energies

Article
Maximum Electrical Power Extraction from Sources by
Load Matching
Sigmund Singer 1 , Shlomi Efrati 2, * , Meir Alon 1 and Doron Shmilovitz 1

1 School of Electrical Engineering, Tel-Aviv University, Tel-Aviv 6997801, Israel; [email protected] (S.S.);
[email protected] (M.A.); [email protected] (D.S.)
2 School of Mechanical Engineering and System Engineering Research Initiative (TAU-SERI),
Tel-Aviv University, Tel-Aviv 6997801, Israel
* Correspondence: [email protected]

Abstract: This paper describes the matching of various loads to sources (including nonlinear ones).
The purpose of matching is to extract the maximum available power from the source. This has
particular importance for renewable sources and energy-harvesting devices, in which unused energy
is just wasted. The main innovations in this paper include (and followed by examples) simplified
calculation of the matching parameter for a controllable load and matching by means of a family of
power-conservative two-port networks, denoted POPI (Pin = Pout), such as transformers, gyrators,
loss-free resistors (LFRs) and series LFRs (SLFRs). An additional innovation described in this paper
is a new, simplified model of an HF power amplifier based on the series LFR concept. This model
predicts that the efficiency of the HF power amplifier operated under the matched-mode condition
can significantly exceed the 50% efficiency limit that is predicted by the conventional model. As
HF power amplifiers drive antennas in transmission and some wireless power transfer (which uses



radiative techniques) systems, it is clear that the operation of such systems in the matched-mode

 condition is not restricted to a 50% efficiency limit.
Citation: Singer, S.; Efrati, S.; Alon,
M.; Shmilovitz, D. Maximum Keywords: maximum power transfer; impedance matching; POPI networks; loss-free resistor (LFR);
Electrical Power Extraction from renewable energy sources; wireless power transfer; HF power amplifier; transmission systems
Sources by Load Matching. Energies
2021, 14, 8025. https://doi.org/
10.3390/en14238025

1. Introduction
Academic Editor: Fengshou Gu
In electrical engineering, the maximum power transfer theorem states that, to obtain
maximum external power from a source with a finite internal resistance, the resistance
Received: 7 October 2021
Accepted: 26 November 2021
of the load must equal the resistance of the source, as viewed from its output terminals.
Published: 1 December 2021
Moritz von Jacobi published the maximum power (transfer) theorem around 1840; it is also
referred to as “Jacobi’s law”.
Publisher’s Note: MDPI stays neutral
The theorem results in maximum power transfer across the circuit but not maximum
with regard to jurisdictional claims in
efficiency. If the resistance of the load is increased beyond the resistance of the source, then
published maps and institutional affil- efficiency is higher because a higher percentage of the source power is transferred to the load,
iations. but the magnitude of the load power is lower because the total circuit resistance increases.
Extraction of the maximum available power is desirable in many systems involving
sources and loads. This is attained by selecting a suitable load or, when possible, by
controlling its parameters so that the source outputs its maximum available power. The
Copyright: © 2021 by the authors.
operation of such systems is referred to as matched-mode operation. Operation in matched
Licensee MDPI, Basel, Switzerland.
mode indicates that the source functions at the maximum power point (denoted MPP). The
This article is an open access article
characteristic of the source determines the MPP voltage and current. In a family of systems
distributed under the terms and that capture ambient energy and convert it into electrical energy, such as photovoltaic
conditions of the Creative Commons systems, wind generators and energy-harvesting devices, operation at MPP also implies
Attribution (CC BY) license (https:// maximum conversion efficiency of the ambient energy. This is because, in most of these
creativecommons.org/licenses/by/ cases, energy (solar irradiation, wind, etc.) cannot be stored in its original form, so the part
4.0/). that is not converted (to the electrical energy) is lost.

Energies 2021, 14, 8025. https://doi.org/10.3390/en14238025 https://www.mdpi.com/journal/energies

Energies 2021, 14, 8025 2 of 19

Systems that operate at MPP, such as those consisting of a voltage source and internal
series resistance (storage battery, for example), result in high losses. It is evident that in such
systems, operating at lower power than the maximum available is practical, preserving it
for short instances in which high peak power is needed.
In this study, various source–load systems in matched-mode conditions were analyzed.
These include linear and nonlinear systems, direct source–load connections and coupling
by means of a power-conservative two-port network, denoted POPI (Pin = Pout), such as
transformers, gyrators and loss-free resistors (LFRs). Various realizations of POPI networks
(by means of switched-mode converters) are reported in [1–7]. Since power electronic
circuits can be operated at high efficiency (up to 98%), using a family of POPI networks to
model circuits is very convenient.
Some application examples of maximum power transfer by means of controlled POPI
networks include photovoltaic systems, battery chargers and maximum energy extraction
(from a decaying source) systems.
It should be noted that our paper is focused on cases in which the power needs of the
load meet the maximum power ability of the source.
In all cases of curtailable loads, and in most cases of non-curtailable loads, the time-
varying power demanded by the load differs from the maximum (time-varying) power that
the source delivers; in order to bridge this difference, a storage element can be applied [8].
In that case, a controlled coupling network (realized by a converter, inverter or combination
of both) is used in order to process the power flow between the source, load and storage so
that the source operates at its maximum power point and fulfills the load requirements.
The combination of storage, load and a coupling network forms a new load (as
viewed from the source) that can extract the maximum available power from the source.
A typical example of an arrangement that includes a curtailable load is a standalone PV
system [9,10]. In this system, the load power demand may vary or even be zero (load
disconnection), whereas, during the entire operation period, the PV array can deliver the
maximum available power to the load/storage system.
An example of a system that includes a non-curtailable load is a PV grid-connected
system [11]. In that case, the grid is viewed from the PV array as a “black box” consisting
of a voltage source, storage and load elements, and it is capable of absorbing all of the
power that the source can deliver. In addition, in this case, matching is achieved by means
of a coupling network consisting of a converter/inverter combination.
The conventional model of an RF power amplifier predicts that its operation in the
matched-mode condition implies a 50% efficiency limit. However, some transmitters (with
an RF power amplifier at the output stage) are known to operate at significantly higher
efficiency (in the matched-mode condition) than is predicted by that model.
In order to resolve this contradiction, we propose a modified model that is based
on the loss-free resistor concept. This model includes all of the properties of the classical
model except the efficiency limit. The model was validated by simulations carried out over
a practical transmission circuit.

2. Matching of Sources by Controllable Loads

A conventional model of a large family of sources, such as batteries, generators,
electromechanical devices, piezoelectric transformers and high-frequency and microwave
transmission systems, consists of a voltage source and series resistance (or series impedance
in the case of sine wave excitation). A voltage source with internal series resistance loaded
with a resistive load is shown in Figure 1.
This model predicts that by loading the source with controlled resistance R L , the
maximum power can be extracted from the source by applying the conventional calculation
method, i.e., ∂R∂P
L
= 0, resulting in R L = R∗L = rS , where R∗L denotes the value of R∗L that
extracts the maximum power from the source.
Energies 2021,14,
Energies2021, 14,8025
x FOR PEER REVIEW 33of
of19
23

rS a

Vs i V RL

b
Figure 1. DC
Figure1. DC source
source with
withinternal resistancerrSs loaded
internalresistance loaded by resistor R
by resistor RLL..
2

PSmax = 14 Er 
S
RL , the


This model predicts that by loading the source
 with controlled resistance

E 
V =V M = a−b 2
maximum power can be extracted from the source by (1)
applying the conventional calcula-
I M = Ia−b = ISC2 

∂P


ISCin≡ RrE = R* = r 
= 0 , resulting
*
RL denotes the value of RL
*

tion method, i.e., L
S L S , where
∂R
where PSmax denotes Lthe maximum source power, and VM and I M are the voltage and
that extracts the maximum power from the source.
current at the maximum power point, respectively.
It should be noted that the load should not 1 Ebe a resistive element only; for example,
2

for a storage battery modeled with an ideal =

PS max voltage 
source u = ub , we have the following:
4 rS

i = i a−b = VSr− u
E 
VM = Va −Vb =−u  
P



Pa−b = i · u = r P 2 u
S
(1)
(2)
∂Pa−b VS −2u I SC 
∂u = I MrS= I a − b =
2∗  VS
= 2
∂Pa−b
= 0 ⇒ V M = U
∂u
E
I SC ≡ 
Inserting VM into the current equation results
r in the following:
S

VS
PS max I M = 2r = ISC
2 VM IM
where denotes the maximum source
S power, and and are the voltage and
(3)
VS2
current at the maximum power point,Pa−bmax = VM IM =
respectively. 4rS
It should be noted that the load should not be a resistive element only; for example,
3. Efficiency of a Source with Internal Resistance in Matched-Mode Operation
for a storage battery modeled with an ideal voltage source u = ub , we have the following:
The efficiency, η, is given by the following:
V −u
ηi = b
= =S v a−b · i a−b
= iaP−out (4)
Pin rp Pin

Pin is the power delivered by VS , includingthe u  power and the internal losses on rS :
V −load
Pa − b = i ⋅ u =  S u
 rp 
Pin = VS i = VS i a−b (2)
(5)
Po∂P= P V −
= 2
v u · i
=
a − b load S a−b a−b
∂u rS
In matched-mode operation:
∂Pa − b V
= 0  VM = U = S
*

ib = I∂Mu = ISC 2 2
Va−b = VM = V2S (6)
Inserting VM into the current
η = Vequation
M IM
= VMresults
= 0.5in
=the
50%following:
VS I M VS

This result is general and independent of the kind of load that enables the source to
operate at the maximum output power.
 VM I M VM
η= = = 0.5 = 50%
VS I M VS

This result is general and independent of the kind of load that enables the source to
operate at the maximum output power.
Energies 2021, 14, 8025 4 of 19

4. AC Source with Internal Impedance

The phasor description of an AC source VS with internal impedance ZS loaded by

4. AC Source with Internal Impedance
Za isThe phasor
shown description
in Figure 2. of an AC source V S with internal impedance ZS loaded by Za
is shown in Figure 2.

ZS a I

Vs V Za

b
Figure 2. AC
Figure 2. AC source
source with
with internal
internal impedance
impedance loaded
loaded by impedance ZZaa..
by impedance

The source and load impedances are given by the following:

The source and load impedances are given by the following:

ZS = rs + jxS
(7)
Za = r a + jx a

It is well known that maximum power transfer, Pmax , is achieved when the load
Za = ZS∗ :
X a = − XS
(8)
r a = rS
In this case, PSmax and the power transfer efficiency, η, are given by the following:
2
VS
PSmax = (9)
4rS

η = 50% (10)

5. Matching General Sources to Controllable Loads

Let us analyze the case in which a source consisting of voltage source E with in-
ternal resistance r P is loaded with a nonlinear load (which approximately describes the
characteristic of an LED string), given by the following:
v
i = ae v0 (11)

where a is controllable. Attempting to find a∗ (the value of a that extracts the maximum
source power P) using the classical method (i.e., ∂P ∂a = 0 ⇒ a∗ = · · · ) is quite a difficult
task. The difficulty increases if the source has nonlinear characteristics (in many cases,
there is no closed-form solution).
A much easier calculation is based on the insight that the maximum power is a
property of the source (of any source), and the suitable load just enables the extraction of
that power, so we can substitute the maximum power point coordinates (VM , I M ) into the
load characteristic in order to obtain the solution with a simpler method. In our case, for
example, we have the following: 
i = IM
(12)
v = VM
VM
−
a = a∗ = I M e V0
(13)
Energies 2021, 14, x FOR PEER REVIEW 6 of 23

Energies 2021, 14, 8025 5 of 19

IM , and to then insert the values into the load i − v curve in order to determine the match-
ing conditions.  
E ISC E
In our source, V
An example is aM =
system ,
2 inI =
M which
2 theI =
SCsource ; therefore, we are
r P characteristics have the following:
given by the following:

E −−β2VE⋅o iS
2
V =ISC (15)
a∗ =S e (14)
2
This equation is a reasonable approximation of a typical fuel cell with a particular
focusIfon
theitssource and the load both have nonlinear characteristics, it is recommended
MPP domain.
to calculate
The load thei −
voltage
v and
curve current
is given byatthe
the MPP (maximum power point), denoted VM
following:
and I M , and to then insert the values into the load i − v curve in order to determine the
u
matching conditions. Vo (16)
An example is a system in which the sourcei = a ⋅ echaracteristics are given by the following:
The source MPP parameters are calculated as 2follows:
VS = E − β · iS (15)
Ps = vs is = E − β ⋅ is is
This equation is a reasonable approximation
( 2
)
of a typical fuel cell with a particular
focus on its MPP domain. ∂P
 E − 3 β ⋅ is = 0
2

The load i − v curve is given by = 0following:

the
∂i s
u
*
i = a · eEVo (16)
is ≡ I M =
The source MPP parameters are calculated 3βas follows: (17)

* 2
PS v=s ≡ = EE− β · i2 iS
vSViSM =


S
3
2
∂iS = 0 ⇒ E − 3β · iS = 0
∂P
q 2 E
iS∗ ≡ I M ≡=PM =3β
Ps max E E
(17)
3 3β
∗
vS ≡ VM = 3 E 2

Inserting VM , IM into the Pload characteristic that the value of a that ex-
q
2 implies
E
Smax ≡ PM = 3 E 3β
*
tracts the maximum power, denoted a is given by:
Inserting VM , I M into the load characteristic implies 2E
that the value of a that extracts
the maximum power, denoted a∗ is given* by: E − 3 vo
a = e (18)
s 3β
E − 3v2E
a∗ = e o (18)
3β
6. Voltage Source Type with a Fuse
6. Voltage
In manySource Type systems,
practical with a Fuse the load is operated at a much lower current than the
short-circuit
In many current;
practicalthe current
systems, theisload
limited by a fuse
is operated at athat
muchis destroyed at currents
lower current than thehigher
short-
circuitIcurrent; the current is limited by a fuse that is destroyed at currents higher than IF
than F (the same can be achieved by an overload relay or by the collapse of some com-
(the same can be achieved by an overload relay or by the collapse of some component in
ponent
the in the
circuit see circuit
Figure see
3). Figure 3).

rS a I

IF
Vs V Load

b
Figure 3. DC
Figure 3. DC source
source with
with internal
internal resistance
resistance and
and aa fuse.
fuse.
 As the maximum current that the source can deliver is limited by IF , this value de-
termines the maximum power point current and voltage as follows:

IM = IF
Energies 2021, 14, 8025 6 of 19
VM = (VS − I F rS ) (19)
Ps max = I F (VS − I F rS )
As the maximum current that the source can deliver is limited by IF , this value
As IF is
determines much
the lower than
maximum thepoint
power shortcurrent
circuit,and I SCvoltage
, this value of power is much lower
as follows:
than that in a case involving only the voltage source and internal resistance (given in
I M = is
Equation (1)) In this case, the efficiency IFgiven by the following:
VM = (VS − IF rS ) (19)
I F (VS − rS I F ) I F rS
η P=Smax = 1
= IF (VS − IF rS ) − (20)
VS I F VS
As IF is much lower than the short circuit, ISC , this value of power is much lower
I F rSin a case involving only the voltage source and internal resistance (given in
than that 1
If E (1)) In, then
Equation this case, the→
Pa−bmax IFVS , andisthe
efficiency given by the
power following:
transfer efficiency is η → 1 .
If the load is controllable, the load parameters that result in maximum power transfer
I (V − r S IF ) I r
η= F S = 1− F S (20)
can be determined by inserting the maximumV I
S F VS coordinates ( IM , VM ), as de-
power point
scribed previously.
If IFErS  1, then Pa−bmax → IF VS , and the power transfer efficiency is η → 1 .
It should be noted that because the fuse (or current limiter) is installed for safety pur-
If the load is controllable, the load parameters that result in maximum power trans-
poses, most practical systems are operated at power levels that are significantly lower
fer can be determined by inserting the maximum power point coordinates (I M , VM ), as
described
than previously.by Ps max .
that determined
It should be noted that because the fuse (or current limiter) is installed for safety
purposes,
7. Matching most practical
by Means ofsystems
Controlled are operated
Two-PortatNetworks
power levels that are significantly lower
than that determined by PSmax .
In most cases, the load is not tunable, so matching can be achieved by applying a
power-conservative
7. Matching by Means two-port element whose
of Controlled Two-Port parameters
Networks can be controlled. There is a fam-
ily of power conservative two-port network elements in which the output power equals
In most cases, the load is not tunable, so matching can be achieved by applying
the input power:
a power-conservative two-port element whose parameters can be controlled. There is a
family of power conservative ( t ) = vO (network
POtwo-port t ) iO ( t ) =elements
Pi ( t ) = vin iwhich
i( t ) i( t )
∀t the output power equals
(21)
the input power:
satisfies Equation (19) Is denoted POPI ( Po ( t ) = Pi ( t ) ) , see
Po (t) = vo (t)io (t) = Pi (t) = vi(t) ii(t) ∀t (21)
A network element that
FigureA4.network element that satisfies Equation (19) Is denoted POPI ( Po (t) = Pi (t)), see
Figure 4.
A Ii IO C

Source Vi POPI VO Load

B D

Control

Figure
Figure4.4.Source–load
Source–loadcoupling
couplingby
bymeans
meansof
ofPOPI
POPInetwork.
network.

Transformers,
Transformers,gyrators
gyratorsand
andloss-free
loss-freeresistors
resistors(LFRs)
(LFRs)are
areamong
amongthe
theelements
elementsin
inthe
the
POPI
POPIfamily
family(Figure
(Figure5).
5).
The governing relations between the input and the output electronic parameters of
the different models are summarized next.
For the LFR: )
i1 = v1/RE (t)
(22)
p2 = p1
For the TVT:     
v1 k(t) 0 v2
= (23)
i1 0 k ( t ) −1 i2
Energies 2021, 14, 8025 7 of 19

For the gyrator, the governing relation is as follows:

    
i1 0 g(t) i2
= (24)
v1 g ( t ) −1 0 v2

Energies 2021, 14, x FOR PEER REVIEW R E (t), k(t) and g(t), respectively.
where, in the above models, the controlled parameters are 8 of 23
The controlled parameter is chosen in a way that accommodates the constraints of the load
and source.

A I1 IO C A I1 IO C A I1 IO C

V1 RE P VO V1 VO VO

B LFR D B TVT D B Gyrator D

Control Control Control

(a) (b) (c)

Figure 5.
Figure Typical models
5. Typical modelsofofPOPI
POPInetworks. (a)(a)
networks. Loss-free resistor
Loss-free (LFR);
resistor (b) time-varying
(LFR); transformer
(b) time-varying trans-
(TVT); (TVT);
former (c) gyrator.
(c) gyrator.

It should
The be noted
governing that between
relations in addition
the to
inputthe and
above-mentioned networks,
the output electronic additional
parameters of
two-port
the networks
different modelsshare the POPI property
are summarized next. and other types of characteristics.
POPI
For thenetworks
LFR: can be realized by a variety of switched-mode converters operated at
high switching frequency. For example, pulse-width modulation (PWM) converters can
t 
realize the transformer; some topologies ofva1 series resonant
RE ( )  [1–3], and a PWM DCM buck-boost
converter and discontinuous
i =
conduction mode (DCM) exhibit gyrator1 characteristics (22)
converter can operate as an LFR [1].
p = p1 
are described next. The first method
As an example, two methods of LFR2 realization
is based on the
For the TVT:continuous control of a PWM converter operated in CCM in such a way
that the input current and voltage parameters obey Ohm’s law, resulting in the creation of
an emulated resistance R E . Then,  v1the  k ( t ) absorbed
 power 0  by v2R E is transferred to the output
terminals, forming the LFR [4].   =  −1    (23)
The “natural” (i.e., non-controlled) 0 k ( t )of the
 i1   realization  i2LFR
 is based on the buck-boost
converter topology operated in DCM (Figure 6a,b). In this converter, the voltage to average
For the gyrator, the governing relation is as follows:
current ratio calculated at the input terminal is as follows:
 in  0
 i1 V g2L( t )  i2 
v Iin= = RE =   (25)
(24)
 1   g ( t ) D1 0· TS v2 
−1 2

where D1 is the duty cycle, TS is the switching time interval and L is the inductance of the
where,
inductor.in the above
As all models,
variables in the
(23)controlled parameters
are constant, RE t , kcharacteristics
are resistive
this imparts t and g t to, there- () () ()
input terminals,
spectively. i.e., emulates
The controlled a resistance,
parameter denoted
is chosen as R
in a way E (Figure
that 6a).
accommodates the constraints
of theThe
loadpower absorbed by the input terminals is transferred to the output (as a result of
and source.
the ideal converter
It should be notedcharacteristics).
that in addition to the above-mentioned networks, additional two-
The realization
port networks share is completed
the by averaging
POPI property thetypes
and other pulsating current at the input by means
of characteristics.
of a low-pass filter, as
POPI networks shown
can in Figure
be realized by a5c.
variety of switched-mode converters operated at
It should befrequency.
high switching noted thatFor
other realizations
example, of POPI networks
pulse-width modulationexist(PWM)
and that in addition
converters to
can
the transformer, gyrator and LFR, there are additional elements in the large
realize the transformer; some topologies of a series resonant converter and discontinuousfamily of POPI
networks that
conduction are characterized
mode (DCM) exhibitbygyrator
different properties. Applications
characteristics [1–3], and a of such DCM
PWM networks
buck-in
powerconverter
boost processingcansystems
operateareas described
an LFR [1].in [6–8,12–16].
As an example, two methods of LFR realization are described next. The first method
is based on the continuous control of a PWM converter operated in CCM in such a way
that the input current and voltage parameters obey Ohm’s law, resulting in the creation
of an emulated resistance R E . Then, the power absorbed by R E is transferred to the out-
put terminals, forming the LFR [4].
The “natural” (i.e., non-controlled) realization of the LFR is based on the buck-boost
converter topology operated in DCM (Figure 6a,b). In this converter, the voltage to aver-
 = RE = 2 (25)
I in D1 ⋅ TS

where D1 is the duty cycle, TS is the switching time interval and L is the inductance
Energies 2021, 14, 8025 of the inductor. As all variables in (23) are constant, this imparts resistive characteristics
8 of 19

to the input terminals, i.e., emulates a resistance, denoted as RE (Figure 6a).

D Vout

Vin L C Load

(a)

Iin A I1 IO C
Ipk

Low
Buck/Boost
V1 Pass VO
DCM
Filter
t
D 1 Ts Ts
P
B LFR D

(b) (c)

(a) Buck-boost
Figure 6. (a) Buck-boost converter
converter topology;
topology; (b)
(b) input
input current
current waveform
waveform indicating
indicating DCM operation;
converter.
(c) LFR realization of the converter.
Energies 2021, 14, x FOR PEER REVIEW 10 of 23
8. Matched-Mode Operation
The power absorbed of input
by the a Photovoltaic
terminalsSystem
is transferred to the output (as a result of
The intersection
the ideal of PV array and typical load characteristics is shown in Figure 7.
converter characteristics).
The realization is completed by averaging the pulsating current at the input by means
I
of a low-pass filter, as shown in Figure 5c.
It should be noted that other
Pmax realizations of POPI networks exist and that in addition
to the transformer, gyrator and LFR, there are additional elements in the large family of
IVP networks that are characterized by different properties. Applications of such net-
POPI
works in power processing systems are described in [6–8,12–16].
1
∀
r
8. Matched-Mode
p Operation of a Photovoltaic System
The intersection of PV array and Vtypical load characteristics is shown in Figure 7.
E be noted that in practical photovoltaic systems, most loads can be described
It should
P
as comprising a voltage source E in series with a resistive element rp. In most cases, the
voltage source is the efficient part of the load, and the resistor is the part with losses. This
description
Pmax applies to a standalone PV system with a storage battery PV system, con-
nectedPto the utility and PV array, or a PV array that powers an electrical motor (for pump-
ing purposes).
A load connected directly
V
to a PV array will very rarely extract the maximum power
VMP VOC
from the array (Figure 7).
Figure 7.
Figure The intersection
7. The intersection of
of PV
PV and
and load
load characteristics
characteristics (consisting
(consisting of
of voltage
voltage source
source and
and resistive
element rrpP) )ininaadirect
element directsource–load
source–loadconnection.
connection.

It should be noted that in practical photovoltaic systems, most loads can be described
In order to extract the maximum available power [17], we can match the load to the
as comprising a voltage source E in series with a resistive element rp . In most cases, the
PV array by means of a POPI two-port element, for example, by means of a transformer,
voltage source is the efficient part of the load, and the resistor is the part with losses.
because the I–V curve of the PV array varies with the insulation, and the load parameters
This description applies to a standalone PV system with a storage battery PV system,
may vary as well, so dynamic matching is necessary. Dynamic matching can be achieved
connected to the utility and PV array, or a PV array that powers an electrical motor (for
by controlling the transfer ratio of the transformer so that the transfer load intersects the
pumping purposes).
maximumA loadpower
connected VM , IM toofathe
pointdirectly PV array
array (Figure
will very8).rarely extract the maximum
The transformer equations power
are as
from
follows:the array (Figure 7).
In order to extract the maximum available power [17], we can match the load to the
PV array by means of a POPI two-port k (t )
vi  element,  vo  by means of a transformer,
0 example,
for
=
 i  varies
 0 with k ( t )   io 
because the I–V curve of the PV array −1 the insulation, and the load parameters
(26)
 i  
may vary as well, so dynamic matching is necessary. Dynamic matching can be achieved

Ii A K:1 C IO

r
 In order to extract the maximum available power [17], we can match the load to the
PV array by means of a POPI two-port element, for example, by means of a transformer,
because the I–V curve of the PV array varies with the insulation, and the load parameters
Energies 2021, 14, 8025 may vary as well, so dynamic matching is necessary. Dynamic matching can be achieved 9 of 19
by controlling the transfer ratio of the transformer so that the transfer load intersects the
maximum power point VM , IM of the array (Figure 8). The transformer equations are as
by controlling the transfer ratio of the transformer so that the transfer load intersects the
follows:
maximum power point VM , I M of the array (Figure 8). The transformer equations are
  i  = ( )
as follows: v k t 0  vo 
vi   k(t)   i vo (26)
 

ii
 ii 
=  0 k−
−10
1 ( t )   o 
io
(26)
0 k (t)

Ii A K:1 C IO

rp
PV Vi VO
Array E

B TVT D
PVarray–load
Figure8.8.PV
Figure array–loadcoupling
couplingby
bymeans
meansofofcontrolled
controlledtransformer.
transformer.
Energies 2021, 14, x FOR PEER REVIEW 11 of 23

Bytransferring
By transferringthe
theload
load to
to the
the input
input terminals
terminals of
ofthe
thetransformer,
transformer,the
theequivalent
equivalentcircuit
cir-
is given by the following (Figure 9):
cuit is given by the following (Figure 9):
IO

rp’
PV Vi
Array E’

Figure
Figure9.9.PV
PVarray
arrayconnected
connectedtotothe
thetransferred
transferredload.
load.

Thecircuit
The circuitequation
equationisisgiven
givenby
bythe
thefollowing:
following:

vivi==iiiirrPp 0'++EE' 0 (27)

(27)

The
Thetransferred resistance rr 0' and
transferredresistance voltageE0 Eare
andvoltage ' are given
given by by
thethe following:
following:

= rrpP ⋅·kk2((tt))
2
rr0 '=
(28)
(28)
= E ·⋅ k ((tt))
EE0 '=

Thecircuit
The circuitbecomes
becomesthe
thefollowing:
following:
2
vvi i== iiiikk2 rrPp ++kE
kE (29)
(29)

As we are interested in operating the circuit at the maximum power point V , I ,

As we are interested in operating the circuit at the maximum power point VMM, IM M ,
replacing the general k with that which enables operation at this power point, denoted
replacing the general
k∗ (t), results k with that which enables operation at this power point, denoted
in the following:
k ( t ) , results in the following:
*

I M r p k2 (t) + k∗ (t) E = VM (30)

I M rp k
2
(t ) + k (t ) E = V
*
M (30)

The solution is as follows:

 * 1  VM  VM 
2
rp 
k ( t ) =  +   + 4 
 2 E
  E  EM  

 (31)
 VM 
 RM = I 
Energies 2021, 14, 8025 10 of 19

The solution is as follows:

 " r  # 
2
 ∗
 k (t) = 1 VM VM r P

E + + 4 RM
 

2 E
Energies 2021, 14, x FOR PEER REVIEW 12 of(31)
23
 

 R = VM 

M IM

9. Matching a PV Array to a Load by Means of LFR

A Ii IO E IO
The matching of a PV array to a load can be achieved by coupling the array to the
load by means of a loss-free resistor (LFR) [18]. As described previously, the LFR is a
power-conservative two-port network (POPI) whose input terminals are buffered from the
P=VOiO
output. The inputRterminal has a resistive
VO characteristic, which is modeled by a resistive
Vi E P
element R E , and the output has a power source with features in the i − v curve. The 12
Energies 2021, 14, x FOR PEER REVIEW power
of 23
that is absorbed by r E is transferred to the output terminals and creates the power source
ii2 ⋅ RE 10).
P =Figure
characteristic (see
VO
B D
A Ii LFR IO E IO

(a) (b)

Figure 10. (a) The LFR model; (b) power source characteristic. P=VOiO
Vi RE P VO

Matching is achieved by controlling RE such that its value is given by the following:
2
P = i ⋅ RE
i
*
RE = RE = RM VO
B LFR D
VM (32)
(a) RM = (b)
IM
Figure10.
Figure (a)The
10.(a) TheLFR
LFRmodel;
model;(b)
(b)power
powersource
sourcecharacteristic.
characteristic.
where VM and IM are the voltage and current of the PV array at MPP, respectively. Equa-
Matching is achieved by controlling R E such that its value is given by the following:
tion (30) implies
Matching is the maximum
achieved by controlling RE such
power transfer to the load.
that its value is given by the following:
R E = R∗E(Figure
When matching with a loss-free resistor = R M 11), we control the value of the em-
(32)
ulated resistance, rE , of the loss-free R RE== VIRMMsuch
*
resistor
M E
= Rthat
M
*
rE = rE , which implies the maxi-
mum power transfer. V (32)
where VM and I M are the voltage and current RM = Mof the PV array at MPP, respectively. Equa-
tion (30) implies the maximum power transfer I M to Vthe load.
*
When matching with a loss-free Rresistor E
= RM =(Figure
M
11), we control the value of(33) the
IM
V
emulated I
resistance, r , of the loss-free resistor such
where M and M are the voltage and current of the PV array at MPP,
E that r E = r ∗ , which implies the
E respectively. Equa-
maximum
An power transfer.
advantage of matching is the fact that the output is buffered from the input ter-
tion (30) implies the maximum power transfer to the load.
minals in the two-port network, so ∗ variationsVdo
load M
not affect the matching.
When matching with a loss-free RresistorE = R M(Figure
= 11), we control the value of the em- (33)
IM
ulated resistance, rE , of the loss-free resistor such that rE = rE , which implies the maxi-
*

A C
mum power transfer.

* VM rp
RE = RM = (33)
IM
PV P RE VO
Array
An advantage of matching is the fact that the output is buffered from the input ter-
E
minals in the two-port network, so load variations do not affect the matching.

B LFR D
A C
Figure11.
Figure PVarray–load
11.PV array–loadsystem
systemmatched
matchedby
bymeans
meansofofLFR.
LFR.

An advantage of matching is the fact that the output is rp buffered from the input
10. Findinginthe
terminals theMatching Parametersoofload
two-port network, the variations
POPI by Load
do notCurrent Measurement
affect the matching.
E P
In the conventional control of aRphotovoltaic
PV VO
system, we determine the instantaneous
Array power of the array by measuring the output voltage and current of the array,
maximum E
as well as their multiplications, and varying the parameters of the matching network to
determine the maximum value. LFR
For a load consisting of a resistance in series with a voltage
B D
Energies 2021, 14, 8025 11 of 19

10. Finding the Matching Parameter of the POPI by Load Current Measurement
In the conventional control of a photovoltaic system, we determine the instantaneous
Energies 2021, 14, x FOR PEER REVIEW
maximum power of the array by measuring the output voltage and current of the13array,
of 23
as well as their multiplications, and varying the parameters of the matching network to
determine the maximum value. For a load consisting of a resistance in series with a voltage
source, the voltage and current are increasing functions of the input power [19]:
vload = vmax  
P = pmax →v = vmax (34)
P = pmax → load iload = imax
iload = imax  (34)

Therefore, coupling a source with nonlinear characteristics to any two ports results
Therefore, coupling a source with nonlinear characteristics to any two ports results in
in a power-conservative two-port network whose internal parameter x can be controlled,
a power-conservative two-port network whose internal parameter x can be controlled, as
as illustrated in Figure 12.
illustrated in Figure 12.
A C

r I
POPI
General With cont. P VO i(load)
Source parameter
X
E
x
x*
B D

Control

(a) (b)

12.(a)(a)AA
Figure12.
Figure general
general source
source coupled
coupled to the
to the loadload by means
by means of POPI
of POPI withwith
cont.cont. X . (b)
parameter,
parameter, X.
The loadload
(b) The current as a as
current function of X
a function of. X.

We do not need to multiply the two measurements VS , IS in order to find the power
and We
thendoapply
not need
the to multiply
load the two its
to determine measurements
maximum value:
VS , IS itin isorder to findtothe
sufficient power
vary the
and then apply
parameter x. the load to determine its maximum value: it is sufficient to vary the pa-
rameter x.
By measuring the load current, when I reaches its peak, we know that the source
By measuring
delivers the load
the maximum current,
available whenThe
power. I reaches its peak,
parameter x thatwe know that
produces this the source
maximum
delivers
power isthe
denoted by X ∗available
maximum . This typepower. The requires
of control parameter x measurements
fewer that produces this maximum
(output current
insteadisof
power denoted by X * . This
input voltage and type
current) and multiplication,
of control requires fewerenabling
measurementsthe construction
(output cur-of
simpler
rent PV of
instead systems. It should
input voltage andbecurrent)
noted that
andthe practical process
multiplication, enabling described above is also
the construction of
suitablePV
simpler forsystems.
a matching network
It should bewith
noted some
thatlosses with constant
the practical processefficiency.
described above is also
suitable for a matching network with some losses with constant efficiency.
11. Maximum Energy Extraction from a Decaying Source
An example
11. Maximum of a situation
Energy requiring
Extraction from a dynamic
Decayingmatching
Source is that of storing the maximum
available energy in loss-free storage. Let
An example of a situation requiring dynamic VS be a pulse sourceisthat
matching thatdecays exponentially
of storing the maxi-
with time, with internal resistance Rs coupled to a loss-free storage system by means of a
mum available
controlled energy in loss-free
power-conservative storage.
network Let VSasbe
(POPI), a pulse
shown in source
Figure that
13. decays exponen-
withVtime,
tially For S ( t ), given by Equation
with internal (33), and
resistance an internal
Rs coupled loss-freerSstorage
to aresistor, , the maximum
system by power
meansis
given by Equation (34).
of a controlled power-conservative network (POPI),t as shown in Figure 13.
vS (t) = E · e− τ (35)
rS A Ii 1 2 E2 − 2t IO C
pmax (t) = VS (t) = e τ (36)
4rS 4rS
IO
In order to transfer the maximum available energy to the storage element, we must
extract
Vs the maximum availableVi power from thePOPIsource during Storage
VOthe entire process:

t Z∞
E2
wmax = pS (t)dt = τ (37)
Decaying Source
8rS
B t =0 D

Control
 For , given
By measuring
S thebyload
Equation (33),
current, and an
when I reaches
internalits resistor,
peak, we S , know
the maximum
that the power
source
delivers the maximum available
is given by Equation (34). power. The parameter x that produces this maximum
power is denoted by X * . This type of control requires t
fewer measurements (output cur-
−
rent instead of input voltage and current) and multiplication, enabling the construction of
v (t ) = E ⋅ e τ (35)
simpler PV systems. It should be noted Sthat the practical process described above is also
Energies 2021, 14, 8025 12 of 19
suitable for a matching network with some losses with 2constant 2t efficiency.
1 2 E −τ
pmax ( t ) = VS ( t ) = e (36)
11. Maximum Energy Extraction from a Decaying 4rS 4rS
Source
For capacitiveofstorage c whose initial voltage vc |t= = vc0is→
is matching 0 (Figure 14),the
we maxi-
Anorder
In example a situation
to transfer the requiring
maximum dynamic
available energy that
to the storage of element,
storing we have
must
the following:
extract the maximum available
mum available energy in loss-free power from the
Zt storage. Let S be
Vsource during the
a pulse source entire process:
that decays exponen-
E2 2 − 2t

wc ( t ) =
tially with time, with internal pSmaxRs
resistance (∞t)coupled − e τ storage system by means
dt = toτEa 1loss-free (38)
of a controlled power-conservative w =
network
t=0 max p s
( t )
(POPI),
8rS=
dt as
8rS
τ
shown in Figure 13. (37)
t =0

For capacitive rstorage

S c whose
A Ii
initial voltage is vc ItO= =Cvc 0 → 0 (Figure 14), we
have the following:
IO

t
POPI E
2
 −
2t
 Storage(38)
Vs wc (Vti ) =  pS max ( t ) dt = τ  1 − e VτO 
t t =0
8rS  
Therefore, the voltage at terminals C − D is given by the following:
Decaying Source B D
2 E 2
−
2t
 E τ  −
2t

vc ( t ) = τ 1 − e τ  = 1 − e
τ
 (39)
C 8rS   2 rS C  
Control

The process is controlled by a time-varying transformer whose transfer ratio is k ( t )

.Figure13.
Figure 13. Decaying
Decaying source
source matched
matchedto
toaageneral
generalstorage
storagesystem
systemby
byaaPOPI
POPInetwork.
network.

rS A Ii K:1 IO C

IO

Vs Vi VO C
t

Decaying Source B TVT D

Control

Figure 14. Decaying source matched to capacitive storage by means of a TVT.

Figure 14. Decaying source matched to capacitive storage by means of a TVT.

k ( t ) is given
Therefore, the by
voltage at terminals C − D is given by the following:
the following:
s rt
2 E2  E − 2t  E − τ  2t

vc (t) = τ 1 −2e τ = e τ 1 − e− τ (39)
Ck8r( tS) = = 2 rS C
v2 ( t ) τ  −
2t
 (40)
The process is controlled by a time-varying transformer  1 − e τ
whose transfer ratio is k(t).
k(t) is given by the following: rS C  
t
E/2 e− τ
k(t) = = r  (40)
v2 ( t ) − 2t

r C 1−e
τ τ
S

When matching with a loss-free resistor, the maximum power energy transfer is
achieved. R E has a constant value that equals rS :

Re = RS ∀t (41)
 When matching with a loss-free resistor, the maximum power energy transfer is
achieved. RE has a constant value that equals rS :
Energies 2021, 14, 8025 13 of 19
Re = R S ∀ t (41)

In a case in which the storage device is a battery with open-circuit voltage, E , and
In a case in which the storage device is a battery with open-circuit voltage, E, and
leakagemodeled
leakage modeledby byparallel
parallelresistance,
resistance, rb (Figure
rb (Figure 15), 15),
it is it
of is of interest
interest to determine
to determine the
the time
time
at at which
which the stored
the stored energyenergy
reachesreaches its maximum.
its maximum. This when
This occurs occursthe
when the maximum
maximum power
of the source
power of the equals
sourcethe leakage
equals power. power.
the leakage

rS A Ii IO C

IO

Vs Vi POPI VO rb E
VO

Decaying Source B D

Control

Figure15.
Figure 15. A
A decaying
decaying source
source coupled
coupledto
toaastorage
storagebattery
batteryby
byaageneral
generalPOPI
POPInetwork.
network.

The timet1t1ininwhich
Thetime whichthe
themaximum
maximum source
source power
power equalsequals thethe
leakage losses
leakage is given
losses by
is given
the following:
by the following: 2 E2 − 2t 2
PSmax − Prb losses ⇒ Er → 4r e τ − Er = 0
b S b
2 2 2t 2 (42)
E E −τ E
 
t1 = Pτ2 ln −4rrPb  → e − = 0
S max Sr lossesb
rb 4 rS rb
After t = t1 , the power loss is greater than the power supplied by the source, so(42) at
t = t1 , the energy that accumulates τ in  therb battery reaches its maximum value:
t1 = ln  
2 t−t4 rS 
R1 2
wstored = PSmax (t)dt − Er t1 =
After t = t1 , the power loss is tgreater
b
=0 than the power supplied by the source, so at
h  2t1  i
E2 τ − rS
t = t1 , the energy that accumulates
= 1 − e τ
rb 8 in the batteryrbreaches − t 1 = its maximum value:
(43)
h  2t1   i
E2 τ − r r
= rb 8 1t −− t e τ − rb 2 ln
S τ
2
b
=
1
E 4rS

= Er τ2 14 t1= 0 − e− τ − rrSrbτ2 ln 4rrb

w h
2stored =  P S max
( 1)
2tt dt − t 1
=
 i
b b S

E τ   rS 
2 2 t1
−
12. Battery Charger =   1 − e  − t1  =
τ

rb  8 
A battery charger can be constructed  rb of a circuit based on the POPI concept.
by means
(43)
E  τ  by a− τDC
In many cases, the charger is supplied  source
rS τ V S rwith  a fuse that is destroyed at
2 2t 1

currents higher than IF . The=typical  1 − e can  − be modeled  =a voltage source Eb with

b
 8 battery ln  by
rb    rb
2  4
series resistance r; in most cases, the voltage drop in r can be neglected r S   (compared to Eb ).
Previous studies have shown E τ 1 
that such  rS τ  rb  
source–load coupling by means of a trans-
2 2 t 1
−
=
former is associated with sensitivity   1 − e [4]
problems τ
 −(of source
ln  and   load currents to voltage
2  4 on a gyrator
variations), so applying a circuitrb based  rbor2LFR 4concept
rS   is preferable.
The coupling circuit controls the source current such that iS ≤ IF . In an extreme case in
which iS equals
12. Battery IF , the highest power transfer is achieved, resulting in the highest charging
Charger
rate. Figure 16 illustrates this process for a gyrator-based charger.
The source current is given by the following:

iS|iS ≤ IF = Yeq VS + Ieq = g2 r + gEb (44)

In most cases, r can be neglected, so (42) can be reduced to the following:

iS|iS ≤ IF ' gEb (45)

Energies 2021, 14, x FOR PEER REVIEW 16 of 23
Energies 2021, 14, 8025 14 of 19

A battery charger can be constructed by means of a circuit based on the POPI concept.
The highest power transfer is achieved when iS = IF , resulting in the fastest battery
In many cases,∗ the charger is supplied by a DC source V with a fuse that is destroyed at
charging, so g that produces this current is given by theS following:
currents higher than I F . The typical battery can be modeled by a voltage source Eb with
I
g∗ = the
series resistance r ; in most cases, Eb voltage drop in r can be neglected (compared
F
to
∗
(46)
Eb ). g = g → PS = PSmax = IF VS
Previous studies have shown that such source–load coupling by means of a trans-
Then, the battery charging current is as follows:
former is associated with sensitivity problems [4] (of source and load currents to voltage
variations), so applying a circuit based on a gyratororVLFR concept is preferable.
S
i ' PSmax/Eb ' IF (47)
The coupling circuit controls V that iS ≤ I F . In an extreme case
the source current such
bmax
b
in which i S equals I F , the highest power transfer is achieved, resulting in the highest
Controlling the value of g such that (45) is satisfied results in the fastest possible
charging rate. Figure 16 illustrates this process for a gyrator-based charger.
charging rate of the battery.
IF IF
A C A
g(t)
r

Vs is ib
Vs is Yeq=g2r Ieq=gEb
Eb

B Gyrator D B

Control
iS ≤ I F

(a) (b)

Figure
Figure 16.
16. (a)
(a) Gyrator-based
Gyrator-based battery
battery charger.
charger. (b) The equivalent circuit.
Energies 2021, 14, x FOR PEER REVIEW
13. Marched-Mode Operation 17 of 23
The source current is givenofbyRF the Power
following:Amplifier
Basic applications of an RF power amplifier include driving to another high-power
2
source, driving a transmitting antenna iS i ≤ I =and S
+ I eq = gmicrowave
YeqVSexciting
F
r + gEb cavity resonators. Among (44)
theseAapplications,
simplified modeldriving of atransmitter
transmission systemare
antennas consisting
the mostofwell a transmitter
known. and antenna
that isIn most
Afed directly
simplified byr the
cases,model can be
RF
of neglected,
apower
transmission sosystem
amplifier (42) canconsisting
output be(the
reduced of to
output the following:
a stage of the and
transmitter transmitter)
antenna
isthat
shown
is fedindirectly
Figure by17a.the
InRFthispower
figure,amplifier
the RF poweroutputamplifier
(the output is modeled by atransmitter)
stage of the “black box”, is
iS i ≤ I  gEb (45)
which
shownincludes
in Figure the arrangement
17a. In this figure, of various
the RF elements
power amplifier(amplifying
S F devices,byresonant
is modeled a “blacktanks,
box”,
LC filters,
which The segments
includes
highest the of transmission
arrangement
power transfer is lines, etc.).
ofachieved
various Thisi “black
elements
when box” extracts
(amplifying
= I F , resultingdevices,DC
in the power
resonant
fastest from
tanks,
battery
aLCDCfilters,
source, E* , and
segments ofconverts
transmission it into high-frequency
lines,
S
etc.). This “black power thatextracts
box” is transferred
DC power to the
froman-a
charging,
tenna.
DC source, soE,gand that produces
converts thishigh-frequency
it into current is given by the
power following:
that is transferred to the antenna.

* IF
g = Antenna ZS c
Eb c
(46)
*
g = g → PS = PS max = I FVS Vs Za
E PRF
Then, the battery charging current is as follows:
d
d
V 
Transmitter Antenna
PS max Equivalent Impedance
ib max   IF  S  (47)
Eb
HF Signal
 Vb 
(a) (b)
Controlling the value of g such that (45) is satisfied results in the fastest possible
Figure17.
Figure (a)The
17.(a) The“black
“blackbox”-based
box”-basedmodel
modelof
oftransmission
transmissionsystem.
system. (b)
(b)Equivalent
Equivalentcircuit
circuit as
as viewed
viewed
charging rate of the battery.
fromthe
from theload
load (antenna).
(antenna).
13. Marched-Mode Operation ofelectrical
The simplest conventional RF Powerequivalent
Amplifiercircuit (Figure 17b) of this model
The simplest conventional electrical equivalent circuit (Figure 17b) of this model con-
consists
Basicofapplications
a high-frequency voltage
of an RF powersource V S , with
amplifier internal
include impedance
driving to anotherZ S high-power
(the electri-
sists
cal of a high-frequency
equivalent of the voltage
“black box” source
output Vas, with
viewed internal
from impedance
the antenna)
source, driving a transmitting antenna and Sexciting microwave cavity resonators.Zand
S
(the
the electrical
antenna
Among
impedance
equivalent Z [20].
of the
these applications,
a “black box”
driving output as antennas
transmitter viewed from are the antenna)
most welland the antenna imped-
known.
ance Z a [20].
As in the case of the previously analyzed AC source, maximum power is extracted
*
when Z S = Z a (Figure 18):

rS = ra
 when Z S = Z a (Figure 18):

rS = ra
(48)
xS = − xa
Energies 2021, 14, 8025 15 of 19
In this case, the efficiency, as calculated previously, is as follows:
2
Pa i ra
η=
= 2
∗ S
i ( ra +AC
As in the case of the previouslyPanalyzed rS ) source, maximum power is extracted
(49)
when Z S = Z a (Figure 18):
rSra==rraS  η = 50%
(48)
xS = − x a
ZS c ZS c

Vs ZS* = Z’a  Matching Za Vs Z’a

Network
ZS* = Z’a
d d
(a) (b)

Figure 18.
Figure Source–load
18.Source–load matching
matching forfor maximum
maximum power
power transfer.
transfer. (a) Principle
(a) Principle circuit.circuit.
(b) The(b) The
equiv-
equivalent
alent load. load.

In this case, the efficiency, as calculated previously, is as follows:

The total conversion efficiency of the transmission system from DC into a high fre-
quency is given by the following:P i2 r a
a
η= = 2
(49)
PS i (r a + r S )
r a = rS ⇒ η = 50%

The total conversion efficiency of the transmission system from DC into a high fre-
quency is given by the following:
P
ηtot = RF (50)
Pdc
where PRF is the high-frequency power delivered to the load (antenna) by the “black box”
(Figure 17a), and Pdc is the DC power delivered by the voltage source E to the “black box”
a − b terminals.
As shown previously (in the case of the AC source with internal impedance loaded by
AC impedance), the theoretical model (Figure 5) predicts an efficiency of 50% in matched-
mode operation. On the other hand, although the measured overall system efficiency is
below this limit in many cases, it often exceeds 50% in other cases. Clearly, efficiencies lower
than this value (even in matched mode) are expected, as practical systems encounter losses
that are not predicted by the theoretical model. However, situations in which practical
systems exhibit efficiencies greater than 50% are surprising; these cannot be explained by
the traditional model.
This contradiction can be resolved by replacing the resistive part of the source
impedance with a series loss-free resistor (SLFR) (Figure 19). The basic LFR (described in
the previous section) is a two-port network with an element, R E , at its input terminals and
an element defined as a power source at its output, whose power equals that absorbed by
the resistance part R E (Figure 19a).
The SLFR concept, introduced by Ivo Barbi [21], is a topological transformation of the
basic LFR model, in which the emulated resistance R E is connected between the input and
the output and the power source, P, in parallel to the input (Figure 19b).
The losses that occur during the operation of a practical system can be modeled
by slightly modifying the SLFR, denoted semi-SLFR. In this case, we assume that only
part of the power that is absorbed by R E is transferred to the input SLFR terminals, i.e.,
P = αPRE (0 ≤ α ≤ 1) (see Figure 19c).
A practical system experiences some losses, so the SLFR model can be modified to
one with some losses by multiplying the power transferred from the other R E to the power
source by an efficiency transfer factor α; i.e., the power that is transferred from R E and
creates the power source (of the SLFR), P, whose power is returned to VS , is as follows:

P = αPRE (0 ≤ α ≤ 1) (51)
 pedance with a series loss-free resistor (SLFR) (Figure 19). The basic LFR (described in the
previous section) is a two-port network with an element, RE , at its input terminals and
an element defined as a power source at its output, whose power equals that absorbed by
Energies 2021, 14, 8025 the resistance part RE (Figure 19a). 16 of 19
The SLFR concept, introduced by Ivo Barbi [21], is a topological transformation of the
basic LFR model, in which the emulated resistance RE is connected between the input
Energies 2021, 14, x FOR PEER REVIEW
and theReplacing
output andthe resistive
the power part
source, P , in parallel
of Z S (Figure 17) withtoa the
semi-SFLR results19b).
input (Figure in the modified
19 of 23
transmission system model shown in Figure 20. The power transfer efficiency
The losses that occur during the operation of a practical system can be modeled by of that circuit
is as follows:
slightly modifying the SLFR, denoted semi-SLFR. In this case, we assume that only part
ηVs ≡ PSα/Pa; i.e., the power that is transferred from (52)
power source
of the power that by an
is efficiency by RE isfactor
absorbed transfer transferred to the input SLFR terminals, i.e.,
P R=
where and
E α
PScreates
is the total
Pload( 0(antenna).
≤ αthe
(high-frequency)
) (see Figure
≤ 1power (of power
source 19c). supplied
the SLFR), by V S , power
P , whose and Pa is
is returned
the powertoabsorbed
VS , is as
by the RE
follows:
1 3 P = α PRE (0 ≤ α
1 ≤ 1) RE 3 (51)

Replacing
RE
the
P
resistive part of ZS (Figure
Load
17) with a semi-SFLR results in the modi-
VS V O V S V P P = PRE O Load
fied transmission system model shown in Figure 20. The power transfer efficiency of that
P =follows:
circuit is as PRE
2 LFR 4 2 SLFR 4
PS
(a)
ηVs ≡ Pa
(b)
(52)

where
1 PS is theR total (high-frequency)
E 3 power supplied by VS , and Pa is the power ab-
I O

sorbed by the load (antenna). P=Const

VS In the
P
matched-mode condition,
P = α PRE V η is given by the following:
Load Vs O

ηVs = 1 VO

2 SEMI-SLFR 4 (2 −α ) (53)
(d)

As α may vary in the range 0 ≤ α ≤1, the power transfer efficiency may vary be-
(c)

Figure
Figure 19.
tween19. LFR
50LFR and
andand SLFR
100% in models.
SLFR (a)
(a)LFR,
LFR,(b)
the matched-mode
models. (b)SLFR, (c)
(c)semi-SLFR
condition.
SLFR, and
semi-SLFR (d)(d)
and power source
power characteristic.
source character-
istic.

A practical system R E = RS
experiences some X
losses,
S so the
c SLFR model can be modified to
one with some losses by multiplying the power transferred from the other RE to the

Xa
P = α PRE
Vs P 0 ≤α ≤1
ra

SEMI-SLFR d

Transmitter Equivalent Circuit Antenna

Figure20.
Figure 20. Simplified
Simplified equivalent
equivalent circuit
circuitof
oftransmission
transmissionsystem
systembased
basedon
onthe
thesemi-SLFR
semi-SLFRconcept.
concept.

In
Asthe
thismatched-mode
model predictscondition, ηVS is given
that the efficiency by internal
of the the following:
model (consisting of VS , ZS

and Za ) can exceed a value of 50%, there

ηVS =is 1no
/(2 contradiction
− α) between that result and(53)
the
total measured efficiency (it is impossible for the total efficiency to be higher than that of
As α may vary in the range 0 ≤ α ≤ 1, the power transfer efficiency may vary between
the internal model). As RF power amplifiers find applications in a variety of systems (com-
50 and 100% in the matched-mode condition.
munication, microwave, wireless power transfer), further understanding their efficiency
As this model predicts that the efficiency of the internal model (consisting of V S ,
limit is important.
Z S and Z a ) can exceed a value of 50%, there is no contradiction between that result and
We do not claim that practical systems actually contain an LFR; we only claim that
the total measured efficiency (it is impossible for the total efficiency to be higher than
this model describes the behavior of an efficient type of such systems operating in
that of the internal model). As RF power amplifiers find applications in a variety of
matched-mode
systems conditions.microwave,
(communication, The fact thatwireless
there arepower
some transfer),
elements that haveunderstanding
further resistive char-
acteristics that do not produce losses
their efficiency limit is important. (in principle only), such as segments of transmission
lines and resonant tanks [22–24] in the RF power amplifier, motivated us to propose the
model described above.

14. Simulation Results of Transmission System

 A practical transmission system consisting of an HF power amplifier and antenna
was designed and simulated (using ADS simulation software) in order to evaluate the
power transfer efficiency (of that system) in the matched-mode condition (Figure 21).
The system was designed to be operated at a frequency of 250 MHz and powered by
Energies 2021, 14, 8025 17 of 19
a 28 V DC voltage source.
Energies 2021, 14, x FOR PEER REVIEW The transmitter was designed to have an output impedance that is purely resistive 20 of 23
(50 Ω), and the load is a variable pure resistive element that represents the antenna.
It
Wewas do found
not claim thatthat
the practical
maximum poweractually
systems was transferred
contain an to LFR;
the load
we onlywhenclaim
the load
that
resistance
thisAmodel was equal
describes to
the the output
behavior impedance
of an efficient of the
type transmitter
of such systems (50 Ω), and
operating the
inreflection
matched-
practical transmission system consisting of an HF power amplifier and antenna
coefficient
mode tended The
conditions. to zero (Figure
fact that there 22).areVoltage
some and current
elements waveforms
that have resistive (in the matched
was designed and simulated (using ADS simulation software) in order to characteristics
evaluate the
condition) at the transistor drain and voltage on the load (antenna) are shown in Figure
power transfer efficiency (of that system) in the matched-mode condition (Figure lines
that do not produce losses (in principle only), such as segments of transmission 21). and
23.
resonant tanks [22–24] in the RF power amplifier, motivated us to propose the model
The system was designed to be operated at a frequency of 250 MHz and powered by
At
described that point,
above. the output RF power was 10.7 W, and the DC supply current and volt-
a 28 V DC voltage source.
age were 0.51 A and 28 V, respectively, so the total efficiency was approximately 75%.
The transmitter was designed to have an output impedance that is purely resistive
14. Simulation
Based on these Results of Transmission
results, it is clear that System
in matched-mode operation, the transfer effi-
(50 Ω), and the load is a variable pure resistive element that represents the antenna.
ciency can be significantly higher than 50%, so our proposed model based on antenna
the loss-free
It was found that the maximum power wasoftransferred
A practical transmission system consisting an HF power to amplifier
the load and when the load was
resistor
designed concept
and is suitable
simulated for a ADS
(using simplified description
simulation software) of real
in transmission
order to evaluatesystem
the behav-
power
resistance was equal to the output impedance of the transmitter (50 Ω), and the reflection
ior.
transfer efficiency (of that system) in the matched-mode condition (Figure 21).
coefficient tended to zero (Figure 22). Voltage and current waveforms (in the matched
condition) at the transistor drain and voltage on the load (antenna) are shown in Figure
23.
28V DC
At that point, the output RF power was Supply 10.7 W, and the DC supply current and volt-
L5=80nH
age were 0.51 A-2.7V and DC 28 V, respectively, so the total efficiency was approximately 75%.
Bias
Based on these results, it isL2=250nH clear that inTLmatched-mode operation,
L4=26.3nH the transfer effi-
C6=220nF

ciency can be significantly higher than 50%, so our proposed model based on the loss-free
C4=9.7pF
resistor concept is suitable for a simplified T2G6000528-Q3 description of real L6=2.9nH
QORVO
transmission system behav-
C3=9.44pF
ior. L1=22nH TL TL
RL=50 Ω
L3=10.4nH C5=15.4pF

C1=220nF
RF Signal
250MHz R1=100 Ω
(sin) 28V DC
Supply
L5=80nH
-2.7V DC
Bias
C2=100pF L4=26.3nH C6=220nF
L2=250nH TL

Practical transmission
Figure 21. Practical Ω load
transmission system with a 50 Ω C4=9.7pFas an antenna.
T2G6000528-Q3 L6=2.9nH
QORVO C3=9.44pF
The system was designed to be operated at a frequency of 250 MHz and powered
L1=22nH TL TL
RL=50 Ω by a
L3=10.4nH C5=15.4pF
28 V DC voltage source.
C1=220nF
RF Signal The transmitter was designed to have an output impedance that is purely resistive
250MHz
(sin) Ω), and the load is a variable pure resistive element that represents the antenna.
R1=100 Ω
(50
It was found that the maximum power was transferred to the load when the load
resistance was equal to the output C2=100pF
impedance of the transmitter (50 Ω), and the reflection
coefficient tended to zero (Figure 22). Voltage and current waveforms (in the matched
Figure 21. Practical
condition) at thetransmission system
transistor drain andwith a 50 Ωon
voltage load
theasload
an antenna.
(antenna) are shown in Figure 23.

Figure 22. The output LFR power as a function of the load resistance (a). The absolute value of the
reflection coefficient Γ as a function of the load resistance (b).

Figure
Figure 22.22. The
The output
output LFRLFR power
power as as a function
a function of of
thethe load
load resistance
resistance (a).(a).
TheThe absolute
absolute value
value of the
of the
coefficient |Γ| as
reflection
reflection coefficient Γ
as aa function
function of
of the
the load
load resistance
resistance (b).
(b).
At that point, the output RF power was 10.7 W, and the DC supply current and voltage
were 0.51 A and 28 V, respectively, so the total efficiency was approximately 75%.
Based on these results, it is clear that in matched-mode operation, the transfer effi-
ciency can be significantly higher than 50%, so our proposed model based on the loss-free
resistor concept is suitable for a simplified description of real transmission system behavior.
Energies 2021,
Energies 14, 14,
2021, x FOR
8025PEER REVIEW 21 of
1823
of 19

Figure 23. The voltage and current waveforms of the transistor drain and of the output (antenna).
Figure 23. The voltage and current waveforms of the transistor drain and of the output (antenna).
15. Conclusions
15. Conclusions
This paper describes the operation of source–load systems in which the load is adjusted
toThis paper such
the source describes themaximum
that the operation available
of source–load
source systems
power isinextracted,
which the load(mode
which is ad- of
justed to the source such that the maximum available source
operation) is called matched-mode operation. A matched-mode operation is achieved by power is extracted, which
(mode of operation)
controlling the loadisparameter
called matched-mode
(if possible) or operation.
by using the A matched-mode operation is
proposed power-conservative
achieved by controlling the load parameter (if possible) or by using
two-port network, such as a gyrator or loss-free resistor (LFR). The “classical” calculation the proposed power-
conservative two-port network, such as a gyrator or loss-free
of the controllable load parameter X (or of the matching network parameter) is based resistor (LFR). The “classi-
cal”
oncalculation
∂Po of the controllable
∗ load parameter X (or of the∗ matching network param-
∂x = 0 ⇒ X = X , where Po is the load power and X is the value of X that results
= Pomax
Pobased
in is ∂P that thisP calculation
eter) on . ItO has= 0 been
 X found
*
= X , where method may be isvery
O is the load power and X
*
the difficult
value of to
apply if the source, ∂ x load or both are nonlinear. Therefore, we propose a simple method
X tothat
derive
resultstheinX∗ P = POmax . It The
parameter. calculation
has been found method
that thisiscalculation
based on the observation
method may bethatverythe
O
maximum power point (MPP) coordinates of the source, that is, the MPP voltage VM
difficult to apply if the source, load or both are nonlinear. Therefore, we propose a simple
and current I M , are characteristics of the source only, so there is no need to go through
*
method∂Pto derive the X parameter. The calculation method is based on the observation
the o = 0 ⇒ X = X ∗ process, because we can simply substitute the coordinates of the
that the∂xmaximum power point (MPP) coordinates of the source, that is, the MPP voltage
voltage and current of the source at maximum power point coordinates VM and I M into
VM the
andload
current IM , areincharacteristics
i − v curve order to obtain of the source
value of X ∗ . so there is no need to go through
only,
∂ P As controllable* loads are quite rare, matching by means of power-conservative two-
= 0  X can
theportOnetworks = Xbe process,
realized by because
meanswe of acan simply
variety substitute the coordinates
of switched-mode of the
converters operated
∂x
at a high switching frequency. As an example, the “natural” realization of a loss-free
voltage
resistorand(LFR)
current of the source at maximum power point coordinates VM and IM into
is presented.
i − v curve
the loadExamples in order toby
of matching obtain
means theofvalue *
theseofnetworks
X . include photovoltaic systems,
As controllable
transferring maximum loads are quite rare,
available energymatching by means source
from a decaying of power-conservative
to storage, and fast two-bat-
port networks
tery chargers. can be realized by means of a variety of switched-mode converters oper-
ated at aAhighsimpleswitching
modified frequency.
model of Asthe
an example,
RF powerthe “natural”
amplifier realization
loss-free of a loss-free
resistor concept is
proposed.
resistor (LFR) The model predicts that, in the matched-mode operation, the 50% efficiency
is presented.
limit impliedof
Examples bymatching
the classical bymodel
meanscan of be significantly
these networksexceeded, so the total efficiency
include photovoltaic systems,of a
system that
transferring includes an
maximum RF power
available amplifier
energy fromata its output stage
decaying source (communication,
to storage, and microwave,
fast bat-
and wireless
tery chargers. power transformer), in which maximum power is extracted, can also exceed
this limit.
A simple modified model of the RF power amplifier loss-free resistor concept is pro-
posed. The model predicts that, in the matched-mode operation, the 50% efficiency limit
Authorby
implied the classical Conceptualization,
Contributions: S.S.; methodology,
model can be significantly S.S. and
exceeded, so D.S.;
the software, M.A.; writing—
total efficiency of a
original draft preparation, S.S.; writing—review and editing, S.E. All authors have read and agreed
to the published version of the manuscript.
Energies 2021, 14, 8025 19 of 19

Funding: This research received no external funding.

Institutional Review Board Statement: Not applicable.
Informed Consent Statement: Not applicable.
Data Availability Statement: The circuit data and the simulation environment SW are included in
the paper.
Conflicts of Interest: The authors declare no conflict of interest.

References
1. Sigmunt, S.; Erickson, R.W. Canonical Modeling of Power Processing Circuits Based on the POPI Concept. IEEE Trans. Power
Electron. 1992, 7, 37–43.
2. Mehrdad, E.; Husain, I.; Bilgiç, M.O. Power Converters as Natural Gyrators. IEEE Trans. Circuits Syst. I Fundam. Theory Appl.
1993, 40, 946–949.
3. Mehrdad, E.; Husain, I.; Bilgic, M.O. Inverse Dual Converter (IDC) for High-Power DC-DC Applications. IEEE Trans. Power
Electron. 1993, 8, 216–223.
4. Sigmunt, S. Realization of Loss-Free Resistive Elements. IEEE Trans. Circuits Syst. 1990, 37, 54–60.
5. Singer, S. Gyrators Application in Power Processing Circuits. IEEE Trans. Ind. Electron. 1987, 3, 313–318. [CrossRef]
6. Cid-Pastor, A.; Martinez-Salamero, L.; El Aroudi, A.; Giral, R.; Calvente, J.; Leyva, R. Synthesis of Loss-Free Resistors Based on
Sliding-Mode Control and Its Applications in Power Processing. Control. Eng. Pract. 2013, 21, 689–699. [CrossRef]
7. Angel, C.; Martinez-Salamero, L.; Alonso, C.; El Aroudi, A.; Valderrama-Blavi, H. Power Distribution Based on Gyrators. IEEE
Trans. Power Electron. 2009, 24, 2907–2909.
8. Sigmund, S.; Beck, Y.; Martinez-Salamero, L. Energy Storage and POPI Networks in Power Processing Systems. IEEE Trans. Power
Electron. 2019, 35, 1425–1439.
9. Hanane, Y.; Aroudam, E. Standalone Photovoltaic System with Maximum Power Point Tracking: Modeling and Simulation. Int. J.
Syst. Dyn. Appl. IJSDA 2018, 7, 94–111.
10. Arghyadip, B.; Bhattacharyya, B.K. A Standalone Solar Photovoltaic Power Generation System with Automated Bidirectional
Battery Charging Discharging Control. In Proceedings of the 2016 2nd International Conference on Advances in Electrical,
Electronics, Information, Communication and Bio-Informatics (AEEICB), Chennai, India, 27–28 February 2016; IEEE: Chennai,
India, 2016; pp. 121–125.
11. Arulkumar, K.; Palanisamy, K.; Vijayakumar, D. Recent Advances and Control Techniques in Grid Connected PV System—A
Review. Int. J. Renew. Energy Res. 2016, 6, 1037–1049.
12. Yoash, L.; Erickson, R.W. High Weighted Efficiency in Single-Phase Solar Inverters by a Variable-Frequency Peak Current
Controller. IEEE Trans. Power Electron. 2015, 31, 248–257.
13. Shlomo, G.; Kuperman, A.; Sitbon, M.; Aharon, I.; Singer, S. Interfacing Renewable Energy Sources for Maximum Power
Transfer—Part I: Statics. Renew. Sustain. Energy Rev. 2014, 31, 501–508.
14. Reham, H.; El Aroudi, A.; Cid-Pastor, A.; Garcia, G.; Olalla, C.; Martinez-Salamero, L. Impedance Matching in Photovoltaic
Systems Using Cascaded Boost Converters and Sliding-Mode Control. IEEE Trans. Power Electron. 2014, 30, 3185–3199.
15. Adria, M.; Vidal-Idiarte, E.; Cid-Pastor, A.; Martínez-Salamero, L. Loss-Free Resistor-Based Power Factor Correction Using a
Semi-Bridgeless Boost Rectifier in Sliding-Mode Control. IEEE Trans. Power Electron. 2014, 30, 5842–5853.
16. Cid-Pastor, A.; Martinez-Salamero, L.; Leyva, R.; Calvente, J.; Giral, R. Design of Photovoltaic-Based Current Sources for
Maximum Power Transfer by Means of Power Gyrators. IET Power Electron. 2011, 4, 674–682. [CrossRef]
17. Alon, K.; Averbukh, M.; Lineykin, S. Maximum Power Point Matching versus Maximum Power Point Tracking for Solar
Generators. Renew. Sustain. Energy Rev. 2013, 19, 11–17.
18. Singer, S.; Giral, R.; Calvente, J.; Leyva, R.; Martinez-Salamero, L.; Naunin, D. Maximum Power Point Tracker Based on a Loss
Free Resistor Topology. In Proceedings of the Fifth European Space Power Conference, Tarragona, Spain, 21–25 September 1998;
Volume 416, p. 65.
19. Shmilovitz, D. On the Control of Photovoltaic Maximum Power Point Tracker via Output Parameters. IEE Proc.-Electr. Power
Appl. 2005, 152, 239–248. [CrossRef]
20. Jussi, R. Power Waves and Conjugate Matching. IEEE Trans. Circuits Syst. II Express Briefs 2008, 55, 92–96.
21. Ivo, B. Series Loss-Free Resistor. Analysis, Realization and Applications. IEEE Trans. Power Electron. 2021, 36, 12857–12866.
22. Singer, S.; Smilovitz, D. Transmission Line-Based Loss-Free Resistor. IEEE Trans. Circuits Syst. I Fundam. Theory Appl. 1994, 41,
120–126. [CrossRef]
23. Ricardo, B.; Olalla, C.; Martinez-Salamero, L. Dynamic Modeling and Control of Self-Oscillating Parallel Resonant Converters
Based on a Variable Structure Systems Approach. IEEE Trans. Power Electron. 2016, 32, 1469–1480.
24. Robert, W.E. Fundamentals of Power Electronics; Chapman & Hall: Chapel Hill, NC, USA, 1997; Chapter 19—Resonant Conversion;
pp. 667–676.