Chapter Three: The Fourier Series

Aga Bayou

Adama Science and Technology University

July 2, 2023

Aga Bayou (ASTU) Signals and Systems Analysis July 2, 2023 1 / 27

Outlines

1 Introduction

2 Defining Equations of Fourier Series

3 Exponential Fourier Spectrum and Concept Of Negative Frequency

4 Properties of Fourier Series

5 The Equivalence of the Trigonometric and Exponential Series

6 Signal Approximation

Aga Bayou (ASTU) Signals and Systems Analysis July 2, 2023 2 / 27

Introduction

The importance of complex exponentials in the study of LTI systems

The response of an LTI system to a complex exponential input is the
same complex exponential with only a change in amplitude
Continuous time: e st −→ H(s)e st
Discrete time : z n −→ H(z)z n
eigenfunction of the system: A signal for which the system output is a
(possibly complex)(H(s) or H(z)) constant times the input
The system’s eigenvalue: the amplitude factor
A continuous time signal x(t) is said to be periodic if there is a
positive non-zero value of T for which

x(t + T ) = x(t) (1)

, for all t

Aga Bayou (ASTU) Signals and Systems Analysis July 2, 2023 3 / 27

Cont. . . .
Any periodic signal can be classified into harmonically related
sinusoids or complex exponential
∞
1 X
x(t) = a0 + (an cosnωt + bn sinnωt) (2)
2
n=1

Or
∞
X 2π
x(t) = ck e jkω0 t ,ω0 = (3)
T0
k=−∞

This decomposed representation is called FOURIER SERIES

Two type of Fourier Series Representation are there. Both are
equivalent to each other
1. Trigonometric Fourier Series
2. Exponential Fourier Series
Aga Bayou (ASTU) Signals and Systems Analysis July 2, 2023 4 / 27
Defining Equations of Fourier Series
A Fourier series may be defined as an expansion of a function in a
series of sines and cosines such as
1
x(t) = a0 + a1 cosωt + a2 cos2ωt + a3 cos3ωt + . . .
2
+ b1 sinωt + b2 sin2ωt + b3 sin3ωt (4)

Or
∞
1 X
x(t) = a0 + (an cosnωt + bn sinnωt) (5)
2
n=1

The coefficients are related to the periodic function x(t)

The terms with the coefficients a1 and b1 together, represent the
fundamental frequency component ω
Likewise, the terms with the coefficients a2 and b2 together, represent
the second harmonic component 2ω, and so on
Aga Bayou (ASTU) Signals and Systems Analysis July 2, 2023 5 / 27
Cont. . . .

NOTE:
The Fourier Series converges to x(t) at all points where x(t) is
continuous
The orthogonal relation for different values of the eigenvalue nω by
choosing the interval [0, 2π]
Z π (
0 : n ̸= m
sinnt sinmtdt =
−π π:n=m
Z π
sinnt cosmtdt = 0
−π
(
π
0 : n ̸= m
Z
cosnt cosmtdt =
−π π:n=m

Aga Bayou (ASTU) Signals and Systems Analysis July 2, 2023 6 / 27

Cont. . . .

By use of these orthogonality, we are able to obtain the coefficients by

multiplying eq 1 or 2 by cosnωt, and then integral from 0 to 2π
Z π
1 1
a0 = x(t)dt
2 2π −π
1 π
Z
an = x(t)cosnωtdt (6)
π −π
1 π
Z
bn = x(t)sinnωtdt
π −π

Aga Bayou (ASTU) Signals and Systems Analysis July 2, 2023 7 / 27

Cont. . . .

Dirichlet’s Conditions
x(t) is absolutely integrable over any, that is
Z T
|x(t)|dt < ∞
0

1. x(t) has a finite number of maxima and minima within any finite
interval of t
2. x(t) has a finite number of discontinuties within any finite interval of t,
and each of these discontinuities are finite
NOTE:
Dirichlet’s conditions are sufficient but not necessary conditions for the
Fourier Series representation

Aga Bayou (ASTU) Signals and Systems Analysis July 2, 2023 8 / 27

Cont. . . .

Exercises: Find the Fourier series of the function

a. x(t) = t, −π < t < π, x(t + 2π) = x(t)
b. x(t) = t(2 , −1 < t < 1, x(t + 2) = x(t)
0, −π < t < 0
c. x(t) =
1, 0 ≤ t < π
−t
d. x(t) = e , −π < t < π
e. x(t) = |t|, −π < t < π
Note (
0 −π < t ≤ 0
i x(t) = t + |t|, on [−π, π], x(t) =
2t 0<t≤π
(
1 + t, −2 < t ≤ 0
ii x(t) = 1 − |t|, on −2 < t ≤ 2, x(t) =
1 − t, 0 < t ≤ 2

Aga Bayou (ASTU) Signals and Systems Analysis July 2, 2023 9 / 27

Cont. . . .

Odd and Even Function

In the Fourier series corresponding to an odd function, only sine terms
can be present.
In the Fourier series corresponding to an even function, only cosine
terms (and possibly a constant which we shall consider a cosine term)
can be present.
Properties of even and a odd function
i (even) × (even) = even
ii (odd) × (odd) = even
iii (even) × (odd) = odd Ra Ra
iv If x(t) is even, then −a x(t)dt = 2 0 x(t)dt
Ra
iv If x(t) is odd, then −a x(t)dt = 0

Aga Bayou (ASTU) Signals and Systems Analysis July 2, 2023 10 / 27

Cont. . . .

Exercise: Find the Fourier coefficients corresponding to the function

and Write the corresponding Fourier series
(
0 −5 < t < 0
a. x(t) =
3 0<t<5
b. x(t) = |t| on −π < t < π
c. x(t) = 2 − t 2 , −2 < t < 2

Aga Bayou (ASTU) Signals and Systems Analysis July 2, 2023 11 / 27

Exponential Fourier Spectrum and Concept Of Negative
Frequency

A periodic signal is analysed interms of Exponential Fourier Series

in the following three stages
1. Representation of periodic signal
2. Amplitude and Phase Spectra of a Periodic Signal
3. Power content of a periodic signal
The complex Exponential Fourier Series representation of a periodic
signal x(t) with fundamental period T0 is given by;
∞
X 2π
x(t) = ck e jkω0 t ,ω0 = (7)
T0
k=−∞

Aga Bayou (ASTU) Signals and Systems Analysis July 2, 2023 12 / 27

Cont. . . .
Where C is known as the Complex Fourier Coefficient and is given by,
Z
1
Ck = x(t)e −jkω0 t dt (8)
T0 T0
R
where T0 denotes the integral over any one period and 0 to T0 or
−T0 /2 to T0 /2 is commonly used for the integration.
The equation 8 derived from multiplying both sides of 7 by e −jω0 t
and integrate over a time period both sides, and finally it becomes
Z T0 Z T0 ∞
X
−jω0 t
x(t)e = Ck e jkω0 t .e −jω0 t dt
0 0 k=−∞
∞
" #T0
X e j(k−1)ω0 t
= Ck (9)
j(k − 1)ω0 t
k=−∞ 0

Aga Bayou (ASTU) Signals and Systems Analysis July 2, 2023 13 / 27

Cont. . . .
When k = 1, the right hand side of (9) evaluated at lower and limit
yields zero
On the other hand, if k = 1,
Z T0
dt = T0
0

Thus;
Z T0
x(t)e −jω0 t = Cl T0
0
Z T0
1
Ck = x(t)e −jω0 t (10)
T0 0

When x(t) is real, then from Eq.(8) it follows that

C−k = Ck∗
Aga Bayou (ASTU) Signals and Systems Analysis July 2, 2023 14 / 27
Cont. . . .

Example
1. Compute the exponential Fourier series for the square waveform of
Figure below. Assume that ω = 1.

2. Compute the first 5 components of the exponential Fourier series for

the waveform of Figure below. Assume that ω = 1.

Aga Bayou (ASTU) Signals and Systems Analysis July 2, 2023 15 / 27

Cont. . . .

Amplitude and Phase Spectra of a Periodic Signal

We can express Complex Fourier Coeffiecient Ck as

Ck = |Ck |e jϕk

A plot of Ck versus the angular frequency ω is called Amplitude

spectrum of the periodic signal x(t) , and
A plot of Φk versus the angular frequency ω is called Phase spectrum
of the periodic signal x(t)
The amplitude and phase spectra are not continuous curves but appear
only at discrete frequencies kω0 so they are discrete frequency spectra
or line spectra
The amplitude spectrum is an even function of ω |Ck | = |Ck |
For real periodic signal x(t), ϕk = −ϕk , phase spectrum is an add
function of 0 for real periodic signal

Aga Bayou (ASTU) Signals and Systems Analysis July 2, 2023 16 / 27

Cont. . . .

Power content of a Periodic Signal

Average Power content of a Periodic Signal is given by;
Z T0
1
P= |x(t)|2 dt (11)
T0 0

If x(t) is represented by the complex exponential Fourier series, then

power becomes
X∞
P= |Ck |2 (12)
k=−∞

This equation is known as Parseval’s Identity or Parseval’s Theorem

Aga Bayou (ASTU) Signals and Systems Analysis July 2, 2023 17 / 27

Properties of Fourier Series

1. Linearity Property:
FS Coefficient FS Coefficient
If x(t) ←−−−−−−→ fxn & y (t) ←−−−−−−→ fyn then linearity property
states that
Fourier series Coefficient
ax(t) + by (t) ←−−−−−−−−−−−−→ afxn + bfyn (13)

2. Time shifting Property:

Fourier series Coefficient
If x(t) ←−−−−−−−−−−−−→ fxn then time shifting property states that
FS Coefficient
x(t − t0 ) ←−−−−−−→ e −jnω0 t0 fxn (14)

3. Frequency shifting Property:

FS Coefficient
If x(t) ←−−−−−−→ fxn then frequency shifting property states that
FS Coefficient
e jnω0 t0 x(t) ←−−−−−−→ e −jnω0 t0 fx(n−n0 ) (15)

Aga Bayou (ASTU) Signals and Systems Analysis July 2, 2023 18 / 27

Cont. . . .

4. Time reversal Property:

FS Coefficient
If x(t) ←−−−−−−→ fxn then time reversal property states that
FS Coefficient
x(−t) ←−−−−−−→ f−xn (16)

5. Time scaling Property:

FS Coefficient
If x(t) ←−−−−−−→ fxn then time scaling property states that
FS Coefficient
x(at) ←−−−−−−→ fxn (17)

Time scaling property changes frequency components from ω0 to nω0

Aga Bayou (ASTU) Signals and Systems Analysis July 2, 2023 19 / 27

Cont. . . .
6. Differentiation and Integration Properties:
FS Coefficient
If x(t) ←−−−−−−→ fxn then differentiation property states that
x(t) FS Coefficient
←−−−−−−→ jnω0 .fxn (18)
dt
and Integration Property states that
Z
FS Coefficient fxn
x(t)dt ←−−−−−−→ (19)
jnω0
7. Multiplication and Convolution Properties:
FS Coefficient FS Coefficient
If x(t) ←−−−−−−→ fxn & y (t) ←−−−−−−→ fyn then multiplication
property states that
FS Coefficient
x(t).y (t) ←−−−−−−→ Tfxn ∗ fyn (20)
and convolution Property states that
FS Coefficient
x(t) ∗ y (t) ←−−−−−−→ Tfxn .fyn (21)
Aga Bayou (ASTU) Signals and Systems Analysis July 2, 2023 20 / 27
Cont. . . .

8. Conjugate and Conjugate Symmetry Properties:

FS Coefficient
If x(t) ←−−−−−−→ fxn then conjugate property states that
FS Coefficient
x ∗ (t) ←−−−−−−→ fxn
∗
(22)

Conjugate symmetry property for real valued time signal states that
∗
fxn = f−xn (23)

Conjugate symmetry property for imaginary valued time signal states

that
∗
fxn = −f−xn (24)

Aga Bayou (ASTU) Signals and Systems Analysis July 2, 2023 21 / 27

The Equivalence of the Trigonometric and Exponential
Series
The trignometric FS
∞
1 X
x(t) = a0 + (an cosnωt + bn sinnωt)
2
n=1
The exponential FS
∞
X
x(t) = cn e jnω0 t
n=−∞

If the FS are equavalent

∞ ∞
1 X X
a0 + (an cosnωt + bn sinnωt) = ck e jkω0 t (25)
2
n=1 k=−∞

How the Cn terms are related to the an and bn ?

Assume x(t) is real, so an and bn are real.
Aga Bayou (ASTU) Signals and Systems Analysis July 2, 2023 22 / 27
Cont. . . .

a0 = c0 is the average of the function of x(t).

likewise if the signal oscillate once in a period T seconds

a1 cosωt + b1 sinωt = c−1 e −jω0 t + c1 e jω0 t

and for two oscillations in T seconds we get

a2 cos2ωt + b2 sin2ωt = c−2 e −j2ω0 t + c2 e j2ω0 t

In general

an cosnωt + bn sinnωt = c−n e −jnω0 t + cn e jnω0 t

c−n the complex conjugate of cn

Aga Bayou (ASTU) Signals and Systems Analysis July 2, 2023 23 / 27

Cont. . . .

By using eulers identities

an bn
cn = −j
2 2
n ̸= 0, with c−n = cn∗

Aga Bayou (ASTU) Signals and Systems Analysis July 2, 2023 24 / 27

Signal Approximation

The reconstructed signal follows the original signal, the more

harmonic are used, the better reconstructed signal.

The region of the jump at t = 0.5, the reconstructed signal is not

exact. the reconstructed signal significantly fluctuates at this position

Aga Bayou (ASTU) Signals and Systems Analysis July 2, 2023 25 / 27

Cont. . . .

In particular, the height of overshooting or undershooting does not

depend on the number of harmonics and is roughly 9% of the jump
height.
However, the duration of the oscillations decreases with the number
of harmonics, eventually leading to a correct approximation in the
limit for infinitely many harmonics

Aga Bayou (ASTU) Signals and Systems Analysis July 2, 2023 26 / 27

 Thank You!!!

Aga Bayou (ASTU) Signals and Systems Analysis July 2, 2023 27 / 27