 Chapter-3: Fourier Series

 Introduction:
 Fourier analysis involves the resolution of certain classes of signals into the
superposition of sines and cosines ( sinusoidal components ) or, equivalently
complex exponential signals.
 The output of a system to a sinewave is another sinewave of the same frequency (
but with a different phase and amplitude) under two conditions:
a) If the system is linear, and
b) If the system is time-invariant.
 Periodic signals have Fourier series representation(Continuous-Time Fourier
Series(CTFS) and Discrete-Time Fourier Series(DTFS))
 Aperiodic (or non-periodic) signals have Fourier transform representation
(Continuous-Time Fourier Transform(CTFT) and Discrete-Time Fourier Transform
(DTFT))

Khemraj koirala,WRC 1
 Fourier Series Definition:
 A general periodic signal or function can be represented as an infinite sum of sine
and cosine function (or, equivalently complex exponential function)which are then
itself periodic. This series of sine and cosine terms (or, complex exponential terms) is
called Fourier series.
 Fourier series can be studied as under:
1) Continuous-Time Fourier Series(CTFS) , and
2) Discrete-Time Fourier Series(DTFS)
1) Continuous-Time Fourier Series(CTFS) :
 It has the following forms:
A. Trigonometric Fourier series
B. Compact trigonometric Fourier series, and
C. Complex exponential Fourier series

Khemraj koirala,WRC 2
A.Trigonometric Fourier Series:
 A CT periodic signal 𝑥 𝑡 have the trigonometric Fourier series representation given
as 𝑥 𝑡 = 𝑎0 + σ∞ 𝑛=1 𝑎𝑛𝑐𝑜𝑠𝑛𝜔0𝑡 + 𝑏𝑛 𝑠𝑖𝑛𝑛𝜔0𝑡 , −∞ < 𝑡 < +∞
 Then we have ....1
Also we can write,
1 𝑇0 1 𝑇0/2 2𝜋
𝑎0 = ‫𝑥 ׬‬ 𝑡 𝑑𝑡 = ‫׬‬ 𝑥 𝑡 𝑑𝑡 , 𝜔0 = is the fundamental
𝑇0 0 𝑇0 −𝑇0/2 𝑇0
frequency.
2
 𝑎𝑛 = ‫𝜔𝑛𝑠𝑜𝑐 𝑡 𝑥 ׬‬0𝑡𝑑𝑡 ….2
𝑇0

2
 𝑏𝑛 = ‫𝜔𝑛𝑛𝑖𝑠 𝑡 𝑥 ׬‬0𝑡𝑑𝑡 ….3
𝑇0

 Note that 𝑏0 = 0 because s𝑖𝑛𝑛𝜔0𝑡 = 0 when 𝑛 = 0.

Khemraj koirala,WRC 3
B) Compact Trigonometric Fourier Series:
 Also known as polar trigonometric Fourier Series.
 Trigonometric Fourier series is
𝑥 𝑡 = 𝑎0 + σ∞𝑛=1 𝑎𝑛𝑐𝑜𝑠𝑛𝜔0𝑡 + 𝑏𝑛 𝑠𝑖𝑛𝑛𝜔0𝑡 , −∞ < 𝑡 < +∞ …..i

𝑎𝑛 𝑏𝑛
𝑥 𝑡 = 𝑎0 + σ∞ 2
𝑛=1 𝑎𝑛 + 𝑏𝑛
2 𝑐𝑜𝑠𝑛𝜔0𝑡 + 𝑠𝑖𝑛𝑛𝜔0𝑡
𝑎𝑛2+𝑏𝑛 2 𝑎𝑛2+𝑏𝑛 2

𝑎𝑛 𝑏𝑛
 Let us define, = 𝑐𝑜𝑠𝜃𝑛 and =−𝑠𝑖𝑛𝜃𝑛 …..ii
𝑎𝑛2+𝑏𝑛 2 𝑎𝑛2+𝑏𝑛 2

𝑏𝑛 −1 𝑏𝑛
So that 𝑡𝑎𝑛𝜃𝑛 = − i.e., 𝜃𝑛 = tan {− }
𝑎𝑛 𝑎𝑛

 Then eq. (ii), 𝑥 𝑡 = 𝑎0 + σ∞ 2 2

𝑛=1 𝑎𝑛 + 𝑏𝑛 (𝑐𝑜𝑠𝑛𝜔0𝑡 𝑐𝑜𝑠𝜃𝑛 − 𝑠𝑖𝑛𝑛𝜔0𝑡 𝑠𝑖𝑛𝜃𝑛)

= 𝑎0 + σ∞ 𝑎𝑛2 +
𝑛=1 Khemraj 𝑏𝑛 2 cos(𝑛𝜔0𝑡 + 𝜃𝑛)
koirala,WRC …..iii4
 Again let us define, 𝑐0 = 𝑎0 and 𝑐𝑛 = 𝑎𝑛2 + 𝑏𝑛 2 then eq. 3 becomes

𝑥 𝑡 = 𝑐0 + σ∞
𝑛=1 𝑐𝑛 cos 𝑛𝜔0𝑡 + 𝜃𝑛 …..iv
 Periodicity of the Trigonometric Fourier Series:
 We have the compact trigonometric Fourier series is
𝑥 𝑡 = 𝑐0 + σ∞ 𝑛=1 𝑐𝑛 cos 𝑛𝜔0𝑡 + 𝜃𝑛

 Then for periodicity, we know 𝑥 𝑡 + 𝑇0 = 𝑥 𝑡 𝑓𝑜𝑟 𝑎𝑙𝑙 𝑡.

 Therefore, 𝑥 𝑡 + 𝑇0 = 𝑐0 + σ∞
𝑛=1 𝑐𝑛 cos 𝑛𝜔0(𝑡 + 𝑇0) + 𝜃𝑛

= 𝑐0 + σ∞
𝑛=1 𝑐𝑛 cos{ 𝑛𝜔0𝑡 + 𝜃𝑛 + 𝑛𝜔0𝑇0}

2𝜋
But 𝑐𝑜𝑠𝑛𝜔0𝑇0 = 𝑐𝑜𝑠𝑛 𝑇0 = 𝑐𝑜𝑠2𝑛𝜋 = 1
𝑇0
Khemraj koirala,WRC 5
 Therefore, 𝑥 𝑡 + 𝑇0 = 𝑐0 + σ∞ 𝑛=1 𝑐𝑛 cos 𝑛𝜔0𝑡 + 𝜃𝑛
= 𝑥 𝑡 for all t.
 Thus, trigonometric Fourier series is periodic function of period 𝑇0 .

C) Complex Exponential Fourier Series:

1) Linear Combinations of Harmonically Related Complex Exponentials:
 The periodic complex exponential is given as …..i
𝑥 𝑡 = 𝑒 𝑗𝜔0𝑡
2π 2π
where 𝑇0 = is the fundamental period and 𝜔0 = is the fundamental
𝜔0 𝑇0
frequency.
 Then harmonically related complex exponential is
2𝜋
𝑗𝑛 𝑇 𝑡
𝑆𝑛(𝑡) = 𝑒 𝑗𝑛𝜔0𝑡
=𝑒 0 …..ii
2π
where, fundamental frequency is 𝑛 𝜔0 =
𝑇0

2𝜋
fundamental period is 𝑇0 =
𝑛 𝜔0
Khemraj Koirala, WRC
6
 Thus the linear combinations of harmonically related complex exponential is

𝑥 𝑡 = σ∞
𝑛=−∞ 𝐷𝑛𝑆𝑛(𝑡)

2𝜋
𝑗𝑛 𝑡
Or 𝑥 𝑡 = σ∞
𝑛=−∞ 𝐷𝑛𝑒
𝑗𝑛𝜔0𝑡 = σ∞ 𝐷
𝑛=−∞ 𝑛 𝑒 𝑇0 …..iii

which is periodic with period 𝑇0 .

 Eq. (iii) is the complex exponential Fourier series and the 𝐷𝑛 are called Fourier
series coefficients. It is also called synthesis equation.

• The terms for 𝑛 = ±1 of 𝐷𝑛 are called fundamental components or first harmonic

components.
• The terms for 𝑛 = ±2 are called second harmonic components.
• The terms for 𝑛 = ±𝑁 are called 𝑁𝑡ℎ harmonic components.

Khemraj koirala,WRC 7
2) Determination of Fourier series Representation of CT Periodic Signal:
 The Exponential Fourier series is
2𝜋
𝑗𝑛 𝑡
𝑥 𝑡 = σ∞ 𝑛=−∞ 𝑛 𝐷 𝑒 𝑗𝑛𝜔0𝑡
= σ ∞
𝐷
𝑛=−∞ 𝑛 𝑒 𝑇0 …..i
 Multiplying both sides of 1 by 𝑒 −𝑗𝑙𝜔0𝑡 , and integrating over the period 𝑇0 , we get
‫ 𝑒 𝑡 𝑥 ׬‬−𝑗𝑙𝜔0𝑡 𝑑𝑡 = ‫{׬‬σ∞ 𝐷
𝑛=−∞ 𝑛 𝑒 𝑗𝑛𝜔0𝑡 } 𝑒 −𝑗𝑙𝜔0𝑡 𝑑𝑡
𝑇0 𝑇0

Interchanging the order of summation and integration in RHS, we have

‫ 𝑒 𝑡 𝑥 ׬‬−𝑗𝑙𝜔0𝑡 𝑑𝑡 = σ∞
𝑛=−∞ 𝐷𝑛 ‫𝑒 ׬‬
𝑗(𝑛−𝑙)𝜔0𝑡 𝑑𝑡 …..ii
𝑇0 𝑇0

 But we know
𝑇 for 𝑛 = 𝑙
‫𝑛(𝑗 𝑒 ׬‬−𝑙)𝜔0𝑡 𝑑𝑡 = ቊ 0
𝑇0 0 for 𝑛 ≠ 𝑙

Khemraj koirala,WRC 8
 Then eq. 2 reduces to
‫ 𝑒)𝑡(𝑥 ׬‬−𝑗𝑙𝜔0𝑡 𝑑𝑡 = 𝑇0𝐷𝑙
𝑇0
1
 Therefore, 𝐷𝑙 =
𝑇0
‫ 𝑒)𝑡(𝑥 ׬‬−𝑗𝑙𝜔0𝑡 𝑑𝑡
𝑇0
1
 In general, 𝐷𝑛 =
𝑇0
‫ 𝑒)𝑡(𝑥 ׬‬−𝑗𝑛𝜔0𝑡 𝑑𝑡 …..iii
𝑇0

Eq. 3 is the required expression for Fourier series coefficients .Eq. (iii) is also called
analysis equation.
 Conversion of Trigonometric Fourier Series into Complex Exponential
Fourier Series:
 𝐷0 = 𝑎0

Khemraj koirala,WRC 9
 1
𝐷𝑛 = 𝑎𝑛 + 𝑗𝑏𝑛
2

1
𝐷−𝑛 = (𝑎𝑛 − 𝑗𝑏𝑛 )
2

 Conversion of Complex Exponential Fourier Series into Trigonometric

Fourier Series:
 𝑎0 = 𝐷0

𝑎𝑛 = 𝐷𝑛 + 𝐷−𝑛

1
𝑏𝑛 = (𝐷𝑛 − 𝐷−𝑛 )
2

Khemraj koirala,WRC 10
• Solved examples:
𝜋
1) Let 𝑥 𝑡 = 1 + 𝑠𝑖𝑛𝜔0𝑡 + 2𝑐𝑜𝑠𝜔0𝑡 + cos(2𝜔0𝑡 + ) , which has fundamental
4
frequency 𝜔0. Determine the Fourier series coefficients.

2) Determine the Fourier series coefficients of the periodic signal in the figure below.

Khemraj koirala,WRC 11
1) Solution:
𝜋
 Given, 𝑥 𝑡 = 1 + 𝑠𝑖𝑛𝜔0𝑡 + 2𝑐𝑜𝑠𝜔0𝑡 + cos(2𝜔0𝑡 + )
4
2𝜋
where, 𝜔0 = is the fundamental frequency.
𝑇0

𝜋 𝜋
𝑗(2𝜔0𝑡+ 4 ) −𝑗(2𝜔0𝑡+ 4 )
𝑒 𝑗𝜔0𝑡 −𝑒 −𝑗𝜔0𝑡 𝑒 𝑗𝜔0𝑡 +𝑒 −𝑗𝜔0𝑡 𝑒 +𝑒
𝑥 𝑡 =1+{ } + 2{ }+{ }
2𝑗 2 2

1 1 1 𝑗 𝜋 𝑗2𝜔 𝑡 1 −𝑗 𝜋 −𝑗2𝜔 𝑡
𝑥 𝑡 =1+ 1 + 𝑒 𝑗𝜔0𝑡 + 1− 𝑒 −𝑗𝜔0𝑡 + 𝑒 4𝑒 0 + 𝑒 4𝑒 0
2𝑗 2𝑗 2 2

1 −𝑗 𝜋 −𝑗2𝜔 𝑡 1 −𝑗𝜔0𝑡 1 𝑗𝜔0𝑡 1 𝑗 𝜋 𝑗2𝜔 𝑡

𝑥 𝑡 = ( 𝑒 4 )𝑒 0 + 1 − 𝑒 +1+ 1 + 𝑒 + ( 𝑒 4 )𝑒 0
2 2𝑗 2𝑗 2

 Comparing this with the eq. 𝑥 𝑡 = σ∞

𝑛=−∞ 𝐷𝑛𝑒
𝑗𝑛𝜔0𝑡

or 𝑥 𝑡 = ⋯ + 𝐷 − 2𝑒 −𝑗2𝜔0𝑡 + 𝐷 − 1Khemraj
𝑒 −𝑗𝜔koirala,WRC
𝑡
0 + 𝐷 + 𝐷 𝑒
0 1
𝑗𝜔0𝑡 + 𝐷 𝑒 𝑗2𝜔0𝑡 +…..
2 12
 We get
1 1
𝐷0 = 1 , 𝐷1 = 1 + , 𝐷−1= 1−
2𝑗 2𝑗
1 𝑗𝜋 1 −𝑗 𝜋
𝐷2 = ( 𝑒 4 ), 𝐷−2= 𝑒 4
2 2

1 𝜋 𝜋 2
= 𝑐𝑜𝑠 + 𝑗𝑠𝑖𝑛 = 1−𝑗
2 4 4 4

1 1 1 2
= { +𝑗 } ×
2 2 2 2

2
= (1 + 𝑗)
4

 And 𝐷𝑛 = 0 𝑓𝑜𝑟 𝑛 > 2.

Khemraj koirala,WRC 13
2) Solution:
 From the given figure the signal 𝑥 𝑡 is defined as
1, 𝑡 < 𝑇1
𝑥 𝑡 =ቐ 𝑇
0, 𝑇1 < 𝑡 <
2

The signal is periodic with fundamental period T and fundamental frequency

2π
𝜔0 = . This is a square wave.
𝑇
Fourier series coefficients:
1
 We have, 𝐷0 = ‫𝑡𝑑 𝑡 𝑥 ׬‬
𝑇
𝑇

1 𝑇1 1
= ‫ 𝑇 𝑡𝑑 ׬‬2𝑇1
𝑇 −𝑇1

2𝑇1
𝐷0 = , 𝑛=0
𝑇
Khemraj koirala,WRC 14
 1
 We know, 𝐷𝑛 =
𝑇
‫ 𝑒)𝑡(𝑥 ׬‬−𝑗𝑛𝜔0𝑡 𝑑𝑡
𝑇
1 𝑇1
= ‫ ׬‬1 𝑒 −𝑗𝑛𝜔0𝑡 𝑑𝑡
𝑇 −𝑇1

1 𝑒 −𝑗𝑛𝜔0𝑡 T1 1 2 𝑒 𝑗𝑛𝜔0𝑇1 −𝑒 −𝑗𝑛𝜔0𝑇1

= = 𝑒 −𝑗𝑛𝜔0𝑇1 − 𝑒 𝑗𝑛𝜔0𝑇1 = ( )
𝑇 −𝑗𝑛𝜔0 -T1 −𝑗𝑛𝜔0𝑇 𝑛𝜔0𝑇 2𝑗

2 2𝜋
= 𝑠𝑖𝑛𝑛𝜔0𝑇1 , 𝑏𝑢𝑡 𝜔0𝑇 = 𝑇 = 2𝜋.
𝑛𝜔0𝑇 𝑇

1
 Therefore, 𝐷𝑛 = sin 𝑛𝜔0𝑇1 , 𝑛 ≠ 0
𝑛𝜋

Khemraj koirala,WRC 15
 2𝜋 𝑇 𝜋
 For 𝑇 = 4𝑇1 , 𝜔0𝑇1 = × = then
𝑇 4 2

2𝑇1 2𝑇1 1
𝐷0 = = = , 𝑛=0
𝑇 4𝑇1 2

𝑛𝜋
sin( 2 )
𝐷𝑛 = , 𝑛≠0
𝑛𝜋

1
𝐷1 = 𝐷−1 = 𝐷2 = 𝐷−2 = 0
𝜋

1
𝐷3 = 𝐷−3 = − 𝐷4 = 𝐷−4 = 0
3𝜋

1
𝐷5 = 𝐷−5 = and so on.
5𝜋

Khemraj koirala,WRC 16
 The plot of 𝐷𝑛 for 𝑇 = 4𝑇1 is shown below.

𝐷𝑛

Khemraj koirala,WRC 17
 Dirichlet Conditions:
 This is a set of conditions which guarantees that a signal 𝑥 𝑡 has its Fourier series
representation except at some isolated values of t for which 𝑥 𝑡 is discontinuous.
 Conditions:
1) Over any one period, 𝑥 𝑡 must be absolutely integrable, i.e.,
‫∞ < 𝑡𝑑 )𝑡(𝑥 ׬‬ .....i
𝑇0

In this case 𝐷𝑛 is also finite.

1 1
𝐷𝑛 = ‫𝑡 𝑥׬‬ 𝑒 −𝑗𝑛𝜔0𝑡 𝑑𝑡 = ‫∞ < 𝑡𝑑 )𝑡(𝑥 ׬‬
𝑇0 𝑇0
𝑇0 𝑇0

Khemraj koirala,WRC 18
2) The signal contains finite number of maxima and minima over any finite
period.
3) The signal 𝑥 𝑡 has finite number of discontinuities over any finite period.

• Examples that violets Dirichlet conditions:

1
1) 𝑥 𝑡 = , 0 < 𝑡 ≤ 1 . This signal is periodic with period 1, but violets condition 1.
𝑡

Khemraj koirala,WRC 19
 2𝜋
2) 𝑥 𝑡 = sin( ) for 0 < 𝑡 ≤ 1. This signal is periodic with period 1 but violets
𝑡
condition 2.

Khemraj koirala,WRC 20
 3)

 This signal is periodic with period 𝑇 = 8 but violets condition3.

Khemraj koirala,WRC 21
 The Fourier Spectrum:
 A periodic signal 𝑥 𝑡 can be specified into two equivalent ways:
1. the time-domain representation where 𝑥 𝑡 is defined as a function of time, and
2. the frequency-domain representation where the signal is defined in terms of its
spectrum. It consists of frequency components 0, ±𝜔0, ±2𝜔0 , ±3𝜔0 .. called
spectrum.

 We know 𝐷𝑛 = 𝐷𝑛 𝑒 𝑗𝜃𝑛 where 𝜃𝑛 = arg {𝐷𝑛} is the phase.

• Amplitude spectrum:
 The plot of amplitudes 𝐷𝑛 versus frequency is called amplitude spectrum.
For real-valued periodic signal 𝐷−𝑛 = 𝐷∗ 𝑛 , where𝐷∗ 𝑛 is the complex
conjugate of 𝐷𝑛 . Then 𝐷−𝑛 = 𝐷𝑛
That is , amplitude spectrum is symmetric or even function of n.

Khemraj koirala,WRC 22
• Phase spectrum:
 The plot of phase 𝜃𝑛 versus frequency gives the phase spectrum.
𝜃−𝑛 = −𝜃𝑛
Phase spectrum is asymmetric or odd function of n.
 Since both the amplitude and phase of 𝐷𝑛 have nonzero values only for discrete
frequencies that are integers, the spectrum (amplitude and phase spectrum) is also
referred as discrete spectrum or line spectrum.

• Note that if we specify the periodic signal 𝑥 𝑡 , we can determine its spectrum;
conversely, if we specify the spectrum, we can determine the corresponding signal.

Khemraj koirala,WRC 23
 Properties of CTFS:
 The CTFS representation consists of a number of properties which may be used to
develop conceptual fundamentals and in reducing the complexity of Fourier series of
many signals.

1) Linerity:
𝐹𝑆
 Let 𝑥 𝑡 𝐷𝑛

𝐹𝑆
and 𝑦 𝑡 𝑌𝑛 , both are periodic with period 𝑇0.

𝐹𝑆
then 𝑧 𝑡 = 𝑎𝑥 𝑡 + b𝑦 𝑡 𝑍𝑛 = 𝑎𝐷𝑛 + 𝑏𝑌𝑛 , …..i
𝑧 𝑡 is also periodic with period 𝑇0.

Khemraj koirala,WRC 24
 Proof:
1
 We know, 𝑍𝑛 =
𝑇0
‫ 𝑒 )𝑡(𝑧 ׬‬−𝑗𝑛𝜔0𝑡 𝑑𝑡
𝑇0
1
=
𝑇0
‫ 𝑡 𝑥𝑎{׬‬+ 𝑏𝑦 𝑡 } 𝑒 −𝑗𝑛𝜔0𝑡 𝑑𝑡
𝑇0
1 1
=𝑎
𝑇0
‫ 𝑒 )𝑡(𝑥 ׬‬−𝑗𝑛𝜔0𝑡 𝑑𝑡 + 𝑏{ 𝑇0
‫ 𝑒 )𝑡(𝑦 ׬‬−𝑗𝑛𝜔0𝑡 𝑑𝑡}
𝑇0 𝑇0

𝑍𝑛 = 𝑎𝐷𝑛 + 𝑏𝑌𝑛

2) Time Shifting:
𝐹𝑆
 Let 𝑥 𝑡 𝐷𝑛
𝐹𝑆
then 𝑦 𝑡 = 𝑥 𝑡 − 𝑡0 ,
𝑌𝑛 = 𝑒 −𝑗𝑛𝜔0𝑡0 𝐷𝑛 …..i

Khemraj koirala,WRC 25
Proof:
1
 We have , 𝑌𝑛 =
𝑇0
‫ 𝑒 )𝑡(𝑦 ׬‬−𝑗𝑛𝜔0𝑡 𝑑𝑡
𝑇0
1
=
𝑇0
‫ 𝑡(𝑥 ׬‬− 𝑡0) 𝑒 −𝑗𝑛𝜔0𝑡 𝑑𝑡
𝑇0

 Let, (𝑡 − 𝑡0) = 𝜏 and 𝜏 it will also range over 𝑇0. Then

1
𝑌𝑛 = ‫ 𝑒 )𝜏(𝑥 ׬‬−𝑗𝑛𝜔0(𝜏+𝑡0) 𝑑𝜏
𝑇0
𝑇0
1
𝑌𝑛 = {
𝑇0
‫ 𝑒 𝜏 𝑥 ׬‬−𝑗𝑛𝜔0𝜏 𝑑𝜏} 𝑒 −𝑗𝑛𝜔0𝑡0 ,
𝑇0

𝑌𝑛 = 𝑒 −𝑗𝑛𝜔0𝑡0 𝐷𝑛

• Note that 𝑌𝑛 = 𝐷𝑛 .Hence, time-shifting does not change the amplitude. It only
changes the phase spectrum.
Khemraj koirala,WRC 26
3) Frequency Shifting:
𝐹𝑆
 Let 𝑥 𝑡 𝐷𝑛
𝐹𝑆
then 𝑒 −𝑗𝑚𝜔0𝑡 𝑥 𝑡 𝐷𝑛−𝑚 …..i
Proof:
1
 We have, 𝐷𝑛 =
𝑇
‫ 𝑒)𝑡(𝑥 ׬‬−𝑗𝑛𝜔0𝑡 𝑑𝑡
𝑇0
1 1
This implies that
𝑇0
‫𝜔𝑚𝑗 𝑒 )𝑡(𝑥{׬‬0𝑡 } 𝑒 −𝑗𝑛𝜔0𝑡 𝑑𝑡 = 𝑇 ‫ 𝑒 )𝑡(𝑥 ׬‬−𝑗(𝑛−𝑚)𝜔0𝑡 𝑑𝑡 = 𝐷𝑛−𝑚
0
𝑇0 𝑇0
𝐹𝑆
 Thus we can write, 𝑒 𝑗𝑚𝜔0𝑡 𝑥 𝑡 𝐷𝑛−𝑚

Khemraj koirala,WRC 27
4) Time reversal:
𝐹𝑆
 If 𝑥 𝑡 𝐷𝑛
𝐹𝑆
then 𝑥 −𝑡 𝐷−𝑛 …..i
Proof:
 We have , 𝑥 𝑡 = σ∞
𝑛=−∞ 𝐷𝑛𝑒
𝑗𝑛𝜔0𝑡
…..ii

then, 𝑥 −𝑡 = σ∞ 𝐷
𝑛=−∞ 𝑛 𝑒 −𝑗𝑛𝜔0𝑡

 Let −𝑛 = 𝑚 in RHS , we get

𝑥 −𝑡 = σ∞
𝑚=−∞ 𝐷−𝑚 𝑒
𝑗𝑚𝜔0𝑡

 In general, 𝑥 −𝑡 = σ∞ 𝐷
𝑛=−∞ −𝑛 𝑒 𝑗𝑛𝜔0𝑡
…..iii

𝐹𝑆
 Therefore 𝑥 −𝑡 𝐷−𝑛
Khemraj koirala,WRC 28
• Time reversal in signal 𝑥 𝑡 dies not change in the period T0 , but results in a time
reversal of the corresponding sequence of Fourier series coefficients.
• Cases:
I. When 𝑥 −𝑡 = 𝑥 𝑡 for all t (even signal),𝐷−𝑛 = 𝐷𝑛.
II. When 𝑥 −𝑡 = - 𝑥 𝑡 for all t (odd signal),𝐷−𝑛 = −𝐷𝑛.

5) Time Scaling:
𝐹𝑆
 If 𝑥 𝑡 𝐷𝑛

𝐹𝑆
then 𝑥 𝑎𝑡 𝑍𝑛 = 𝐷𝑛 …..i
 That is, this property does not change the Fourier series coefficients but change the
fundamental period (and thus fundamental frequency).
Proof:
 Let 𝑥 𝑡 = σ∞ 𝑛=−∞ 𝐷𝑛𝑒
𝑗𝑛𝜔0𝑡

𝑥 𝑎𝑡 = σ∞
𝑛=−∞ 𝑛𝑍 𝑒
Khemraj 𝑗𝑛(𝑎𝜔
Koirala, WRC0)𝑡 29
 1 1
where 𝐷𝑛 =
𝑇
‫ 𝑒)𝑡(𝑥 ׬‬−𝑗𝑛𝜔0𝑡 𝑑𝑡 and 𝑍𝑛 = 𝑇 ′ ‫ 𝑒)𝑡𝑎(𝑥 ׬‬−𝑗𝑛(𝑎𝜔0)𝑡 𝑑𝑡
0
𝑇0 𝑇0’
2𝜋 2𝜋 𝑇0
 Let 𝜔0′ = 𝑎𝜔0 then 𝑇0’ = = = ,
𝜔0′ 𝑎𝜔0 𝑎

1
 Therefore , 𝑍𝑛 = ‫׬‬ 𝑥(𝑎𝑡)𝑒 −𝑗𝑛(𝑎𝜔0)𝑡 𝑑𝑡
(𝑇0/𝑎) 𝑇 /a
0

1
 Put 𝑎𝑡 = 𝜏 𝑡ℎ𝑒𝑛 𝑑𝑡 = 𝑑𝜏
𝑎

1
 Therefore, 𝑍𝑛 = ‫׬‬ 𝑥(𝜏) 𝑒 −𝑗𝑛𝜔0𝜏 𝑑𝜏 = 𝐷𝑛
𝑇0 𝑇
0

 Hence , 𝑍𝑛 = 𝐷𝑛

Khemraj koirala,WRC 30
6) Multiplication:

𝐹𝑆 𝐹𝑆
 If 𝑥 𝑡 𝐷𝑛 and 𝑦 𝑡 𝑌𝑛
𝐹𝑆
then 𝑧 𝑡 = 𝑥 𝑡 . 𝑦 𝑡 𝑍𝑛 = σ∞ 𝑙=−∞ 𝐷𝑙 𝑌𝑛−𝑙 …..i
where 𝑥 𝑡 , 𝑦 𝑡 𝑎𝑛𝑑 𝑧 𝑡 are periodic with period T0.
Proof:
 Given 𝑥 𝑡 = σ∞𝑛=−∞ 𝐷𝑛𝑒
𝑗𝑛𝜔0𝑡

𝑦 𝑡 = σ∞ 𝑌
𝑛=−∞ 𝑛 𝑒 𝑗𝑛𝜔0𝑡

then 𝑥 𝑡 . 𝑦 𝑡 = {σ∞ 𝐷
𝑙=−∞ 𝑙 𝑒 𝑗𝑙𝜔0𝑡 } {σ∞
𝑚=−∞ 𝑌 𝑚 𝑒 𝑗𝑚𝜔0𝑡 }

= σ∞ 𝐷
𝑙=−∞ 𝑙 {σ ∞
𝑚=−∞ 𝑌 𝑚 𝑒 𝑗 𝑙+𝑚 𝜔0𝑡 } …..ii
 Put 𝑙 + 𝑚 = 𝑛 in eq.i , then 𝑚 = 𝑛 − 𝑙
when 𝑚 = −∞, 𝑛 = −∞ and when 𝑚 = ∞, 𝑛 = ∞
Khemraj koirala,WRC 31
 Then eq.1 becomes
𝑥 𝑡 . 𝑦 𝑡 = σ∞ 𝐷 σ ∞ 𝑌
𝑙=−∞ 𝑙 𝑛=−∞ 𝑛−𝑙 𝑒 𝑗𝑛𝜔0𝑡

∞
i.e., 𝑥 𝑡 . 𝑦 𝑡 = σ∞ {σ 𝐷 𝑌
𝑛=−∞ 𝑙=−∞ 𝑙 𝑛−𝑙 }𝑒 𝑗𝑛𝜔0𝑡

 If 𝑍𝑛 are the Fourier series coefficients of 𝑥 𝑡 . 𝑦 𝑡 then we have

𝑥 𝑡 . 𝑦 𝑡 = σ∞ 𝑍
𝑛=−∞ 𝑛 𝑒 𝑗𝑛𝜔0𝑡 …..iii

where, 𝑍𝑛 = σ∞
𝑙=−∞ 𝐷𝑙 𝑌𝑛−𝑙

• Note that 𝑍𝑛 is convolution of 𝐷𝑛 and 𝑌𝑛 .

Khemraj koirala,WRC 32
7) Conjugation and Conjugate Symmetry:
 Taking the complex conjugate of a periodic signal 𝑥 𝑡 has the effect of complex
conjugation and time reversal on the corresponding Fourier series coefficients.
 That is, if
𝐹𝑆
𝑥 𝑡 𝐷𝑛

𝐹𝑆
then 𝑥 𝑡∗
𝐷∗ _𝑛 …..i
Proof:
 Let, 𝑥 𝑡 = σ∞ 𝑛=−∞ 𝐷𝑛 𝑒
𝑗𝑛𝜔0𝑡
…..ii
 Taking complex conjugate on both sides ,we get
∞ 𝑗𝑛𝜔 𝑡 ∗
𝑥 𝑡 = σ𝑛=−∞ 𝐷𝑛 𝑒
∗
0 = σ∞𝑛=−∞ 𝐷 𝑛 𝑒
∗ −𝑗𝑛𝜔0𝑡

 Put 𝑛 = −𝑛, we obtain

𝑥∗ 𝑡 = σ∞ 𝑛=−∞ 𝐷 _𝑛 𝑒
∗ 𝑗𝑛𝜔0𝑡 …..iii

𝐹𝑆
 Therefore , 𝑥 𝑡
∗
𝐷 ∗
_𝑛
Khemraj koirala,WRC 33
Cases:
i. If 𝒙 𝒕 is real:
 Here, 𝑥 𝑡 = 𝑥∗ 𝑡 then 𝐷𝑛 = 𝐷∗ _𝑛 , i.e., 𝐷_𝑛 = 𝐷∗ 𝑛

 Hence 𝐷𝑛 = 𝐷_𝑛
ii. If 𝒙 𝒕 is real and even:
 If 𝑥 𝑡 is real 𝐷𝑛 =
𝐷∗
_𝑛

 If 𝑥 𝑡 is even 𝐷𝑛 =
𝐷_𝑛

Then 𝐷_𝑛 = 𝐷∗ _𝑛 , i.e., 𝐷𝑛 = 𝐷∗ 𝑛

 Which shows that 𝐷𝑛 is purely real and contains no imaginary part.
 In this case 𝐷0 is real.

Khemraj koirala,WRC 34
iii. If 𝒙 𝒕 is real and odd:
 If 𝑥 𝑡 is real 𝐷𝑛 =
𝐷∗ _𝑛

 If 𝑥 𝑡 is odd 𝐷𝑛 =
−𝐷_𝑛

 Then −𝐷_𝑛 = 𝐷∗ _𝑛 ,i.e., 𝐷∗ 𝑛 = −𝐷𝑛

 In this case , Fourier series coefficients are purely imaginary and odd.
 Also 𝐷0 = 0

Khemraj koirala,WRC 35
8) Convolution:
𝐹𝑆 𝐹𝑆
 If 𝑥 𝑡 𝐷𝑛 , and 𝑧 𝑡 𝑍𝑛
then the convolution of 𝑥 𝑡 and 𝑧 𝑡 over a period 𝑇0, results in a signal 𝑦 𝑡 ,
𝐹𝑆
such that 𝑦 𝑡 =𝑥 𝑡 ∗𝑧 𝑡 𝑌𝑛 = 𝑇0𝐷𝑛𝑍𝑛 . …..i

Proof:
 The convolution(for periodic signals) is defined as
𝑇
𝑦 𝑡 = ‫׬‬0 0 𝑥 𝜏 𝑧 𝑡 − 𝜏 𝑑𝜏
𝐹𝑆
 From time shifting property, we have 𝑧 𝑡 − 𝜏 𝑍𝑛 𝑒 −𝑗𝑛𝜔0𝜏 , thus
𝑇0 𝐹𝑆 𝑇0 𝑇 0 1 𝑇0
−𝑗𝑛𝜔0𝜏 𝑑𝜏 𝑇0𝑍𝑛 { ‫׬‬0 𝑥(𝜏)𝑒 −𝑗𝑛𝜔0𝜏 𝑑𝜏}
‫׬‬0 𝑥 𝜏 𝑧(𝑡 − 𝜏) 𝑑𝜏 ‫׬‬
𝑇0 0
𝑥(𝜏) 𝑍𝑛 𝑒 =
𝑇0
= 𝑇0𝑍𝑛 𝐷𝑛
𝑇0 𝐹𝑆
 Hence, ‫׬‬0 𝑥 𝜏 𝑧(𝑡 − 𝜏) 𝑑𝜏 𝑇0 𝑍𝑛𝐷𝑛

Khemraj koirala,WRC 36
9) Parseval’s Relation for CT Periodic Signals:
 It states that the total average power of a periodic signal 𝑥 𝑡 is equal to the sum of
the average power of its harmonic components (or Fourier components).

1
 Mathematically, 𝑃𝑎𝑣 =
𝑇0
‫ )𝑡(𝑥 ׬‬2 𝑑𝑡 = σ∞
𝑛=−∞ 𝐷𝑛
2 , …..i
𝑇0
1
where, ‫׬‬ 𝑥(𝑡) 2 𝑑𝑡 is the total average power of the periodic signal 𝑥 𝑡 .
𝑇0
𝑇0
Proof:
1
 𝑃𝑎𝑣 = ‫ )𝑡(𝑥 ׬‬2 𝑑𝑡
𝑇0 𝑇
0

𝑥 𝑡 = σ∞ 𝐷
𝑛=−∞ 𝑛 𝑒 𝑗𝑛𝜔0𝑡
and𝑥∗
𝑡 = σ ∞
𝑛=−∞ 𝑛 𝑒
𝐷∗ −𝑗𝑛𝜔0𝑡

𝑥∗ 𝑡 is the complex conjugate of 𝑥 𝑡 .

Khemraj koirala,WRC 37
 Then we have,
1 1
𝑃𝑎𝑣 =
𝑇0
‫ )𝑡(𝑥 ׬‬2 𝑑𝑡 = 𝑇0
‫ 𝑡 𝑥 ׬‬. 𝑥∗ 𝑡 𝑑𝑡
𝑇0 𝑇0
1
=
𝑇0
‫ 𝑡 𝑥 ׬‬. { σ∞ 𝐷∗
𝑛=−∞ 𝑛 𝑒 −𝑗𝑛𝜔0𝑡 } 𝑑𝑡

𝑇0
∞ 1
= σ𝑛=−∞ 𝐷 𝑛 {
∗
𝑇0
‫ 𝑒 𝑡 𝑥 ׬‬−𝑗𝑛𝜔0𝑡 𝑑𝑡}
𝑇0
= σ∞
𝑛=−∞ 𝑛 𝐷𝑛
𝐷∗

= σ∞
𝑛=−∞ 𝐷𝑛
2

 Where 𝐷𝑛 2 is the average power in the nth harmonic (or Fourier series) component.
The power spectrum is the symmetric spectrum.

Khemraj koirala,WRC 38
 Note:
a. A spectrum showing the relation of average power to the corresponding frequency
component(harmonics) is called power density spectrum.

Khemraj koirala,WRC 39