Fig. 10·89). In this way, a co-phasal
surface is formed at the opening of the parabolic antenna
and its radiation is h,ghly directional.
In Ca~tesian coordinates, the parabolic surface is defined
by the equation (the origin of the coordinates coincides
with the apex of the para~oloid)
x2+y2=4fz,
and in spherical ones, by the
equation (the origin of the coordinates
2
coincides with the focus of the I
paraboloid)
,• 2f I 2H
Q= 1-t....;__cos---::·a . I
J...-,
Parabolic antennas can be of the • I
short-focus or long-focus kind. In a
short-focus antenna, the focus is inside F1g. 10-89. Profile of
a parabolic antenna:
the antenna (/<~). in a long-focus /-opening; 2-apex of
the parabola.
one, the focus is outside the antenna
(f>if>· The limit value is the case f={-. when the focus
coincides with the opening of the antenna.
"fo calculate the radiation of parabolic antennas, use is
made of the equivalent surface electric and magnetic
current method. Two methods of integration can be used:
1) over the surface which passes over the shadow side of the
paraboloid and over its opening, and 2) over the surface
which passes over the shadow and illuminated sides of the
paraboloid. To simplify the problem, in both cases, the
radiation of the electric surface currents distributed over
the shadow side of the paraboloid is neglected and onlY the
radiation of the surface currents dfstributed over the
irradiated part of surface is taken into account.
In the first case, the irradiated part of the surface is
plane, the equivalent electric and magnetic surface
currents are regarded as co-phasal and the distribution of
their amplitudes assumed proportional to the quantity
26• 409
fF( 0), where p is th~ distance from the focus to the
paraboloid and F( 6), the amplitude directional charac-
teristic of the radiator. Thus, in this case, the calcuJation
of the radiation of a parabolic antenna is reduced to the
caJcuJation of the radiation of a co-phasal circular ~surface
with a corresponding distribution of the amplitudes of
the equivalent surface currents. • _
In the second case, the irradiated part of the surface
coincides with the ~urface of the paraboloid, so that ·the
equivalent surface magnetic currents on this surface equal
zero and the surface electric currents are real currents
induced on the irradiated side of the paraboloid. The electric
surface current is approximately defined as double the
value of the component (tangential to the surface) of the
magnetic field intensity of the incident wave from the radi-
ator. In this case, the distribution of the amplitudes and
phases of the induced currents is a more complex function
of the coordinates than in the first case and, in addition,
the direction of the currents changes from point to point
on the surface of the paraboloid. Consequently, in the
second case, the calculation of the radiation of a parabolic
antenna is effectea by integrating the radiation of the
electric currents induced on the paraboloid.
Both methods of calculating the radiation of a parabolic
antenna are approximate, since the distribution of the
surface currents is deterntined approximately. However,
the first method is less accurate since, because of the finite
dimensions of the paraboloid, the rays reflected from the
reflector and progressing towards the opening are not exactly
parallel and this is not lakt!U into account in the calculation.
On the other hand, the first method is less cumbersome than
the second one. •
Let us examine the picture of the distribution of the
currents induced by the radiator on the irradiated side of the
paraboloid. Assume the radiator to be an elementary electric
dipole directed along the x-axis with a plane counter-reflector
in the shape of a disc. If the counter-reftector is approxi·
mately replaced by the mirror image of the dipole, the
phase centre of the radiator can be regarded as coinciding
with the centre of the counter-reflector, so that this point
should be caused to coincide with the focus of the parKbuloid
(Fig. 10-90). Then, the maenetic fi~ld i¥ensity of the
" I ...

'r
radiator on the paraboloid surface in spherical coordinates
wi 11 be ex pressed as: ,·
e-itp
HffJ (8, ffJ)= F (8, cp) .
e.
In rectangular coordinates, the magnetic field intensity
has two components: H1 and ff1 • The vector of the surface
electric current density is defined as J"= (H, n] where n,
is the vector equal to unity and perpendicular to the surface
of the ·paraboloid. Since n has all three components,
there are all three components :r
of the electric current: J~,
J: and J; . Accordingly, the
picture of the electric field 2R
distribution projected on the z
xy-plane looks as is shown in
Fig. 10-91. In the case of the
long-focus systen1, the electric §
current components J! have
the same direction in aJI four Fig. J0-90. Parabolic reflector:
quadrants, and the direction 1-electrlc dipole; 2-metal disc.
of the current components JZ
changes from quadrant to quadrant. In the case of the
short-focus system, current nodes are present in the
paraboloid, situated at the points of intersection of the axis
of the electric dipole of the radiator with the surface of
the paraboloid; and the picture of the current distribution
on the xy-plane differs (on part of the surface) from that
in the case of the long-focus paraboloid. Now, in certain parts
of the reftecti>r, the electric current component parallel to
the x-axis is of opposite direction relatively to the same
component in the main part of the reflector.
The radiation characteristic of a parabolic antenna is
defined in this way, in the two principal planes: xz and yz,
by the current components J~, with the radiation of the
current components J: being cancelled out in these planes.
The polarisation of the radiated field lt linear. In the other
planes passing through the z-axis, radiation occurs too,
J:
defined by the current components and, as a consequence,
there appears a transverse (relatively to the principal one)
polarisation of the radiated field. The total field is ellipti-
cally polarised. The transverse polarisation, called cross
polarisation, is parasitic and its presence decreases the
antenna directive gain.
The level of cross polarisation is determined by the
curvature of the paraboloid surface: the smaller the radius
of the curvature of the reflector, at a prescribed diameter
2R, i.e., the less long- focussed the re Hector, the higher

•
(a) (IJ)
Fig. 10·91. Distribution of the electric current In a
paraboloid:
a-long-f ocus reflector; 6-•hort· focua reflector,

the level of the c-urrent components 1:

relatively to the
components J~, the larger the field with a parasitic polari-
sation.
In the short-focus paraboloid, due to the appearance of
current nodes (in the case of a dipole radiator), apart from
an increased cross polarisation, the additional undesirable
radiation of the zones of the reflector appears, in which
the current components parallel to the x-axis are of the
opposite direction. These zones of the paraboloid redure
the radiation in the main direction and increase the side
lobes of the directional diagram.
The directional characteristics of a parabolic antenna irradi-
ated by a dipole with a counter-reflector calculated from
the current distribution in the reflector, is defined as:
(6) = 1. 48 aJ, (a) I o (b)- bJo (a) J 1 (b) + 0.52 J 1 (b)
I a 1 -b1 b
+
+ 0,5 bJ. (1.5a) 11 ~~>.~:.·~~(b) J.(l.5a) t (l0-76)
where
R
a=3.5 2f • b=kR s~ e, .
 R Is the radius of the paraboloid opening;
8, the angle with the paraboloid axis;
J,. (x), a Bessel function of the n-th order.
This expression defines the directional diagram of the
antenna in the xz-plane (£-plane). To define the directional
diagram in the yz-plane (H-plane) the (+) sig n in front of
the last item of this express;.on should be replaced by the
(-) sign.
In ~he limit case of a long-focus reflector, \Vhen
i t 0, the expression (10-76) tak es the following rorm:
f (9) = 2 11 :b). (10-77)
The expression (10-77) corresponds to a uni forr n dirstri-
bution of the field amplitude at the paraboloid opening.
Indeed, for small values f(SJ
of the argument bJ 1 (b) Ad f ;;::r--r-,--.,--~.,.....~.---..........
~ 0.5 sin b and the ex- aa~--
pression ( 10-77) becomes na
sin b . " t--~Mr
f (6) = whtch corre-
b , a4 ...-~~~~~r+--f-+--+-~+--~
sponds to the radiation a2 ~---~.....__..~~
characteristic of an ideal
plane antenna. 0 2 3 8 I
Fig. 10-92 represents 6 kHsin8
the directional diagram I
of a paraboloid, calcu- 06~
lated from ( 10-76) for :ez-oJol.
Q6~~~ ~~~~~~~
various values of RI/. It Q4 t---t--~~
is seen from , this figure
that when the ratio RI! 112 t--lf- ---t-- 11-.., .;
decreases, the maj or lobe 1 l J 4
of the directional dia- 5 6
gram narrow~ down, Fig. 10-92. Directional diag~ams of a
whe reas the stde lobes parabolic antenna.
increase, this being con-
nected with the increase of the uniformity of distribution
of the field amplitude at the opening~of the antenna.
It should be noted that the cited directional diagrams are
characteristic also in the case of a radiator of another type.
Indeed, in the case of any radiator of sufficiently small
dimensions, the phase centre of which coincides with the
paraboloid focus, the field at the opening of the paraboloid
40'1
will be co-phasa1, and if the directional diagram of the
radiator under consideration is identical to the directional
diagram of a dipole radiator with a counter-reflector. the
directional diagrams of\ the paraboloid will likewise be
identical.
The directive gain of a parabolic antenna can be expressed
as:
D = 4nsk
A• ,, s=n R•·

where k 1 is the utilisation factor of the surface of the

reflector. It depends on the Jaw governing the distribution of
the field amplitudes at the opening of the paraboloid. For a
radiator consisting of an elementary dipole with a counter-
reflector, the quantity k, can be expressed as:

k =5.15 [o.423J (a.s;, )+0.26 ~ r•

1 1

This quantity is represented graphically in Fig. 10-93.

lt is seen from the graph that the utilisation factor of the
reflector has an optimum
f.O .... value. which occurs when
A; RI/= 1.3 and equals kt =
IJ8 =0.83. This optimum occurs
?' ........

0.6 ~ ' as a result of the action

of two opposed factors.
Q4 In the case of smalJ values
of R{f, the re Hector is irradi-
Q2 ated uniformly which
o Rff leads to a narrowing of the
Q6 Q8 tO t2 !4 t6 t8 2 n1ajor lobe of the direction-
a) diagram, but, at the
Fig. 10-93. Curve of utilisation
factor of reOector surf ace.
same tt.Ole, 8 consa·der-
• able part of the energy
emitted by the radiator does not reach the reftector
(Hawing over its edge), which leads to a lowering of the ,
directive gain. On the other hand, in the case of large
values of RI/, almost all the energy of the radiator reaches
the reflector but, at the same tinie, the edges of the reftector
are only very slightly irradiated, which leads to a widening
of the directional diagram and, once again. to a lowering:of
the directive gain. In the case of an opti~um value of lhe
ratio Rlf, the field intensity at the edce ol~ the opening. of
••I
the r~tlec.tor constitutes approximately 1/ 1 of the field
intenstty tn the centre of the opening. ,. .
The above-mentioned optimum relatively to k 1 occurs for
any radiator, but if the directional diagram of the radi·
a tor under consideration is sharper· than in the case of a
dipole radiator, the optimum value of k 1 occurs when the
ratio Rlf is smaller than 1.3. Conversely, for a wider
directional diagram of the radiator, this optimum occurs
for a ratio Rlf larger than 1.3.
Let us· now define the width of the dirt'ctional diagram of
the paraboloid corresponding to the maximum directive gain
of the antenna (when Rlf= 1.3). lt follows from the graphs
in Fig. 10-92 that the width of the directional diagratn at
half-power in the E-vector plane can be defined as:
iR.
•

29.,. ::::: 75°

and in the H-vector plane, as:

29.,. ~ 70° iR .
If a lower level of the side lobes is required than in the
case when Rlf= 1.3, one has to be content, at a prescribed
diameter of the opening, with a wider directional diagram and
smaller directive gain of th~ antenna, choosing the ratio
Rlf equal, for example, to two.
Note that the value of the side lobes in real conditions
is larger than can be defined from the graphs in Fig. 10-92,
because the above-cited expression (10-76) takes no account
of a series of, factors: a) the branching off of the electric
currents to the shadow surface of the reflector; b) the effect
of cross polarisation; c) the overflow of part of the energy
of the radiator over the edges of the reflector; d) the shadows
set up by the radiator itself which is located on the path of
the rays reflected from the re fleet or.
Various devices can be used as radiators of parabolic
antennas. The choice of the type of radiator often depends on
wJ'lether a coaxiaJ line or a waveguide i-s used as a feed line
for the radiator. Thus, when an electric dipole with a counter-
reftector is fed by a coaxial line, use can be made of the
dipole feed circuit shown in Figs. 10-3, b or 10-3, c. In that
case, the dupole is placed at a distance of approximately A/4
from the cQuuter-reftector and the diameter of the counter·
«»
reflector is taken equal to approximately 0.8A. In another
case, a horn is utilised as the radiator, fed by a rectangu·
Jar waveguide. In that case, in order to prevent the horn
from shadowing the radiation of the reflector, as well as
to reduce the effect the reflector has on the horn, frequent
use is made of a displaced horn shown in Fig. 10-94. Fig.

f f.
~
2

3 3
3 4
Fig. 10-94. Horn Fig. 10-95. Slot racli- Fig. 10-96. Dipole
radiator: ator: radiator:
/-reflector; 2- /-slots; 2-resonator: ·'- /-dipoles; 2-metal
horn; 3-wave- waveguldc; 4-retlector. plate; 3-wavegulde; 4-
guide. reflector.

10-95 represents a twin-slot radiator in which slots of half·

wave-length are located on a rectangular resonator ied by a
narrowing rectangular waveguide. The distances between the
slots of the radia1or are chosen such as to obtain the de-
sirable directional diagram. Fig. 10-96
sa represents a dipole radiator with a
dipole reflector fed by a rectangular
waveguide. Both dipoles are secured
on a metal plate and excited by the
open end of the waveguide. Their
2 I lengths are chosen such that the energy
maximum sheuld be radiated by them
Fig. 10-97. Oj pole
towards the parabolic re Hector. Fig.
radiator with a plane 10-97 shows a dipole radiator with a
counter-reflector: plane counter-refteclor fed by a coaxial
1-dlpole; 2-quarter-
wave sleeve; a-coaxial
line. The segment of coaxial IineA./4
line; 4-counter-reflector; long between the dipole and the
s- reflector. counter-reflector plays the role of a
metal insulator.
The purpose of the metal sleeve in the dipole radiator in
Fig. 10-97 is to prevent the electric currents from branching
oft to the outer surface of the coaxial line. However, the
currents on the coaxial line cannot be eliminated altogether
and, in the reeion between the metal sleeve fiDd the counter·
410
reflector, they are of considerable magnitude. The radia-
tion of these parasitic currents B is superimposed on that of
the dipole currents A (Fig. 10-98) and this causes the wave
front to rotate by a certain angle (dotted line) at the opening
of the paraboloid. As a consequence. the direction of the
radiation maximum of the antenna is deviated by a certain
angle from its axis. When the radiator rotates around its axis,
the directional diagran1 of the antenna describes a cone, the
minimum of radiation occurring along the antenna axis.
This rotation is made use of in radar for target location. in
which case, the distortion of the phase wave front at the
opening of a parabolic reflector turns out to be a de!\irable
I factor.
Wow For rotation or scanning
rtronf purposes of the directional dia-
l
A

.
Fig. 10·98. Distorti on of Fig. J0-99. Displacernent of
the antenna radiatio n radiato r out of focus:
phase front. 1-(ocua; 2-radlat or.

gram, the phase centre of the radiator is often intentionally

removed at a certain distance A from the focus as shown in
Fig. 10-99. At the same time. the rays reflected from the
reflector are deviated to the opposite side from the antenna
axis and form with the antenna axis an angJea , approximately
defined by the expression tan a= t .Consequently, the wave
front and the maximum radiation direction of the antenna
rotate by an angle ex. Apart from rotating, the wave front
becomes also distorted, which leads to a widening of the
directional diagram and an increase (from the inner side) of
the side lobes. However, when the displacement A is small,
the distortions are insignificant, so that, in practice. such
displacements of the radiator are used for scannine the
411
directional diagrams within the Jimits of Its double or triple
width.
Let us now consider the reaction of the reflector on the
radiator. As was mentioned earlier, part of the energy re-
ftected from the reflector returns to the radiator, since it is
on the path of the rays reflected from the reftector. If the
radiator was matched to the feed line in the absence of a
reflector, the introduction of a reflect~ leads to the ap-
pearance in the feed line of a ttdlected wave with a reflection
coefficient which can be evaluated from the expression .

IPI= {~.
where P, is the power reflected from the reflector and
received by the radiator;
P,, the power radiated by the radiator.
If we designate by £, the electric field intensity of
the wave reflected from the reflector and incident on the
radiator, then in accordance with Paragraph 6-4, the power
received by the radiator is: ,
>.,•E: DTJ
....
P' = n• 960 ' (6-32)
where D is the radiator directive gain;
TJ, the radiator efficiency.
The field intensity of the reflected wave of the radiator
E, can be regarded as approximately equal to the field
intensity of the wave radiated by the radiator at the apex
of the paraboloid £ 0 , which is expr~d as:
_ Jr60P 0Dtt
E. - f .
Subst it ut ing this expression into (6-32), we obtain for the
reflection coefficient
IPI==~·
To eliminate the effect of the reflector In the line it can be
tuned up by introducing into it a complementary reflection
cancelling the reftection caused by the. eftect of the reftec-
tor. However, this can only be done on one fixed frequency.
When the frequency of the oscillations changes, the.pbase,of
the coefficient of the retlection caused by' t,~e~retlector under·
·~

412
goes a sharp change, due to the fact that the ratio of the
focal distance to the wave-length f/'J.. is much higher than
unity. This leads to a narrowing of the radiator pass-band.
Conseq.uently, special measures have to .be adopted in order
to decrease the magnitude of the coefficient of reflection
from the reflector, i.e.. to decrease the reaction of the
reflector on the radiator.
One of the ways of decreasing the effect of the reflector
on the radiator is to remove the radiator from the antenna
field (see Fig. 10-94). Another way is to set up a supple-
mentary reflection of the energy fron1 the reflector, which

Fig.IO-lOO.Installatlon fig. 10-101. Quarter-

of a supplementary wave lattice on a
disc reflector: reflector.
1-dtsc;· 2- reflector.

leads to a decrease of the magnitude of the field of the ra-

diator reflected wave (Fig. 10-100).
To this end. a metal disc is placed at a distance of ap-
proximately a quarter of a wave-length from the apex of the
reflector. The electric current induced in the disc sets up in
the radiator a -supplementary field opposed in phase to the
main field reflected from the reflector. The proper choice
of the diameter of the disc considerably reduces the radiator
total field, thereby substantially eliminating the effect the
reflector has on the radiator. ...
This effect can also be eliminated by installing a
quarter-wave lattice on the reflector (Fig. 10-101). The
lattice consists of parallel metal plates A/4 wide placed on
the irradiated side of the paraboloid at a distance s apart,
small in comparison with the wave-length. The plates form
411
an angle of 45° together with the vector Elnc of the field
incident from the radiator. The field reflected from the
reflector rotates the plane of the polarisation by 90° rela-
tively to the incident field. so that the reflected field does
not affect the radiator.
As regards the accuracy to size of the parabolic reflector,
it can be determined from Fig. 10-102. Suppose dimension
lpcose
.....,.\
-
-- .

~--f-....._-

Fig. 10-102. Calculating Fig. 10-103. Reflector in

the accuracy of design of the shape of a parabolic
the reflector. cylinder.

~Q is projected on the surface of the reflector; the distortion

of the fie Id phase at the opening of the re flee tor is then~
'I>, =kA Q(l +cos B),
n
and if we regard \I',= 8 , as the permissible field phase
distortion, the permissible inaccuracy of execution of the
reflector is expressed as:
1 -
AQ= 16 (1 +cos 8) •
•
It can be seen from this expression that the accuracy to size
of the reflector in the central part should be greater than
at the edges of the reflector. Furthermore, the edges of the
reftector are less intensely excited, which reduces the effect
the inaccuracy to size of the edges of the reflector has on the
antenna directional diagram.
Let us now pass on to the investigation of parabolic
cylinders. Reflectors in the shape of parabolic cylinders find
particularly frequent application when the object Is to
obtain directional diagrams of diifereJJt ':Width in mutually
. '
414
 •
perpendicular planes. The radiator of the parabolic cylinder
(Fig. 10-103) placed on the focal line is a co-phasal linear
antenna of length L. It can consist of, for example, half-wave
dipole shown in Fig. 10-103. The radiator sets up a cylindri-
cal wave front and because the reflector profile in the xz-
plane has a parabolic form, a plane wave front is set up at
the antenna opening. At the same time, the distribution
of the field amplitudes in they-axis direction at the opening
is uniform, and, in the x-axis direction, it depends on the
directional diagram of the radiator in the xz-plane. In this
way, the directional diagram of the parabolic cylinder in the
yz-plane is determined by the dimension L and, in the xz-
plane, by the dimension 2R. The directional diagrams can
be given different forms in two mutually perpendicular
planes by the proper choice .of L and R. The dimension
2R should be chosen such that the reflector should capture
the main part of the energy radiated by the radiator.
Linear radiators can be of the most varied forms, such as
waveguide slot antennas, sectoral horns, arrays of half-wave
dipoles excited by a rectangular
waveguide, etc. •

The linear radiator often consists 4

of a parabolic cylinder bounded by
parallel planes (segment-parabolic
antenna) and, in turn, excited l>y
the open end of a rectangular wave-
guide (Fig. 10-104).
The distance between the plates(~) Fig. 10-104. Segment-
of a segment-parabolic antenna is parabolic antenna.
chosen such as to preclude the propa-
gation of higher-order modes. As a rule, theE-vector of the ex-
cited field is chosen perpendicular to the parallel planes of the
radiator. The dimension !1 should then be smaller than A/2.
In that case, the velocity of propagation of the wave (of TEM
mode) between the planes equals that of light. Sometimes,
the E-vector is chosen parallel to the planes of the radiator,
in which case, the dimension ll is defined from the
condition .1<A<2A and an H,, mod~ is propagated in
the inter-plane space. In the latter case, the accuracy to
size of the distance 4 should be greater than in the first
case.

41S
 • <

10-14. Other Reflector Antennu •

' '

The present paragraph will be devoted to a brief descrl~>:

'

tion of a number of other reflector antennas. For further'

details, the teader is referred to the. special' literature.
To begin with, let us describe. the reflector antenna shown
In Fig. 10-105. The antenna consists of a cylindrital reflector
of special form 1, forming a directional diagram in. the
vertical plane, and of a linear radiator 2 in the shape of a
1

--t--

'
L---

z
F~. 10·106:
Fig. 10-105. Cylindrical reflector Re ector cross
of special form. section.
parabolic cylirrder with bounding planes, forming a dir~ction­
al· diagram·~ in the horizontal plane.
As shown in Fig. 10-106, the lower part .of the reflector
is of a nearly parabolic shape and sets up a parallel beam· of
reflected r~.ys, whereas the upper part has a nearly .sph~rical
shape and sets up dispersive reflected rays. The dtrect1onal
diagram of the an-
tenna in the vertical
plane is approximately-
expressed as f ( O)b
~ = csc 9 and has the
r•hiJfCIJ ·form of a vertical fan.
Fig. 10-107. Directional diagram of a of this ;.. Antenn~
ground radar. kind, used, in par-·
ticular, in ground
radars ensuring radiation of in-coming or out-going aircraft
w1th equal intensity (Fig. 10-107).
Spherical reflectors find application in radar technique.
When the radiator is situated at the point f=-}, the spberi·
caJ reftector operates just as a Pfrab<?Jic reflector. ·
. ,· ~
'
'tl
'418
 Indeed, from a comparison of the equation of the curve of
the circumference (Fig. 10-1 08) I· •

x• + (R-z) = R11

or
x•=2Rz ( 1- 2'R)
with the equation of a parabola (x2=4fz), it is seen that
part of ~he circumference in the vicinity of the apex (2~~1)
can be regarded as part of a parabola with a focus lying at
the point f=R/2. Consequently, a z
spherical reflector irradiated by a point
source enables to transform a spheri-
cal wave in a nearly plane one and,
thereby, ensure the formation of narrow
directional diagrams. At the same
time, by displacing the radiator along z
a circumference of ra'dius R/2, such
a reflector enables to scan the direction-
•

al diagram of the antenna within

the limits of a very narrow angle.
The property of spherical reflectors
is utilised for obtaining a highly
directional radiation, the directional
diagram of which has a 360° scanning Fig. 10-108. Spheric4l
reflector:
angle. We heve in mind the heJical- 1-clrcumference; 2-
spherical antenna [66 ], which consists centreof theelrcumference;
3- focus of the parabola.
of a closed sphere (Fig. 10-109), rnade
of current-conJucting strips lying
close to one another at an angle of 45° to the meridians of
the sphere. The radiator situated inside the sphere sets up

Wew from IJBhlnd View from oiJovs

Pig. 10·109. Helical-spherical antenna.

27-2111 411
In the wave incident uron the reflector an electric field
intensity vector paralle to the current-conducting strips
of the reftector. The part of the surface of the sphere thus
irradiated entirely reflects the incident wave towards the
radiator, forming a parallel beam which, on the diametrically
opposed parts of the reflector, meets the current-conducting
strips orientated towards the electric field vector at
an angle of goo and pass freely throufUt the reflector. The
radiator, which lies half-way between the centre of the
sphere and its surface, rotates in the horizontal plane, this
leading to the rotation of the directional diagram in that
plane. When it is displaced along the circumference in the
vertical plane the directional diagram
can scan in the vertical plane within
2 a certain angle.
The spherical shape of the reflector
is very convenient for making large
collapsible air reflectors enabling
the rapid unfolding and folding
of the antenna. The antenna consists
of a canvas envelope carrying current-
conducting strips. The advantage
of such an antenna consists in that
f the canvas envelope reproduces a
spherical surface with the utmost ease
Fig. 10-110. Horn .. and accuracy.
parabolic antenna: Horn-parabolic antennas find appli-
1-horn; 2-parabollc aur- cation in radio-relay lines (Fig.
fac:e; a-opening
10-110); they are a successful cotnbi-
nation of a pyramidal horn with part of
a parabolic surface. The focus of the parabolic surface
coincides with the phase centre of the horn and the
~l~ctrotnagnetic waves incident from ijle horn upon the
parabolic reflector are reflected towards the opening
without returning to the horn.
Horn-parabolic antennas are characterised by a small level
of the side lobes and reverse radiation and are, therefore,
suitable for radio-relay lines in which receiving antennas
are combined with transmitting ones. Thus, horn-parabolic
antennas operating on a certain radio-relay line 167] and
standing side by side, have a transient attenuation of the
order of 90 db.
The flate angles of the horns are taken equal to 30·40°,
418
and continuous transitions of the hy.perbolic type of a length
of the order of (10-:-15) Aare inserted betw,en the inputs of
the horns and the feed waveguides. This lowers the coefficient
of reflection from the horn in the wavegui de down to 1-2 ~h.
Plane reflectors find applica- z
tion in radio-relay lines;
they are placed on top of a
mast at an angle of 45° to the
horizon and rotate · by an
angle of· 90° a beam set up by
a highly directional antenna
at the foot of the mast (Fig.
10-111). The advantage of de-
vices of this kind consists in
tbat long feed lines are no-
longer needed when the
antennas are placed on top of
a mast. As1 shown by theo-
retical calculations [68 ], the
efficiency factor of plane
reflectors can be higher than
unity. In this case, the efficien- Fig. 10-111. Antenna system of
cy factor is the ratio of the pow- radio-relay line:
er P1 received by the receiver 1-anten na: 2-plane reflector.
when using a plane reflector
(Fig. 10-112) to the power P2 received by the receiver

I • P,

...........
I

Fig. 10:112. Delerminlng the efficiency of a plane

reflector:
J -antenn a; 1-teflector.

419
when no reflector is present and the antenna is located
at the place where the plane reflector was situated. The
calculated curves are given in Fig. 10-113, where the
quantity k2 =10 log~ is plotted on the Y-axis and the
quantity :-!.
on the ~-axis; the curves are given for a
series of values of Rla.
The curves represented in Fig. 10-1 r8 apply to the case
when the distance between the reflector and the antenna d is
large in comparison with the output opening of the antenna
R and the reflector a (R~a~O.l76). The calculation has
been effected for the case when the field distribution at the
output opening of the
6~~-~~~~~----~ antenna during trans-
5~-+~~~~~~~~ mission is defined by
R the function
4~~~~~~0 ~~~
3 -~0~~~~~ ;•
2~~-+~~ ~r-~~~
h (s>= 1-o.ss4 71
1
0
11
-f
~-2
-3
-4 3
-5
-6
-?0 O.f Q2• Q3 Q4 Q5 118 Q9 tO
f_ Ad
k-4al
Fig. 10-! 13. Efficiency of a plane Fig. 10-114. Double·
reflector. mirror antenna.

where l=Ria and s=rla (r being the radial distance from

the centre of the output opening of the antenna).
It can be seen from the curves that the plane reflector is
capable of amplifying the received power up to 4 times.
Apart from the antennas described above, doubl~·rnirror
antennas find application in radar. Their.,lllf.in advantage Is
•
410
the possibility of reducing the size of antenna systems. An
antenna of this kind is shown in Fig. 10-114. The antenna
consists of a parabolic reflector 1, an hy,perbolic reflector 2
and a radiator 3. The radiator is placed near the apex oJ
the parabolic reflector; the waves from the reflector are inci·
dent upon the hyperbolic reflector and are reflected towards
the parabolic reflector which forms a plane wave. The hyper·
bolic reflector gives rise to a certain shadowing of the
reflected. rays. but its dimensions are taken much sn1aller
than those of the parabolic reflector, so that the shadowing
is insignificant.

•
 CHAPTER ELEVEN
Short-Wave Antennas
11-1. Classification of Antennas
The term short-wave antennas applies to antennas used in
the 10 to 100 m wave range and, in particular, from 15 to
60 m. They are used in telegraph and telephone communi-
cations, phototelegraph and long-distance broadcasting. The
antennas are also used for special purposes in local communi-
cations.
Short-wave antennas must satisfy a number of specific
requirements arising from the conditions of propagation of
radio waves. Indeed, three of four waves are necessary, as a
rule, to ensure round-the-clock radio communications: one
for day-time, one for night-time and one or two for the inter-
mediate periods. The wave-lengths are shortened in summer
and lengthened in winter as requ~red. Furthermore, all
operating waves are shortened during the years of maximum
solar activity and lengthened during those of minimum solar
activity. This requires the use of antennas of the multiple-
tuned type. .
Furthermore, long-distance communications on short waves
rely on space (ionospheric) waves which, in their path, are
being reftected one or more times from the ionospheric layers.
of the atmosphere and reach the Teception point at a differ-
ent angles to the horizon; since ionospheric conditions
change in time, so do the angles of incidence of the waves.
Accordingly, the width of the antenna directional diagram in
the vertical plane must satisfy the following requiremen.t:
it should not be too narrow·, otherwise the sienat wltl
ftuctuate and eventually fade Ollt. ..!. l~·
I
1
1
'

~~~
 On being reflected from the upper layers of the lono·
sphere, radio waves are deviated from tht;ir regular path
(deviation of the ray) by up to several degrees, so that the
antenna directional diagram in the hori2ontal plane should
not be too narrow either .
.Experiment~ I d.~ta show that, as a rule, the optimum
wtdth of the dtrecttonal diagram amounts to several tens of
degrees in the vertical plane, and 10 to 15° in the horizontal
plane.
Due to the existence of a round-the-world echo the
radiation diagrams should be unidirectional and, owing to
interference caused by thunderstorms, it is desirable to use
antennas, the directional diagrams of .:which have small
side lobes. As regards the plane of polarisation of the electro·
rnagnetic field, it undergoes a rotation during its reflection
from the ionosphere, so that, in practice, antennas with
a horizontal as well as vertical polarisation are used. How-
ever, transmitting antennas with vertical dipoles produce a
strong radiation along the earth, giving rise to considerable
losses in the earth. Hence, horizontal antennas are usually
preferred~ ·
Horizontal antennas are also preferred for reception
because various noises have, on the whole, a vertical field
polarisation.
The antennas used on short waves are of the simple as
well as multiple types. The antennas are further subdivided
i.nto tuned antennas, destined to operate on one fixed
frequency, and multiple-tuned ones, destined to operate in
a twofold or threefold wave range.
Simple antennas comprise symmetrical horizontal dipoles,
multiple-tuned dipoles, corner dipoles, etc. Multiple an-
tennas comprise co-phasi.l horizontal antennas, travelling·
wave antennas and rhombic antennas.
To begin with, we shall discuss simple antennas of the
tuned type.

11-2. Simple Dipoles of the Tuned Type

First we shall deal with the symmetrical horizontal
dipole shown in Fig. 11-1. The dipole is made of a copper or
bimetal wire 3 to 5 mm In diameter and is secured on two
supportins~ masts. It can be of any length within tbe limits
411
of 0.25 ~J:. ~0.64, which ensures a constant direction to
the tnaxhnum radiation in the horizontal plane, forming
\Vith the axis of the dipole a 90° angle. The dipole is in-
stalled at a ht>ight H of the order of-} -:-A, in which case, the
maxiJnum radiation direction in the vertical plane forms
with the horizon an angle of 30-15°.
The dipole is fed by a twin feeder with a-wave impedance W
of the order of 600 ohms, which is tuned on a travelling
wave by means of, as a rule, an inductive stub. ·
1• 2t •1

•
Fig. 11-1. Symrnetrical horizon dipole:
/-Insulator; 2-feeder.

To avoid the distortion of the directional diagram,

insulators are inserted in the guys of the masts, forming
segments of less than quarter-wave-length, so that the electric
currents induced in the guys have a small magnitude and
their radiation is insignificant.
The length of the dipole is often taken so that f==0.25
(half-wave dipole) and, at the same time, the dipole is shunt-
fed, as was shown in Fig. 10·1, b. . -
Simple short-wave antennas of the tuned type are some-
times made lo0p-like, in accordance with the circuit shown
in Fig. 10-1, c. In that case, a travelling-wave ratio equal to
0.5 is set up in the twin feeder with a wave impedance
W=600 ohms.

11-3. Simple Dipoles of the Multiple-Tuned Type

· As indicated above, the operating waves of radio communi·
cations have to be frequently changed. This has led to the
use ·of multiple-tuned antennas. The multiple-tuned dipole
suggested by S. I. Nadenenko is of tl}js k,ipd (Fig. 11-2).
-· >
~

4H
 Th.e dip~ le is secur~d to two ma.sts, Just ~s a simple sym-
metrlcal dt pole and 1s made of thtn w1res of diameter 2r=3
to 5 mm and number n=B to 12, forming a cylinder of a dia·
meter of the order of 1 m. It is destined for operation in
a wave range satisfying the condition 0.25=;;;;;; { :s;;;;0.64 and
is fed by a four-wire feeder \vith a wave in1pedance of the
order of W= 300 ohms.

Fig. 11-2. Multiple-tuned Nadenenko dipole .

•

The wave impedance of the dipole can be expressed as:

W = 120 (ln 21
Vequlv
- 1) , (11-1)

where is the equivalent radius of th~ dipole related

Qequlv
to the radius of the cylinder Q forn1ed by the wires as
follows:
nI nr
Qequ lv= Q } Q•

The wave impedance of the dipole is of the order of 300 ohms

and the input resistance curves of the dipole are considers·
bly smoothed out (Figs_ 2-13 and 2-14). The travelling-wave
ratio jn the feeder in the operating wave range lies within
the interval K~0.25+0.70, which enables to operate
I

without specially tuning the feeder on a travel ling wave and,

consequently, to pass rapidly from one operating wave to
another.
The power gain of a multiple-tuned dipole (with the
influence of the earth taken into account) is found to equal
l l
s==4.24 when y =0.2 5 and s=6. 9 whe~ T=0 .5.
The multiple-tuned dipole can also consist of a biconical
antenna (Fig. 11-3). When Q~lm, a dipole of this kind
becomes multiple-tuned on short waves. Its wave in1pedance
is defined .as: ·
· W= 120ln(coti-) (11-2)

425
 or, for small values of the angle ,, approximately by the
expression W= 120 In ( = 120 In !)
In this dipole. the!.
gradual decrease of the diameter of the dipole towards the
'feed points serves to match the input resistance of the
antenna to the wave impedance of the feeder.
--l .,

Fig. 11-3. Biconical dipole.

t 1-4. Corner Antenna

Just 1ike the dipoles discussed above, the corner antenna is
symmetrical (Fig. 11-4). It was designed by A. A. Pistol-
kors with the object of obtaining a multidirectional
radiation in the horizontal plane. The antenna consists
of mutually perpendicular wires and is horizontal to the
earth.
The directional diagrams of the corner antenna in the
horizontal plane ~an be expressed as:
) = [cos (kl cos cp)-cos kl _cos (kl sin cp)-cos kl] +
'( q> sin q> cos cp
+i r sin (kl_ CO~ cp)_-COS cp Sin kl_ • sin (kl sin cp)-sln cp sin kl] • ( 11 _3)
sin cp cos 'P

l
r
• .,
~
....
.,.._-l· •I
Fig.ll-4. The Pistol- Fig.. 11-5. Directioil'at diagrams of a corner.
kors corner antenna. antenna.

They are shown in Fig. 11-5 and, in the wave range of

.f = 0.5 + 0.64, are close to non-directional ones.
Corner antennas are sometimes made of thick cylinders of
the Nadenenko dipole type and. in tbi~·~1'•Y_'*iensure .the.oJ)er·
ation in the wave r~nae. without re~UNDJ of the feeders.
,.
 The angle cz of corner antennas can even be less than 90°,
in which case the directional diagrams of the antenna differ
from the ones shown in Fig. 11-5. Corner/antennas of this
kind with an angle a= 70° were used for radio communications
on the first artificial satellite of the earth. As we know, for
operating on the waves A= 15tn and A=7.5m, the lengths of
the dipoles were l=2.9m and l=2.4m.

11-5. Multiple Short-Wave Antennas

of the Tuned Type
The main circuits of antennas of the tuned type consist of
a number of half-wave dipoles forming a co-phasal array. The
distances between the dipoles in the antenba array are usualJy
chosen equal to half a wave-length. Fig. 10-22 shows the
circuit of the most commonly used co-phasal antenna
utilised in the short-wave range. Fig. 11-6 shows the way the
co-phasal antenna is secured on metal supporting masts.

Fi~. 11·6. Fastening of co·phasal antenna.

Co-phasal horizontai antennas are generally used with a

reflector which presents exactly the same aspect as the an-
tenna proper. The reflector lies at a di-stance d=(0.22+0.25) A.
from the antenna. In most cases, a passive reflector is
used, in which a twin line (stub) approximately half a wave·
length long, which serves for tuning up the reflector, is con-
nected to the points c of the distributing feeders (Fig. 11.:.7, a).
In an active reflector, the points c of the distributive feeder
(Fig. 11-7, b) are connected to the feeder of the antenna by
means of phasing circuits. The phasing circuits are chosen
such that the current in the reflector should be equal to the
current in the antenna and precede it in phase by 90•.
421
 The antenna directional diagrams in the horizontal
plane depend on the number of dipoles in one tier of the
antenna system (m) and on the type of tuning of the
reftector. They can be calculated from the expression
_ OOn/pcos ( -i-cos8) sin ( m-i-cos8) X
Eh- r0 sin e I ( n
sn !os 9)
2
X Vl+mi + 2m, cos ( '1'- ~d sin a)' 2
(11-4)
where d is the distance between the antenna and the
rf:'ftector;
f

I I

d=(I121Ji0.Z5)A
. I 1/
c l c I

• .
(aJ {b)
Fig. 11-7. -Antenna circuit with a reflector:
a-passive reflector: 1- .t unJna stub. IJ-actlve reflector: 1- phasing circuit.

m,, the ratio of the current amplitude in the re-

flector to the current amplitude in the antenna;
1J', the angle by which the current in the reflector
precedfs the current in the antenna
For an active reflector m1 = 1, 1j)==90° and d= {- , the
directional diagram obtained is unidirectional with a zero
of radiation when 8=270° and with radiation maximum when
6=90°. In the case of a passive reflector,. it is impossible
to obtain simultaneously m1 =1 and 9=90°, hence the
radiation in the direction 8=270° cannot be cancelled
corn pletel y.
The directional diagrams of a co-pha5al antenna in the
vertical plane depend on the number of tiers in the antenna
(n) and on the height at which it is fastened above the
surface of the earth .. As a rule, the first tier of the antenna
is at a height h= {- +-}; there is no point in having it
lower because, in that case, due to the inftuepce
,,
of the'
finite
428
conductivity of the earth, the antenna may radiate in the
direction of the axis of the dipoles. If the-· earth is regarded
as an ideal conductor, which is the case in engineering
practice, the directional diagram in the vertical plane is
ex pressed as:
n
120m I sin n 2 sin A 2 h
~~

sin ( ~ sin A )
0 1
.v ' '

x V 1 +m~ + 2m, cos ( 'Jl- 2

~d cos tJ.) , (11-5)

where hmcan is the distance between the surface of the earth

and the centre of the antenna system;
ll, the angle between the horizon and the direction
to the point of observation of the field.
It is evident that along the earth surface (L1 =0°) we get
radiation 11ero, since the currents of the mirror image of
the antenna are of opposite direction relatively to the
antenna currents.
The antenna directive and power gain can be caJculated
frorn the expressions
Elrl Dn
D= o .•
30/ 0I R11 ' B= 1.64 •

Here, the radiation resistance Rr. can be calculated from

the expressions of Chapter Three, taking account of the
influence of the earth. The antenna directive gain T) is usually
taken equal to unity, since the antenna radiation resistance
is much larger than the resistance of the losses in the
antenna.
Table 11-1 gives the values of the parameters of the
antennas utilised in practice (CH-'!.-
m
R designates a co-phasal
horizontal antenna with a reflector of n tier and m dipoles
in a tier).
Apart from co .. phasal horizontal antennas for operating
on one fixed wave, co-phasal horizontal antennas can be
used for operatin.g on two fixed waves. The antenna circuit
of this kind is shown in Fig. 11-8. Since the distance from
the transmitter to any of the dipoles is the same, the
dipoles are found to be co-phasal when operating on any wave-
length. Co11sequently. the radiation maximum maintains its
429
 Table 11-1

.er •- ...•
I
.a
...
~
11 .~
- ...
c:;-
I

•u --i
••
c-., -·~
e-= ....
U)

='"0
-=
u .Q
ea 0 c: ....

Antenna
.c
-:..
11
• --•.a.
...
::o.:
.,-
uN-
...... '
::
•• :f
;~-a
-=c
0
.a
• -~
11 E •
_c::
-ocu
"'CJ..C:.c •
-a.!-
_c..
-~
.... o.i-
-
-0 c:- uCl c_ ....
>.

~
.. c:
-:,\U

:c<
.,
....., e ~'
u
V
....
·-
Q
;'
0
Q.
.!•
c~..
< ...
.c ·-ea ;.
-eGJa~
"'CJ
·- ea a
... - c:
~ bOg..::
c:, -"'·
.c«~·-

_.....
.,: c
a= >.5
tl-

•
2
CH-R
2 0. 75 )., 35.3 21.5 17 65 26
•
2
CHJ6R 0. 75 A 230-280 140-170 17 9 26
4
CH-R
4 I. 75 A 167.5 102 8 35 9.5
6
CHj6R 1. 75 )., 740-820 450-500 7 9 10
•

direction and the -antenna is utilised as a multiple-tuned

antenna when the condition 0.25:s;;;;-f-:s;;;;o.64 is observed. In
f---A,,Q---4
--. ~

. (Q)

{/J)

Fig. 11-8. Tuning circuit of antenna on multiple waves:

d-reflector tunl111 clrcult; b-anteona tunlna circuit.

practice, it is found convenient to tune the antenna on tWo

multiple waves Ao and 2Ao. At the same t.irnc:r~thedimenslons
of the antenna are chosen su~b tbat ··d~O('zoAo and 2l==~o~.
,•

480 •
 The electric parameters of the antenna on multiple waves
are given in Table 11·2 and the circuits of the simultaneous
tuning of the retlector and feeder on two waves, in Fig. 11·8.
The reftector Is tuned on the wave 2Ao by displacing the
short-circuiting bridge M along the tuning stub and, on the
wave Ao, by displacing the other stub, which, on the wave
Ao gives a short-circuit and on the wave 2Ao, an infinitely
large resistance (Fig. 11-8, a)
The feeder is tuned on a travelling wave, first on the wave
2Ac,, and, after the tuning stub of lengta xs has been secured,
the feeder is tuned on a travelling wave on the wave 1 0 •
At the same time, instead of an inductive stub of length x2,
a combined stub is used, as shown in Fig. 11-8, b.
In this way, tuning is automatic both on the Ao and 2Ao
waves of reflector as well as. feeder, no additional tuning
being necessary when passing from one operating wave to
another.
Table 11·2
,
.c
be
c:
-•...
.c
•
I...
u ---
··-
c:- ..
>.
.,_ -
-
•

c••
....
~
•

-ICl
u
u
-...
..c:a
E"
='Q
E.!:
-... .
. 0
cc:
-.....
2
-
-
tJ
--a
ou-=
--.!
... -
u-a
~ ... o..:. ~t-
Antenna •iJ
tiD
Ill-
.8
•1:1
-•...
"'
a<~
...,.c:
0 -.c- - -
c: •
---'=e oe- I
"'g
">I
-ac::•
-.c
·--...
c: • oE..,.
(V

u
g.
~
-&•
·-
cu e
c:
Ill
...
•
~
0 ~
o.2
.,-
-·
cca ·- ..
.c
-:!uW
...
"C tiD c: "C
.c ....
-~ ....
"C till c: .,
-ea&~:~

0 I :C.ae g. < .. ., ~--· c:

'~-
~--a
., a.-

2 Ao 21.5 17 65 24
MCH2R 2Ac; 0. 75A0 8.25 38 84 60
I
36.5 17 35 24
2
J: I
I

MCHTR 0.751. 8.8 40 60 85

70.7 17 18 22
MCHsR
2
J: 0. 751, 13.3 40 32 75

IJ-6. Mlltlple Short-Wave Antennas

Type of the Multiple-TJ1ned
As was indicated before, the multiple co-phasal antennas
discussed earlier are antennas of the tuned type, because
they enable to operate either on one fixed wave or on two
fixed waves (antenna on multiple waves).
431
 .In the case of an antenna of the first kind, when operating
on waves differing from the nominal one (A.o), the equality
of the current amplitudes and phases at various tiers
of the antenna is upset, which 1eads to a distortion of the
directional diagram in the vertical plane. The directional
properties undergo an insignificant deterioration only in a
narrow range (of the order of a few per cent) in the vicinity:
of the nominal (rated) wave; moreo~r, the larger the
number of tiers, the more narrow this range.
0.41,

Q5A,

To trrJnsmlffr
Fig. 11-9. Sectio n of co-phasal multip le-tun ed
anten na.

In the case of an antenna on multiple \vaves, the antenna

is matched to the feeder I ine and the re Hector tuned auto-
matically only on two fixed waves.
Recently, a complex horizontal cg-phasal four-tier antenna
has been designed, which can operate in a continuous
wave range with a ratio of 1 : 2.5 [69 ]. The circuit of one
section of an antenna of this kind is sliown in Fig. 11-9.
As can be seen from the figure, the antenna is fed in exactly
the same way as in the case of an antenna on multiple waves .
and, as a consequence, the phase of excitation of the dipoles
is not frequency dependent. The distance between the stages
is taken equal to 0.5Ao.
To obtain a sufficiently high travelling-wave ratio in
the feeder, quarter-wave step-up transformers are used,
thereby ensuring satisfactory matching of the antenna to
the feeder in a wide wave range: the travelline;wave ratio does
';.

432
not fall below 0.5 in the range (0.9-:-1.7)A0 and below 0.3
In the range (0. 7-+ 1. 9) A.0 • .
The use of quarter-wave set-up transformers enables also
to utilise feed lines of relatively high wave impedances.
In the absence of set-up transforrners, one would have to
use feed lines of low wave impedances, which complicates
the manufacture of vertical feeders.
The inter-tier connections are accomplished by means of
a feeder line with a wave impedance Wt =550 ohms. The
first two-stage quarter-wave transformer consists of two
feeder segments of length 0.25A-o each with a wave im-
pedance W2=396 ohms and Ws=550 ohms. The second
two-stage transformer consists of two ,quarter-wave feeder
segments with a wave impedance W,= 366 ohrns and
W5=480 ohms. The antenna pass-band can be irnproved
by using dipoles of low wave impedance. A dipole of this
kind consists either of three wires 4 to 6 mm in diameter,
lying at th~ apices of an equilateral triangle of side 0.035 Ao
or of four wires lying at the apices of a quadrangle.
The antenna can be of two kinds: with a tunable or an
aperiodical reflector. In the first case, the· reflector system
is entirely analogous to the antenna systen1 and situated
at a distance d=0.264 A0 from it.
In the second case, the reflector, consisting of a grid of
horizontal wires (copper or bin1etal) 6.6 mm in diameter
spaced 0.073 A0 apart, is at a distance d=0.23 A0 from the
antenna. The height of the grid is approximately equal
to 2 A-0 and its width L=L0 +0.18 A0 , where L0 is the width
of the antenna system.
The antenna consists of two identical sections connected
by a distributive feeder (W6=550 ohms). The main feeder
(W7 =275 ohms) is connected to the inter-section distribu-
tive feeder in its centre. Displacing the connection points
of the main feeder towards one or the other section leads
to the rot-ation of the directional diagrarn of the antenna by
a certain angle (8 to 20°) .
.,.
11-7. Rhombic Antenna
Rhombic antennas have been in use as receiving short-
wave antennas since the end of the twenties. As transmit-
ting antennas, they have been used since the middle of the
thirt)es. ·
28 -2122 433
 The antenna consists of four wires of length I each and
has the form of a rhombus. An absorbing resis~ance R.,
equal to the wave impedance of the antenna, is connected
to one of the acute angles of the rhombus, the feeder being
connected to the other one. Thus, a travelling wave is
formed in the antenna and travels towards the absorbing
resistance.
In a wire with a travelling-current ve, the radiation
maximum forms with the axis of the wire an angle smaller
than 90° (Figs. 4-10 and 4-11), furthermore the angle of

Fig. 11-10. Circuit of rhombic antenna.

maximum radiation is a function of the ratio of the length ol

the wire to the wave-length {- • By disposing the wires at
a certain angle to one another, the maximum total radiation
can be coincided with the bisectrix of this angle; furthermore,
this maximum radiation is found to be directed towards the
absorbing resistance.
The antenna lies horizontally above the earth and its
directional diagram in the horizonJal plane is expressed as
E _ 2401 0 coscp • [~(l-s·n 1, _A))] X
h- ro 1-sin cp s tn 2 l \ cp t'

x Sin [~(1 sin (cp+P»]. (11-6)

where I 0 is the current at the beiinning of the antenna;

p, the angle between the antenna axis and the direction
towards the point of observation;
l, the length of the side of the rhomb;
q>, half the obtuse angle of the rhomb.
The directional diagrams calcula~ .from this . e
are shown in Fig. 11-11 for 1==21.·'.(1==6~
. 1
and for lt r4A
f84
and c:p=65°. Th e directional diagrams are characterised by
the possession of relatively large side lobes.
The directional diagram in the vertical plane, with the
influence of the earth taken into account (assumed to be
ideally conducting) is expressed as:
E _ 4801 0 cos cp • 1
v- r0 1-c os ~ sfn <p stn X

•
x [~ (1 -·c os L1 sin q>)] sin (kh sin &) , (11-7)
120• ao• 10 ·
•

240.
zro•l t•ZA ;rp=65" 21o· aoo•
t•"4A;rp=65·
Fig . 11 .. 11. Dir ect ion al diagran1s of a rho mb ic antenna in the
hor izo nta l plane~ ·

where f:J. is the angle between the horizon and the direction
towards the point of observation;
h, the· he igh t of the antenna above the earth.
Fig. 11·12 shows the directional dia gra ms of a rhombic
antenna for l=2 i.., h=0.5A,, q>=65° and l= 4A , h='A, •

•
•

ts o· --l•Z_,A;I;J•Q5A-- -- -- tr mo· ~C :: :: :: .:_n•

; rp= -~1...,..•4......
A; II•A ; rp•
Fig. 11-12. Dir ect ion al diagrams of a rhombic antenna in vertical
pla ne.

cp~65°, calculated from this expression. We see that the

radiation along the earth is zero and the angle of elevation
of the radial ma xin1um is a function of the ratio hf'A.. In
that plane, as can be seen from the diagrams, relatively
large side _l.obes occur.
43S
 It can be seen from the direct·o tal diagram in the
vertical and horizontal planes that the direction of the
maximum radiation and the unidirectivity are maintained
in a wide wave range.
The radiation resistance of a rhombic antenna was calcu-
lated by the induced emf method, no account being taken
of the attenuation of the current in the wires, by. Li-Chan
(70 ]. It was found to equal R'E, = 780 ~hms for l=4 'A and
h= 'A.
An approximate calculation (no account taken of· the
reciprocal influence of the wires and the influence of the
earth) of the radiation resistance of a rhombic antenna can
be obtained from the expression
R'E,~240 [ C-1+ln2kl--Ct2kl+
. sin 2kl]
kl , (11-8)
2

where C=0.5772 ...

The calculation obtained by this expression yields the
value of 800 to 500 ohms for the radiation resistance in
l
the wave range (0.8-+- 2.5) A0 , where A0 = 4 .
Part of the eneygy fed to the rhombic antenna is lost in
the absorbing resistance. Hence, the antenna efficiency is
less than unity. If we take into consideration the fact that
thP. current amplitude changes along the wire as a resu It of
the exponential radiation I x= I "e-~x, the antenna efficiency
can be determined from the balance of the powers. Indeed,
the power fed to the antenna is:
J" '
P,n=2ow.
At the same time. it is assumed that the antenna input
resistance equals its wave impedance. rhe power radiated
by the antenna equals
- J:e-~~'R
Pz-2 "' l:
and the power lost in the absorbing resistance is:
1
P\O:JSel = 2
:
e -63'W
t '

at the same time, the absorbing resistance is taken to be

equal to the wave impedance of tQ.e an~eMa. Substituting
~

436
the cited expressions into the condition P 1 = Pz
,. n
+P ,
we obtain: o••
1

Rr,
w =2 sh 2Pl.
In the case of small attenuations, we get approxin1ately
sh 2Pl ~ 2p1, the attenuation constant being then
ex pressed as:
( 11-9)
The efficiency factor of the rhon1bic antenna is expressed as:
p I~
__;_ e-·~'-1-
, -
.• - P I --
in
W -- e-t.~t
~ . (Il-l 0)
Experimental data give the following values for the
efficiency factor of a rhombic antenna: on the long wave of
the operating range, T}=-0.5 and on the short wave of the
range, f)= 0. 75. Thus, from 25 to 50 ~o of the po\ver is
spent on heating the absorbing resistance. If the absorbing
r~sistance is removed (Ra=O or Ra=oo), the efficiency
factor becomes close to unity But then, a reflected wave
is set up in the antenna, which gives rise to a radiation
in the opposite direction, thereby lowering the directivity
of the antenna. The antenna input resistance will then be
frequency dependent.
The antenna directive gain can be expressed as:
E" r 1
D= vo
30/: Rz 0
and if we take into account that Rto=Rr. e-·~ 1 = W (l-e-·~ 1),
on substituting ( 11-7) into this expression, we obtain the
following calculating expression for the directive gain:
e-•~ cos• <p
D= 7,680 W (1-e- 4)1) (l-cos ~sin cp)1 X
x sin• ~(1-cos A sin cp)J sin 1 (kh sin A). (11-ll)
~

Now, let us· examine the question of the choice of the main
dimensions of a rhombic antenna The antenna will have
its maximum radiation in a given direction when the
separatE' factors in (11-11) assume their maximum value The
factor sin• (kh sin 6) will have its maximum value for
487
kh sin !J. 0 = ~ , from which we derive the optimum height
of the antenna above the earth ·
h= 4s~ A 0
0 , (11-12)
where A0 is the calculated (main) wave-length; the ·angle A
should be defined from the condition of the propagation of
cos• cp ....
radio waves. The factor (l -cos A• s1n cp )• Is at a maximum
for the condition
sin q>= sin !l 0 • (11-13)
f"'rom this we derive the value of half the obtuse angle of
the rhombus. And finally, the maximum of the factor
sin' [¥ (1-cos !J. sin q> >]
0

occurs on the condition

~(1-cos!J.,sinq>)= ~.
From this we derive the length of the side of the rhomb
- l= 2 (1-cO:A. sin cp) • (11-14)
A. 0 is generally referred to as the optimum wave-length of
the antenna. Calculations show that when ll 0 = 15°, we get
the following data for the antenna: h= A0 , <p = 75° and
l = 7.5. However, in practice, the following dimensions are
taken for the antenna: h== 'A 0 , q>=65° and l= 4A. 0 • In that
case, the power gain is c.lown only by 15% and, at the
sante time, the size of the antenna is considerably reduced.
Above, the wave impedance of the antenna W was
assumed to be a constant quantity. However, it increases as
we move towards the obtuse angle, ·since the distance
t-etween the wires of the rhombus increases. To maintain
a constant wave impedance of the antenna along its length,
the rhombus is usually formed of two wires diverging
towards the obtuse angle, as shown in Fig. 11-13. The wave
mped~nce of the antenna is then expressed as:
2J
W==276logV~ui' ·(11-15)
where s is the distance between, the diverging sides Of the
•\ rhombus in any section;, : · J~
'
' ..
~
 H, the distance between the wires, forming a side
of the rhombus; I

d, the diameter of the wires.

In practice, the value H of the obtuse angle of the rhombus
is taken of the order of 2 m. Then, as shown by experience,
the travelling-wave ratio in the feeder is found to be close to
unity in the operating wave range.
•

Fig. 11·13. General view of a rhombac antenna.

In the case of transmitting antennas, the absorbing

resistance is a twin-wire iron line, because a rather high
power has to be dissipated in that resistance.
The length of the iron line is chosen so that the
intensity of the reflected wave at its input should be much
lower than the intensity of the incident wave. If we desig·
nate by p the ratio of these intensities, the length of the
line will be defined as:
p=e-s'l·
where the attenuation constant in the iron line is
'

expressed as:
f'= :; , R = · ~ oa
1
5 5 1
V I'~ [ohms/m],
wbere r is the radius of the wire, in mm;
I' , the relative permittivity (for sbort waves 11, ~ 80);
· Q; the resistivity, in ohms· m;
1, the wave-length, In m:
W1, the wave impedance of the iron line.
If W==100 ohms, r== 1 mm, p==O.l, the necessary length
of the iron.. line for the wave range "- == 15+45 m is l==350 m.
-
The absorbing line is usually placed on wooden posts under
the antenna, as shown in Fig. 11-13.
In practice, to increase the directivity, the rhombic
antennas are connected in parallel. Fig. 11-14 shows a
so-caJled duplex
rhombic antenna
designed by G. Z.
1
-Eisenberg. The
d distance between
1 the diagonals of the
T rho m bus is taken
equal to d=O.B Ao.
In that case, the
Fig. 11 -14. Double rhombic antenna.
antenna directional
diagrams in the
horizontal plane narrow down and the relative magnitude
of the side lobes decreases. This is defined by the factor
2 cos c: sin ~) ,
which should be introduced into the expression (11-6).
Calculations an{i experiments show that the coefficient
of amplification of a double rhombus is about twice that of
a single rhotnbus.
Table 11-8

c:0 0• .-
-
c:
c:
•
....
:<J
ftlcu ..·- ·- -
0 '
v"""C.
.,oeo.~
...!.
0
~

Antenna
~

..>.ea
--
-• oo .. --ea
a .!o
-., ., c: .,c:
... .c:
-u
•
-..,..
u
..; -e.,
tll - - L. ~
ea.,
.,.bl ...... _.,
0
.c: "'- .,
r•

... bD "a ., be~ ...

c·-
0 c:_ c -IQC:bl
Cl. <o- Ec:as~

Single
l
108 50 9.6 1 16 13.6 0.75
rhombus 104 48 13 20 17 0.75
l =4A0 , 38 15 28 31 32 0.65
q>=65°

Double 0.7~ 218 100 9.6 14 13.6 0,75

rhombus ).. 208 95 13 18 17 0.75
1=4 10 , 2'A'0 55 25 28 29 32 0.75
'
cp=65°

HO
 Table 11-3 gives the values of the electric parameters of
a single and double rhombic antennas. ,
In table 11-3, 2p.,, and 2~~ 1 • represent the width of the
directional diagram at a quarter of the radiated power.

11-8. Travelling-Wave Antenna

The rhombic antenna described in the preceding para-
graph is used both as a transmitting and a receiving short-
wave antenna.
The short-wave travelling-wave antenna discussed below
is used only as a receiving antenna. It consists of an array
of symmetrical dipoles connected through capacitances of
T Direction of orriw l
L of wove

To receiver

l d •

1
Fig. 11-15. Circu it of a travelling-wave antenna.

small value to a twin-wire gathering line (Fig. 11-15).

The length of each half of a symmetrical dipole l is taken
less than a quarter of a wave-length. The distance between
neighbouring dipoles is taken of the order of one-tenth of a
wave-length. At one end, the twin-wire line is loaded with
an absorbing r,esistance equal to its wave irnpedance. At the
other end, the line is connected to the receiver input, the
resistance of which is also taken equal to the wave impedance
of the line. Due to the presence of capacitors (uncoupling
elements), the input resistance of the dipoles increase
and they shunt the gathering line but insignificantly.
The electromagnetic wave in the field of which the
antenna lies induces in the symmetrical dipoles emf's equal
in. amplitude but with a phase shift between neighbouring
dipoles equal to angle kd cos p, where Pis the angle between
the direction of incidence of the wave and the antenna axis.
Current waves arise in the twin-wire line due to the action of
these elementary emf's, which are propagated from each
dipole to both ends of the twin-wire line. Since at both ends,
44J
the line is loaded on resistances equal to its wave impedance
and the input resistances of the dipoles are considerably
higher than the wave impedance of the line, these current
waves are hardly at all reflected along their path of propa-
gation and are fully absorbed by the end resistances. The
influence of the dipoles, the input resistance of which is of
a capacitive nature and which are close to one another rela-
tively to the wave-length manifests itselfJn a decrease of the
phase velocity of propagation of the current waves along the
line.
If we assume that the velocity of propagation of the
current waves along the line equals the velocity of light, the
elementary currfnt waves at the receiver input when the
electromagnetic wave comes from the absorbing resistance
(P =0) should reinforce one another, being added up ari thtneti-
cal ly. At the same time, the value of the total field at the
receiver input will be proportional to the number of dipoles
N. It might seem that the antenna could be taken as long as
desired and the longer, the better. However, due to the in flu-
ence of the dipoles, the phase velocity of the current waves
along the line is aJways less than the velocity of electro-
magnetic waves in space, so that the elementary current waves
fron1 each of the dl poles at the receiver input differ in phase
by a certain small angle and are added up geometrically.
At preset phase velocity and wave-length, the elementary
current waves from the first and last dipoles at a certain,
definite (optimum) length of the antenna, are in anti phase.
The total current at the receiver input is then at a maximum
and any further increase or decrease of the length of the
antenna causes the current to faJI. In practice, the optimum
length of the antenna is found to equal approximately three
to five wave-lengths. In just the same way, at prescribed
wave-length and length of antenna, there is an optimum phase
velocity for which the currents at the receiver input from the
first and last dipoles are in antiphase.
When the electromagnetic wave is propagated from other
directions, in particular from that end of the line to which
the receiver is connected, the elementary current waves at
the receiver input are also added up geometrically but
with a large phase shift. Hence, they will cancel instead
of intensifying one another and the total current will,
generally speaking, be all the closer to zero as the direction
of reception is closer to the antenna axis ftom its receiving
442
end. As a result, the directional diagran1 of the antenna has
the form shown in Fig. 11-16, a.
Travelling-wave antennas are multiple-tuned. and retain a
directional diagram of approximately the same shape as that
shown in Fig. 11-16, in a 2+2.5-foJd wave range. With
respect to long waves, this range has the limitation that
the antenna loses its directivity and receives less power
because the input resistance of the dipoles becomes very
large. With respect to short waves, this range has the
limitation that the dipoles reach resonance, as a result of
which, the phase velocity of the wave is considerably lowered
and the reception from the main direction decreases.
As can be seen from the diagrams in Fig. 11-16, the an-
tenna has a considerable directive gain. However, due to the

(D) (b)
Fig. 11-16. Directional diagram of a travelling-wave antenna:
a-horizontal plane; b-vertlcal plane.

presence of an absorbing resistance and the weak coupling

of the dipoles with the line, the antenna efficiency is low and
the directive gain is found to be of the order of unity. For
this reason, travelling-wave antennas are only used for
reception. The relatively small size of the side lobes of the
directional diagram makes the antenna quite suitable for
short-wave reception, especially when the ratio of the power
of the signal to that of the interference is more important
than the absolute power of the signal.
The antenna is usually secured in a horizontal position
o~ approximately 17m high supports (so as to lower the cost
of production) and consists of two separate arrays: one for
riight-time and for day-time duty. The night-time array con-
sists of one array of dipoles and that of day-time-of two
single arrays of dipoles connected in parallel. As indicated
earlier, in the day-time system. the dipoles are coupled with
the line t~rough capacitances and, in the night-time array,
448
the coupling is conductive. The antenna has the following
characteristics:
a) Day-time array
Length of array L=87 .5 m.
Length of dipoles 1=3.9 m.
Distance between dipoles d=2.5 m.
Number of dipoles N=36. - ·
Capacitanre of coupling capacttors C = 6.5-8 pf.
Operating wavE' range ).,= 15 + 35 m.

b) Ntght-time array

Length of array L~87.5 m.

Length of dipoles l=B m
Distance between dipoles d=2.5 m.
Operating wave range "-=35 + 60 m.

Let us now investigate the mathematical side of the

problem. Let the antenna have N dipoles. Let P be the angle
between the antenna axis and the direction of incidence
of the electromagnetic wave in the antenna plane (Fig. 11-15).
Taking the phase of the field intensity in the vicinity of dipole
nun1ber 0 (near the receiver) as equal to zero, the expres·
sion for the emf induced by the electromagnetic wave in the
n-th dipole will be:
E = E0 'A. cos (kl sin ~)-cos kl tnkd cos p (11-16)
n n sin kl cos ~ e ·
'

The current at the receiver input caused by the emf equals:

-
I = E 0 ), cos (kl sin P)-cos kl X
n (W A +2ZA) n stn kl cos~
X etnkd cos 1J- tn t ltd, • (11-17)
where ZA is the input resistance of each of the dipoles,
with the resistance of the coupling capacitors
taken into account;
WA• the wave impedance of the line, with the inftu·
ence of the dipoles taken into account;
~=-t. the ratio of the velocity of propagation of
the wave in space to the phase velocity of the
wave propagation in the linea.• · ~
The total current in the receiver from all the dipoles
equals: .
, E0 A cos (kl sin p)~cos Id
I rec= ( W ) X
ZA + A
2
A.
2n sin kl cos .,

sin [ N:d (cos p- ~)]

(11-18)
· X sin [ kd (cos~-;)
] •
2
•

(11-18) expresses the current in the load depending on the

incidence angle of thP wave P and defines the antenna
directional diagram in the horizontal pta·ne (in the plane
of the antenna).
When two parallel systetns are connected to a common
matched load, as in the day-time antenna, the expression
(11-18) should be multiplied by the quantity
, 2 cos (~sin ~) ,
where D is the distance bet\\ e~n the axes of the arrays.
Note that in (11-18), ZA is expressed as:·
ZA=Rl'.o-iWwcotkl-i ~, (11-19)
Ca> sd

where Rr.o is the radiation resistance of a symmetrical dipole

related to the point of its connection to the
line;
W'w, the wave impedance of the dipoles;
Csd, the capacitance of two series connected
condensers in each symmetrical dipole.
The expres3ion defining the directional diagram in the
vertical plane passing through the antenna axis is:
N kd ]
, E01 (1-cos kl) r
sin - 2 (cos A-;)
I rec = ( ZA W )
+ A n sin kl sin
[kd (cos A-;) ] X
2 2
x sin (kh sin A), r
(11-20)
where h is the height of the antenna above the earth;
A, the angle of elevation of the ray above the
horizon.
Fig. 11-16, b shows the directional diagram of the
antenna i.~ the vertical plane on the wave ). = 15 cm.
445
 Let us now define the conditions of maximum reception
of the travelling-wave antenna. When the radio waves come
from the main direction <P =0), in accordance with (11-18),
the current in the receiver becomes equal to zero on
condition that:
Nkd
2
(1-;)= n,
whence we see that if the phase velocity of the wave in
the antenna satisfies the condition '-1= ;d'
there il! no
reception from the main direction. The maximum of re-
ception from the main direction will occur on condition that:
N kd n
-2(l-;)= 2'
from which we get the condition of optimum reception:
A A
£-1=2Nd~2L' (11-21)
where L is the length of the antenna.
If, for example, L=5A (which is what we get in
practice for the short waves of the operating range), then
;-1=0.1 and v=0.9v1• The length of the dipoles and
the value of the "tapacitance of the coupling capacitors
should be chosen such that the velocity of the wave
propagation should not be less than that value.
When the length of the dipoles is l < ~ (which is what
we get in practice), the reactive part of the input resistance
of the dipoles is of a capacitive nature and equals:
1
med = Wdcotkl,
-
where cd is the equivalent capacitance of the dipoles.
The maximum shunting effect of the. dipoles occurs on
a wave for which l = ~ . That is why it is advisable to
define the necessary capacitance of the coupling capacitors,
for precise) y that case.
The velocity of propagation of the electromagnetic wave
in a twin line is expressed as:
!=JIL,C,.
"'• inductance, in by/m;
where L 1 is the distributed
c,. the distributed capacitan~&.i~ in .fJm.
'f
,
448
 .
If the wave-length is resonant for the dipoles, the
velocity of propagation of the wave in , the line is
ex pressed as:
-tJ= y'L, (c.+~).
where Cc is the coupling capacitance, in f;
d, the distance between two neighbouring dipoles,
•
1n m.
The ratio of these velocities is:
C I-t- Cc
d-
--
t-~--
V - ' cl • (11-22)
For a twin line
JO' .
C, = 0
[pfjm], (11-23)
361n-
i • '
where D is the distance between the wires;
r, the radius of the wires.
The necessary capacitance of the coupling capacitors
can be defined from (11-22) and (11·23). Assuming, for
example, that D=90 mm, r = 1.5 mm, E==l.l, we get
Cc~3.6 pf and, consequently, each of the two series connected
capacitors should have a capacitance of the order of 7.2 pf.
In the case of the above-mentioned dimensions of the line
(D =90 mm and r= 1.5 mm), the wave impedance of the line
equals 500 ohms. Since the coupling capacitors increase the
distributed capacitance between the wires, the wave
impedance of the line decreases. It depends on the wave-
length. and, in, the case of the above-cited exam pie, it equa Is
approximately 400 ohms.
Consequently, the absorbing resistance should be 400
ohms. In the ease of two arrays connected in parallel, the
receiver input resistance should be 200 ohrns. If the antenna
is connected to the receiver through a feeder, the wave
impedance of the latter should be 200 ohms, which is the
case in practice. .,
The line of the night-time antenna array is usually of
four wires, with a wave impedance of approximately 200
ohms. A plane view of the antenna is shown in Fig. 11-17.
The electric parameters of the antenna array are given
in Table 11-4.
441
 ~·«JOOhms Ho= /NJOo/11118 Ra=400tJhms
•
~ p
.[J - - ~) - 0 ..
~
.. , ~
... • -
loo
•

:::
-
- -.. ) -
~
•- ~
... - - - -
•
• - - • ~
... - - -
- - ..., - ..
q
- .. ~ -
I""

-
~
-- ..
- ~
.
I"" - •
.. ..I
~

4
..
~
- ...
-. -
I"" -
4

..... I

-
I I

-
...
- ... '
...
-- - .... ~

-
- - - ... -
~ ..
I"" I""

< - .. -
...
~

....
-
-
* -
... )
-
--, - -
-..
.. -
~
~

-
.... - -.. t c .. (

-
c
- • c
- -- -
4
I""
•
~
•
~ - ...
•
- - ~
~ -
...""' ,
I
~
- ""'
~
W=~r)
~

d ~
I""
~

W=«JOjJhms W•400o/7ms ohms

- ' "

' I W=200ohms lb receiver

70 rscelver •

Fig. 11-17. Travelling-wave antenna.

As follows from the travelling·wave antenna theory inves-

tigated in Paragraph 4-5, the phase velocity of waves in
antennas of this kind should be inferior to the velocity of
radio waves in free space.
Table 11-4

-·-c.
' •
0 -.2." - -
c• •
c
. tail

...c
c·~
., ..
c:.c
......
~

-.. - -·-
c:: • ... a
c::<2
.5!:>. .....
u ... c.
uOC>~
• -
- ...- ilu- ..
-- ... -•
..
Cu
o-
.. ....... a 0
u
...
.c= > !:.c u ~

tlo
c:
...
cu
c•
_.,
IIIU

..
OJ:
...
~

-
.,
..- -...
I)
~
·-.,
-u.., c..,
-s::
0 ..we..,
oc: ... -- -· ..... .,·i t'
"'C::C
oec..,
.., c: ~
.... <1
oa~
11
u
uo --~ .a.
I
,c., .!i:lt ..
,1:1 & •
~ u .c-
•
Et
-.o>
-o
".a
::c~ -
f
0 ~~
ra.u
<-a
o-
'G ....
;:;s~
&
-··- - i "0 till s::
iS-~-
"a C."'G
...
CsJ

, 0. 7A0 80 2.75 16 24
40
18 0.0565
2A0 0. 7A0 20 0.28 25 32 0.023
~

148
However, tn the case of the short-wave travelling-wave
antenna investigated above, the use of capacitors as coupling
elements for dipoles with a distributive lirie is not indis-
pensable. These elements can consist of, for example, active
resistances. The use of inductive coupling elements 1s
inadvisable because, in that case, the phase velocity in
the line throughout the operating wave range exceeds the
velocity of radio waves in free space, as a result the
directive properties of the antenna are considerably in1paired.
The use of active resistances as coupling elements enables
to obtain a considerable improvement of the antenna
parameters in comparison with those of antennas which have
capacitors as coupling elements [71 ). ,
This can be explained as follows. The directive gain of a
travelling-wave antenna (as of any other antenna) decreases
as the wave-length increases. Hence, it may be particularly
desirable to increase the directive gain in the long·wave
tnargin of the range. Here, the input resistance of the
dipoles (without coupling element) is of a capacitive nature
and of a magnitude which enables to obtain in the line a
regimen close to optimum type of operation (the phase shift
between the currents of the first and last dipoles amounts
to 180-230°). However, when a capacitive coupling element
is being used, the equivalent capacitance of the dipole turns
out to be quite small and the magnitude of the phase velocity
in the line exceeds the optimum value, which results
in a deterioration of the directive properties. In the case
of an active coupling resistance, the dipole input resistance
ensures by itself a velocity close to the optimu1n in the long-
wave margin of the range.
Since the capacitive coupling imP.edance increases linearly
with an increase of the wave, this causes a decrease of
the antenna efficiency approximately proportional to the
square of the wave-length. The antenna power gain decreases
correspondingly. If the resistance of the coupling element
was not frequency dependent, the decrease of power gain
would be proportion~! to the ratio ~: (h~re, L is the length
of the antenna). Thus, in the case of a travelling-wave an-
tenna with coupling capacitors, an increase in wave-length
causes a decrease of the power gain approximately proportion-
al to the fourth power of the wave-length. But in the case
of an active coupling resistance, the fall of the power gain
29-2122 449
is considerably slower. Calculations as well as experiments
show that the efficiency of the antenna with resistances
(double-array system) lies between 20 and 60%, whereas
the efficiency of an antenna with capacitors is 0.5 to 20%.
In both antennas, the efficiency decreases as the wave-length
increases. It has been established that the optimum
magnitude of the active coupling resistances (for antennas
consisting of two arrays) is 180-200 ohws.
Antennas of this kind have recently found application at
radio centres of the Soviet Union. .
11-9. Diversity Radio Reception
On short waves, as a rule, several beams arrive at various
angles to the horizon. Due to the insta bi Ii ty of the ionised
layers of the atmosphere, the difference of path of the beams
changes in titne causing weakening of fading out of the
signals and, in telegraphy, to the disappearance of isolated
letters or even whole words.
The routine way of checking fading is through diversity
reception. It has been experimentally established that fading
is not identical in two even relatively close spots. At any
given instant, antennas situated at these spots give rise
to signals of different magnitude; moreover, full fading out
of the signals does not occur simultaneously for different
antennas.
To eliminate fading, two or three receiving antennas are
installed on the premises of the radio centre, approximately
300 m apart.
It has been shown by experin1ents that the best results in
this respect are obtained when the antennas are situated
along the I ine of arrival of the \\'ave. This is due to thr fact
that a displacement towards the arrival of the rays leads to
a more pronounced interference of the rays falling at
different angles to the horizon. •
In telegraphy, three separate receivers are used, one for
each antenna, and the signals are aggregated following
rectification at the receiver's outputs.'
Three receivers are likewise used in telephony, but the
signal delivered to the telephone comes from that receiver
which picks up the clearest signals at the given moment, the
other receivers being switched out. The object is to avoid
noises from the receivers which, at the given moment,
deliver weak signals.
 CHAPTER TWELVE
Medium- and Long-Wave Antennas
12-1. Classification of Antennas
Taking into consideration the conditions of propagation
of medium and long wavt-'s, i.e., waves of frorn 100 rn upwards,
the antennas used on these waves rnust have a vertical polari-
sation of the electromagnetic field. In most cases, the dirtc·
tional diagrams in the horizontal plane should represent
circumferences; directional antennas can also be used for
broadcasting, chiefly in a preset direction, or for radio com-
munications between two points, or to eliminate inter-
ference. On long waves, the maximum radiation of antennas
should be directed along the surface of the earth. From the
point of view of their directional properties, long-wave
antennas are relatively sirnple.
Long-wave antennas comprise antennas of the wire type
(T -antennas. ~.-antennas, umbrella antennas, etc.) and of
the tower or mast type (antennas on insulators, shunt-fed
antennas, mast antennas of upper feed, etc.). For reception,
antennas of the frame type (simple or screened) are used, as
well as antennas of the goniometrica I type.
Let us begin with an investigation of T- and L-antennas.

12-2. T- and L-antennas #

T- and L-antennas find wide application in broadcasting

and radio communications, on medium and long waves. The
general view of these antennas is shown in Fig. 12-1. As a
rule, they are secured on two lOO to 250 m high masts spaced
100 to 250 m apart and represent a plane net of horizontal
29* 451
and vertical wires. The antenna array has two to sixteen
wires spaced 1 to 1.5 m apart. The copper wires are usually
5 to 8 mm in diameter. The supporting masts are secured by

~--o-~
Fig. 12-1. General view of T• and L.. antennas.

means of several tiers of guys into which insulators are

inserted to eliminate the influence of the guys on the antenna
radiation. The antenna downJeads directly connect the radio
trahsmitter; hence there is no need for the feed lines.
The electric current distribution along the antenna is
shown in Fig. 12-2. Since the geometrical dimensions of the
.......-lh~

lv

Fig. 12-2. Current distribution in the antenna .

•

antenna are small as compared with the operating wave-

length, the influence of the earth adds to the radiation of the
vertical part of the antenna and weakens the radiation of the
horizontal part. The horizontal part of the antenna is
destined to increase the capacitance of the antenna and ensure
a more uniform current distribution along its vertical part.
The coupling of the transmitter with the antenna downlead
can be inductive, conductive or capacitive. Apart from a
coupling circuit, a circuit for tuning the an1enna in resonance
is also inserted into the antenna; the circuit acts as an
inductance coil in case of lengthening and as a condenser
in case of shortening. Examples of the antenna coupling and
tuning are shown in Fig. 12-3. .. (
451
 The antennas are fastened to masts (of wood or metal) by
means of insulators and are fixed in position by means
of cables through pulleys secured to the top of the sup-
porting masts .

of Wh~of
nods voltage ontlnode
Fig. 12-3. Coupling of antenna \\'ith trans-
, 111 it t er.

To reduce losses into the earth, the lower end of the

antenna downlead is connected to the earthing through
tuning and coupling
circuits. Tl
T• and L-antennas find o I w,
frequent application
among radio hours. Sn1all
+I
in size they are often l"
installed on the roofs of
houses. T
z
To simplify calcula-
tions of T- and L-anten-
nas, in particular the ea)- Fig. 12-4. Calculating· the equivalent
culation of the current length of the antenna.
and voltage distribution
along the antenna, the horizontal part of the antenna is
replaced by an equivalent length b of the same wave
impedance as the vertical part (Fig. 12-4). The current and
voltage distribution in the antenna is expressed as:
sink (lv +b-z)
1• I o sin k (lv + b) ;
= (12-1)
U _ U cos k (lv+b-z)
, - o cosk (lv+b) '

461
where I 0 and U 0 are the current and voltage at the base
of the antenna;
b, the equivalent length of the horizontal part of
the antenna;
lv, the length of the vertical part (downlead) of the
antenna.
The equivalent length b is chosen such that the current
and voltage distribution in the vertical-part should retnain
the same as in the antenna with a horizontal part. It also
means that the input resistance of the horizontal part of
the antenna is equal to the input resistance of a vertical
segment of equivalent length.
From (12-1) one can define the current and voltage at
the upper end of the vertical part of the antenna:
sin kb
lzv = I o sin +
k (I v b) ;
( 12-2)
cos kb
U lv = V o cos k (ly +b) .
Along the horizontal part of an L-antenna, the current
and voltage are distributed as
.. I = I sin k (lh- x) •
~ lv sin klh '
U =V cos k (/h-x)
x lv cos klh •

Along the horizontal part of a T -antenna, the current and

voltage are distributed as

I
sink(~2 ~x
I-·
x- 2
IV_ -- - - - - ·
i klh '
sn
2

cos k e;·-X).
Ux= Ulv kl -'
cos b
2
where Lh ~~ the length of the horizontal part of the antenna;
x, the distance along the horizontal part, measured
from the connection point of the vertical part of
the antenna
Knowing the current and voltage at the ba~e of the
antenna, as well as the equivalent length of the horizontal
••

454
part, one can calculate the current and voltage at any
point of the antenna. /
Let us define the equivalent length of the horizontal
part of the antenna. For an L-antenna, the resistance of the
horizontal part at the point where the vertical part is
connected to it, equaJs:
Xh=- iW h cot klh,
where Wh is the wave impedance of the horizontal part of
the antenna.
Let us replace the horizontal part of the antenna bv an
equivalent wire of length b with a wave impedance equal
to the wave impedance W v of the vertical part of the
antenna, so that its input resistance should remain the
same as before ·
Xh=- iW v cot kb.
This leads ~s to the equivalent Jength of the horizontal
part of the L-antenna
:h
cotkb= cot klh.
V
(12-3)
For aT-anten na, the resistance of each arm of the horizontal
part equals- iW h cot k~h . The resistance of the horizontal
part loading the vertical part of the antenna equals:
X . wh
t 2
t klh
CO 2
h=- I

The equivalent length of the horizontal part of aT-anten na

is then ex pressed as:
CO tkb wh
= 2Wv CO
tklb
. (12-4)
2
The above expressions enable the easy calculation of th~
reactive part of the input resistance of the antenna T- anh
L-antennas are usually applied on long waves, for whic
the length of the antenna is small relatively to the oper·
ating wave-length. That is why the reactive part of the input
resistance of the antenna can be calculated by means of
the expression valid in the ea se of a sinusoidal distri·
bution,
Xv=- iW .cot k (lv +b). (12-5)
To make th e calculation and definition of the character of
the reactive part of the antenna input resistance easier the
455
concept of the natural wave-length of the antenna is
usually introduced, by which one understands the longest
wave for which the antenna is tuned in resonance without
the introduction into the antenna of a tuning resistance.
The natural wave-length A0 is defined from the condition
k 0 (lv+b)=T,

where k.=i;, and b is a function df the wave-length.

For an unloaded antenna, the natural wave-length
equals
· A0 =4lv.
Due to the fact that k 0lv=T-k 0 b, we obtain the
following ratios for the definition of the natural wave·
length from the expressions (12-3) and (12-4)
w
k lv=w:cot k lh;
tan 0 0
(12-6)
wh t kolh
t an k ol v = 2W
V
co 2 •

The equations (12-6) are transcendental and are solved

graphically. .
Consequently, in the general case, the natural wave-
length does not equal the quadruple value of the wave-length
and depends on the ratio of the wave impedance of the
vertical and horizontal parts of the antenna.
Let us now deal with antenna tl)ning. On operating
waves differing from the natural wave-length, the antenna
has to be tuned in resonance by inserting a reactive resistance
at its base. Furthermore, we should have the ratio
Xv+Xt=O,
whete X1 is the antenna tuning resistance.
There are two possible cases. Let Aoper> A0 ; then, in
accordance with (12-5). the antenna represents a capacitive
resistance. In that case, to be tuned in resonance, the·
antenna has to be lengthened, i.e., an inductance
lengthening coil inserted at its base, the resistance of
which should equal
(i)Ltn&tb= Wvcot k(lv+ b). (12-7)
Now let "-onr <A,. In that case, the electric length of the
antenna wi 1 be greater than 90° . The amenna represents
''

156
an inductive resistance and the insertion of a shortening
capacitance is necessary to tune it in reso}1ance
1
WC
abrtn
=- w V COt k (LV -t- b). (12-8)

The expressions (12-7) and (12-8) are utilised to define the

tuning elements of the antenna.
Let us pass on to the determination of the radiation
resistance of T- and L-antennas. In the case of antennas
that are short in comparison with the wave-length, in the
present case, of antennas operating on waves larger than
the natural wave-length, they can be regarded as dipoles
and the radiation resistance calculated frorn the expression

Rr.o = 160n ~
1
(
Aheff )
, ( 12-9)
opcr

where heff represents the effective height of the antenna,

which is calculated from an equivalent surface of current
in the antenna and expressed as:
h = .!_cos kb-cos k (lv -t- b) •
( 12-10)
eff k sin k (lv +b)

In practice, the calculations based on the expressions ( 12-9)

and (12-10) are iound to be sufficiently accurate.
As regards the calculation of the directional diagram
of T· and L-antennas, the radiation of the horizontal part
in cotnparison with the radiation of the vertical part is
usually negl~ted. The radiation of the antennas in the
horizontal plane is then uniform and, in the vertical plane,
the directional diagrams are the same as for dipole
antennas.
When operating on waves shorter than the natural
wave-length, a more accurate expression is to be pre-
ferred, viz.,
601 0 X
E= r 0 sin k (lv +b)
.,
cos kbcos(klvcos9)-cos9 stnkbsin(klvcos9)-cosk (lv +b) (l 2·ll)
X sin 0 '

where 9 is the angle between the antenna axis and the

direction towards the point of observation.
451
 Note that the full power fed to the antenna can be
characterised by the expression
,.
P A= T (Rio+ R,osses),
where R1ossps is the resistance of the losses.
In that case, the efficiency of tb.e antenna can be
ex pressed as~
Rr.o
11=-=---~­
R to + R losses

The resistance of the losses 1s mainly determined by the

power lost on heating the earth. The expression generally
used for calculating the losses in the earth on long waves
is M Shu leik in's etnpir1cal ex press ion
••
R losses= A AoPer
Ao ( 12-12)

The coefficient A is defined by the electric paran1eters (e, o)

of the soil and the earthing system of the antenna. It
changes w:th in ra4her large limits: frorn 0.5 to 7 When the
earthing is properly done, the quantity A does not exceed
2.5 ohms (for '"opt"r =A0 ).
Thus, the expression for the full active resistance of
the antenna is
1
Rin= 160n ( heft ) + A Aopt•r
1
( 12-13)
Aoper Ao

At a certa1n wave-lenllth, R,n is at a minimum. When

measuring the minimum input resistanc~ .,r the given antenna
and calculating th~ radiation resistance, the quantity A can
be def1ned from (12-13), after which, the resistance of the
losses can be determined from (12-12).
The earthing system often consists of a wire net buried
at a small depth. The wires capture. the displacen1ent
currents branching off from the antenna and terminating on
the earth and, thereby, result in a decrease of the losses in
the earth (Fig. 12·5). The larger the number ot wires of the
earthing ~ystem and the greater their length, the sma Iter
the part of the current branching off from the antPnna and
flowing directly through the earth. the ~m a ller the energy
losses in the earth and the higher the ef.f~ciency.
458
 The power fed to the antenna being prescribed. the current
at the base of the antenna equals: .
I=
0
V + 2PA
R :Z::o R losses •
The voltage at the , _________
base of the antenna is expressed as:
U o= I o l (Rio+ R,oss~s) +· X~n •
1 1

The voltages and current at the upper ends of the antenna

can be defined from ( 12-2) .
..-...--..-- ..
I
/ ""'
//
"'
_,- -- ' '
',~
....... \
\
I I /,. ", \\ \
,,,
I I I
/ '\
'
\' \ \

l i l tI I
' ' \
\\\11\II
~' '-~~~
- ........--
----
..... ,_,,
__ , /

'
......____ ~

(Q) (/))
Fig. 12-5. Earthing syste1n of an L-aritenna:
a-central earthing: h-radlal earthing

The wave impedance of the vertical and horizontal parts

of the antenna of the above expressions can be defined
through the sta1 ic capacitance of the antenna by means of
the expression
w- ao - Ctcmfcm) •
'fhe distributed capacitance C, measured in centimetres per
centimetre of length of the vertical or horizontal part, are
usually calculated by the approximate Howe n1ethod, ac-
cording to which, the static charge is assumed to be distributed
uniformly over the surface of the wire. It is then found
possible tc' calculate the potential in the conductor, the po·
tential which changes along the conductor. Taking the mean
value of the potential all along the length of the conductor
and dividing by it the fuJJ charge of the wire, the capacitance
of the \\'ire and then, the capacitance per unit length of
the wire can be defined. To calculate the capacitances of
T· or L-antennas, the potentials induced on each conductor
by the other conductors of the net and its mirror in1age in
the earth have to be taken into account.
459
 Let us now consider the pass-band of the antennas. Due
to the fact that the geometrical dimensions of T- and L-
antennas are small in comparison with the wave-length, these
antennas are highly directional. The problem of 1he transmis-
sion of telephone or telegraph signals without distortion
is therefore an important one and the antenna pass-band has
to be defined in order to evaluate its possibilities in this
respect. -
When the operating waves are considerably larger .than
the natural wave-length (Aoper ~ A. 0 ), the expression (12-5)
which defines the reactive part of the antenna input
resistance can be expressed app~9ximately as follows:
· Wv . 1
X In~ - l k (lv+b) = - t wCA' ( 12-14)

where CA is the total capacitance of the antenna array, in

farads.
This expression indicates that when the antenna is
lengthened, the reactive energy accumulated in the antenna
is mainly concentrated in the electric field. Then, the quality
of the antenna circuit, for the tuning of which a lengthening
coi I L1ngth is inserted, can be approximately defined as the
ratio of the expression (12-14) to the active resistance of the
antenna, under resonant condition

(12-15)

Hence, in the case of lengthening, the pass-band of T-

or L-aQtcnnas is expressed as:

(12-16)

Thus, the pass-band of the antenna will be all the

larger as the electric length of the antenna and the radiation
resistance are larger and the antenna wave impedance is
smaller When the antenna is shortened, the expressions
(12-15) and (12-16) are no longer valid. In that case, the
tuning of the antenna is obtained through the insertion into
it of a shortening capacitor C•hrtn• and the coupling with the
generator, through the insertion of an io.ductance coil L.
Thus, the reactive part of the general· ·input irupedance of
460
the antenna, i.e., of the impedance of an array with tuning
elements, is expressed as:
X,n=i(J)L+.l(l) d ahrtn
-iWvcotk(i.+b) .
On the resonant frequency, this in1pedance equals zero:
X 1 n(~r)=i(J)rL+.l(l)rdshrtn- -iW.cotkr(l.+b) =O. (12·17)
On frequencies close to the resonant one, it can be
approximately represented by two terms of a Taylor. series:
,
Xin (ID)~ Xin (ror)+ dWXtn (wr), (12-18)
where Aw is the increm~nt of angular frequency.
Taking (12-17) into account and performing the differ-
entiation indicated in (12-18). we obtain:
. ((I)) ,.,.,
X ln ~
A
u(J)
I dXtn ((a))
dw
I _
w=ro -
r
. .ll [L +
=t (J) 2
I + wC
V IV b ·+ ]
sin1 k r (lv +b) •
wrCslutn

But according to (12-17)

1
1 =L- !•cotkr(l.-1-b).
(l)r Csh rtn r

So that we obtain
X 10 (ro) ~ i2l1w X [ L + W v 2kr <lv 1-· b)-
1
sin 2kr (lv +b) ]
4wr sin k, (lv +b) •
Designating the second item of this expression by L.,
sometimes called the antenna effective inductance, the an-
tenna input impedance on frequencies close to the resonant
one, is written as:
Zin ~ Rin+ i2Aro (L ·+- Le>· • ( 12-19)
Thus, the reactive part of the antenna input impedance
is proportional to the frequency increment. The active part
of the input resistance can be regarded as constant within
the pass-band. We know that the antenna pass-band is the
band on the limit frequencies of whicfi the current in the
antenna decreases by V2 in comparison with the current at
resonance. It can be shown that the frequency increment
(Aro) for which such a decrease of current occurs, is defined
by assuming that in the expression (12-19), the active
461
and reactive parts of the input_ resistance are equal. The
pass-band can then be found from the equality Rin = 2~IDmax
(L + L,) which yields the expression
2/\Ctlmax Rln (12-20)
Wr = Wr (L + Le) ·
This is the expression generally used for calculating
shortened antennas. However, ttJe pass-band is often
simply defined from the curves of the antenna input impe·
dance plotted as a function of the frequency.

12-3. U-Antennas and Antennas with Multiple Downleads

Apart from the 1~- and L-antennas described above, U-
antennas and antennas with multiple downleads are also
used in the long-wave range.
The circuit of a U-antenna is shown in Fig. 12-6. The
energy from the transmitter is fed to one of the downleads,
a reactance being connected
to the second downlead in
order to obtain a co-phasal
distribution of the current in
the downleads of the antenna
I I shown in Fig. 12-6.
A U-antenna in actual fact
is two L-antennas, so both
' are calculated by the same
1. d •I methods. However, due to
Fig. 12-6. U-antenna. the reciprocal influence of the
downleads, the radiation
resistance of the antenna downlead increases. ln can be
expressed as:
•
•

where Rr.t is the induced resistance, defined from the

expression
_ i!_ R (cos kd _ (l-k tP) sfn kd]
REl - 2 Eo k•d•
2
(12-21)
k'd' '
where is the radiation resistance defined from the
Rl;o
expression (12-9);
d, the distance between the down leads of the antenna.
The expression (12-21) is obtained by the induced emf
method. The field of one of the downleads in the vicinity•

462
of the other one is defined by replacing the downleads by
dipoles of corresponding effective heights ..
As a result of the increase of the radiation resistance
of the antenna in comparison with an L-antenna, its efficiency
and pass-band are higher. In addition, we get a certain
d irectivity in the horizontal plane, expressed as:
E = 2£•• cos ( ~ cos cp ) ,
where E 0 , is the field Intensity set up by one downlead;
cp, the angle between the plane of the antenna and
the direction towards the reference point.
An antenna with multiple downleads. proposed by
Alexandersen is shown in Fig. 12-7.
The energy from the transmitter is fed to the middle down-
lead of the antenna. Reactances are connected to two other
downleads, their
magnitudes being
chosen in guch a
way as to obtain a
co-phasal oscilla-
tion of the currents
in aiJ the down·
leads. The current '"~-
~
distribution in the
tuned .antenna is 1• d • I• d • I
shown in Fig. 1~-7. Fig. 12-7. Antenna with multiple down-
The antenna can be leads.
regarded as con-
sisting of three T-antennas, so that its calculation is reduced
to the calculation of a T-antenna. The only difference is
the calculation of the radiation resistance of the downlead,
which can be defined from the expression
Rr. = Rr.o + 2Rr.!,
where Rr.1 is the radiation resistance induced by the
neighbouring down lead, defined from the expression (12-21).
To calculate the resistance induced in the extreme down·
leads, d in the expression ( 12-21) should be replaced by
2d. The radiation impedance of the antenna related to the
current in the downlead to which the energy is fed, is the
sum of all the natural and induced resistances, because the
currents in all the downleads are co-phasal. This concerns
also U-antennas.
463
 The directional diagram in the horizontal plane of an
antenna with two downJeads is expressed as:
E= EOl ll -l- 2 cos (kd cos cp)].
Just as in the case of U-antennas, the efficiency and pass-
band of the antenna are higher than in the case of a T -antenna.
The antenna is coupled directly to the transmitter by
introducing one of the downleads in.to the transmitter
building, but the antenna can also be fed by means of a feeder.

12-4. Mast Antennas

The disadvantages ofT- and L-antennas and other antennas
of this type are: a) the need for two or more masts and
b) the distortion of the directional diagrams caused by the
influence of the supporting cables.
Consequently, mast antennas are used on medium waves
(especially in the 300 to 800 m range). Fig. 12-8 shows three

{0) (b) (C)

Fig. 12-8. Mast antennas:
11-mast on an insulator: b-shunt-fed m~Jft; c-mast fed from the top.

methods of excitation of the masts: at the base of the mast,

which requires a supporting insulator, by· means of a shunt,
and by feeding the antenna from the top.
From the electrical point of view, a ma~t antenna on an
insulator (Fig. 12-8, a) represents ,a base-fed vertical rod.·
It is calculated like an asymmetrical dipole without a load
at the upper end. The current and voltage distribution in
the antenna are calculated from the expressions •

I = I sin 1:· (l-z) U U cos k (l- z)

z o c:in hi ' J = O ro~ bl '

where l is the length of the dipole (Figt 1~-9).

464
 •

The current and voltage at the antenna base are

defined frorn the ex press ions
I=,/ 2PA - .
~ Jl Rr.o + Rtosses
The input resistance of the antenna placed on an insulator
can be calculated from the expressions
I
sh 2Pl - ~ sln2kl T lu
R1n=WA ch2~L-cos2kl;
sin 2kl + ~ sh 2Pl T
z I
I I
X1n=- W A eh 2~1-cos 2kl- • I
In these expressions, the attenuation is
defined from the expressions · Fig. 12-9. Current
and volt age d istri-
Al R,l Rl Rr,a bution In an asyrn-
t' = W A ' 1 = sin 2kl nletrical dipole.
' l - 2kl

and the wave impedance of the antenna, from the ex-

•
press ton

where r is the mean radius of the mast cross section.

For antennas that are short rei at ively to the wave-length
( ~ <0.2 ). simpler expressions can be used:
Rza
RIn= sin2 kl , X I n = - wA cot kl.
The values cif the radiation resistances of mast antennas
are given in Table 12·1.
Table 12·1

.. -
an 0
.,... .., ,....,
1/1 et
•
an
•
Q
c...
•
11)
C'll
•
Q
~
•
11)
~
•
...•
Q

•
0
11)
•
&I)
en
•
;,
tD
•
~
• '"=-•
C) Q Q 0 0 Q 0 C) 0

f
•
Q 0
=
•
REa ohms 3.2 6.5 18 36.5 60 84 100 H16 99.5,83 63.5 46.5 42

For short antennas. the radiation resistance related to the

current at the base can be calculated from the expressions
( 12-9) and ( 12-10) for b == 0.
J0--2122 465
 Table 12-2
' I
Symbols Parameters Mast W A =200 ohms Fine wfre W A =500 ohms

-
kl Electric length of antenna 180° 45° 180° 45°
t

Rza Radiation resistances 99.5 ohms 3.2 ohms 99.5 ohms 3.2 ohms

Pl Attenuation 0.497 0.044 0.199 0.0176

Rin Input resistance 435 ohms 6.5 ohms 2,045 ohms 6.5 ohms

X in Input reactance 0 200 ohms 0 500 ohms

-{ Rrn + x~: Input imnedance

•
435 ohms 200 ohms 2.045 opms 500 ohms

io Current at base 15.2 A 124 A 7 A 124 A

u, Voltage at base 6,600 V 24,800 V 14,300 V G2,000 V

u. Voltage at the antenna end 5,850 V 35,000 V 14,000 V 88,000 V

.
 The earthing systetn of mast antennas usually consists of
60 to 120 radial earthed wires 0.3 A long .. ·In the case of an
adequate earthing system, the resistance of the losses of mast
antennas of this kind related to the current antinode usually
atnounts to approximately 5 to 10 ohms and the antenna ef-
ficiency is about 80 to 90~,6. From the electrical point of view,
1nast antennas have a definite advantage over thin dipoles.
Let us investigate the type of operation of two antennas:
a mast .antenna. (WA=- 200 ohms) and a \Vire antenna (W ..
=500 ohms) wtthout a load at the upper end, operating in
,=
I l
a range from A = 0.125 to A = 0.5. Let the power fed to
the antenna equal 100 kW and the efficiency, for sitnplicity,
equal 100%.
The calculated data are given in Table 12-2.
It can be seen frorn Table 12-2 that the voJtage of rnast
antennas at the lower as well as the upper ends is con-
siderably low than that of wire antennas. This simp-
lifies the operating conditions of the insulators and ena-
bles to accommodate a larger power into n1ast antennas
than into wire antennas for the sarne brea'kdown voJtages.
Due to the difficulty of installing and maintaining masts
placed on insulators, shunt-fed earthed mast antennas are
preferred (see Fig. 12-8, b). The feed Iine is usually con·
nected to the mast at a height t!QUal to •I 1 to •I 10 of the height
of the mast. The lower and upper parts of the antenna are
connected in paraJleJ relatively to the feed line and, if
the antenna is short in cornparison with the wave-length, the
lower part of the antenna represents an inductance and the
current on it has a cosinal distribution with the antinode
at the earthe'd end, whereas the upper part represents a
capacitance and the current on it has a sinal distribution
(Fig. 12-10). At the connection point of the feeder, the
antenna current has a finite discontinuity and the directions
of the current in the upper and lower parts of the antenna
coincide. The equivalent circuit of the antenna is shown in
the same Fig. 12-10.
Fig. 12-8, c represents the mast antenna designed by G.
Eisenberg. The circuit of this antenna showing the direction
of the conduction and displacement currents is given in
Fig. 12-11, a. The antenna is fed at its upper end by means of
a coaxial line (formed by a vertical wire inside the body of
the mast).
30* 467
 The antenna equivalent circuit is shown in Fig. 12-11. b.
Since the antenna is fed from the top, a current antinode is
always formed at the earthed end of the mast antenna.
The advantages of such a mast antenna consist 1) in a more
uniform distribution of its current as compared with a mast
antenna on an insulator, the current
distribution being all the more uniform
as the electric leng.th of the antenna is
a

Fig. 12-10. Current distribu tion in a shunt-fe d

antenna and its equivalent circuit.

/t ,,
I I I ,, \ \
I' I
I I I
_..., '\ \
I I \ I ' \ I
I I I
I I ' I I
I I
I I I I
I
(Q) (/J)
Fig. 12-11. Mast antenna fed from the top:
a-anten na circuit. fa •quivale• , t cl rcul t

shorter, and 2) in the absence of a supporting insulator.

Consequently, mast antennas fed from the top have a higher
radiation resistance and, as shown by calculations and
experiments, a higher efficiency.
Part of the upper row of guys is utilised as a capacitive
load at the upper end of the mast, as shown in Fig. 12-8, c.
The feed wire runs inside the mast in a special container
or without it and is fixed by means of supporting or stick
insula tors.
Now let us consider the utilisation of mast antennas as
anti-fading broadcasting antennas. At night, medium waves, ''

468
especially in the 200 to 600 m range, are propagated in two
ways: along the su·rface of the earth, as ,in day-time, and
by being reflected from the upper, ionised layers of the at-
mosphere. Being supt'rposed on one another, the surface wave
and the wave reflected from the ionosphere interfere and
cause a distortion and deterioration of the signals.
There are three zones of audibility in the night-time. In
the first zone, which is in the vicinity of the transmitter,
the field intensity of the reflected wave is much smaller
than the field intensity of the surface wave. In the second
zone, which is at a sop1ewhat greater distance (80 to 160 kn1),
the signal of the surface wave gradually weakens and the
signal of the reflected wave grows more powerful, the two
signals having the san1e value at a certain distance. In
the third zone, which is stilt further out, the signa) of the
reflected wave grows st i 11 n1ore pOwPrf u I and rcn1a ins
constant up to rather large distances; as for the signal of
the surface,wave, it is so weak that the intf•rference between
these two signals is practically non·existent.
To rnove away the second zone, known as the zone of n<'ar
fading, the use of so-caJJed anti-fading antennas is nt'cessary.
Consequently the zone of reliable reception can he widened
only through the use of antennas of an appropriate design.
The directional diagram of the antenna should be such
that the maximum energy be radiated along the surface of
the earth and no radiation occur at high angle~ (45 to fi5°)
to the horizon. It is not found possible, in practice, to get
rid of the reflected wave, but one can appreciably weaken
it and amplify the surface wave.
ThP maximum•
value of the field of the surface wave is
obtained when the antenna has a height 1=0.64A (kl =230°).
However, at that height, there occurs a considerable radi-
ation at high angles to the horizon and, as a result, the
reflected wave has a considerable field at angles of 45 to 55°.
An antenna of length l == 0.53 A. (kl == 190°) gives a con·
siderably smaller field intensity at these angles, although,
in that case, the field intensity "f the ground (surface) wave
decreases by approximately 10%. A

Thus, the best anti-fading antenna has a. height l==0.53A.

The desirable directional diagram can also be obtained in
the case of antennas with a load at the top. In that case.
the optimum results are obtained when the equivalent
length of the antenna equa·ls approximately 190°. This is
469
 particularly valuable in the case of anti-fading antennas
operating on Jong waves. Moreaver, the desirable directional
diagrams can be obtained with sectionalise~ antennas, i.e.,
antennas into which inductance coils are inserted (~t
points higher than the middle). The equivalent length of
such antennas should also be approximately 190°.
It should be pointed out that antennas of a high wave
imp,edance have satisfactory directional diagrams because,
in that case, the current in the node is cTose to zero and the
directional diagram c~t the desirable angles to the horizon
(45 to 55°) has deep minimums.
The current distribution in anti-fading antennas has the
aspect shown in Fig. 12-12. In all the three above-examined
cases (antennas without a Joad at the top, with a load at the
top, and sectionalised antennas) the current node occurs at
the lower end of the antenna. Such a current distribution
corresponds to a srnall radiation of the antenna in a 45 to
55° direction of the angles and, in the second zone, raises the
ratio of the strength of the signal of the surface wave to that
of the signal of the reflected wave .
.Fig. 12-13 shows the curves of the field intensity of the
surface and reflected waves depending on the distance on a
wave A = 300 n1, set up by a short mast antenna and a mast
antenna of length l =-= 0.53 '"· It can be seen that the use
of an anti-fading antenna moves away the zone of 100~~
fading by approxin1ately two times. ·
Because they cannot be installed at a distance smaiJer
than the height of the mast from the radio station building,
mast antennas are fed by tneans of a coaxial cable. I~oreover,
an antenna hut is erected in the vicinity of the mast; it houses
the circuits for tuning the antenna i-n resonance and coupling
it with the cable. The coupling of the antenna with the
cable is chosen in such a way that a travelling wave should
be set up in the latter. •
The tuning circuits of mast antennas and their coupling
with the cable are shown in Fig. 12-14. It represents the
circuits with an inductive. a conductive and a capacitive
coupling. l...et us investigate in detail the circuit with a
capacitive coupling. The antenna should be tuned in reso-
nance so that the condition
Xtn+Xt+Xc:=O
should be observed.
410
 I
I l I

'J>JI»»»J.>.'inJ>>»». ''lJ.>))jjffHJ»J»m
(a) (/J} (C)
Fig. 12-12. Current distribution in anti~fading
antennas:
a-kl= 190°; /J-k(l+b)=: 19oc•; c-klc= 190°

WOr---~--~----~--~----~--------
80rr--·~--~----+----+----~--~--~
60
40
~ 20
1
~
m~-~~---+--~~--4----+-·---~---~
ao~---- '+---+---+--·-+---~-__,~---~~---~
l
~
6.0
4.0 1------11
+----+H_egion of tOO foding+-----+-----r
e Hefleclecl wove
·S: 2.0
~tO

l
·~

·~
Q8
Q8
Q4 1 - - - ;.._-t---___;:
Q9.__
~ Q2~~~~~~
u:
o.9JJ~~~~
a06~~~~~
aDf~~~~~~~~~~~~~~~~~-~~~

80 100 240 320 400 480

IJI.sloflli', km
Fig. 12·13. Intensity of the tieid set up by an anti-fading
antenna. The power of the transmitter is 50 kW, the
conductivity of the soil 0'=5.10-• ohms/m. directional
antenna l=0.53). (the mast is earthed): non . directional
antenna: short earthed mast.
 Under this condition, the resistance inserted into the
ircuit of the feeder equals:
x•
Rtns == Rcln .
The feeder is then loadt'd on that resistance plus the re-
sistance of the coupling capacitor. To back off the latter, an
inductive resistance is
1.1 Ct in~ted in para lie!_
into the feeder; it is· •

defined from the con-

,.,..,~~"'""""""~.,..,..,~,...,.,.,..,~......,.,..,.....-....JIIIIIIP d it ion
(0) Xc=O. Xcomp +
To obtain a travelling
A wave in the feeder,
"~~~ ;;;; ~;;; ;t;~ I~~L E condit
one hasionto satisfy the
~

Rlns= W f•
The resistance of the
coupling capacitor is
...
defined fron1 the above-
.... ~~~~~~~=.,.,.,.1:~.,.,.,.-J·l£ cited expressions
Xc=V W,Rc (12-22)
(C)
Fig. 12-14. Tuning cfrcufts of mast an- as well 85 the tuning
tennas: resistance of the an-
a-circui t with an Inductive coupling : h-cJrcul t tenna Xt and the COrn-
with a conduct lve coupiJng: c-ctrcul1 with pensat ion resistance
a capaclt IVt' coupJiuK X
romp•
In the case of a transmitter of low power, the feeder is
not tuned on a travelling wave. In that·case, the antenna
is connected to the feeder directly, without transition
devices and the tuning of the antenna-feeder system is
effected in the buildin g of the transmitter. ~-.,.\

12-5. Slot Antennas on Low Supports

Mast antennas and tower antennas, which are widely used
in the medium-wave range, have a height of the order of
100 to 250 m. However, the use of high mas~s and towers Is
412
undes_irable for a number of reasons. The m3in ones are:
the h1gh cost of the antennas and the considerable technical
difficu lties attending their design (for example. the
problems concerning the insulation of the base and guys of
the mast are quite complicated).
The replacement of high mast antennas (of the order of
0.16-+-0.2A., not lower) by low ones, results in a lowering of
the antenna efficiency, due to an increase of the losses into
14- ---- -2/J .

.,._-2a-~

'

K,
,
•

Fig. 12-15. Slot antenna on low support s.

the earth caused by high current densities occurring ;n the

vicinity of the antenna base.
In this connection, a medium ..wave antenna on low
supports of adequate efficiency has been devised in the
Soviet Union (74 J.
The principle of the antenna is similar to that of the
cylindrical dipole excited by a transverse slot used on
centimetre and decimetre waves. However, the dipole has
been considerably modified in order to make it suitable for
operation on medium waves.
The antenna consists of a cylindrical wire net fed by a
ring wire feeder (distributive feeder) (Fig. 12-15). Vertical
wires forming a cylinder are connected to one of the wires
of the ring feeder, the other wire being earthed. If we sup·
plement this device by Its mirror image, the distributive
feeder becomes a ring slot exciting a cylindrical dipole.
413
 As in the case of a usual wire antenna, the vertical part
of the dipole is supplemented by a horizontal part con-
sisting of a number of radial wires of radius b. This widens
the pass-band of the antenna and, by reducing the current
density in the earth in the vicinity of the dipole, reduces
the losses.
The feeder from the transmitter is connected to the points
cd of the distributive (ring) feeder, which are chosen so as
to ensure matching of the antenna to tlie feeder. To tune
the antenna in resonance, i.e., in order that the reactive
component of the input resistance should be zero at the point
cd, the short-circuited stubs 1 and 2 are used, the length of
which is regulated by means of the short circuiters Kt and K2 •
Such a construction enables to decrease the density of the
currents arising in the earth close to the antenna, because,
instead of gathering at one point at the base of the antenna
as in the case of ordinary antennas, the currents flow over
the whole periphery of the slot. Owing to the decrease of
the current density in the earth, the antenna efficiency for
quite a smaiJ height of the vertical part and simple earthing
( 120 radial wires) is of the same order as that of high mast
antennas. The efficiency is further improved through the
use of tuning elen1ents, of segments of long lines instead
of the lutnped inductances and capacitances generally used.
The antenna, which is destined for operation on 200 to
600 m, has the following geometrical dimensions:
height of the vertical part h=22 m;
radius of the cylinder a=24 m;
radius of the horizontal part b=67 .5 m;
length of the tuning stubs l =30 m.

12-6. Wave Antennas •

The wave antenna designed by Beveridge is utilised for
the a·eception of signals from long-wave radio stations. l t
consists of a wire half a wave to several wave-lengths long,
directed towards the transmitter and suspended at a height
of up to 8 m above the earth. The end of the wire towards the
transmitter is earthed through a resistance equal to the
wave impedance of the antenna. The otht'r end of the wire is
connected to the receiver, the \Vave impedanc~ of which is
also equal to the wave impedance of the antenna tl·tg. 12-16).
4i4
 The antenna responds to the horizontal component of the
intensity of the electric field caused by t~~ finite conduc-
tivity of the soil. As shown in Fig. 12-16," the front of the
wave propagated along the surface of the parth is inclined
towards the direction of motion of the wave. When the wave
comes from the direction of the absorbing resistance to the
receiver, the horizontal component of the electric field
intensity vector is directed along the axis of the \Vi re. When
the wave comes from other directions, it is the projection
of the horizontal con1ponent of thr electric field intensity
vector that acts along the axis of tlu~ wire. The projection
is equal to the product of the horizontal con1ponent of the
\<'2 1

R=W
I

Fig. 12-16. Wave antenna:

/-antenna: 2-wave front.

electric field intensity vector by the cosine of the angJe

between the antenna and the direction towards the
transmitter.
As in the case of short-wavP travelling-wave antennas, the
current in the receiver is at its n1aximu1n when the direction
of motion of the wave coincides with the direction
from the absorbing resistance to the receiver along the
wire axis. If the phase velocity of propagation of the current
wave along the wire were equal to the velocity of the
incident electromagnetic wave, the current would be all the
higher as the wire is longer. However, due to the fact that
the reverse conductor for the current is the earth, the conduc-
tivity and permittivity of which differ frorn those of air,
the phase velocity of propagation of the current wave along
the wire is inferior to the velocity of propagation of the
incident e1ectromagnetic wave. For this reason, the current
maximum in the receiver occurs at a c~rtain definite ratio
of l/1...
When the wave comes from the direction opposed to that
of maximum response, the current maximum occurs in the
absorbing resistance, the current in the receiver itself being
at a minimum or even zero. This is due to the fact that the
415
elementary current waves, excited in the wire as a result
of the action of the elementary emf's induced by the wave
in the various regions of t9! wire,
will cancel instead of reijnforcing
one another.
Thus, the directional diagra1n ef
the antenna has the aspect shown in
. t· 1 Fig. 12-17.
Fig. l 2· 17 · Darec •ana It is clear tfiat the directional
dia~ram of wave antenna.
diagram of a wave antenna is
characterised by the factor
cos 9 . [ kl
F (8)= ~-cos 8 stn 2 l
(;-cos P> ,
where p is the angle between the axis of the wire and the
direction of arrival of the wave;
L, the length of the antenna;
~=~.the ratio of the velocity of propagati on of
V
the electron1agnetic wave to the velocity of propa-
gation of the current wave. .
The Beveridge antenna is made use of as a reception antenna
to eliminate atmqspheric interference on long waves, as
well as interference from stations of neighbouring frequen-
cies. As a rule, the antenna is not utilised for transmission
on account of its low efficiency.
To ex plain the existence of important losses in the antenna
as well as to illustrate the directivity of the antenna from

----~ -· .,_.., ~ ~, / I
, ,__ _._L7
2 •
Fig. 12-18. Wave antenna used for transmission:
1-earthlng; 2-currents In the earth

the point of view of transmission, let us examine the circuit

of Fig. 12-18, similar to Fig. 12-16 but with a transmitter
instead of a receiver. Due to the fact that the wire is loaded
on a resistance equa I to its wave impedance, there arises in
it a current wave travelling towards the load. The displace-
ment currents branching off from each element of the wire
are transformed in the earth into conduction currents and
,
416
polarisation currents. Due to the finite conductivity of the
earth, the currents penetrate down to a certain depth and have
vertical as well as horizontal components: The horizontal
component of the current in the earth, which is equal to the
current in the antenna, gives no radiation of electromagnetic
energy, because its radiation is cancelled by the radiation
of the current antenna. The radiation component is the verti-
cal component of the current in the earth. In the direction
along the antenna axis, this current has approximately the
same density, but its phase changes from point to point as
e-lkx, where£= v1 , kx is the distance in degrees from the
V
transmitter to the point under consideration on the wire.
It is clear that the maximum of radiation'of this current is in
.- • •

•
circuit

..
Fig. 12-19. Circuit of wave anteuna with varaable
direction of zero reception.

the direction of the absorbing resistance along the antenna

axis and the directional diagram has the shape aJr~ady shown
in Fig. 12-17.
As a result of the radiation as well as losses in the earth,
the current wave is attenuated in the course of its travel to
the far end of the antenna. Elementary calculations would
show that the' radiation resistance of the antenna forms an
insignificant quantity of the overall resistance of the losses
in the antenna, so that the efficiency of the antenna is low.
The antenna is therefore not used for transmission.
In practice, it is found necessary to regulate the direction
of the zero of reception, so that use is often made of the wave
antenna circuit shown in Fig. 12-19. The antenna represents
a horizontal twin line. The electromagnetic waves arriving
fto1n the main direction induce in both wires electric
currents (dotted arrows) of equal magnitude, which escape
into the earth through the primary of the transformer at
the far end. Furthermore, these currents induce in the secon-
dary of the transformer an emf under ·the influence of which,
471
antiphase currents (solid lines) arise in the wires. T~e emf
induced by these currents in the winding 1 is fed to the
phasing circuit. The wave of the current (co-phasal) moving
towards the receiver is led into the earth and induces in
the winding 2 an emf which is likewise fed to the phasing
circuit. This emf has a phase shift reiatively to the emf in
the winding 1 and differs from it in amplitude.
Thus, by regulating the ratio of these two emf's in the
phasing circuit, one can obtain a zero 01 the signal at the
input of the receiver for a wave arriving from any unde-
sirable direction and, in so doing, get rid of the interference.

12-7. Frame Antennas

A frame antenna in its simplest form is shown in Fig. 12-20.
The frame constitutes a directive system with two zero
reception directions. The dimensions of the frame relatively
d

d<<).
tE 11 h<<A
, -
' 11 ' f
liJ receiver .......-tf._.._..,..
Fig. 12-20. Frame Fig. J2.. 21. Calculating
antenna. the directional diagram
of a frame antenna.

to the wave-length are usual1y small and it can be regarded

as a magnetic dipole. _
Let us consider the frame as a vertical reception system.
Let the vector of the electric field intensity of the wave be
parallel to the sides h of the frame and let the direction of
the electromagnetic wave form with the plane of the frame
an angle cp (Fig. 12-21). Let E be the electric field intensity
in the centre of the frame. Then, an emf
kd
1 - coscp
G. • .• Eh e 2
I

will be induced in the vertical side of the frame 1, and

an emf
ktJ
IJ.=£he -1.-cOsfP

478
will be induced in the vertical side of the frame 2. The
emf's act in opposite directions, so that. the total emf
giving rise to the current in the frame and determining the
difference of potential at the capacitor (and input of the
receiver), equals:
G1=1l, ~ll.=i2Eh sin (~cos cp). (12-23)
If, as is usua 11 y t ht' case, the d i n1ensi ons of the frame
are small in comparison with the \vave-length. we can
approximately write:
cC 1=iEh kd cos~· (12-24)
The expression (12-24) shows 1) that the directional dia-
granl of the frame is the same as that of a dipole ("figure
of eight") and 2) that the emf induced in the frame has a
phase shift of 90° relatively to the intensity of the field
of the wave.
The effective height of the frame is understood to mean
the ratio of the induced emf to the field intensity when
the energy is received from a direction lying in the plant'
of the frame (cp=O),
hcrt=hkd.
If the trame consists of n windings connected in series, the
emf induced in the frame increases by n times. so that the
effective height of such a frame is:
2nns
herf = A. , (12-25)

where s = hd is the area of the frame.

The radiat'ion resistance of the frame, which is small
in comparison with the wave-length and distant from the
earth, can be calculated from the expression R1: =80:t x
1

X ( ~! r, which, taking account of (12-25), leads to the

expression
R:tr=3t,2oo(r.r. o2-2s>
"
which, \Vhen n =1, coincides with the expression (1-19). To
take an example, let us suppose that A = 1,000 m, n = 10,
s=4 m2 • Rr.1 will then be found to equal 5.10- 1 ohms.
Thus, the radiation resistance of a frame antenna on long
waves is very small and it is difficult to cause the resistance
479
of the losses of the frame to be equally small. Hence, the
efficiency of a frame antenna is very low and it is not
utilised for transmission on long waves in the version
described above. being only used for reception.
Note that the obtention of a directional diagram of the
frame antenna in the shape of a pure "figure of eight', de-
scribed by the ex press ion ( 12-24) requires a strict electrical
synunetry of the frame systen1. This QJeans, in particular.
that the input of the recPivPr should also be symmetrical,
otherwise the frame will receive also a single-wire system:
? the directional diagram of a vertical
dipole (circumference) will be superim-
posed on the directional diagram of the
1 frame and the resulting directional
diagram of the frame antenna wi 11 be
distorted. This undesirable effect of
2 the frame is referred to as the
antenna effect.
Frame antennas are made use of in
measuring apparatus (comparators),
ro ~ceiver as reception antennas of long-wave
stations with the object of eliminating
Fig. 12·22. Screened interference stations and directive
frame:
1-lrame; 2-screen; 3- noise, in radiogoniometrical instal-
gap lations determining the direction
of the radio station received, etc.
The frame is generally mounted on a wooden or other non-
metal framework and placed in a vertical position so as to
be able to rotate it around the vertical axis. Frames of
this kind are often enclosed within a circular screen with
a gap at the top (Fig. 12 22). l\t the same time, the recefver
is enclosed in the same screen as the frame, so that an
antenna of this kind is. in essence, sym!lletrical. with good
receiving qualities. Indeed, an electromagnetic wave coming
towards the fran1e axis induces db the screen currents. which
are so distributed that no difference of potential is set up a~
the gap of the screen and the inner region of the screen, i.e .•
the ·franle is not excited. When the electromagnetic wave
comes from other directions, a difference of potential is set
up at the gap of the screen and the frame is excited. The
antenna effect of the frame is thus eliminated and the
directional diagram suffers no distortion (there are directions
of zero reception).
480
 12-8. Radiogoniometers
A radiogoniometer represents two stationary external
frames (Fig. 12-23) situated at an angle of 90° to one another.
The frames are usually triangular and are st~cured on one
mast, approximately 70 nt high. ThP frames are connrctrd to
the receiver by means of
two twin lines. Each line is
connected to one of the two f
coi Is of the goniometer (to
the stator). These coi Is are
disposed at an angle of 90° J
to one another and coupled 4
with a third rnobiJe coil
ea lied the searcher. Frum
the searcher, the vo Itage
induced by the stationary J
fran1es is fed ,to the input Fig. 12-23. Radiogonionu.-.ter:
of the receiver • /,2-fraOJc'i; 3-fN:der; .J-gnnlnrnf!te r:
5- 5Ca re her
Let the direction of I
arrival of the wave which is /
2 ][
being received form with
the frame 1 an angle r.p _ ......._..ffJ_
(Fig. 12-24) and let the f
current induced in the
stator I equal I 0 'Nhen the Fig. 12-24. Explaining the emf
angle cp=O. Then, when the induced in the radiogoniorneter.
wave arrives from any other
direction, the current induced in the stator I wilJ equal:

and in the stator If,

/
1
=/ 0 sincp.
Let the mutual induction between the stator and the rotor
equal M when their coupling is at its maxirnum. Then, if
the angle between them equals a, the emf induced in the
searcher by the stator I wi 11 be: •
<CI = I owM cos «p cos a.
and the emf induced by the second stator in the searcher
will be:

31-2122
481
The total emf will equal:
cfJ= I 0 IDM cos (cp-cx). (12-27)
It is seen from ( 12-27) that when a is constant and cp is
variable, the maximum of reception will occur for q>=a.
In this way, the directional diagram of an antenna of this
kind does not differ from that of an ordinary frame antenna but
the maximum reception direction is detertuined by the position
of the searcher. The rotation of the coH of the searcher
at the output of the receiver gives rise to the same effect
as in the case of a rotating frame antenna.
If to the fran1e system described above, we add a vertical
antenna, we can obtain a device which gives a unidirectional
(cardioid) reception (Fig. 12-25).

c,
TiJ receiver
Fig. 12-25. Circuit of cardioid reception:
I, ~-frames~ 3-vcrtlcal antenna; 4-adding circuit; 5-
gonlometcr.

From the antenna and goniometer, the voltage is fed to

the adding circuit. By regulating the ratios of the emf's
introduced into the adding circuit by the antenna and the
gonion1eter, one can obtain a total-emf, which will change as:•

cC= I 0roM [1 +cos (<p -a)]. ( 12-28)

•
The phasing of the emf is effected by the capacitor Cs and the
control of the amplitude, by changing the coupling between
the coils Ls and L 2 • The capacitor C2 serves for tuning the
adding circuit in resonance.
The directional diagram of the antenna system, expressed
by (12-28) has the shape of a cardioid. A diagram of this
kind presents considerable advantages over the directional
diagran1 of a frame. It enables to define a unique direction
towards the radio station which is being received, a~ well as
to eliminate interference coming front a direction lying within
,

482
an angle of 90 to 270°. without any substantial weakening
of the signa] received. ,
Note that it is impossible to obtain a totally zero signal
either in a frame system (goniometer) or in a system ~iving
a cardioid reception. The ratio of the n1aximurn to the mini-
mum reception, kno\vn as the anti-interference factor, does
not, as a rule, exceed a few tens. This is due to the fact that,
even in the event of full symn1etry of the systen1, it can only
be tuned on one frequency, say the carrier frequf•ncy. On the
side frequenctes of the signal received, the system is found to
be somewhat out of phase, so that no zero of rec<.~ption can
be obtained. ·
Furthermore, in the medium wave range, at night, apart
from the ground ray, a sky ray also reaches the point of
reception. It is not found po-;stble to cancel the act ion of these
t\vo rays in the frequency band and the anti-interference
factor of the antenna is thereby lo\\'cred.
 Appendix I
Table of Values R, 2

d
0.0 0.5 l.O 1.5 2.0 2.5 3.0 3.5
h

0.0 +73 29 -12. 36 + 4 . 08 -1.77 -+I 18 -0 75 +0.42 -0.33

0.5 +26 40 -11.80 -t-8. 83 -5.75 +a. 76 -2.79 +1.86 -1.54
1.0 -4 065 -0.78 +:3. 56 -6.26 -1-6.05 -5 67 +4.51 -3.94
1.5 -t-1. 78 +0.80 -2.92 +1.96 +0.16 -2.40 +3.24 -3~76
2.0 -0.96 -1.00 -t-1. 13 +0.56 -2.55 +2.74 -2.07 +·0. 74
2.5 +O 58 1-0.45 -0.42 -0.96 ·+
l. 59 -0.28 -1.59 -t-2 66
3.0 -0.43 -0.30 +0.13 -t-0. 85 -0 45 -0.10 +I. 74 -1.03

-- - --
dl I
I
5.0 5.5 6.0 6.5 7.0 7.5
4.0 -1.5
h I I
I
•

0.0 -t-0. 21 ·-0. 18 -t-0. 15 -0.12 +0.12 -0.10 +0.06 -0.03

0.5 +1.08 -0.85 -t-0. 69 -0.57 +0.51 -0.45 +0.36 -0.30
1. 0 +3.08 -2.50 -~-2. l 0 -1 80 +1.56 -1.18 +1.14 -1.00
1. 5 -+-3. 68 -3.40 +3.14 -2.90 +2.61 -2.31 +2.06 -1.86
2.0 -i-0.51 -1.30 +1 82 -2 24 +2.28 -2.29 +2.26 -2.14
2.5 -2.49 +2.00 -1.35 +0.49 -0.06 -0.45 +0.85 -1.03
3.0 -0.09 +1.12 -1.87 +1. 77 -2.02 +I.
71 -1.32 +0.66

484
 App endi x 11

Table of Resistances
h=O
---··---
II ~-X--,-~
rl Rn A x,2 (/ Ru
I
'\ ~
•
12 I w

' --- ' - '

'-·-
0.00 +73. J +42.5 0. 10
I
+6i. 3 I
I

I
-t- 7. 5 '
0.2 1.4 . 2. 4 I 6. 1
0.02 -f-72.9 -f- 35. I 0. J 2 -+·64. 91 -t-1.4
0 6 7.3 2.9 5.8
0.04 +72.3 +27.8 0. 14 +62.0 -4.4

0.06 +71 .0
I .3
+20.8
7.0
0. J6 -f-58. 8 I
3.2 I-9.8 5.4

I .6 6.8 I 3.6 4.9

0.0~ ~-69.4 + 14 . .0 0. 18 +55.2 -14.7
2.1 6.5 3.8 I 4.5
3.8 4.5 I. 0 1 .8
0.20 +51.4 -19.2 I. 12 +13.4 -f-9.5
4.0 4.0 0.8 I. 9
0.22 +47 .4 -23.2 I • J4 +14.2 +7.6
4.3 3.6 '
0.6 1. 8
0.24 +43.1 -26.8 1.16 +14 .18 +5.8
4.6 3.0 0.4 I. 9
0.26 +38.5 -29.8 1.18 +15.2 +3.9
4.5 2.6 • 0.0 2.0
0.28 +34.0 -32.4 1. 20 +15.2 +1.9
4.7 2.0 0.0 1. 8
0.30 +29.3 -34.4 1.22 +15.2 +O.l
4.7 1. 6 0.3 1. 9
0.32 +24.6 -36.0 1. 24 +14.9 -1.8
485
 Continued

d R,a 4
I X 11 A d 4 X 12 4
l Rta

4.6 I .I 0.3 1. 7
0.34 -J- 20 0 -37.1 1.26 + 14. 3. -3.5
4.8 0.6 0.8 1. 6
0.36 +15.2 -37.7 1.28 +13.5 -5.1
4.6 0.1 0.9 I. 6
0.38 ··f-10.6 -37.8 1 .30 -t 12.6 -6.7
4.4 0.3 1. 1 1.4
0.40 -f-6.2 -37.5 1.32 -f-11. 5 -8.1
4.2 0.8 I. 2 l. 2
0.42 -~- 2. 0 -36.7 1. 34 +10.3 -9.3
4.0 1. 1 1. 4 1. 1
0.44 -2.0 -35.6 1.36 +8.9 -10.4
3.8 1. 7 I. 4 0.8
0.46 -5.8 -33.9 1. 38 -t7 .5 -11.2
3.6 1. 8 1. 5 0.7
0.48 -9.4 -32.1
~
1. 40 +6.0 -11.9
3. I 2.2 I. 6 0.5
0.50 -12.5 -29.9 1.42 +4.4 -12.4
2.9 2.4 1. 6 0.2
0.52 -15.4 -27.5 1. 44 +2.8 -12.6
2.5 2.6 1. 6 0.1
0.54 -17.9 -24.9 1. 46 +1.2 -12 7
2.2 ' 2. 9 1. 6 0. l
0.56 -20.1 -22.0 I .48 -0.4 -12.6
1. 8 3.0 1. 4 0.3
0.58 -21.9 -19.0 1.50 -1.8 -12.3
1. 4 3. 1 I I .6 0.5
0.60 -2~i.3 -15.9 1. 52 -3.4 -11.8
1. I 3 2 I .4 0.6
0.62 -24.4 -12.7 1. 54 -4.8 -11.2
0.6 3.2 1. 2 0.8
0.64 -25.0 -9.5 1. 56 -6.0 -10.4
0.3 3. 1 1. 1 0.9
0.66 -25.3 -6.4 1. 58 -7.1 -9.5
0.0 3.1 1.0 I .1
0,68 -25.3 -3.3 1. 60 -8.1 -8.4
0.4 3. 1 0.9 1.2
486
 Continued
I
t! R,, A Xn 6 d R,2 A X I'J 6
1
0.70 -24.9 -0.2 1. 62 -9.0 -7.2
0.7 2.8 0.8 I I .3
0.72 -24.2 +2.6 1.64 -9.8 -5.9
1. 1 2.7 0.5 ] .2
0.74 -23.1 +5.3 1.66 -10.3 -4.7
1. 3 2.6 0.4 I. 4
0.76 -21.8 +7.9 I .68 -10.7 -3.~l
1.5 2.3 0.2 I .3
0.78 -20.3 +10.2 1. 70 -10.9 -2.0
1. 8 2.0. 0.0 1. 4
0.80 -18.5 +12.2 1. 72 -10.9 -0.6
1. 9 1. 8 0.2 I .4
0.82 -16.6 • +14.0 1. 74 -10.7 +0.8
2. 1 1. 6 0.2 1. 2
0.84 -14.5 +15.6 1.76 -10.5 +2.0
2.3 1. 3 0.5 I. 3
0.86 -12.2 +16 9 1. 78 -10 0 +3.3
2.4 1. 0 0.6 1. I
0.88 -9.8 +17.9 1.80 -9.4 +4.4
2.3 0.6 0.7 1.I
0.90 -7.5 +18.5 1.82 -8.7 +5.5
2.4 0.5 0.8 1.0
0.92 -5.1 +19.0 1.84 -7.9 +6.5
2.4 0 .I 0.9 0.9
0.94 -2.7 +19.1 1.86 -7.0 +7.4
I

2.2 0.2 1. 1 0.6

0.96 -0.5 -1-18.9 1.88 -5.9 -f-8.0
2.3 0.4 1. 1 0.7
0.98 +1.8 +18.5 1.90 -4.8 +8.7
2.2 • 0.8 1.2 0.4
1.00 +4.0 +17. 7 1. 92 -3.6 +9.1
2.0 0.9 1. 2 0.3
1. 02 +6.0 +16.8 1. 94 -2.5 +9.4
1.8 1• 1 1.2 0.1
1.04 +7.8 +15. 7 1.96 -1.3 +9.5
I. 7 1.2 I .2 0.0
1.06 +9.5 +14.5 1. 98 -0.1 +9.5
1.5 1.6 1.2 0.1
487
 Continued

d R,, t\ Xu L\ d Rn L\ X 12 li

1. 08 + 11 .0 +12.9 2.00 +1.1 +9.4

I. 4 1. 6 0.7 0.5
1. 10 -J-12.4 + 11.3 .. ~

1.0 I. 8
2.00 + 1.1 +9.4 2.90 -3.4 +5.6 •

1. I 0.3 0.7 0.3

2 02 1-2.2 -f-9. I 2.92 -2.7 +5.9
1.1 0.5 0.8 0.3
2.04 +3.3 +R.6 2.94 -1.9 +6.2
1. 0 0.6 0.8 0.1
2 06 -1-4.3 +8.0 2. 96 -1 I +6.3
0.9 0 6 '
0.8 0.1
2.08 +52 +7.4 2.98 -03 • +6.4
0.9 0.7 0.8 0.1
2. 10 +6.1 +6.7 3.00 -1-0.5 +6.3
0 7 0.9 0.8 0.1
2. J2 -1-6 8 +5.8 3.02 + 1.3 +6.2
0.6 - I .0 0.7 0.3
2.14 +7.4 +4.8 3.04 +2.0 +5.9
0.5 0.9 0.7 0.3
2.16 +1 .9 +3.9 3.06 +2.7 +56
0.3 1. 0 0.7 0.4
2.18 +8.2 +2.9 3.08 +3.4 +5.2
0.2 I .I 0.6 0.5
2.20 +8.4 +J .8 3.10 +4.0 +4. 7
0. I 1 I - 0.5 0 6
2.22 +8.5 +O. 7 3. 12 +4.5 +4 .I
•
0. J I. 0 0.4 0.6
2.24 +8.4 -0.3 3. 14 +4.9 • +3.5
o. 1 I 1. I 0 4 0.7
2.26 +8.3 -1.4 3. 16 +5.3 +2.8'
0.3 0.9 0.3 0.6
2.28 +B.O -2.3 3. 18 +5.6 +2 2
0.4 I. 0 0.1 0.8
2.30 +7.6 -3.3 3 20 +5.7 -f-1 . 4
0.6 0.9 0.2 0.7
2.32 +i.O -4.2 3.22 +5.9 +0. 71
0.7 0.8 0.0 0.8
 Cont1nu td

~ I
d Ru 6 x., A d R ., 1\ X I'J A

-
2.34 +6.3 -5.0 3.24 +5.9 -0. J
0.7 0.7 0. I 0.7
2.36 +5.6 -5.7 3.26 +5.8 -0.8
0.7 0.6 0.2 0 7
2.38 +4.9 -6.3 3.28 -t 5. 6 -1.5
0 9 0.5 0.3 0 6
2 40 +4.0 -6.8 3.30 +5.3 -2.1
0.9 0.4 0.3 0.7
2.42 +3.1 -7.2 3.32 +5.Cl -2.8
0.9 0.2 0.5 0.6
2.44 +2.21 -7.4 3.34 +4.5 -3.4
1.0 0.2 0 4 0.5
2.46 +1.2 -7,6 3.36 +4.1 •
-3 9

I.0 0.0 0.5 0.5

2. 48 +0.2 -7.6 3.38 +3 6 -4.4

0.9 0. I 0.7 0.3

2 50 -0.7 -7.5 3.40 +2.9 -4.7

0.9 0.2 0.6 0.3

2.52 --1.6 -7.3 3.42 +2.3 -5.0
0.9 0.2 0.6 0.3
2.54 -2.5 -7.1 3.44 +I. 7 -5.3
'
0 8 0.5 0.7 0 .I
2.56 -3.3 -6.6 3.46 +1.0 -5.4
0.8 0.5 0.7 0. I
2.58 -4.1 -6.1 3.48 +0.3 -5.5
0.7 0.6 0.7 UI
2.60 -4.8 -55 3.50 -0.4 -S.4
..
0.6 0.7 0.7 u. 1
2.62 -5.4 -4.8 3.52 -1.1 -5.3
•
0.5 0.7 0.6 0.2
2.64 -5.9 -4.1 3.54 -1.7 -5.1

489
 Continued

d R., A Xu A d R12 A X ·2 &

0.4 0.8 0.6 0.3

2.66 -6.3 -3.3 3.56 -2.3 -4.8

2.68 -6.7
0.4
-2.5
0.8
3.58 -2.9
- 0.6
-4.5
0.3
•
0.2 0.9 0.5 0.4
2.70 -6.9 -1.6 3.60 -3.4 -4.1
o.o 0.9 0.4 0.5
2.72 -6.9 -0.7 3.62 -3.8 -3.6
0.0 0.8 0.4 0.5
2.74 -6.9 +O.l 3.64 -4.2 -3.1
0.1 0.9 0.4 0.6
2.76 -6.5 +1.0 3.66 -4.6 -2.5
0.3 0.9 0.2 0.6
2.78 -6.5 +1.9 3.68 -4.8 -1.9
0.2 ... 0.7 0.2 0.6
2.80 -6.3 +2.6 3. 70 -5.0 -1.3
0.5 0.7 0.1 0.6
2.82 -5.8 +3.3 3.72 -5.1 -0.7
0.5 0.7 0.0 0.7
2.84 -5.3 +4.0 3.74 -5.1 0.0
0.5 0.6 0.1 0.6
2.86 -4.8 +4.ti 3.76 -5.0 +0.6
0.7 0.5 0.1 0.6
2.88 -4.1 +5.1 3.78 -4.9 +1.2
0.7 0.5 0.2 0.6

0.2 0.6 0.5 0.3

3.80 -4.7 +1.8 3.92 -2.1 +4.4
0.3 0.5 0.6 0.2
3.82 -4.4 +2.3 3.94 -1.5 +4.6

490
 Contint~ed

0.3 0.6 0.5 0 .I

3.84 -4.1 +2.9 3.96 -1.0 +4.7
0.5 0.4 0.6 0.1
3.86 -3.6 +3.3 3.98 -0.4 +4.R
•
0.4 0.5 0.6 0.1
3.88 -3.2 +3.8 4.00 +0.2 +4.7
0.6 0.3
3.90 -2.6 +4.1
0.5

h=0.5
d I__R_
_,. 12
.L
J ~ I X·~ , A I d ' R 12 I A I X 12 I A
I. 7
0.00 +26 .4 +20.2 0.60 -14.1 +0.4
0. 1 3.8 . 0. 2 I. 7
0.02 +26 .3 +16.4 0.62 -13.9 +2.1
0.4 3.5 0.4 1.6
0.04 +25 .9 +12.9 0.64 -13.5 -t-3. 7
0.5 3.5 0.5 1. 5
0.06 +25 .4 +9.4 0.66 -13.0 -+5.2
0.8 3.2 0.8 1. 5
0.08 +24 .6 +6.2 0.68 -12.2 +6.7
1. 1 3.1 0.9 1.4
0.10 +23 .5 +3.1 0.70 -11.3 +8.1
' 1. 2 2.9 ] .0 l .2
0.12 ~j-22 .3 +0.2 0.72 -10.3 +9.3
1.4 2.6 1. 2 1. 1
0.14 +20 .9 -2.4 0.74 -9.1 +10.4
1. 6 2.4 I. 4 0.9
0.16 +19 .3 -4.8 0.76 -7.7 +11.3
1. 7 2.2 I. 3 0.7
0.18 +17 .6 -7.0 0.78 -64
,.. +12.0
1. 9 I. 9 I. 5 0.6
0.20 +15 .7 -8.9 0.80 -4.9 +12.6
I. 9 I. 7 I. 6 0.4
0.22 +13 .8 -10.6 0.82 -3.3 +13.0
2. 1 1. 3 1. 4 0.3
I
491
 Continued

,, R ,., 1\ x., tJ rJ R ,, A Xn &

-
0.24 + 11.7 -11.9 0.84 -1.9 +13.3
2. I 1. J 1.6 0.0
0.26 +9.6 -13.0 0.86 -0.3 +13.3
.....
2.2 0.9 1. 6 0.1
0.28 +7.4 -13.9 0.88 +1.3 +13.2 •

2.2 0.6 1.5 0.4

0.30 +5.2 -14.5 0.90 +2.8 +12.8
2. 1 0.3 1. 4 0.4
0.32 +3.1 -14.8 0.92 +4.2 +12.4
2.0 0.2 1. 4 0.7
0.34 +I. 1 -15.0 0.94 +5.6 +11. 7
2. I 0.3 1.2 0.8
0.36 -1.0 -14.7 0.96 +6.8 +10.9
2.0 0.4 1.2 0.9
0.38 -3.0 -14.3 0.98 +8.0 +10 0
1. 9 0.6 1. 0 1.0
0.40 -4.9 -13 7 1.00 +9.0 +8.9
l. 7 - 0.8 0.9 1.2
0.42 -6.6 -12.9 1.02 +9.9 +7. 7
I. 6 1. ] 0.8 1.2
0.44 -8.2 -11.8 1. 04 +JO. 7 +6.5
1. 4 1. 2 0.6 1.4
0.46 -9.6 -10.6 1.06 +11.3 +5.1
1. 2 I. 3 0.4 I 3
0.4~ -10.8 -9.3 1. 08 +11. 7 +3.8
I. 1 J. 4 - 0 3 I 4
0.50 -11.9 -7.9 I. 10 +12.0 +2.4
0.8 I. 6 0,1 1.5
0.52 -12.7 -6 3 1.12 +12. t +0.9
0.7 1. 6 0.1 1.4
0.54 -13.4 -4.7 I. 14 +12.0 -0.5
C.4 1. 7 0.2 1. 4
0.56 -13.8 -3.0 I. 16 + 11.8 -1.9
0.3 1. 7 0.4 1.3
0.58 -14.1 -1.3 1.18 +11.4 -3.2
0.0 1. 7 0.6 1.3
0.6 1.3 0.~ 0.9
1 20 +10.8 --4.5 2.12 +7.5 +2.9
.

492
 Continu ed

d R,, l!i X,1 ~ d Ru A X I'J ~

0.7 1. 2 0.3 1. 0
.
I. 22 +10.1 -5.7 2. 14 +7.8 +1.9
0.8 I. 2 0.1 0.9
1.74 +9.3 -69 2. 16 -t- 7. 9 -t-1.0
0.9 0.9 o.o 1.0
1 26 +8.4 -7.8 2 18 +7.9 +0.0
1. 0 0~8 0. 1 I. 0
I 28 +7.4 -8.6 2.20 +1 .8 -1.0
1. J 0.7 0.3 0.9
) .30 +6.3 -9.3 •
2.22 "+-7. 5 -1.9
1.2 0.6 0.3 0.8
I. 32 +·5. 1 -9.9 2.24 +7.2 -2.7
J
1.3 0.5 0.4 0.9
I. 34 +3.8 -10.4 2.26 +6.8 -3.6
l. 3 0.3 0.6 0.7
1. 36 +2.5 -10.7 2.28 +6.2 • -4.3
1. 3 0. 1 0.6 0.8
1.38 +1.2 -10.8 2 30 -t-5. 6 -5.1
1. 3 0.0 0.7 0.6
1.40 -0.1 -10.8 2.32 +4.9 --5.7
1. 2 0.2 0.8 0.5
1. 42 -1.3 -10.6 2.34 +4.1 -62
1. 3 0.3 0.8 0.4
1.44 -2.6 -10.3 2.36 +3.3 -6.6
I. I 0.5 0,.8 0.3
•
1.46 -3.7 -9.8 • 2.38 +2.5 -6.9
I.1 0.6 0.9 0.2
1.48 -4.8 -9 2 2.40 +1.6 -71
I. 0 0.7 0.9 0.1
1.50 -5.8 -8.5 2.42 +0. 7 -7.2
1. 0 0.8 0.9 0.0
1.52 -6.8 -7.7 2.44 -02 -1.2
0.8 0.9 , 0.8 0.1
1.54 ..-7.6 -6.8 2.46 -1.0 -7.t
0.7 1.0 0.9 0.3
1.56 -8.3 -5.8 2.48 -1.9 -6.8
0.5 1• I 0.8 0.3
1.58 -8.8 -4.7 2.50 -2.7 -6.5

498
 Continued

d R,., ll x,, 4 d R •., I ll x•., 4

0.4 1. 1
' 0.7 0.4
1.60 -9.2 -3.6 2.52 -3.4 -6.1
0.3 1. 2 0.7 0.5
1.62 -9.5 -2.4 ?.54 -4.r -5.6
0.2 1. 1 0.6 0.6
•

1.64 -9.7 -1.3 2.56 -4.7 -5.0

0.0 1. 2 0.6 0.6
1.66 -9.7 -0.1 2.58 -5.3 -4.4
0.2 1. 1 0.4 0.7
1.68 -9.5 -1-1.0 2.60 -5.7 -3.7
0.2 1. 2 0.4 0.8
1. 70 -9.3 +2.2 2.62 -6.1 -2.9
0.4 1. 0 0.3 0.8
]. 72 -8.9 +3.2 2.64 -6.4 -2.1
0.6 I. 1 0. 1 0.8
). 74 -8.3 +4.3 2.66 -6.5 -1.3
0.6 0.9 0. 1 0.8
1. 76 -7.7 't-5.2 2.68 -6.6 -0.5
0.7 0.8 0.0 0.8
I. 78 -7.0 +6.0 2.70 -6.6 +0.3
0.9 0.7 0.2 0.8
1.80 -6.1 +6.7 2.72 -6.4 +1.1
0.9 0.7 0.2 0.8
1.82 -5.2 +7.4 2.74 -6.2 +i.9
0.9 0.5 0.3 0.7
1.84 -4.3 -J-7. 9 2. 76- -5.9 +2 6
1. 0 0.4 0.4 0.7
1.86 -3.3 +8.3 2.78 -5.5 +3.3
•
1. 1 0.2 0.5 0.6
1.88 -2.2 +8.5 2.80 -5.0 +3.9
1.0 0.2 0.5 0.6
1. 90 -1.2 +8.7 2.82 -4.5 +4.5
I.1 0.0 0.6 0.5
1. 92 -0.1 +8.7 2.84 -3.9 +5.0
1. 0 0 .I 0.7 0.4
1.94 +0.9 +8.6 2.86 -3.2 +5.4
1. 1 0.3 0.7 0.3
1.96 +2.0 +8.3 2.88 -2.6 +5.7
-

494
 Continued
,'
d R l'J ll Xn ll d R •., ll X •., A

1. 0 0.3 0.7 0.2

1. 98 +3.0 +8.0 2.90 -J .8 +5.9
0.9 0.5 0.8 0.2
2.00 +3.9 +7.5 2.92 -1.0 +6.1
0.8 0.6 0.7 0.0
2.02 +4.7 +6.9 2.94 -0.3 -f-6. 1
0.8 0.7 0.8 0.0
2.04 +5.5 +6.2 2.96 +0.5 -t-6.1
0.6 0.7 0.8 0.2
2.06 -~6.1 -t-5. 5 2.98 +1.3 +5.9
0.6 0.8 0.6 0.2
2.08 +6.7 -f-4.7 3.00 +1.9 +5.7
0.5 0.9· 0.7 0.3
2.10 +7.2 0.3 +3.8 0.9 3.02 +2.6 0.6 -t-5. 4 0.4
3.04 0.6 0.4 0.5
-4.5
0.3
+3.2 ~
+5.0 3.54 -2.6
0.6 0.5 0.6 0.4
3.06 +3.8 +4.5 3.56 -3.2 -4.1
0.6 0.6 0.4 0.4
3.08 +4.4 -t-3 9 3.58 -3.6 • -3.7
0.3 0.5 0.4 0.5
3.10 +4.7 +3.4 3.60 -4.0 -3.2
0.4 0.6 0.3 0.6
3.12 +5.1 +2.8 3.62 -4.3 -26
0.3 0.7 0.3 0.5
3.14 +5.4 -J-2 .1 3.64 -4.6 -2.1
0.2 0.7 0.2 0.6
3.16 -t-5. 6 +1.4 3.66 -4.8 -1.5
0.1 0.6 0. 1 0.6
3.18 +5.7 +0.8 3.68 -4.9 -0.9
0. I 0.8 0.0 0.7
3.20 +5.6 I
0.0 3.70 -4.9 -0.2
0.0 0.7 0.0 0.6
3.22 +5.6 -0.7 3.72 -4.9 +0.4
0.1 0.6 0.1 0.6
3.24 +5.5 -1.3 3.74 -4.8 +1.0
0.3 0.7 0.2 0.5
3.26 +5.2 -2.0 3.76 -4.6 +1.6
0.3 0.6 0.2 0.5
3.28 +4.9 -2.6 3.78 -4.4 +2.1
0.4 0.6 0.3 0.5
3.30 +4.5 -3.2 3.80 -4.1 +2.6
0.4 0.5 0.4 0.5
3.'32 +4.1 -3.7 3.82 -3.7 +3.t
0.6 0.4 0.4 0.4
3.34 +3.5 -4.1 3.84 -3.3 +3.5
0.5 0.4 0.5 0.4

495
 Continu ed

d Ru A x.., t\ d R., t\
X'' 4

3.36 +3.0 -4.5 3.86 -2.8 +3.9

0.6 0.3 0.5 0.3
3.38 +2.4
0.6
-4.8
0.3 I
3.88 ...
-2.3
0.6
+4.2
0.2
3.40 +1.8 -5.1 3.90 -1.7 +4.4
0.7 0. I 0.5 '0.2
3.42 +1. I -5.2 3.92 -1.2 +4.6
0.6 0.1 0.6 0.0
3.44 +0.5 -5.3 3.94 -0.6 +4.6
0.7 0.0 0.6 0. I
3.46 -0.2 -5.3 3.96 0.0 +4 7
0.6 0. I 0.6 0 I
3.48 -0.8 -5.2 3.98 +O 6 +4.6
0.7 0.2 0.5 0 .I
3.50 -1.5 -5.0 4.00 +1.1 +4.5
06 0.2
3.52 -2.1 -4.8
0.5 0.3

h=l.O
•

fl Rn ~ X •:a A d R12 {\ Xu '~

I
0.0 0.2
0.00 -4.1 -0.7 0.20 -4.0 +0.5
0.0 0.0 0.1 0.2
0 02 -4 I -0.7 0.22 -3.9 +0.7
0 0 0.0 0.1 0.2
0.04 -4 I -0 7 0.24 -3.8 +0.9
0.0 0 I 0.0 0.3
0.06 -4 J -·0.6 0.26 -3.8 +1.2
0.0 0 I • 0.2 0.3
0.08 -4 I -05 0.28 -3.6 +1~5
o.o 0. I 0.1 0.3
0.10 -4 1 -0.4 0.30 -S.S +1.8
0.0 0.1 0.2 0.2
0 12 -4.1 -0.3 0.32 -3.3 +2.0
0.0 0.2 . 0.2 0.3
0.14 -4.1 -0.1 0.34 -3.1 +2.3
0.0 0.2 0.2 0.3
0.16 -4.1 +0.1 0.36 -2.9 +2.6
0.1 0.2 0.2 0.2
0. IH -4.0 +0.3 0.38 -2.7 +2.8

496
 Continued

d R,2 ~ X 12 ~ d R11 4 X •., 6

0 0 0.2 0.3 0.3

0.3 0.3 0.5 0.6
0.40 -2.4 +3.1 1.32 -5.0 -3.9
0.2 0.2 0.4 0.5
0.42 ~2.2 +3.3 I 34 -54 -34
0.3 0.3 0 4 0.5
0.44 -1.9 +3.6 I. 36 -5.8 -2 9
0.3 0.2 0.3 0.7
0.46 -1.6 +3.8 1.38 -61 -2 2
0.3 0.2 0.2 0.6
•
0.48 -1.3 +4.0 1.40 -6.3 -1.6
0.5 0. 1 0. I 0.7
0.50 -Q • 8 t I +4.1 1. 42 -6.4 -09
0.5 0. 1 0. I 0.7
0.52 -0.3 +4.2 I .44 -6.5 -02
0.4 0.0 •
0.0 0.7
0.54 -t·O. 1 +4.2 1.46 -6.5 +0.5
0.5 0.0 0. I 0.7
0.56 +0.6 +4.2 1.48 -6.4 +I 2
0.4 0.1 0.2 0.7
0.58 +1.0 +4.3 1 50 -6.2 +1.9
0.5 0.1 0.2 0.6
0.60 +1.5 +4.2 1.52 -6.0 +2.5
0.4 0. 1 0.3 0.7
0.62 +1.9 +4.1 I .54 -5.7 +3.2
'0.5 0.2 0.4 0 6
0.64 -1-2.4 +3.9 1.56 -5.3 +3.8
0.4 0.2 0.5 0.5
0.66 +2.~ +3.7 1.58 -4.8 +4.3
0.4 0.2 0.5 0.5
0.68 +3.2 +3.5 1.60 -4.3 +4.8
0.4 0.3 0 5 0.5
0.70 +3.6 +3.2 I. 62 -3.8
... +5.3
0.4 0.4 0.6 0.4
0.72 +4.0 +2.8 1.64 -3.2 +5.7
0.3 0.3 07 0.3
0.74 +4.3 +2.5 1.66 -2.5 +6.0
0.3 0.4 0.6 0.2

32-2122 491
 C'ontinued

d Rat 6 XII 6 tl R" 6 X,s 6

0.76 +4.6
0.3
+2.1
0.5
I. 68 -1.9
0.7
+6.2 '
'

0.2
0.78 +4.9 +I.6 1. 70 -1.2.. • +6.4
0.2 0.5 0.7 0.1
0.80 +5.1 +1.1 1. 72 -5.0 +6.5 '

0.1 0.5 0.8 0.0

0.82 +5.2 +0.6 I. 74 +0.3 +6.5
0.1 0.5 0.7 0.1
0.84 +5.3 +O.l 1. 76 +1 . 0 +6.4
0.1 0.6 0.7 0.2
0.86 -t-5. 4 -0.5 1. 78 +I. 7 +6.2
0.0 0.5 0.7 0.2
0.88 +5.4 -1.0 1. 80 +2.4 +6.0
0.1 0.6 0.6 0.3
0.90 +5.3 -1.6 1.82 +3.0 +5.7
0.1 0.5 0.6 0.4
0.92 +5.2 -::-2.1 1.84 +3.6 +5.3
0.2 0.6 0.6 0.5
0.94 +5.0 -2.7 1.86 +4.2 +4.8
0.3 0.5 0.5 0.5
0.96 +4.7 -3.2 I. 88 +4.7 +4.3
0.3 0.5 0.4 0.5
0 .98 +4.4 -3.7 1. 90 +5.1 +3.8
0.3 0.5 0.4 0.6
1.00 +4.1 -4.2 1. 92 +5.5 +3.2
-
0.4 0.4 0.3 0.7
1.02 +3.7 -4.6 1. 94 +5.8 +2.5
0.5 0.4 • 0. 2 0.7
1.04 +3.2 -5.0 1.96 +6.0 +1.8
0.5 0.4 0.2 0.7
1.06 +2.7 -5.4 1.98 +6.2 +1.1
0.6 0.3 0.1 0.7
1.08 +2.1 -5.7 2.00 +6.3 +0.4
0.6 0.2 0.0 0.7
l.l 0 +1.5 -5.9 2.02 +6.3 -0.3
0.6 0.2 0.2 0.7
1.12 -t-0.9 -6.1 2.04 +6.1 -1.0
0.6 0.0 0.' 1 0.7
 Continu ed

I ~ I
d Rst 6. XII 6 d R,.,
l
X 12 I 6

1.14 +0.3 -6.1 2.06 +6.1 -1.7

0.7 0. 1 0. 3 0.6
1.16 -0.4 -6.2 2.08 +5.7 -2.3
0.6 0. 1 0. 3 0.7
1.18 -1.0 -6.1 2.10 +5.4 -3.0
0.6 0.1 0. 4 0.6
1.20 -1.6 -6.0 2.12 +5.0 -3.6
0.7 0.2 0 .4 0.5
1. 22 -2.3 -5.8 2. 14 +4.6 -4.]
0.6 0. I 0 .6 0.5
1.24 -2.9 -5.7 2.16 +4.0 -4.6
0.6 0.4 0 .5 0.4
. 2.18 -5.0
1.26 -3.5 -5.3 +3.5
0.6 0.4 0. 6 0.3
1.28 -4.1 -4.9 2. 20 . +2. 9 -5.3
0.4 0.4 ·o. 7 0.3
1.30 -4.5 -4.5 2.22 +2.2 -5.6
0.5 0.6 0 .6 0.2
0.6 0.2 0.1 0.5
2.24 -4-1.6 -5.8 3.12 +5.0 -1.0
0.7 0. 1 0. 2 0.6
2.26 +0.9 -5.9 3.14 +4.81 -1.7
I

0.7 0.1 0 .3 0.6

2.28 +0.2 I -6 . 0 3.16 +4.5 -2.2
I 0.7 0.1 0 .3 0.5
2.30 -0.5
I

-5.9 3.18 +4.2 -2.71

0.7 0.1 0. 4 0.5
2.32 -1.2 -5.8 3.20 +3.81 -3.2
0.7 0.2 0. 4 0.4
2.84 -1.9 -5.6 3.22 -1-3.4 -3.6
0.6 0.2 0 5 0.4
2.36 -2.5 -5.4 3.24 -~2. 9 -4.0
0.6 0.4 , 0. 0.3
2.38 --3.1 -5.0 3.26 +2.4 -4.5
0.5 0.4 0. 0.2
2.40 -3.6 -4.6 3.28 +1.9 -45
. I
0.5 0. 5 0.2
'

l l
0.5
I
S2*
499
 Continued

d Rat A Xn A d R t2 A X a2 A

2.42 -4.1
0.4
-4.1
0.5
3.30 +1.4
0.6
--4.71 0.1
2.44 -4 5 -3.6 3.32 +0.8.. 1- -4.8
0.4 o.u 0.6 0.0
2.46 -4.9 -3.1 3.34 +0.2 -4.8 •

0.3 0.7 0.6 0 ..0

2.48 -5.2 -2.4 3.36 -0.4 -4.8
0.2 0.6 0.5 0. 1
2.50 -5.4 -1.8 3.38 -0.9 -4.7
0.2 0.6 0.6 0.2
2.52 -5.6 -1.2 3.40 -1.5 -4.5
0.1 0.7 0.5 0.2
2.54 -5.7 -0.5 3.42 -2.0 -4.3
0.0 0.7 0.5 0.3
2.56 -5.7 +0.2 3.44 -2.5 -4.0
0.1 0.6 0.5 0.4
2.58 -5.6 +O.B 3.46 -3.0 -3.6
0.2 0.7 0.4 0.3
2.60 -5.4 +1.5 3.48 -3.4 -3.3
0.2 0.6 0.3 0.5
2.62 -5.2 +2.1 3.50 -3.7 -2.8
0.3 0.6 0.3 0.5
2.64 -4.9 +2.7 3.52 -4.0 -2.3
0.4 0,5 0.3 0.5
2.66 -4.5 -t-3. 2 3.54 -4.3 -1.8
0.5 0.5
I
0 .I 0.5
2.70 -3.6 +4.2 3.56 -4.4 -1.3
0.5 0.3 • 0 .I 0.6
2.72 -3.1 +4.5 3.58 -4.5 -0.7
0"6 0.4 0.1 0.5
2.74 -2.5 +4.9 3.60 -4.6 -0.2
0.6 0.2 0 .I 0.6
2.76 -1.9 -1-5. 1 3.62 -4.5 0.4
0.6 0.2 0.0 0.5
2.78 -1.3 +5.3 3.64 -4.5 0.9
0.6 0.1 0.2 0.5
I

100
 Conttnued

l
d d
'
Ru I A
----~--~--~--~--~~~---~~--~--~---

2.80 -0.7 +5.4 3.66 -4.3 14

0.7 0 (l 0.3 0.5
2.82 0.0 +5.4 3.70 -3.8 2.4
0.6 0. I 0.3 0.4
2.84 +0.6 +5.3 3 72 --3.5 2.8
0.6 0 .I 0.4 0.4
2.86 +1.2 +5.2 3.74 -3.1 3.2
0.6 0.2 0.4 0.4
2.88 +1.8 +5.0 3.76 -2 7 3.6
0.6 0.3 0.5 0.3
2.90 +2.4 +4.7 3.78 -2.2 3.9
0.5 0.3 0.5 0.2
2.92 +2.9 +4.4 3.80 -1.7 4. I
•

0.5 0.4 0.5 0.1

2.94 +3.4 -t-4. 0 3.82 -1.2 4.2
0.4 0.4 0.6 0 1
2.96 +3.8 +3.6 3.84 -0.7 4.3
0.4 0.5 0.6 0.0
2.98 -t-4.2 -1-3. 1 3.86 -0.1 4.3
0.3 0.6 0.5 0.0
3.00 +4.5 +2.5 3.88 +0.4 4.3
,
u.a 0. f, 0.5 0. I
3.02 +4.8 +2.0 3.90 +0.9 4.2
0.7 0.5 0. I
0.2
3.92 +1.4 4. 1
3.04 +5.0 + 1.3
0.5 0.5 0.2
0.1 3.94 +1.9 3.9
3.06 +5.1 +0.8 0.3
0.7 0.4
0.1 3.96 +2t3 3.6
+O.l 0.3
3.08 +5.2 0.6 0.4
3.98 +2.7 3.3
0.1 -0.5
0.5 0.4 0.4
3.10 +5.1 4.00 +3.1 +2.9
0.1
501
 Continued
h= 1.5

d Rat A x,. A d R,'j A X a1 A

-

1.8 0.4
0.0 +I. 7 +·0.2 2.1 -1.9. • -3.8
0.0 0.1 1. 7 1.4
0.1 +1.7 +0.1 2.2 -3.6 -2.4 •

0.0 0.2 0.7 2.2

0.2 +1.7 -0.1 2.3 -4.3 -0 2
0.1 0.4 0.5 2.2
0.3 +1.6 --0.5 2.4 -3.8 +2.0
0.2 0.4 1 .6 1. 7
0.4 +1.4 -0.9 2.5 -2.2 +3. 7
'
0.3 0.5 2.2 0.6
0.5 +1.1 -1.4 2.6 0.0 +4.3
0.6 0.4 2.3 0.7
0.6 +0.5 -1.8 2.7 -t-2.3 +3.6
0.9 0.2 1. 5 1. 7
0.7 -0.4 :.......2.0
...
2.8 +3.8 +1.9
0.9 0 .I 0.3 2.3
0.8 -1.3 -1.9 2.9 +4.1 -o.4
0.8 0.6 0.8 2.2
0.9 -2.1 -1.3 3.0 -i- 3. 3 -2.6
0.6 0.1 1. 9 I. 3
1. 0 -2.7 -0.3 3. 1 +1.4 -~.9
0.1 I. 3 2¥3 0.1
1. 1 -2.8 +1.0 3.2 -0.9 -4.0
-
0.6 1.2 2.3 1.2
1.2 -2.2 +2.2 3.3 -2.9 -2.8
I •1 1.0 • l. 0 2.0
1.3 -1.1 +3.2 3.4 -3.9 -0.8
1.5 0.3 0.2 2.3
1.4 +0.4 +3.5 3.5 -3.7 +1.5
1. 7 0.4 1.5 I. 7
I. 5 +2.1 +3.1 3.6 -2.2 +3.2
1.3 1.3 2.1 0.7
1.6 +3.4 +1.8 3.7 -o.l +3.9
0.6 1.8 2 .I 0.7
1. 7 +4.0 0.0 3.8 +2.0 +3.2
'

502
 Continut
L·
R,, x,.. R,.,
d A A d
I A x" 4

0.4 1.9 1. 4 1. 7
1.8 +3.6 -1.9 3.9 +3.4 +1.5
I. 4 1. 6 0.2 2 .I
I. 9 +2.2 -3.5 4.0 +3.6 -0.6
•
2.3 0.7
2.0 -0.1 -4.2

h=2.0
•

d R,, fl X 12 6 d R,, 6
X'' A

I
'
0.7 0.5
0.0 -1.0 -0.1 1.5 +0.6 -2.1
0.0 0.0 . 1. 0 0.2
0.1 -1.0 -0.1 1. 6 -0.~4 -2.3·
0.1 0.1 • 1.0 0.3
0.2 -0.9 0.0 1. 7 -1.4 -2.0
0.0 0.2 0.8 0.7
0.3 -0.9 +0.2 1.8 -2.2 -1.3
0.0 0.2 0.4 1.0
0.4 -0.9 +0.4 1. 9 -2.6 -0.3
0.2 0.2 0.1 1.3
0.5 -0.7 +0.6 2.0 -2.5 +1.0
'
0.2 0.2 0.6 1.1
0.6 -0.5 +0.8 2.1 -1.9 +2.1
0.3 0.2 1.1 0.7
0.7 -0.2 +1.0 2.2 -0.8 +2.8
0.4 0.2 1. 4 0.1
0.8 +0.2 +1.2 2.3 +0.6 +2.9
0.5 0.1 1. 4 0.6
0.9 +0.7 +1.1 2.4 +2·.0 +2.3
0.4 0.2 0.9 1.2
1.0 +1.1 +0.9 2.5 +2.9 +1.1
0.4 0.4 0.2 1.5
1.1 +1.5 +0.5 2.6 +3.1 -0.4
0.2 0.7 0.6 1.5
508
 Continued

•
d kn ~ Xu A d R" 6 X 12 4

- I
1.2 +1. 7 -0.2 2.7
. +2.5 -1.9
0.0 0.7 1. 2 1. 0
1.3 +1. 7 -0.9 2.8 +1.3. 1- -2.9
0.4 0.7 I .7 0.3
I. 4 +1.3 -1.6 2.9 -0.4 -3.2 •

0.7 0.5 1. 5 0.6

1.5 0.6 I .5 0.8
3.0 -1.9 -2.6 3 6 +2.1 +2.3
1. I I .3 I. 0 1. 4
3. J -3.0 -1.3 3.7 +3.1 +0.9
0.2 I .8 0.0 1.8
3.2 -3.2 +0.5 3.8 +3.1 -0.9
0.7 1.5 1. 0 1.5
3.3 -2.5 +2.0 3.9 +2.1 -2.4
I. 4 I. 0 l. 6 0.7
3 4 -1.1 +3.0 4.0 +0.5 -3.1
1 7 ... 0.1
3.5 +0.6 +3.1
1.5 0.8

h=2. 5

d R,a ~ X 12 ~ d R,t A X 11 A
-- -

0.3 0.7
0.0 +O 6 +0.1 2. I +1.9 • +0.1
0.0 0. I 0.1 0.8
0.1 +0.6 0.0 2.2 +1.8 -0.7
0.0 0.0 0.4 0.8
0.2 +0.6 0.0 2.3 +1.4 -1.5
0.0 0 .I 0.7 0.5
0.3 +0.6 -0.1 2.4 +0.7 -2.0
0.1 0.1 I. 0 0.1
0.4 +0.5 -0.2 2.5 -0.3 -2.1
•'
00 0.] .
I 0.9
'
0.2
504
 Continuea
--
I/ A,,
•

II
d Rn
•
£\
•
Xn 6 d
•
R 11 I ti
4

0.5 +0.5 -0.3 2.6 -1.2 -1.9

0.1 0.2 0.8 0.7
0.6 +O 4 -0.5 2.7 -2.0 -1 2
0.1 0.1 0.3 1.0
0.7 +0.3 -0.6 2.8 -2.3 -0.2
0.2 0. 1 0. I l. 2
0.8 +0.1 -0.7 2.9 -2.2 +1 0
0.2 0. 1 0.3 0.9
0.9 -0.1 -0.8 3.0 -1.4 +I 9
•
0.3 0. I 1. 0 0.5
1. 0 -0 4 -0 . 7 3.1 -0.4 -t-2. 4
0.3 0. ] I .2 0.0
I.I -0.7 I
-0.6 3.2 -1-0. 8 -f-2.4
0.3 0.2 1. 1 0.8
1. 2 -1.0 -0.4 3.3 +1.9 +1.6
0. 1 0.4 0.5 I .0
1.3 -1.] 0.0 3.4 +2.4 +0.6
0.0 0.4 0.0 ' 1.3
1. 4 -1.1 +0.4 3.5 -f-2. 4 -0.7
0. 1 0.5 0.7 I .3
1.5 -1.0 +0.9 3.6 +1. 7 -1 8
0.3 0:3 l.I 0.7
I. 6 -0.7 +1.2 3.7 +0.6 -2.5
0.5 0.3 l. 3 00
l. 7 . -0.2 I +1.5 3.8 -0.7 -25
0.6 0.2 l .2 () 7
~

1.8 +0.4 +1.7 3.9 -1.9 -1.8

0.7 0.4 0.6 1. 2
1. 9 +1.1 -t- I . 3 4.0 -2.5 -0.6
0.5 0.5
2.0 +1.6 +0.8 I
I
, II
•
I
 Continued
h=3.0

t1 R., ll XII 6 d R,, 6 Xt, 6

I
0.0 0.1
0.0 -0.5 0.0 0.3 -0.5.. +0.1
le
0.0 0.0 0.0 0.0
0.1 -0.5 0.0 0.4 -0.5 +O.I •

0.0 0.0 0.1 0.1

0.2 -0.5 0.0 0.5 -0.4 +0.2
0.0 0.1 0.1 0.1
0.1 0.1 0.1 0.6
0.6 -0.3 +0.3 2.4 -1.4 +0.6
0.1 0.1 0.4 0.6
0.7 -0.2 +0.4 2.5 -1.0 +1.2
0.1 0.1 0.4 0.3
0.8 -0.1 +0.5 2.6 -0.5 +1.5
0.2 0.0 0.7 0.1
0.9 +0.1 +0.5 2.7 +0.2 +1.6
0.1 0.1 0.7 0.2
1.0 -t-0.2 -F0.6 2.8 +0.9 +1.4
0. 1 0.1 0.6 0.4
1.1 +0.3 +0.5 2.9 +1.5 +1.0
0.2 0.1 0.3 0.8
1.2 +0.5 +0.4 3.0 +1.8 +0.2
0.2 0.1 0.1 0.8
1.3 +0.7 +0.3 3.1 +1.7 -0.6
0.1 0.2 0.4 0.7
1. 4 +0.8 +0.1 3.2 --t-1.3 -1.0
0.1 0.3 0.7 0.5
1.5 +0.9 -0.2 3.3 +0.6 • -1.8
0.1 0.3 0.9 0.1
1. 6 +0.8 -0.5 3.4 -0.3 -1.9
0.2 0.3 0.9 0.3
1. 7 +0.6 -0.8 3.5 -1.2 -1.6
0.3 0.2 0.6 0.7
1.8 +0.3 -1.0 3.6 -1.8 -0.9
0.4 0.1 0.2 1.0
1.9 -0.1 -1.1 3.7 -2.0 +0. J
0.4 0.0 0.3 1.0

606
 Conti nued
- I ,-
Rn !!. R •., !!. A
d ~
x" d
I X l'l

2.0 -0.5 -1.1 3.8 -1.7 +1.1

0.4 0.2 0.6 0.7
2.1 -0.9 -0.9 3.9 -1.1 +1.8
0.4 0.4 1.0 0.3
2.2 _:1.3 -0.5 4.0 -0.1 -t-2.1
0.2 0.5
2.3 -1.5 0.0
0.1 0.6

h=3.5

d Rn ' L\ X 1., !!. d Rn l1 x.~ fi

I •
0.4 0. 0
0.0 +0.3 0.0 2. I +0.3 +0.9
0.0 0.0 0.3 0.1
0.1 +0.3 0.0 2.2 +0.6 +0.8
0.0 0.0 0.2 0.2
0.2 +0.3 0.0 2.3 +0.8 +0.6
0.0 0.1 0.2 0.3
0.3 +0.3 -0.1 2.4 +1.0 +0.3
0.0 0.0 0.1 0.5
0.4 +0.3 '
-0.1 2.5 +1.1 -0.2
0.0 0.0 0.1 0.4
0.5 +0.3 -0.1 2.6 +1.0 -0.6
0.0 0.1 0.3 0.4
0.6 +0.3 -0.2 2.7 +0.7 -1.0
0.1 0.1 0.4 0.3
0.7 +0.2 -0.3 2.8 +0.3 -1.3
0.1 0.0 0.5 0.1
0.8 +O.l -0.3 2.9 -0."2 -1.2
0.0 0.1 0.5 0.0
0.9 +0.1 -0.4 3.0 -0.7 -1.2
0. I 0.0 0.5 0.4
1. 0 0.0 -0.4 3.1 -1.2 -0.8

607
 Conttnued

d Ra2 4 X 12 4 d R 1., A X" ~

I
0.1 0.0 0.2 0.6
1. I -0. J -0.4 3.2 -1.4 -0.2
0.2 0.0 ~
~ 0.0 0.7
1. 2 -0.3 -0.4 3.3 -1.4 +0.5
• 0.1 0. I 0.3 o.s
1. 3 -0.4 -0.3 3.4 -1.1 +1.0
0.1 0.1 0.5 0.4
1.4 -0.5 -0.2 3.5 -0.6 +1.4
0. I 0. I 0.7 0.2
I. 5 -0 6 -0.1 3.6 +O.l +1.6
0.0 0.2 0.7 0.2
1. 6 -0.6 +O.l 3.7 +0.8 +1.4
0.0 0.2 0.5 0.5
1. 7 -0.6 -t-0.3 3.8 +1.3 +0.9
0.1 0.2 0.3 0.7
1.8 -0.5 +0.5 3.9 +1.6 +0.2
0.2 0.2 0.0 0.8
1. 9 -0.3 -to.7 4.0 +1.6 -0.6
0.2 0.2
2.0 -0.1 -f-0.9
I