EMJ26103

Signal and Linear Systems

LECTURE 4
Fourier Series

Prepared by:
Dr. Abdul Halim Ismail
Lecturer: Dr. Nurul Syahirah Binti Khalid

Faculty of Electrical Engineering

& Technology
Contents
1) Frequency Domain

2) Introduction to Fourier Series

3) Orthogonal Set of Function

4) Trigonometric Form of Fourier Series

5) Exponential Fourier Series

6) GIBBS Phenomenon

7) Parseval’s Theorem

FacultySchool
of Electrical Engineering
of Mechatronic Engineering
Universiti Malaysia Perlis (UniMAP)
& Technology 2
Frequency Domain
§ Signals are represented as superposition of complex
sinusoids which leads to a useful expression for the system
output and provide a characterization of signals and systems.

§ Example in music, the orchestra is a superposition of sounds

generated by different equipment having different frequency
range such as string, base, violin and etc.
§ Another example is the choir team, where difference people
having difference voice pitch (freq.)

§ A time-domain graph shows how a signal changes over time,

whereas a frequency-domain graph shows how much of the
signal lies within each given frequency band over a range of
frequencies.

FacultySchool
of Electrical Engineering
of Mechatronic Engineering
Universiti Malaysia Perlis (UniMAP)
& Technology 3
Frequency Domain

§ Imagine you are recording a signal, for example music that consists of the
singer, guitar, drums and keyboard. Then loud noise suddenly appears.

x 1(t) Singer Gain, a Noise, x n

x 2(t)
Guitar Gain, b

+ y(t)
x 3(t) Drum Gain, c

x 4(t)
Keyboard Gain, d y(t) = ?

FacultySchool
of Electrical Engineering
of Mechatronic Engineering
Universiti Malaysia Perlis (UniMAP)
& Technology 4
§ Because of the loud signal xn, the musical recording is contaminated by
the high-frequency noise, which is added to the output signal y(t).

"($) = &'(($) + *'+($) + ,'-($) + .'/($) + '0($)

§ Removing the noise in time-domain may be difficult due to that fact that
the noise also lies respectively with time.

§ The signals can be transformed into frequency domain, where the

required signal (music recording) is in low frequencies, while the loud
noise is in high frequencies.

§ Hence, a filter can be design in frequency domain, and the loud noise can
be significantly removed.

FacultySchool
of Electrical Engineering
of Mechatronic Engineering
Universiti Malaysia Perlis (UniMAP)
& Technology 5
Remove the noise by
using appropriate filter,
i.e. low-pass filter

FacultySchool
of Electrical Engineering
of Mechatronic Engineering
Universiti Malaysia Perlis (UniMAP)
& Technology 6
(a) Composite of signals of low and high freq.
(b) Plotting the signal in time-domain
(c) Plotting the signal in freq-domain

FacultySchool
of Electrical Engineering
of Mechatronic Engineering
Universiti Malaysia Perlis (UniMAP)
& Technology 7
Fourier Series
Introduction

1. The representation of a periodic signal in terms of complex

exponentials, or equivalently in terms of sine and cosine
waveforms, leads to the Fourier series that are used
extensively used in all fields of science and engineering.

2. The Fourier series is named after the French physicist Jean

Baptiste Fourier (1768-1830) –who was the first to suggest –
periodic signals could be represented by a sum of sinusoids.

3. We learn how to decompose periodic

signals into their frequency components.

FacultySchool
of Electrical Engineering
of Mechatronic Engineering
Universiti Malaysia Perlis (UniMAP)
& Technology 8
Fourier Series
Introduction

ò ò
sin = - cos; cos = sin
d d
sin t = cos t ; cos t = - sin t
dt dt
1
òt sin ntdt = 2 (sin nt - nt cos nt )
n
1
òt cos ntdt = 2 (cos nt + nt sin nt )
n
FacultySchool
of Electrical Engineering
of Mechatronic Engineering
Universiti Malaysia Perlis (UniMAP)
& Technology 9
Fourier Series
Introduction

1.Cosine Function
The function f(x) = cos x is an even function. That is, it is
symmetrical about the vertical axis.
We have: cos (-x) = cos x

-
òp cos q = 0

FacultySchool
of Electrical Engineering
of Mechatronic Engineering
Universiti Malaysia Perlis (UniMAP)
& Technology 10
Fourier Series
Introduction

2. Sine Function
The function f(x) = sin x is an odd function. That is, it is
symmetrical about the origin.
We have: sin (-x) = -sin x

-
òp sin q = 0

FacultySchool
of Electrical Engineering
of Mechatronic Engineering
Universiti Malaysia Perlis (UniMAP)
& Technology 11
3. Multiples of sine and cosine curve.
Consider the function y = sin x

From the graph,

sin np = 0 for n = 0, 1, 2, 3, …
(2n - 1)p
sin = (-1) n +1 for n = 0, 1, 2, 3, …
2

FacultySchool
of Electrical Engineering
of Mechatronic Engineering
Universiti Malaysia Perlis (UniMAP)
& Technology
y = cos x

cos(2np ) = 1 for n = 0, 1, 2, 3, …
cos[(2n - 1)p ] = -1 for n = 0, 1, 2, 3, …
cos(np ) = (-1) n for n = 0, 1, 2, 3, …

FacultySchool
of Electrical Engineering
of Mechatronic Engineering
Universiti Malaysia Perlis (UniMAP)
& Technology
Orthogonal Set of Functions

A set of functions {φ1(x), φ2(x), … } is an orthogonal set of

functions on the interval [a, b] if any two functions in the
set are orthogonal to each other.

ò
(f n , f m ) = f n ( x)f m ( x)dx = 0, n ¹ m
a

FacultySchool
of Electrical Engineering
of Mechatronic Engineering
Universiti Malaysia Perlis (UniMAP)
& Technology 14
Orthogonal Set of Functions
Sinusoidal Prove

Show that:
{1, cos x, sin x, cos 2 x, sin 2 x,...} form an orthogonal set from [-π, π].

To prove orthogonal set of function, let separate the set into:

(1, cos nx) = 0

(1, sin nx) = 0
(cos(nx), cos(mx)) = 0, n ¹ m
(sin(nx), sin( mx)) = 0, n ¹ m
(cos(nx), sin( mx)) = 0,

FacultySchool
of Electrical Engineering
of Mechatronic Engineering
Universiti Malaysia Perlis (UniMAP)
& Technology 15
Orthogonal Set of Functions
Sinusoidal Prove

1st case: (1, cos nx) = 0

p

òp1• cos nxdx

-
p
é1 ù
= ê sin nx ú
ën û -p
1
= [ 0 - 0] = 0 is proven!
n

FacultySchool
of Electrical Engineering
of Mechatronic Engineering
Universiti Malaysia Perlis (UniMAP)
& Technology 16
Orthogonal Set of Functions
Sinusoidal Prove

2nd case: (1, sin(nx)) = 0

is proven!

FacultySchool
of Electrical Engineering
of Mechatronic Engineering
Universiti Malaysia Perlis (UniMAP)
& Technology 17
Orthogonal Set of Functions
Sinusoidal Prove

3rd case: (sin(nx), sin(mx)) = 0

p
1
-
òp sin( nx) sin( mx)dx - - > * sin q sin f =
2
[cos(q - f ) - cos(q + f )]

p p
1 1
=
2 òp
-
cos(nx - mx) - cos(nx + mx)dx =
2
-
òp cos(n - m) x - cos(n + m)xdx
p
1é 1 1 ù
= ê sin( n - m) x - sin( n + m) x ú
2 ën - m n+m û -p
1é 1 1 ù
= sin( n - m )p - sin( n + m )p
2 êë n - m n+m ú
û
é 1 1 ù
-ê sin( n - m)(-p ) - sin( n + m)(-p )ú
ën - m n+m û
=0
is proven!
FacultySchool
of Electrical Engineering
of Mechatronic Engineering
Universiti Malaysia Perlis (UniMAP)
& Technology 18
TRIGONOMETRIC FORM OF FOURIER SERIES
Periodic Functions

§ A sinusoidal signal, is a periodic signal with period.

§ The sum of two sinusoids is periodic provided that their

frequencies are integral multiples of a fundamental frequency.

§ We can show that a signal x(t), is a sum of sine and cosine

functions whose frequencies are integral multiples of, is
periodic signal.

§ A function f(q) is periodic if it is defined for all real q, and

there is a repetitive time period (fundamental period), T such
that f(q+T) = f(q).

FacultySchool
of Electrical Engineering
of Mechatronic Engineering
Universiti Malaysia Perlis (UniMAP)
& Technology 19
 f (q )

0
q

f (q )

0
q
f (q )

0
q

FacultySchool
of Electrical Engineering
of Mechatronic Engineering
Universiti Malaysia Perlis (UniMAP)
& Technology
FOURIER SERIES

Let the signal x(t) be:

x(t ) = a0 + a1 cos w0 t + a 2 cos 2w0 t + ... + a k cos kw0 t

+ b1 sin w0t + b2 sin 2w0t + ... + bk sin kw0t

i.e.
k
x(t ) = a0 + åa
n =1
n cos w0 nt + bn sin nw0 t

FacultySchool
of Electrical Engineering
of Mechatronic Engineering
Universiti Malaysia Perlis (UniMAP)
& Technology 21
FOURIER SERIES

For the signal x(t) to be periodic, it must satisfy the condition

x(t) = x(t + T) for all t, i.e.
k
x(t + T ) = a0 + åa
n =1
n cos w0 n(t + T ) + bn sin w0 n(t + T )

k
2p 2p
= a0 + å
n =1
a n cos w0 n(t +
w0
) + bn sin w0 n(t +
w0
)

k
= a0 + åa
n =1
n cos(w0 nt + 2np ) + bn sin(w0 nt + 2np )

k
= a0 + åa
n =1
n cosw0 nt + bn sin w0 nt

= x(t )

FacultySchool
of Electrical Engineering
of Mechatronic Engineering
Universiti Malaysia Perlis (UniMAP)
& Technology 22
FOURIER SERIES

§ This proves that the signal x(t), which is summation of

sine and cosine functions of frequencies 0, ω0, 2ω0,…,
kω0 is a periodic signal x(t) with period T.

§ The infinite series of sine and cosine terms of frequencies

0, ω0, 2ω0,…, kω0 is known as trigonometric form of
Fourier series and can be written as:
¥
x(t ) = åa
n =0
n cos nw0 t + bn sin nw0 t

FacultySchool
of Electrical Engineering
of Mechatronic Engineering
Universiti Malaysia Perlis (UniMAP)
& Technology 23
FOURIER SERIES

§ Simplifying the previous Fourier series equation by

separating the initial value at n=0 give;
¥
x(t ) = a0 + åa
n =1
n cos nw0 t + bn sin nw0 t

§ The constants a0, a1,…, an, b0, b1,…, bn, are called Fourier
coefficients.

§ Therefore, we need to derive and determine these

coefficients.

FacultySchool
of Electrical Engineering
of Mechatronic Engineering
Universiti Malaysia Perlis (UniMAP)
& Technology 24
FOURIER SERIES
Evaluation of the Fourier Coefficients of Trigonometry F.S.

§ Given the Trigonometry F.S. as below, we need to determine all

the coefficients.
¥
x(t ) = a0 + åa
n =1
n cos nw0 t + bn sin nw0 t

§ To determine a0:
Integrate both sides of the func. over one period (t0 to t0 + T);
t 0 +T t 0 +T t 0 +T
é¥ ù
ò x(t )dt = a ò dt + ò å
t0
0
t0 t0
ê [a n cos nw0 t + bn sin nw0 t ]dt ú
êë n =1 úû

¥ t 0 +T ¥ t 0 +T

= a0T + å a ò cos nw tdt +å b ò sin nw tdt

n =1
n 0
n =1
n 0
t0 t0

FacultySchool
of Electrical Engineering
of Mechatronic Engineering
Universiti Malaysia Perlis (UniMAP)
& Technology 25
FOURIER SERIES
Evaluation of the Fourier Coefficients of Trigonometry F.S.

§ Since
t 0 +T t 0 +T

ò cos nw tdt = 0
t0
0 and ò sin nw tdt = 0
t0
0

(the net areas of sinusoids over complete periods are zero for any
nonzero integer n and any time t0.)

§ Thus, we could write a0 as;

t 0 +T t 0 +T
1
ò x(t )dt = a0T or à a0 =
T ò x(t )dt
t0
t0

FacultySchool
of Electrical Engineering
of Mechatronic Engineering
Universiti Malaysia Perlis (UniMAP)
& Technology 26
FOURIER SERIES
Evaluation of the Fourier Coefficients of Trigonometry F.S.

§ To determine an and bn:

The orthogonality and sinusoidal properties as the following
can be utilized;
t 0 +T ìï 0, m¹n
1)
ò
cos nw0 t cos mw0 tdt = í T
ï
, m=n¹0
t0 î2

t 0 +T ìï 0, m¹n
ò
2) sin nw0t sin mw0tdt = í T
ï
, m=n¹0
t0 î2

t 0 +T
3) , for all m and n.
ò
t0
sin nw0 t cos mw0 tdt = 0

FacultySchool
of Electrical Engineering
of Mechatronic Engineering
Universiti Malaysia Perlis (UniMAP)
& Technology 27
FOURIER SERIES
Evaluation of the Fourier Coefficients of Trigonometry F.S.

§ To find an :
multiply the Trigo. F.S. by cos mω0t and integrate over
one period. That is,
t 0 +T t 0 +T ¥ t 0 +T

ò x(t ) cos mw dt = a ò cos mw dt + å a ò

t0
0 0
t0
0
n =1
n
t0
cos nw0 t cos mw0 tdt

¥ t 0 +T

+ åb ò
n =1
n sin nw0 t cos mw0 tdt
t0

FacultySchool
of Electrical Engineering
of Mechatronic Engineering
Universiti Malaysia Perlis (UniMAP)
& Technology 28
FOURIER SERIES
Evaluation of the Fourier Coefficients of Trigonometry F.S.

§ To first and third integrals are equal zero and the second
integral is equal to T/2 when m = n: T
0
t 0 +T t 0 +T ¥ t 0 +T 2
ò x(t ) cos mw dt = a ò cos mw dt + å a ò
t0
0 0
t0
0
n =1
n
t0
cos nw0 t cos mw0 tdt

¥ t 0 +T 0
+ åb ò
n =1
n sin nw0 t cos mw0 tdt
t 0 +T
t0 2
§ Therefore:
an =
T ò x(t ) cos nw dt
t0
0

t 0 +T t 0 +T
T or am = 2
ò x(t ) cos mw0 dt = a m
2 T ò x(t ) cos mw dt
t0
0
t0

FacultySchool
of Electrical Engineering
of Mechatronic Engineering
Universiti Malaysia Perlis (UniMAP)
& Technology 29
FOURIER SERIES
Evaluation of the Fourier Coefficients of Trigonometry F.S.

§ To find bn :
multiply the Trigo. F.S. by sin mω0t and integrate over
one period. Then,
t 0 +T t 0 +T ¥ t 0 +T

ò x(t ) sin mw dt = a ò sin mw dt + å a ò

t0
0 0
t0
0
n =1
n
t0
cos nw0 t sin mw0 tdt

¥ t 0 +T

+ åb ò
n =1
n sin nw0 t sin smw0 tdt
t0

FacultySchool
of Electrical Engineering
of Mechatronic Engineering
Universiti Malaysia Perlis (UniMAP)
& Technology 30
FOURIER SERIES
Evaluation of the Fourier Coefficients of Trigonometry F.S.

§ The first and second integrals are zero, and the third integral is
equal to T/2 when m = n. Thus,
t 0 +T t 0 +T 0 t 0 +T 0
¥

ò x(t ) sin mw dt = a ò sin mw dt + å a ò

t0
0 0
t0
0
n =1
n
t0
cos nw0 t sin mw0 tdt

t 0 +T T
¥

åb ò
2
+ n sin nw0 t sin smw0 tdt
n =1 t0 t 0 +T
2
bn =
T ò x(t ) sin nw tdt 0
§ Thus: t0

t 0 +T t 0 +T
T or 2
ò x(t ) sin mw0 dt = bm
2
bm =
T ò x(t ) sin mw tdt
t0
0
t0
*a0, an and bn are called trigonometric Fourier series coefficients.
FacultySchool
of Electrical Engineering
of Mechatronic Engineering
Universiti Malaysia Perlis (UniMAP)
& Technology 31
FOURIER SERIES
Evaluation of the Fourier Coefficients of Trigonometry F.S.:
SUMMARY

§ To find a0 1
t 0 +T
Integrate both sides, solving for a0 =
T ò x(t )dt
t0

§ To find an
Multiply both sides with cosmω0t and then integrate both sides,
solving for t +T 0
2
an =
T
t0
ò x(t ) cos nw dt
0

§ To find bn
Multiply both sides with sinmω0t and then integrate both sides,
solving for t 0 +T
2
bn =
T ò x(t ) sin nw tdt
t0
0

FacultySchool
of Electrical Engineering
of Mechatronic Engineering
Universiti Malaysia Perlis (UniMAP)
& Technology 32
FOURIER SERIES
EXAMPLE (1)

Find the Fourier series expansion of the half wave rectified sine wave as
shown in figure below.
x(t)

----- A ----

-2π -π 0 π 2π 3π

FacultySchool
of Electrical Engineering
of Mechatronic Engineering
Universiti Malaysia Perlis (UniMAP)
& Technology 33
FOURIER SERIES
EXAMPLE (1) : Solution

The periodic waveform shown in Figure with period 2 π is half of a

sine wave with period 2π.

ì 2p
ï A sin wt = A sin t = A sin t ,0 £ t £ p
x(t ) = í 2p
ïî 0 , p £ t £ 2p

Now the fundamental period: T = 2π.

2p 2p
Fundamental frequency: w0 = = =1
T 2p

FacultySchool
of Electrical Engineering
of Mechatronic Engineering
Universiti Malaysia Perlis (UniMAP)
& Technology 34
FOURIER SERIES
EXAMPLE (1) : Solution

Let: t 0 = 0, t 0 + T = T = 2p
T 2p 2p
1 1 1
a0 =
T ò
0
x(t )dt =
2p ò
0
x(t )dt =
2p ò A sin tdt
0

A A 2A A
= [- cos t ]0 = [- (cos p - cos 0)] = =
p
2p 2p 2p p
Thus,
A
a0 =
p

FacultySchool
of Electrical Engineering
of Mechatronic Engineering
Universiti Malaysia Perlis (UniMAP)
& Technology 35
FOURIER SERIES
EXAMPLE (1) : Solution

T 2p
2 2
an =
T ò
x(t ) cos nw0 tdt =
0
2p ò x(t ) cos ntdt
0
p p
1 A
pò ò
= A sin t cos ntdt = A sin t cos ntdt
p
0 0
p p
A A é cos(1 + n)t cos(1 - n)t ù
=
2p ò
0
sin(1 + n)t + sin(1 - n)t =
2p ê- 1 + n - 1 - n ú
ë û0
p
A é cos(1 + n)p - cos 0 cos(1 - n)p - cos 0 ù
=- ê + ú
2p ë 1+ n 1- n û0

A ìïé (-1) ( n +1) - 1 ù (-1) ( n -1) - 1 üï

=- íê ú+ ý
2p ïîêë 1 + n úû 1- n ïþ

FacultySchool
of Electrical Engineering
of Mechatronic Engineering
Universiti Malaysia Perlis (UniMAP)
& Technology 36
FOURIER SERIES
EXAMPLE (1) : Solution

A é 1-1 1-1 ù
For odd n: an = - ê + ú =0
2p ë1 + n 1 - n û

A é - 1 - 1 - 1 - 1ù A é 2 2 ù 2A
For even n: an = - ê + ú = ê + ú =-
2p ë 1 + n 1 - n û 2p ë1 + n 1 - n û p (n 2 - 1)

Therefore:
2A
an = - (for even n)
p (n 2 - 1)

FacultySchool
of Electrical Engineering
of Mechatronic Engineering
Universiti Malaysia Perlis (UniMAP)
& Technology 37
FOURIER SERIES
EXAMPLE (1) : Solution
T 2p
2 2
bn =
T ò
x(t ) sin nw0 tdt =
0
2p ò x(t ) sin ntdt
0
p p
1 A
pò ò
= A sin t sin ntdt = sin t sin ntdt
p
0p 0
A
=
2p ò [cos(n - 1)t - cos(n + 1)t ]dt
0
p
A é sin( n - 1)t sin( n + 1)t ù
= ê -
2p ë n - 1 n + 1 úû 0

This is zero for all values of n except for n = 1.

For n = 1:
A
b1 =
2p
FacultySchool
of Electrical Engineering
of Mechatronic Engineering
Universiti Malaysia Perlis (UniMAP)
& Technology 38
FOURIER SERIES
EXAMPLE (1) : Solution

Therefore, the trigonometric Fourier series is:

¥
x(t ) = a0 + åa
n =1
n cos nw0 t + bn sin nw0 t
¥
= a0 + b1 sin t + åa
n =1
n cos nt

¥
AA 2A
= +
p 2p
sin t -
n = even p
å
( n 2
- 1)
cos nt

FacultySchool
of Electrical Engineering
of Mechatronic Engineering
Universiti Malaysia Perlis (UniMAP)
& Technology 39
FOURIER SERIES
EXAMPLE (2)

Obtain the trigonometric Fourier series for the waveform

shown in figure below:

x(t)

A
-----
-----

-4π -3π -2π -π 0 π 2π 3

π

FacultySchool
of Electrical Engineering
of Mechatronic Engineering
Universiti Malaysia Perlis (UniMAP)
& Technology 40
FOURIER SERIES
EXAMPLE (2): Solution

The waveform shown in Figure is periodic with a period T = 2π.

Let: t 0 = 0, t 0 + T = 2p

2p 2p
Then, Fundamental frequency: w0 = = =1
T 2p

The waveform is described by:

ìæ A ö
ïç ÷t , 0 £ t £ p
x(t ) = íè p ø
ïî 0, p £ t £ 2p

FacultySchool
of Electrical Engineering
of Mechatronic Engineering
Universiti Malaysia Perlis (UniMAP)
& Technology 41
FOURIER SERIES
EXAMPLE (2): Solution

Let: t 0 = 0, t 0 + T = T = 2p
T 2p 2p
1 1 1 æ Aö
a0 =
T ò
0
x(t )dt =
2p ò
0
x(t )dt =
2p ò
0
ç ÷tdt
èp ø
p
A ét 2ù
A
= 2 ê ú =
2p êë 2 úû 4
0

Thus,
A
a0 =
4

FacultySchool
of Electrical Engineering
of Mechatronic Engineering
Universiti Malaysia Perlis (UniMAP)
& Technology 42
FOURIER SERIES
EXAMPLE (2): Solution

T
2
an =
T ò
x(t ) cos nw0 dt
0p p
2 A A
=
2p òp
0
t cos ntdt =
p 2 ò t cos ntdt
0

é p p ù p
A é t sin nt ù sin nt A æ 0 - 0 cos nt ö
= 2 êê ú
p êë ë n û 0 0 n
- ò dt ú = 2
ú p è n
û
ç + 2 ÷
n ø0
A
= 2 (cos np - cos 0 )p0
p n
Thus, ì 2A
ï- , for odd n
an = í p 2 n 2
ïî 0, for even n

FacultySchool
of Electrical Engineering
of Mechatronic Engineering
Universiti Malaysia Perlis (UniMAP)
& Technology 43
FOURIER SERIES
EXAMPLE (2): Solution
T 2p
2 2
bn =
T ò
x(t ) sin nw0 tdt =
0
2p ò x(t ) sin ntdt
0
p p
1 A A
= ò
p p
t sin ntdt =
0
p ò t sin ntdt
2
0

A é é t (- cos nt ) ù p p - cos nt ù
=
p2
êê
êë
ë
n úò -
û0 0 n
dt ú
ú
û
é p
A cos np æ sin nt ö ù A A
= ê- p + ç 2 ÷ ú = - 2 cos np = (-1) n +1
p2 êë n è n ø 0 úû p np
ì A
ï np , For odd n
bn = í ,
A
ï-
î np For even n

FacultySchool
of Electrical Engineering
of Mechatronic Engineering
Universiti Malaysia Perlis (UniMAP)
& Technology 44
FOURIER SERIES
EXAMPLE (2): Solution

The trigonometric Fourier series is:

¥
x(t ) = a0 + åa
n =1
n cos nw0 t + bn sin nw0 t
¥ ¥
A 2A cos nt A sin nt
= - 2
4 p n =odd n 2 å+
p å (-1)
n =1
n +1
n
A 2A é 1 1 ù
= - 2 êcos t + 2 cos 3t + 2 cos 5t + ...ú
4 p ë 3 5 û
Aé 1 1 1 ù
+ êsin t - sin 2t + sin 3t - sin 4t + ...ú
pë 2 3 4 û

FacultySchool
of Electrical Engineering
of Mechatronic Engineering
Universiti Malaysia Perlis (UniMAP)
& Technology 45
EXPONENTIAL FOURIER SERIES
Complex Exponentials

§ The trigonometry Fourier Series contains sinusoidal element, that

can be represented in complex exponentials.

§ Recall back the trigonometry Euler’s identities;

e jnw0t = cos nw0t + j sin nw0t

e - jnw0t = cos nw0t - j sin nw0t

cos nw0t = e
2
(
1 jnw0t - jnw0t
+e )
sin nw0t =
2j
e (
1 jnw0t - jnw0t
-e
j
) (
= - e jnw0t - e - jnw0t
2
)

FacultySchool
of Electrical Engineering
of Mechatronic Engineering
Universiti Malaysia Perlis (UniMAP)
& Technology 46
EXPONENTIAL FOURIER SERIES
Complex Exponentials

§ The set of complex exponential functions:

{ , n = 0, ±1, ±2,!}
e jnw0t

forms a closed orthogonal set over an interval (t0 , t0 + T)

where T=(2π/ω0 ) for any value of t0 , and therefore it can be
used as a Fourier series.

§ Using Euler’ identity, we can write:

é e j ( nw0t +q n ) + e - j ( nw0t +q n ) ù
An cos(nw0 t + q n ) = An ê ú
êë 2 úû

FacultySchool
of Electrical Engineering
of Mechatronic Engineering
Universiti Malaysia Perlis (UniMAP)
& Technology 47
EXPONENTIAL FOURIER SERIES
Complex Exponentials

§ Substituting this in the definition of cosine Fourier

representation, we obtain:

¥
An j ( nw0t +qn ) - j ( nw0t +qn )
x(t ) = A0 + å éëe +e ùû
n =1 2
¥
An jnw0t jqn
= A0 + å éëe e + e - jnw0t e - jqn ùû
n =1 2
¥
æ An jnw0t jqn ö ¥ æ An - jnw0t - jqn ö
= A0 + å ç e e ÷ + å ç e e ÷
n =1 è 2 ø n=1 è 2 ø
¥
æ An jqn ö jnw0t ¥ æ An j ( -qn ) ö - jnw0t
= A0 + å ç e ÷ e + åç e ÷e
n =1 è 2 ø n =1 è 2 ø

FacultySchool
of Electrical Engineering
of Mechatronic Engineering
Universiti Malaysia Perlis (UniMAP)
& Technology 48
EXPONENTIAL FOURIER SERIES
Complex Exponentials

§ Letting n = -k in the second summation, we have

¥ -¥
æ An jq n ö jnw0t æ Ak jq k ö jkw0t
x(t ) = A0 + å ç
n =1 è
2
e ÷e
ø
+ å ç
k = -1è
2
e ÷e
ø

§ On comparing the above equations for x(t), we get

An = Ak ; (-q n ) = q k , n > 0
,k<0

FacultySchool
of Electrical Engineering
of Mechatronic Engineering
Universiti Malaysia Perlis (UniMAP)
& Technology 49
EXPONENTIAL FOURIER SERIES
Complex Exponentials

§ Let us define
An jq n
C = A0 ; C n = e ,n > 0
2
§ Thus
¥ -¥
An jq n jnw0t An jq n jnw0t
x(t ) = A0 + å
n =1
2
e e + å
n = -1
2
e e

¥
i.e x(t ) = å
n = -¥
C n e jnw0t

This is known as exponential form of Fourier series. The

above equation is called the synthesis equation.

FacultySchool
of Electrical Engineering
of Mechatronic Engineering
Universiti Malaysia Perlis (UniMAP)
& Technology 50
EXPONENTIAL FOURIER SERIES
Complex Exponentials

§ So the exponential series from cosine series is:

C = A0
An jq n
Cn = e
2

FacultySchool
of Electrical Engineering
of Mechatronic Engineering
Universiti Malaysia Perlis (UniMAP)
& Technology 51
EXPONENTIAL FOURIER SERIES
Complex Exponentials: Evaluating the Coefficients

We have: ¥
2p
x(t ) = åC e
n = -¥
n
jnw0t
w0 =
T

Multiplying both sides by e - jkw0t and integrating over one period,

t 0 +T t 0 +T
æ ¥ ö - jkw t
ò x(t )e - jkw0t dt = ò å ç ÷e
jn w t
Cn e 0 0
dt
ç ÷
t0 t0 è n = -¥ ø
¥ t 0 +T

= å
n = -¥
Cn ò e jnw0t e - jkw0t dt
t0

FacultySchool
of Electrical Engineering
of Mechatronic Engineering 52
Universiti Malaysia Perlis (UniMAP)
& Technology
EXPONENTIAL FOURIER SERIES
Complex Exponentials: Evaluating the Coefficients

We know
t 0 +T
ì 0, k ¹ n
òe
jnw0t - jkw0t
e dt = í
t0 îT , k = n
t 0 +T t 0 +T
1
ò ò
- jkw0t
Thus, x(t )e dt = TC k or Ck = x(t )e - jkw0t dt
T
t0 t0

t 0 +T t 0 +T
or 1 1
ò x(t )e ò x(t )dt
- jnw0t
Cn = dt C0 = A0 =
T T
t0 t0

,where C n are the Fourier coefficients of the exponential series.

FacultySchool
of Electrical Engineering
of Mechatronic Engineering
Universiti Malaysia Perlis (UniMAP)
& Technology 53
EXPONENTIAL FOURIER SERIES
Complex Exponentials

q The Fourier coefficients of x(t) have only a discrete

spectrum because values of Cn exist only for discrete
values of n.
q As it represents a complex spectrum, it has both
magnitude and phase spectra.
q Note:
a. The magnitude line spectrum is always an even function
of n
b. The phase line spectrum is always an odd function of n.

FacultySchool
of Electrical Engineering
of Mechatronic Engineering
Universiti Malaysia Perlis (UniMAP)
& Technology 54
Trigonometric Fourier Series from
EXPONENTIAL Fourier Series
§ The complex exponential Fourier series is given by:

¥
x(t ) = åCe
n =-¥
n
jnw0t

-1 ¥
= C0 + å Cn e
n =-¥
jnw0t
+ å Cne jnw0t
n =1
¥
= C0 + å ( C- ne - jnw0t + Cne jnw0t )
n =1
¥
= C0 + å C- n ( cos nw0t - j sin nw0t ) + Cn (cos nw0t + j sin nw0t )
n =1
¥
= C0 + å ( Cn + C- n ) cos nw0t + j ( Cn - C- n ) sin nw0t
n =1

FacultySchool
of Electrical Engineering
of Mechatronic Engineering
Universiti Malaysia Perlis (UniMAP)
& Technology 55
Trigonometric Fourier Series from
EXPONENTIAL Fourier Series
§ Comparing with standard trigonometric Fourier series:

¥
x(t ) = a0 + åa
n =1
n cos nw0 t + bn sin nw0 t

§ We get the conversion of exponential to Trigonometric:

a0 = C0
an = Cn + C- n
bn = j (C n - C - n )

FacultySchool
of Electrical Engineering
of Mechatronic Engineering
Universiti Malaysia Perlis (UniMAP)
& Technology 56
Trigonometric Fourier Series from
EXPONENTIAL Fourier Series
§ From exponential series, we know:
T T
1 1
Cn = ò x(t )e - jnw0t dt = ò x(t ) [ cos nw0t - j sin nw0t ]dt
T0 T0
1æ 2 ö 1
T
= ç ò x(t )cos nw0tdt - j sin nw0tdt ÷ = [ an - jbn ]
2èT 0 ø 2
T T
1 1
C- n = ò x(t )e jnw0t dt = ò x(t ) [ cos nw0t + j sin nw0t ]dt
T0 T0
1æ 2 ö 1
T
= ç ò x(t )cos nw0tdt + j sin nw0tdt ÷ = [ an + jbn ]
2èT 0 ø 2
T
1
C0 = ò x(t )dt = a0
T0

FacultySchool
of Electrical Engineering
of Mechatronic Engineering
Universiti Malaysia Perlis (UniMAP)
& Technology 57
Trigonometric Fourier Series from
EXPONENTIAL Fourier Series
§ Then, we can obtain the coefficients conversion from
trigonometric form to exponential form of Fourier
Series,
C0 = a0
1
C n = (a n - jbn )
2
1
C - n = (a n + jbn )
2

FacultySchool
of Electrical Engineering
of Mechatronic Engineering
Universiti Malaysia Perlis (UniMAP)
& Technology 58
EXPONENTIAL FOURIER SERIES
EXAMPLE (1)

Obtain the exponential Fourier series for the waveform

shown below. Also draw the frequency spectrum.

FacultySchool
of Electrical Engineering
of Mechatronic Engineering
Universiti Malaysia Perlis (UniMAP)
& Technology 59
Solution:
The periodic waveform shown, with a period T = 2π can be
expressed as:

ì A, 0£t £p
x(t ) = í
î- A, 0 £ t £ 2p

Let t 0 = 0, t 0 + T = 2p
2p 2p
And fundamental frequency: w0 = = =1
T 2p

FacultySchool
of Electrical Engineering
of Mechatronic Engineering
Universiti Malaysia Perlis (UniMAP)
& Technology
 Solution:

Exponential Fourier Series:

æ p 2p ö A
[ ]
T
1 1 ç
C0 = ò
x(t )dt = Adt + - Adt ÷ =
ò ò (t )p0 - (t )p2p = 0
T 2p ç ÷ 2p
0 è0 p ø
T
1
Cn =
T ò
0
x(t )e - jnw0t dt
p 2p
æ
1 ç
p 2p ö A éæ e - jnt ö æe ö - jnt ù
= A(t )e - jnt dt + A(t )e - jnt dt ÷ =
ò ò êç ÷ -ç ÷ ú
2p ç ÷ 2p êç - jn ÷ ç - jn ÷ ú
è0 p ø ëè ø0 è øp û

FacultySchool
of Electrical Engineering
of Mechatronic Engineering
Universiti Malaysia Perlis (UniMAP)
& Technology
Cont.…
=-
A
jn 2p êë
(
é e - jnp - e 0 ) - (e
p
0
- j 2pn
-e )p
- jnp 2p ù
úû

=-
A
jn 2p
[ n
][
(-1) - 1 - 1 - (-1) = - j
2A
np
n
]
ì 2A
ï- j , for _ odd _ n
Cn = í np
ïî 0, for _ even _ n
¥ ¥
2 A jnt
x(t ) = C0 + åC e
n = -¥
n
jnw0t
= å
n = -¥
- j e , for _ odd _ n
np
n¹0
2A 2A 2A
C0 = 0, C1 = C -1 = , C3 = C - 3 = , C5 = C - 5 =
p 3p 5p

FacultySchool
of Electrical Engineering
of Mechatronic Engineering
Universiti Malaysia Perlis (UniMAP)
& Technology
Cont.…
2A 2A 2A
C0 = 0, C1 = C -1 = , C3 = C - 3 = , C5 = C - 5 =
p 3p 5p

2A 2A
p Cn p
2A 2A
2A 3p 3p
2A
5p
5p
•
- - - -2 -1 0 1 2 3 4 5
5 4 3

FacultySchool
of Electrical Engineering
of Mechatronic Engineering
Universiti Malaysia Perlis (UniMAP)
& Technology
EXPONENTIAL FOURIER SERIES
EXAMPLE (2)

Obtain the exponential Fourier series for the waveform

shown below. Also draw the frequency spectrum.

FacultySchool
of Electrical Engineering
of Mechatronic Engineering
Universiti Malaysia Perlis (UniMAP)
& Technology 64
Solution:
The waveform showncan be expressed over one period (0 to π)
as:

2p
x(t ) = A sin wt where w= =1
2p
Because it is part of a sine wave with period = 2 π.
x(t ) = A sin t 0£t £p
The full wave rectified sine wave is periodic with T = π.

t 0 = 0, t 0 + T = 0 + p = p 2p 2p
w0 = = =2
T p

FacultySchool
of Electrical Engineering
of Mechatronic Engineering
Universiti Malaysia Perlis (UniMAP)
& Technology
Exponential Fourier Series:
¥ ¥
x(t ) = åC e
n = -¥
n
jnw0t
= åC e
n = -¥
n
j 2 nt

T p p
1 1 A
ò ò ò
- jnw0t
Cn = x(t )e dt = A sin te - j 2 nt dt = sin te - j 2 nt dt
T p p
0 0 0
p æp ö
A æç e jt - e - jt ö - j 2 nt A ç e j (1- 2 n )t - e - j (1+ 2 n )t ÷dt
= ò
p çè 2 j
÷e
÷
ø
dt =
j 2p çò ÷
0 è0 ø
p
A æ e j (1- 2 n )t e - j (1+ 2 n )t ö
= ç - ÷
j 2p ç j (1 - 2n) - j (1 + 2n) ÷
è ø0
A æ e j (1- 2 n )p - e 0 e - j (1+ 2 n )p - e 0 ö
= ç - ÷
j 2p ç j (1 - 2n) - j (1 + 2n) ÷
è ø

FacultySchool
of Electrical Engineering
of Mechatronic Engineering
Universiti Malaysia Perlis (UniMAP)
& Technology
Cont.…
2A
Cn =
p (1 - 4n 2 )
T p
1 1 A A
C0 = ò x(t )dt = ò A sin tdt = [- cos t ]p0 = [- (cos p - cos 0)] = 2 A
T p p p p
0 0
¥ ¥
2A 2A 2A æ e j 2 nt ö
x(t ) = å p (1 - 4n
n = -¥
2
)
e j 2 nt
=
p
+ å
ç
p n = -¥ çè 1 - 4n 2
÷
÷
ø
n¹0

2A 2A 2A
C0 = C2 = C-2 = =-
p p (1 - 4(2)2 ) 15p
2A 2A 2A 2A
C1 = C -1 = =- C3 = C - 3 = =-
p (1 - 4(1)2 ) 3p p (1 - 4(3)2 ) 35p

FacultySchool
of Electrical Engineering
of Mechatronic Engineering
Universiti Malaysia Perlis (UniMAP)
& Technology
 2A
Cont.… Cn
p

- - - -2 -1 0 1 2 3 4 5
5 4 2A
3
2A
- -
35p 35p
2A 2A
- -
15p 15p
2A 2A
- -
3p 3p

The frequency spectrum

FacultySchool
of Electrical Engineering
of Mechatronic Engineering
Universiti Malaysia Perlis (UniMAP)
& Technology
GIBBS Phenomenon

J. Willard Gibbs (1899), discovered that for a periodic signal with

discontinuities, if the signal is reconstructed by adding the Fourier
series, overshoots appear around the edges.

FacultySchool
of Electrical Engineering
of Mechatronic Engineering
Universiti Malaysia Perlis (UniMAP)
& Technology 69
GIBBS Phenomenon

§ These overshoots decay outwards in a damped oscillatory

manner away from edges. This is called Gibbs phenomenon.

§ The overshoots at the discontinuity according to the Gibbs

are found to be around 9 percent of the height of discontinuity
irrespective of the number of terms in the Fourier series.
§ It has also been observed that as more number of terms in the
series are added, the frequency increases, overshoots, get
sharper, but the adjoining oscillation amplitude reduces.
§ That is error between the original signal x(t) and the truncated
signal xn (t) reduced except at the edges as n increases.
§ Thus, the truncated Fourier series approaches x(t) as the
number of terms in approximation increases.

FacultySchool
of Electrical Engineering
of Mechatronic Engineering
Universiti Malaysia Perlis (UniMAP)
& Technology 70
3 harmonics signal 9 harmonics signal

21 harmonics signal 45 harmonics signal

71
FacultySchool
of Electrical Engineering
of Mechatronic Engineering
Universiti Malaysia Perlis (UniMAP)
& Technology
Parseval’s Theorem

FS
If, x1 (t ) ¬¾®
¾ C n , [ for _ complex _ x1 (t )]
FS
and, x2 (t ) ¬¾®
¾ Dn , [ for _ complex _ x2 (t )]

Then, the Parseval’s relation states that,

t 0 +T ¥
1
T ò x(t )1 x(t ) 2 dt = å
n = -¥
C n Dn* [for complex x1 (t)
t0 and x2 (t)]

And, Parseval’s identity states that,

t 0 +T ¥
1
ò å
2
x(t )1 x(t ) 2 dt = Cn Þ if _ x(t )1 = x(t ) 2 = x(t )
T n = -¥
t0

FacultySchool
of Electrical Engineering
of Mechatronic Engineering
Universiti Malaysia Perlis (UniMAP)
& Technology
Parseval’s Theorem

Proof: Parseval’s relation,

t 0 +T t 0 +T

ò å( )
1 1 æ ¥ ö *
T ò
t0
x1 (t ) x2* (t )dt =
T
t0
ç
ç
è n = -¥
C n e jnw0t ÷ x2 (t )dt
÷
ø

Interchanging the order of integration and summation in the RHS,

t 0 +T ¥ é t 0 +T ù
1 1
T òx1 (t ) x2* (t )dt = å
n = -¥
Cn ê
êT ò x2* (t )e jnw0t dt ú
ú
t0 ë t 0 û
*
¥ é 0 t + T ù ¥
1
= å
n = -¥
Cn ê
êT ò x2 (t )e - jn w 0 t
dt ú =
ú å
n = -¥
C n [Dn ]*
ë t 0 û

FacultySchool
of Electrical Engineering
of Mechatronic Engineering
Universiti Malaysia Perlis (UniMAP)
& Technology
Parseval’s Theorem

Therefore,
t 0 +T ¥
1
T ò x1 (t ) x2* (t )dt = å
n = -¥
C n Dn* Proved.
t0

Parseval’s Identity:
If x1 (t)=x2 (t)=x(t), then the relation changes to

t 0 +T ¥
1
T ò x(t ) x * (t )dt = å
n = -¥
C n C n*
t0

FacultySchool
of Electrical Engineering
of Mechatronic Engineering
Universiti Malaysia Perlis (UniMAP)
& Technology
Parseval’s Theorem

Since,

2 2
x(t ) = x1 (t ) x2* (t ) _ and _ C n = C n C n*

Substitute this value,

t 0 +T ¥
1
ò å
2 2
x(t ) dt = Cn Proved.
T n = -¥
t0

FacultySchool
of Electrical Engineering
of Mechatronic Engineering
Universiti Malaysia Perlis (UniMAP)
& Technology