Chapter

Continuous-Time
Fourier Series 3
Introduction
As we have seen in the previous chapter, we can obtain the response of a linear system to an arbitrary input
by representing it in terms of basic signals. The specific signals used were the shifted -functions. Often, it is convenient
to choose a set of orthogonal waveforms as the basic signals. There are several reasons for doing this. First, it is
mathematically convenient to represent an arbitrary signal as a weighted sum of orthogonal waveforms, since many
of the calculations involving signals are simplified by using such a representation. Second, it is possible to visualize
the signal as a vector in orthogonal coordinate system, with the orthogonal waveforms being coordinates. Finally,
representations in terms of orthogonal basis functions provides a convenient means of solving for the response of
linear systems to arbitrary inputs.
For periodic signals, a convenient choice for an orthogonal basis is the set of harmonically related complex
exponential. The choice of these waveforms is appropriate since such complex exponentials are periodic, are relatively
easy to manipulate mathematically, and yield results that the have a meaningful physical interpretation. The
representation of a non-sinusoidal periodic signal in terms of complex exponentials, or equivalently; in terms of sine
and cosine waveforms, leads to the Fourier series. Or, in other words, the Fourier series as a mathematical tool that
allows the representation of any periodic signal as the sum of harmonically related sinusoids.
So we saw that, it is mathematically convenient to represent a signal in terms of orthogonal signals. Another
reason for doing this is:
when we represent any signal x(t) as a weighted sum of a basic signal then it is easier to find output of that
signal when it is applied as input to an LTI system. The basic signal is eC1t where s1 is any constant. Suppose eC1t is
input to the LTI system.

C1t
Input =e LTI System y(t)

Fig. 3.1 : e s1t is input to system

So, y(t) = input  h(t) ...(1)

= eC1t  h(t)

C ( t  )
=  h ( ) e 1 d


GATE MASTERS PUBLICATION

 106 Signals & Systems


 y(t) = eC1 t  h() e
 C1
d {Here C1 is a constant} ...(2)


Now if we define a new function H(s) where ‘s’ is variable,


 st
H(s) =  h(t ) e d ...(3)


In the above equation ‘s’ is variable, and


H (s ) s  C = C1 t
1  h(t ) e dt


So, equation (2) can be modified as,

y(t) = eC1 t  H (C1 ) ...(4)

Thus if we define x(t) as a weighted sum of eCk t for k = 1, 2, 3... then output of LTI system for x(t) as input
can be easily determined.

That is if, x(t) =  ak eCkt


System
Then, output of eCk t  H ( s) s  C
k

System
Output of ak eCk t  ak eCk t  H ( s)
s  Ck

 
System
  ak eCk t   ak  eCk t  H ( s)
s  Ck
k   k  

Study Note

• Thus when input to an LTI system is of the form eCt where C is a constant, then output is eCt  H ( s) .
s C

• eCt is eigen function of the continuous time LTI system and H ( s ) s  C is eigen value of this function

3.1 Fourier Series Representation of Continuous Time Periodic Signal

Let x(t) is a periodic signal with period T, if Ck represent exponential Fourier series coefficient then,

x(t) =  Ck exp ( j o kt ) ...(5)
k  

1
Ck =  x (t ) e  j o kt dt ...(6)
T t T 

2
In above equations o  , and t = <T> means integration over one period of x(t).
T
If we analyze equation (5) we can see that,
x(t) = ....+ C–2 exp(–j2ot) + C–1 exp(–jot) + C0 + C1 exp(jot) + C2 exp(2jot) + ....

GATE MASTERS PUBLICATION

 Continuous-Time Fourier Series 107

So, C0 : Shows the dc content in signal x(t).

C1/C–1 : Show content of o/–o that is fundamental frequency in signal x(t).
C2/C–2 : Show content of 2o/–2o that is content of 2nd harmonic in signal x(t).
CN/C–N : Show content of No/–No, that is content of Nth harmonic in signal x(t).
Again analyzing equation (5) we can see that,
x(t) = ...+ C–2 exp(–2jot) + C–1 exp(–jot) + C0
+ C1 exp(jot) + C2 exp(2jot) + ....
Since we know that, exp(jt) = cos(t) + j sin(t) ...(6)
Now using equation (6), x(t) can be written as,
x(t) = ...+ C–2[cos(2ot) – j sin(2ot)] + C–1[cosot – j sinot]
+ C0 + C1[cos(ot) + j sin(ot)] + C2[cos(2ot) + j sin(2ot)] +....
 x(t) = C0 + cos(ot) [C1 + C–1] + cos(2ot) [C2 + C–2] + ....
+ sin(ot) [j(C1 – C–1)] + sin(2ot) [j(C2 – C–2)] + ....
 
 x(t) = C0   cos( k o t ) [Ck  C k ]   sin( k o t ) [ j (Ck  C k )] ...(7)
k 1 k 1

The above equation is written in simplified form as,

 
x(t) = C0   ak cos( k o t )   bk sin( k o t ) ...(8)
k 1 k 1

The equation (8) represented x(t) in terms of sine and cosine terms. We can see that, ak are coefficient of
cosine terms and bk are coefficient of sine terms.
Also, ak = Ck + C–k ...(9)
bk = j[Ck – C–k] ...(10)

Study Note
• C0 is the average value of signal x(t),

1
C0 =  x(t ) dt
T tT 

• If in any question if we have to find ak or bk then best method is to find Ck using equations (6) and use (9) and (10) to find
answer.
• To find ak or bk we have direct relations also,
2
ak =  x (t ) cos( k  o t ) dt
T t  T 

2
bk =  x (t ) sin( k  o t ) dt
T t  T 

Let us assume that signal x(t) is real.

Since, x(t) is a real signal.
 x(t) = x(t) ...(11)
Now let us assume exponential Fourier series coefficient of x(t) are Ck, thus
1
Ck =  x (t ) e  j o kt dt ...(12)
T T 

GATE MASTERS PUBLICATION

 108 Signals & Systems

Let, the exponential Fourier series coefficient of x(t) are Bk, thus
1
Bk =  x* (t ) e  jo kt dt ...(13)
T T 

Taking conjugate both side in equation (12) we get

1
Ck =  x  (t ) e j o kt dt ...(14)
T T 

Putting k = –k in equation (14) we get,

1
C k =  x * (t ) e  j o kt dt ...(15)
T T 

By comparing equation (15) and (12) we get

Bk = C k

Thus if Ck are exponential Fourier coefficients of x(t) then exponential Fourier coefficient of x(t) will be C k .
And if x(t) is real i.e. x(t) = x(t) then
 Ck = C k ...(16)

REMEMBER If x(t) is real then,

Ck = C k and Ck  Ck
That is exponential Fourier series coefficient are conjugate symmetric.

Now, for real signal x(t) the trigonometric Fourier series coefficient will be
ak = Ck  Ck  Ck  Ck  2 Real{Ck } ...(17)

bk = {Ck  Ck }  j{Ck  Ck }  2 Img {Ck } ...(18)

REMEMBER For real valued signal x(t)

ak = 2 times Real part of Ck
bk = –2 times Imaginary part of Ck

Since any signal real periodic signal x(t) can be written as sum of it’s even and odd parts that is xe(t) and xo(t).
If exponential Fourier series coefficients of x(t) are Ck then lets find out contribution due to xe(t) and xo(t).
 x(t) = xe(t) + xo(t) [Since x(t) is real xe(t) and x0(t) will also be real]
1
and Ck =  x (t ) e  j o kt dt
T T 

1
=  ( xe (t )  xo (t )) e  jo kt dt
T T 

1 1
=  xe (t ) e  j o kt dt   xo (t ) e  j o kt dt
T T 
T T 

= Exponential Fourier series coefficient of xe(t) +

Exponential Fourier series coefficient of xo(t)
The exponential Fourier series coefficient of xe(t) are Ek

GATE MASTERS PUBLICATION

 Continuous-Time Fourier Series 109

1
 Ek =  xe (t ) e  j o kt dt
T T 

1
=  xe (t ) [cos(o kt )  j (sin o kt )] dt
T T 

In above equation xe(t) is real and even.

So, xe(t) cos(okt) will also be real and even and xe(t) sin(okt) will be real and odd.

1  
 Ek =   xe (t ) cos(o kt ) dt  j  xe (t ) sin(o kt ) dt 
T T  
 T 

Since xe(t) sin(okt) is odd it’s integration over the period will be zero.
1
 Ek =  x e(t ) cos( 0 kt ) dt,
T T

since xe(t) and cos(okt) both are real so Ek will also be real.
The exponential Fourier series coefficient of xo(t) are Ok,
1
 Ok =  xo (t ) e  jo kt dt
T T 

1
=  xo (t ) (cos(o kt )  j sin(o kt )) dt
T T 

Here, xo(t) cos(okt) will be odd and xo(t) sin(okt) will be even.
1
 Ok =  j  xo (t ) sin(  o kt ) dt
T T 

Since xo(t) and sin(okt) both are real so Ok is imaginary.

Thus, Ck = [Ek + Ok]
Here, Ek is real and Ok is imaginary.
 Real {Ck} = Exponential Fourier series coefficient of xe(t) ...(19)
and jImaginary{Ck} = Exponential Fourier series coefficient of x0(t) ...(20)

REMEMBER If x(t) is real and Ck are its exponential Fourier series coefficients then Real {Ck} i.e. real part of Ck are
exponential Fourier series coefficients of xe(t) and j Img{Ck} are exponential Fourier series coefficient
of x0(t).

3.2 Convergence of Fourier Series

Any periodic signal, i.e., one for which x(t) = x(t + T), can be expressed by a Fourier series provided that,
1. if it is discontinuous, there are a finite number of discontinuities in the period T;
2. it has a finite average value over the period T;
3. it has a finite number of positive and negative maxima in the period T.
When these Dirichlet condition are satisfied, the Fourier series exist. These conditions are sufficient but not
necessary.

GATE MASTERS PUBLICATION

110 Signals & Systems

Example 3.1

Find the exponential Fourier series coefficient of the signal x(t).

x(t)

1.0

t
–8 –6 –4 –2 2 4 6 8

Solution 3.1
We can see that x(t) is periodic with period and T = 8. To find Ck
4
1  j o kt
Ck =  x (t ) e dt
8 4

2 
o = 
8 4
4 2
1  jk  / 4t 1  jk  / 4t
 Ck =  x(t ) e dt   1 e dt
8 4
8 2

 k 
k k sin  
1 j j  2 
Ck =  [e 2 e 2 ]
j 2k k
We can find C0 using above expression and applying L-Hospital rule, or C0 is average value of signal x(t).
4 2
1 1 1
So, C0 =  x (t ) dt   1 dt  2
8 4
8 2

The plot of Ck is
Ck

0.5

1/

–3
k
–2 –1 0 1 2 3

–1/3 –1/3

Study Note
We can see that using Fourier series we find Ck of a signal, Ck are Fourier series coefficient which has value only at integer
value of k, so Ck is a discrete signal.Generally Ck is non-periodic, thus we say that Fourier series convert a continuous
periodic signal x(t) into discrete non-periodic signal Ck.

GATE MASTERS PUBLICATION

 Continuous-Time Fourier Series 111

Example 3.2

The periodic signal x(t) is real and Ck are it’s exponential Fourier series coefficient. If value of C1 = 1 + 3j,
C2 = 0.5 + 0.5j then the value of C–1 and C–2 will be _______ and ______.
Solution 3.2
Since, x(t) is real so Ck will be conjugate symmetric.
 Ck = C k or Ck  Ck

 C–1 = C1  1  3 j

 C–2 = C2  0.5  0.5 j

Example 3.3

The real periodic signal x(t) has exponential Fourier series coeffici ent, Ck  0.5  3 j . Then what is
k k
exponential Fourier series coefficient of even and odd part of x(t).
Solution 3.3
We know that real part of Ck correspond to exponential Fourier series coefficient of even part of x(t). That is 0.5/
k.
3
And for odd part of x(t) the exponential Fourier series coefficient are j Imaginary part of Ck. That is j .
k

3.3 Polar Form Representation of Fourier Series

We have already seen exponential and trigonometric representation of Fourier series, the trigonometric
representation can be modified into polar form.
Case-1: From equation (8),

  
x(t) =  C0   ak cos( k o t )  bk sin( k o t ) 
 k 1 

If we assume that, dk = ak2  bk2 ...(21)

ak = dk cosk ...(22)
bk = –dk sink ...(23)

So, x(t) = C0   d k cos  k cos( k o t )  d k sin  k sin( k o t )
k 1


= C0   d k cos( k o t   k ) ...(25)
k 1

The equation (25) represent polar form of Fourier series of signal x(t). In this expression C0 show dc content
in x(t), dk show the amplitude of kth harmonic in x(t).
Case-2: The polar form can also be written in another form.

GATE MASTERS PUBLICATION

 112 Signals & Systems

Let, dk = a k2  bk2
ak = dk sink ...(26)
bk = dk cosk ...(27)
a 
k = tan1  k  ...(28)
 bk 
So, equation (8) can be written as,

x(t) = C0   ( d k sin  k cos( k o t )  d k cos  k sin k o t )
k 1


= C0   d k sin( k o t   k ) ...(29)
k 1

So equation (25) and (29) both show polar form representation of Fourier series.

3.4 Symmetry Condition

1. x(t) is an even signal i.e. x(t) = x(–t)
Let us assume that Ck are exponential Fourier series coefficient of signal x(t). Thus,
1
Ck =  x (t ) e  jo kt dt ...(30)
T T 

Now, let Bk are exponential Fourier series coefficient of x(–t), Thus

1
Bk =  x(t ) e  jo kt dt ...(31)
T T 

Now replacing t by –t in equation (30) we get

1
Ck =  x( t ) e j0kt dt
T T

and replacing k by –k in above equation we get

1
C–k =  x(t ) e  j0 kt dt ...(32)
T T 

So comparing (32) and (31) we get

Bk = C–k
That is if Ck are exponential Fourier series coefficient of x(t) then C–k will be exponential Fourier series
coefficient of x(–t).
Since, x(t) = x(–t) (signal is even)
 Ck = C–k ...(33)

In the above case we took x(t) as even and we found Ck are also even. Now using equation (9) and (10)
ak = Ck + C–k = 2Ck
bk = j[Ck – C–k] = 0
Thus for even signal in trignometric fourier series representation only cosine terms will be present and sine

GATE MASTERS PUBLICATION

 Continuous-Time Fourier Series 113

terms will be absent.

Note: If x(t) is even then Ck is also even and trignometric fourier series representation has only cosine terms.

Now, let us assume x(t) is real and even and Ck are it’s exponential Fourier series coefficient, then Ck is even
and conjugate symmetric that is
Ck = C–k and Ck  C k

 C k = C–k
Thus if x(t) is real and even then Ck is real and even.

Note: If x(t) is real and even then Ck is also real and even.

When x(t) is real and even then using equation (17) and (18) we get
ak = 2 Real {Ck} = 2 Ck
bk = 2Img {Ck} = 0
Thus when x(t) is real and even then ak = 2 Ck and bk = 0, that is trigonometric representation of Fourier
series of x(t) only cosine terms exist. Also the coefficient of cosine terms ak are also real and even.

Remember: If x(t) is real and even then it’s trigonometric representation has only cosine terms and coefficients of cosine
term is also real and even.

2. x(t) is odd signal, i.e., x(t)

If x(t) is an odd signal and Ck is exponential Fourier series coefficient of x(t).
Since, x(t) = –x(–t), thus
Ck = –C–k ...(34)
Thus if x(t) odd then exponential Fourier series coefficients are also odd.

Note: For odd signals C0 = –C0, that is C0 = 0. That is for odd signals average value of signal = 0 always.

In the above case we took x(t) as odd and we found Ck are also odd. Now using equation (9) and (10)
ak = Ck + C–k = 0
bk = j[Ck – C–k] = j2Ck
Thus for odd signal in trignometric fourier series representation only sine terms will be present and cosine
terms will be absent.

Note: If x(t) is odd then Ck is also odd and trignometric fourier series representation has only sine terms.

Now let us assume that signal is real and odd

Ck = –C–k and Ck  C k
 C k = –C–k
Thus C–k is purely imaginary, that is Ck is also imaginary.
Thus when x(t) is real and odd then Ck is odd and imaginary.
Now if we use equation (17) and (18) for a real and odd signal then,
ak = 2 real{Ck} = 0

GATE MASTERS PUBLICATION

 114 Signals & Systems

bk = –2 Img {Ck} = –2 Img{Ck]

So in trigonometric Fourier series representation of real and odd x(t) there are no cosine terms and only sine
terms are present.

REMEMBER • For odd signal, Ck is also odd.

• For real and odd signal, Ck is imaginary and odd.
• For real and odd signal, only sine terms exist in trigonometric representation of x(t).

 T
3. Half wave symmetric, i.e., x( t )   x  t  
 2
That is if we shift the signal right or left by period/2, then we get –x(t). Some example of signals having such
symmetry are shown in Fig. 3.2.
x (t) x (t)

t t
0 T 0 T

Fig. 3.2 : Waveforms with half-wave symmetry

Let x(t) has exponential Fourier series coefficient Ck,
1
Ck =  x (t ) e  jo kt dt ...(35)
T T 

Now if x(t + T/2) is created by shifting x(t) then it will also be periodic with period T and let Bk are its
exponential Fourier series coefficient. Thus,
1  T
Bk =  x  t   e  jo kt dt ...(36)
T T  
2

T
Let, t =
2
 T
1  jo k   
2
 Bk =  x ( ) e  d
T T 

1
=  x() e  jo k  e jo kT /2 d 
T T 

 1 
 jo k  j kT /2
=   x (  ) e d  e o
T
 T  

2
Using equation (35) and o 
T
 Bk = Ck ejk

GATE MASTERS PUBLICATION

 Continuous-Time Fourier Series 115

 T  T
So, exponential Fourier series of x  t   are Ck ejk and since x(t )   x  t   , Thus
 2   2

Ck = –Ck ejk
 Ck + Ck ejk = 0
 Ck(1 + ejk) = 0
• If k is even then 2 Ck = 0, thus for half wave symmetric signal x(t), Ck = 0 for even value of k.
• If k is odd then 0 = 0, thus Ck can be non-zero for odd value of k.

Study Note
• For half wave symmetric signal, that is x(t) = –x(t ± T/2) the Fourier series coefficient Ck = 0 for even values of k and
non-zero for odd value of k.
• Thus half wave symmetric signal has only odd harmonics and even harmonics are zero. The trigonometric representation
will also have only odd harmonics of sine and cosine terms.

3.4 Quarter Wave Symmetry

A periodic function possesses a quarter-wave symmetry if
(i) it has either odd or even symmetry and
(ii) it has half-wave symmetry.
A waveform that satisfies the above conditions and if it is shifted towards right or left by T/4 period gives a
new waveform with a half-wave symmetry, but the even (odd) symmetry of the original unshifted waveform has
changed into odd (even) symmetry, is said to be having a quarter-wave symmetry.
If x(t) is real and even and half-wave symmetric:
Then Ck is real and even, for even values of k it is zero.
So, ak = 2 Ck and it will be zero for even value of k
and bk = 0
Thus x(t) will have only odd harmonics of cosine terms.
If x(t) is real and odd and half wave symmetric:
Then Ck is imaginary and odd, for even values of k it is zero.
So, ak = 0
bk = –2 Img{Ck} and it will be zero for even values of k
Thus x(t) will have only odd harmonics of sine terms.

3.5 Properties of Continuous Time Fourier Series

3.5.1 Linearity
If x(t) and y(t) denote periodic signals with period T and
Exponential Fourier
x(t ) 
Series coefficient
 Ck

GATE MASTERS PUBLICATION

116 Signals & Systems

Exponential Fourier
y(t ) 
Series coefficient
 Bk

Exponential Fourier
Then, ax(t )  by(t ) 
Series coefficient
 aCk  bBk

Example 3.4

Find the exponential Fourier series coefficient of signal x(t).

x (t)

5.0

t
–2 – –/2 0 /2  2

Solution 3.4
So we can see that signal x(t) is periodic with period 2,
 2 
1   j kt 1  jkt
Ck =  x(t ) e 2  dt   x(t ) e dt
2  2 


/2  
1 1 j k j k
=  5 e jkt dt = [5( e 2  e 2 )]
2  / 2
2 (  jk )

 5   
Ck =  sin  k  
  k  2 

Example 3.5

Find the exponential Fourier series coefficient of y(t).

y (t)

5/2

t
– 0  +

–5/2

Solution 3.5
5
We can see that, y(t) =  x(t )   {x(t) is signal in example 3.4}
 2

GATE MASTERS PUBLICATION

 Continuous-Time Fourier Series 117

5
So, y(t) = x(t ) 
2
The Fourier series coefficient of y(t) (Bk) = The Fourier series coefficient of x(t) (Ck) –
The Fourier series coefficient of (5/2)
 Bk = Ck – Fourier series coefficient of (5/2)
Since, signal 5/2 is a dc signal let it’s Fourier series coefficient are Dk
5 / 2 , k  0
so, Dk = 
 0, k0

C0  5 / 2 , k  0
So, Bk = 
 Ck , k 0

5  k 
Since, Ck = sin  
k  2 
C0 can be obtained by L-Hospital rule.
5
 C0 =
2

 0, k0

 Bk =  5  k 
 k sin  2  , k  0
  
So, y(t) has Fourier series coefficient equal to Bk.
In this example we used linearity property.

REMEMBER • Analyzing Example 3.4 and 3.5 we can see that x(t) is not looking like half-wave symmetric but if
we creat a signal y(t) = x(t) –5/2 then y(t) is a half-wave symmetric signal. Thus y(t) will have only
odd harmonics and using linearity property we can also say that x(t) will have only odd harmonics
because x(t) and y(t) have only difference in their dc values as:
{Bk = Ck – Fourier series coefficient of dc signal (5/2)}
• Thus if in any question we have a signal x(t) and we have to find harmonics of the signal then we can
shift the signal on amplitude axis i.e. we can form signal (x(t) ± a) to check that signal is half-wave
symmetric or not. If (x(t) ± a) is half-wave symmetric then x(t) will also have only odd harmonics.

3.5.2 Time Shifting

When a time shift is applied to a period signal x(t), the period ‘T’ of the signal is preserved. If
Exponential Fourier
x(t ) 
Series coefficient
 Ck

Exponential Fourier
then, y(t) = x(t  to ) 
Series coefficient
 Bk  Ck exp ( jo kto )

Thus the magnitude of coefficient remain same but the phase changes.
Proof:
1
Since, Ck =  x(t ) e jo kt dt
T T 

Let, t – t0 = 

GATE MASTERS PUBLICATION

 118 Signals & Systems

1
and Bk =  x(t  to ) e jo kt dt
T T 

Let, t – to = 
1
So, Bk =  x() e  jo k (   to ) d 
T T 

 
Bk =  1 x() e  jo k  d   e jo kto

 T T  
 
 Bk = Ck exp(–jokto)

3.5.3 Frequency Shifting

Exponential Fourier
If x(t) is a periodic signal, and x(t ) 
Series coefficient
 Ck ,

Exponential Fourier
then, y(t) = x(t ) e j0mt 
Series coefficient
 Bk  Ck  m

Proof: By definition,

x(t) =  Ck exp( jo kt )
k  

 
and y(t) =  Bk exp( j o kt ) =  Ck  m exp( j o kt )
k   k 

Let, k–m = 

 y(t) =  C exp( jo (  m) t )
  


=  C exp( jo t ) exp( jo mt )
  

= x(t) exp (jomt)

Thus, y(t) = ( x(t ) e jo mt )

REMEMBER In case of time shifting,

x(t )  Ck , x(t  t0 )  Ck e j0kt0
The shifting in time goes with same sign in fourier series.
In case of frequency shifting,

x(t )  Ck , x(t ) e j0 mt  Ck  m

The shifting in frequency goes with apposite sign in time domain.

3.5.4 Time Reversal

GATE MASTERS PUBLICATION

 Continuous-Time Fourier Series 119

Exponential Fourier
If x(t) is periodic and x(t ) 
Series coefficient
 Ck .

Exponential Fourier
Then, x( t ) 
Series coefficient
 C k

This we have already produced.

3.5.5 Time Scaling

Exponential Fourier
If x(t) is a periodic signal, x(t ) 
Series coefficient
 Ck .

Exponential Fourier
Then, y(t) = x(at ) 
Series coefficient
 Bk  Ck

Proof: x(t) is periodic with period T, then period of x(at) will be T/a. So,
T
1  2 
Ck =  x(t ) e jo kt dt ,  o  
T 0  T 

T /a
1   2 
and Bk =  x(t ) e  jo kt dt ,  o  
T /a 0  T /a

T /a
1
=  x(t ) e jao kt dt
T /a 0

Let, at = 
T
1  jo k 
 Bk =
T  x ( ) e d
0

 Bk = Ck
The Fourier series coefficient has not changed but the Fourier series representation will change. We can see
that,

x(t) =  Ck exp( j 0 kt )
k  

the representation will be

Ck Ck

C0 C0
C–1 C–1
C–2 C1 C–2 C1
C–3 C2 C–3 C2
C3 C3
k 0 20 30 
–3 –2 –1 0 1 2 3 –30 –20 –0

So we can see that, x(t) has C0 at  = 0, C1 at  = 20, C–1 at  = –0, C2 at  = 20 and so on.
Now, for x(at) the value of coefficient are same. Thus,
 
x(at) =  Bk exp( j 0 kt )   Ck exp( j 0 kt )
k   k  

GATE MASTERS PUBLICATION

 120 Signals & Systems

Bk Bk

C0 C0
C–1 C–1
C–2 C1 C–2 C1
C–3 C2 C–3 C2
C3 C3
k 
0 1 2 3 –30 –20 –0 0 20 30

3.5.6 Periodic Convolution

If x(t) and y(t) are two periodic signal with period T and
Exponential Fourier
x(t ) 
Series coefficient
 Ck
Exponential Fourier
y(t ) 
Series coefficient
 Bk
then if z(t) is periodic convolution of x(t) and y(t), that is
1
z(t) =  x() y (t  ) d   x(t ) * y (t ) ...(36)
T T 
If Zk are fourier series coefficient of z(t) then
Exponential Fourier
z(t ) 
Series coefficient
 Zk  Ck Bk
Proof:
1
Since, Zk =  z (t ) e  jo kt dt
T T 

1 1 
=    x() y (t  ) d   e jo kt dt
T T
T    T 



 
 
1 1 
Zk =  x ()   y (t  ) e  jo kt dt  d 
T T   T T  
  
 j 0 k

 Bk e 

We know that, y (t )  Bk , y (t  )  Bk e jo k 

1
 Zk =  x( ) Bk e jo k  d 
T T 

 Zk = Bk Ck ...(37)

Study Note
Periodic convolution in time domain lead to multiplication in frequency domain,
Exponential Fourier
x(t )  y(t ) 
Series coefficient
 Bk Ck

3.5.7 Multiplication in Time Domain

If x(t) and y(t) are periodic with period T and

GATE MASTERS PUBLICATION

 Continuous-Time Fourier Series 121

Exponential Fourier
x(t ) 
Series coefficient
 Ck

Exponential Fourier
y (t ) 
Series coefficient
 Bk


Exponential Fourier
then, Z(t) = x(t ) y (t ) 
Series coefficient
 Zk   Cn Bk  n
n  

Thus, Zk is discrete time convolution of Ck and Bk.

Proof:
1
Zk =  z (t ) e  jo kt dt
T T 


z(t) = x(t ) y (t )   Z k exp ( jo kt )
k  

 
Thus, z(t) =  Cn exp ( j o nt )   Bm exp ( jo mt )
n   m  

Let, k = n + m, m = k – n
 
 z(t) =   Cn Bk  n exp ( jo kt )
k   n 


 Zk =  Cn Bk  n ...(38)
n  

Study Note
Multiplication in time domain lead to convolution in frequency domain

Exponential Fourier
x(t )  y(t ) 
Series coefficient
  Cn Bk  n  Ck  Bk
n  

3.5.8 Differentiation
Exponential Fourier
If x(t) is a periodic signal and x(t ) 
Series coefficient
 Ck .

dx(t ) Exponential Fourier

Then, y(t) = 
Series coefficient
 Bk  jo k Ck
dt
Proof:

From definition, x(t) =  Ck exp( j o kt )
k  

Now, if we differentiate both side in above equation with respect to ‘t’ we get

dx(t )
 y (t ) =  j o k Ck exp ( jo kt )
dt k  


=  [ jo k Ck ] exp ( jo kt )
k  

GATE MASTERS PUBLICATION

 122 Signals & Systems

So, exponential Fourier series coefficient of y(t) are

Bk = [jok Ck]

Study Note

dx(t )
If y (t )  then average value of y(t) will always be zero because differentiation removes dc component.
dt

3.5.9 Integration
Exponential Fourier
If x(t) is any periodic signal, and x(t ) 
Series Coefficient
 Ck ,

t
Exponential Fourier
then, y(t) =  x() d  
Series Coefficient
 Bk  Ck


Ck
Bk =
jo k


Proof: By definition, x(t ) =  Ck exp( jo kt )
k  

then, integrating above equation we get

t 
Ck
y(t) =  x ( ) d    exp( jo kt )
 k  j o k

Study Note
t Ck
• So we can see that if signal x(t) has Fourier series coefficient Ck then will have Fourier series .
 x ( ) d  j o k


Thus integration attenuates the magnitude of the high frequency component of the signal.
• Since high frequency components are main cause of sharp details thus integration make signal smooth.

Example 3.6

Find the Fourier series coefficient of signal z ( t ).

z(t)

–/2
t
– /2  3/2 2

Solution 3.6

GATE MASTERS PUBLICATION

 Continuous-Time Fourier Series 123

1
It will be very difficult to find out integration of  z (t ) exp( jo kt ) dt . Now let us define a new signal w(t).
T T 

d (t)
w(t) = z(t )
dt
4 10/
w(t) = y(t ),

t
– 

[y(t) is signal in example (3.5)]

–10/
4
So Fourier series, Wk = Bk


 0, k0

=  20  k 
 2 sin  2  , k  0
 k  

Wk
Using the differentiation property, Zk =
jo k
since period = 2, so 0 = 1
Wk
 Zk =
jk
Wk 20  k 
So, Zk =  2 2 sin   ; when k  0
jk j k  2 
and Z0 is average value of Z(t) = 0

 0, k0

 Zk =  20 sin( k  / 2)
 , k0
 j 2 k 2

Study Note

d
Suppose we have to find Fourier series coefficient of any signal z(t) and we do this by finding w (t )  z (t ), then average
dt

Wk
value of w(t) will be zero always even if average value of z(t)  0. So if we define Zk  then value of Zk will be correct
j 0 k
for all values of k except k = 0. So to find Z0 we simply find average value of Z(t).

Example 3.7

Find the exponential Fourier series coefficient of signal a(t).

GATE MASTERS PUBLICATION

 124 Signals & Systems

a (t)

10

t
–/2 /2 3/2 5/2 7/2

Solution 3.7
d
Again we find, b(t) = a(t )
dt
b(t) = w(t) (of example 3.6)
So, Bk = Wk
 0, k0

=  20  k 
 2 sin  2  , k  0
 k  

Bk
and Ak = , k0
j 0 k

20  k 
= sin  , k  0
j 2 k 2  2 
and A0 = Average value of a(t)
1
=  a(t ) dt
2  2 

A0 = 5
 5, k0

 Ak =  20 sin  k  , k  0
 j 2 k 2  
  2 

3.5.10 Conjugation in Time Domain

If x(t) is a periodic signal with period T, if
Exponential Fourier
x(t ) 
Series coefficient
 Ck

Exponential Fourier
then, y(t )  x (t ) 
Series coefficient
 Bk  C k

Proof: By definition,
1
Ck =  x(t ) e jo kt dt ...(39)
T T 

1
and Bk =  x (t ) e  jo kt dt ...(40)
T T 

GATE MASTERS PUBLICATION

 Continuous-Time Fourier Series 125

Conjugating both side in equation (39),

1
Ck =  x (t ) e  jo kt dt
T T 

Now replacing k by –k we get

1
Ck =  x (t ) e  jo kt dt ...(41)
T T 

so using (40) and (41) we get

Bk = C k
Thus if x(t) is real, that is
x(t) = x(t)
 Ck = C k [Fourier series coefficient are conjugate symmetric]
if x(t) is imaginary, that is
x(t) = –x(t)
 Ck = C k [Fourier series coefficient are conjugate anti-symmetric]

3.5.11 Parseval’s Theorem for Power Signals

Since all periodic signals are power signal, if x(t) is a periodic signal and
Exponential Fourier
x(t ) 
Series coefficient
 Ck

T 
1 2 2
then, Power of x(t) =
T
x(t ) dt   Ck
0 k  

The Parseval’s relation state that total average power in a periodic signal equals the sum of average power in
all of it’s harmonics.
Proof: By definition,

x(t) =  Ck exp ( jo kt )
k  

2
The power of Ck exp(jokt) is Ck and exp(j0kt) are orthogonal for different values of k. So


1 2 2
T  x(t ) dt =  Ck
T  k  

Example 3.8

Suppose we are given the following information about a signal x(t):

1. x(t) is real and odd.
2. x(t) is periodic with period T = 2 and has Fourier coefficients Xn.
3. Xn = 0 for n  1.

GATE MASTERS PUBLICATION

 126 Signals & Systems

2
1 2
4.
2  x( t ) dt  1.
0

Solution 3.8
Since, x(t) is real and odd (clue 1), its Fourier series coefficients are purely imaginary and odd. Therefore,
Xn = –X–n and X0 = 0
Also, since it is given that, Xn = 0 for n  1, the only unknown Fourier series coefficients are X1 and X–1. Using
Parseval’s relation,
T 
1 2 2

T 
x(t ) dt =  Xn
0 n  

for the given signal we have,

2 
1 2 2

2  x(t ) dt =  Xn 1
0 n  1

2 2
X1  X 1 =1
2
2 X1 =1

2 1
X1 =
2
1 1
Therefore, X 1 =  X1  j or X1 =  X1   j
2 2
two possible signals which satisfy the given condition are,
1 1
x1(t) = j e j t  j e j t
2 2

1  e j t  e j t 
= j 2j 
    2 sin( t )
2  2j 

1 1
x2(t) =  j e jt  j e jt
2 2

1  e j t  e j t 
= j 2j 
    2 sin( t )
2  2j 

3.6 Finding Signal from It’s Fourier Series Coefficient

If exponential Fourier series coefficients Ck are know then signal can be obtained using formulae,

x(t) =  Ck exp( j 0 kt )
k  

Here, Ck = Ck e jk

where Ck is magnitude of Ck and k is phase.

GATE MASTERS PUBLICATION

 Continuous-Time Fourier Series 127

Example 3.9

A continous-time periodic signal x(t) is real-valued and has a fundamental period T0 = 8. The non zero
Fourier series coefficients(Xn) for x(t) are
X 1 = X–1 = 2, X 3 = X *-3 = 4j
Find x(t).

Solution 3.9
The signal x(t) is periodic with fundamental frequency o = 2/8 = /4, we have


x(t) =  X n e jnot
n  

= X1 e jot  X 1 e jot  X 3 e j 3o t  X 3 e3 jot

= 2 e j  / 4t  2 e j  / 4t  4 j e j 3 / 4t  4 j e j 3 / 4t

 j t 
j t   j3 t 
 j3 t 
e 4  e 4  e 4  e 4     3 
= 4   8   4cos  t   8sin  t 
 2   2j   4   4 
   

 3 
= 4 cos  t   8 cos  t  
4
   4 2

Example 3.10

If signal x(t) is periodic with period 2 and it’s Fourier series coefficient Xk  (1 / 3) k . Then the expression
of x(t) is _______ .
Solution 3.10

x(t) =  X k exp( jo kt )
k  

2
Here, o = 1
2
 k
1
 x(t) =   3  exp ( jkt )
k  

 k 1 k
1 1
=   3  exp( jkt )     exp( jkt )
k 0 k    3 

1 (3 exp( jt ))1
= 
exp( jk ) 1
1 1
3 3 exp( jt )

GATE MASTERS PUBLICATION

 128 Signals & Systems

3 1 9exp( jt )  3  3  exp( jt )
=  =
3  exp( jt ) 3 exp( jt )  1 (3  exp( jt )) (3exp( jt )  1)

8 exp( jt )
 x(t) =
10 exp( jt )  3 exp(2 jt )  3

3.7 To find Fourier Series Coefficient of Signal x(t)

Case-1: If signal x(t) is a sum of sine, cosine or complex exponentials then to find the exponential Fourier
series coefficients we should follow this procedure.
Step-1: First of all find the period of signal x(t), if in case x(t) is non-periodic then Fourier series coefficient
cannot be found. If signal is periodic then move to step-2.
Step-2: Since, period of x(t) is T and o = 2/T. Now expand each sine and cosine in terms of complex
exponentials,

e j  e j e j  e j
sin = and cos  
2j 2
Step-3: Now x(t) has only complex exponential terms, Since,

x(t) =  Ck exp( jo kt )
k  

Compare each term in x(t) with Ck exp(jokt) to find value of k and Ck.
Case-2: For rest of the cases we can easily find Fourier series coefficient Ck using formulae,
1
Ck =  x(t ) exp(  j o kt ) dt
T T 

Example 3.11

For the continuous-time periodic signal,

 2   5 
x(t) = 2  cos  t   4sin  t 
 3   3 
Determine the fundamental frequency o and the Fourier series coefficients Xn such that,

x(t) =  X n e jnot
n  

Solution 3.11

2   5 
Consider the given signal, x(t) = 2  cos  t   4sin  t 
 3   3 

 j 2 t 2
j t   j 5 t 5
j t 
e 3 e 3  e 3 e 3 
= 2   4 
 2   2j 
   

GATE MASTERS PUBLICATION

 Continuous-Time Fourier Series 129

   
1 j2 3t 1  j2 3t 2 j53t 2  j53t
= 2 e  e  e  e
2 2 j j

   
1 j2 3t 1  j2 3t 5 t 5 j t
x(t) = 2  e  e 2je 3 2je 3
2 2

Comparing with, x(t) = X0  X2 e j 20t  X2 e j 20t  X5 e j 50t  X5 e j 50t

The non zero Fourier series coefficients are
1
X 0 = 2, X2  X2  , X5  X5  2 j
2

Example 3.12

If Fourier series coefficients of x(t) are Ck then find Fourier series coefficient of
(a) x(t – to) + x(t + to) (b) Even {x(t)}

d2 x( t )
(c) Real {x(t)} (d)
dt 2
(e) x(3t – 1)
Solution 3.12
(a) y(t) = x(t – to) + x(t + to)

If y(t )  Bk
Using time shifting property,
Bk = Ck exp(–jokto) + Ck exp(jokto)
Bk = 2 cos(o kto) Ck
x(t )  x( t )
(b) y(t) = Even {x(t )} 
2
If y(t )  Bk

Ck  C k
then, Bk =
2

x(t )  x (t )
(c) y(t) = Real {x(t)} =
2
If y(t )  Bk

Ck  C k
then, Bk =
2

d2
(d) y(t) = x(t )
dt 2
If y(t )  Bk
then, Bk = (jok) ( j0k) Ck

GATE MASTERS PUBLICATION

130 Signals & Systems

2 2
= o k Ck
(e) y(t) = x(3t – 1)
If y(t )  Bk
If period of x(t) is T then that of y(t) is T/3, so if fundamental frequency of x(t) is o then that of y(t) is
o =3o
Now let, z(t) = x(3t)
so, Zk = Ck because scaling do not effect Fourier series coefficient
1
Now, y(t) = z  t  
 3 
2
So, Bk =
T 0

1
= Z k exp   j 3o k  
 3
Bk = Ck exp(–jok)

Example 3.13

Suppose we are given the following information about a signal x(t).

1. x(t) is real.
2. x(t) is periodic with period T = 6 and has Fourier coefficients Xn.
3. Xn = 0 for n = 0 and n > 2.
4. x(t) = –x(t – 3).
3
1 2 1
5.  x( t ) dt  .
6 3
2

6. X1 is a positive real number.

Show that x(t) = A cos(Bt + C), and determine the values of the constants A, B and C.
Solution 3.13
The given signal x(t) is periodic with period T = 6 and fundamental frequency o = 2/6 = /3. Since x(t) is real,
Xn  X n . Since Xn = 0 for n = 0 and n > 2, the only unknown Fourier series coefficients are X1, X–1 ,X2, and X–

2. Also x(t) is half wave symmetric so X2, and X–2 will be zero
x(t) = X 1 e jot  X 1 e  jo t
Since, X1 is a positive real number, X1 = X–1,
 e jot  e jot 
x(t) = 2 X1  
 2
 
x(t) = 2 X1 cos(o t )
 t 
x(t) = 2 X1 cos  
3
Now using Parseval’s relation on clue 5, we get

GATE MASTERS PUBLICATION

 Continuous-Time Fourier Series 131

T/2 
1 2 2
 x(t ) dt =  Xn
T T / 2
n  

3
1
x(t ) dt = X1 2  X1 2  1
2
6  2
3

2 1
2 X1 =
2
1
X1 =
2
Since, X1 is positive, we have X1 = X–1 = 1/2.

 t 
Therefore, x(t) = cos  
3

Example 3.14

Consider the periodic impulse train  T0 ( t ) shown in figure and defined by


 T0 ( t )   ( t  kT0 )
k  

(a) Determine the complex exponential Fourier series of  T0 ( t ).

(b) Determine the trigonometric Fourier series of  T0 ( t ).

T0 (t)

(t) (t – T0)

t
–T0 0 T0 2T0

Solution 3.14

2
(a) Let, T0 (t ) =  Ck e jk ot , o 
k   T0

Since (t) is involved, we obtain

T0 / 2
1 1
Ck =  (t ) e jk ot dt 
T0 T0 / 2
T0


Hence, we get T0 (t ) =  (t  kT0 )
k  


1 2
=  e jk ot , o 
T0 k   T0


a0
(b) Let, T0 (t ) =   (ak cos k o t  bk sin k o t )
2 k 1

GATE MASTERS PUBLICATION

132 Signals & Systems

2
o =
T0

Since T0 (t ) is even bk = 0 and ak is equal to 2Ck

2
ak = 2Ck = T
0

 2
1 2
Thus, we get T0 (t ) =   cos k ot , o 
T0 T0 k 1
T0

Study Note
• We can see that in case of impulse train the Fourier series coefficient are constant Ck = 1/T0, in case of square wave

1
Fourier series coefficient are C k  inversely proportional to k, in case of triangular wave Fourier series coefficients
k

1
are Ck  inversely proportional to k2.
k2
• Thus in impulse train all harmonics have equal contribution, in square wave higher harmonics have less contribution
because value of Ck decrease with increase in k, in triangular wave decrease Ck is more with increase in k so higher
harmonics have much less contribution.

Example 3.15

Find the exponential Fourier series coefficient of signal x(t).

x(t)

t
– 0  2 3

Solution 3.15
dx(t )
Let, y(t) =
dt

y (t)
A (t –)
A (t)

t
– 0  2 3

–A/

GATE MASTERS PUBLICATION

 Continuous-Time Fourier Series 133


A
So, y(t) = 

  A (t  k )
k  

The Fourier series coefficient of y(t) = The Fourier series coefficient of dc signal (–A/) +
The Fourier series coefficient of impulse train
A
The Fourier series coefficient of impulse train is Ik 

and Fourier series coefficient of dc signal is DCk,
 A
 , k  0
DCk =  
 0, k  0

So, Y k = Ik + DCk
 0, k  0

Yk =  A
  , k  0

d
Since, y(t) = x(t )
dt
Yk
So, Xk = for k  0
jo k

A
= , k0
j o k
Since time period = , o = 2
A
 Xk = , k0
j 2k
A
X 0 = Average value of x(t) =
2
 A
 2 , k 0
So, Xk = 
 A , k0
 j 2k

Example 3.16

Find the exponential Fourier series coefficient of x(t)


 m  2m 
x(t) =  t    t  
m    3  3 

Solution 3.16
Let us plot signal x(t),

GATE MASTERS PUBLICATION

134 Signals & Systems

x (t)
2(t –2/3)
2(t)

(t –1/3) (t –1)

t
0 1/3 2/3 1

So we can see that period is 2/3 and in one period there are two impulses.
2/3
1
Ck =  x ( t ) e  j  o kt dt
2/3 0

2
o =  3
2/3
2/3
3   1    j 3 kt 3
 Ck =   2 (t )    t    e dt = 2  e j 3k(1 / 3) 
2 0   3   2

3  3 
Ck = 2  e j k   3  e j k 
2  2 

Example 3.17

Consider the signal x(t), the power of x(t) till its second harmonic is _____ W.
x (t)

15

t
–10 –5 –3 0 3 5 10

Solution 3.17
First of all lets find exponential Fourier series coefficient of x(t), Ck
1
Ck =  x (t ) e  jo kt dt
T T 

5 2 5 2
1 j
10
kt 1 j
10
kt
=  x(t ) e dt   15 e dt
10 5
10 5

2 2 3 3
15  j 3k  j 3k 15 j k j k
= [e 10  e 10 ] = [e 5 e 5 ]
 j 2k 2 j k

 15  3  
Ck =  sin  k  
 k  5 

2
2
The power in x(t) till 2nd harmonic is  Ck .
2

GATE MASTERS PUBLICATION

 Continuous-Time Fourier Series 135

 C0 = 9,
15  3 
C1 = C–1 = sin  
  5 

15  6 
C2 = C2  sin  
2  5 
2
2
 Power fill 2nd harmonic =  Ck  126.16 Watt
2

Example 3.18

Find the magnitude of 2nd harmonic in the signal x(t) = 3 sin(4t + 30°) – 4 cos(12t + 60°).
Solution 3.18
Since, x(t) = 3 sin(4t + 30°) – 4 cos(12t + 60°)
Let, time period of x(t) is T.
2 
Period of 3 sin(4t + 30°) is T1 =  .
4 2
2 
Period of 4 cos(12t + 60°) is T2   .
4 6
T1 3
So, is rational , LCM of denominator or ratio =1.
T2 1


 T = T1 
2
 o = 4
Thus if we expand x(t) we get

3 j 4t j 30 3  j 4t  j 30 4 j12t  j 60 4  j12t j 60

x(t) = e e  e e  e e  e e
2j 2j 2 2

Now if we compare each term with Ck e jo kt i.e. Ck e j 4kt we can see that x(t) has terms with k = ±1 and k = ±3
and not for k = ±2. Thus x(t) has no 2nd harmonic. So answer is 0.

Example 3.19

Which harmonics will be present in these signals.

x (t)
1
A

(a) t
–T 0 T 2T

–A

GATE MASTERS PUBLICATION

136 Signals & Systems

x 2(t)
A

(b)
t
–T/2 0 T/2 T

d
(c) y 1( t )  x (t )
dt 1
d
(d) y 2 ( t )  x (t )
dt 2
Solution 3.19
(a) Signal x1(t) is odd, thus it will have only sine terms in trigonometric representation of Fourier series.
(b) If we shift x2(t) on amplitude axis by –A/2 then we can see that it is half wave symmetric also. So x2(t) is
even and it will have only odd harmonics except C0 as it has non zero average value of signal. Thus x2(t)
will have odd cosine terms.
dx1 (t )
(c) Since x1(t) was odd so y1(t) = , which will be even thus y1(t) will have only cosine terms in trigonometric
dt
Fourier series representation.

d
(d) y 2 (t )  x (t ), since x2(t) was even so y2(t) will be odd. Thus y2(t) will have only odd harmonics of sine
dt 2
terms and zero dc value.

Example 3.20

Determine the time signal corresponding to the magnitude and phase spectra shown in figure with 0 = .
2 2
Cn
1 1

n
–4 –3 –2 –1 0 1 2 3 4

/4
/8
Cn

–4
n
–3 –2 –1 0 1 2 4

–/8
–/4

Solution 3.20
From figure, we have

C3 = 1   1 e j/4
4

GATE MASTERS PUBLICATION

 Continuous-Time Fourier Series 137

C–3 = 1    1 e j /4
4

C4 = 2   2 e j /8
8

C–4 = 2   2 e  j  /8
8
By definition, the exponential Fourier series is

x(t) =  Cn e jn0t
n  

= 2 e  j  /8 e  j 4ot  1e j  /4 e j 3o t  e  j  /4 e j 3ot  2 e j /8 e j 4ot

= 2{ e j[( /8)  4t ]  e j[( /8)  4t ]}  { e j[( /4)  3t ]  e j[( /4) 3t ] }

 
= 4cos  4t    2cos  3t  
 8   4 

Example 3.21

Find the time domain signal whose Fourier series coefficient is given by,
Cn = j(n – 1) – j(n + 1) + (n – 3) + (n + 3), o = 
Solution 3.21
 
We have, x(t) =  Cn e jnot   Cn e jnt
n   n  

= jejt – je–jt + ej3t + e–j3t

= 2 cos3t – 2 sint

Example 3.22

For the signal shown in figure obtain the Fourier series representation using time differentiation property.
x (t)

–T/2 T/2
t
–T –/2 0 /2 T

Solution 3.22
Differentiating x(t) with respect to time, we get the signal y(t) as shown in figure.

GATE MASTERS PUBLICATION

138 Signals & Systems

y (t) = [dx (t)/ dt]

/2
t
–2T –3T/2 –T –T/2 –/2 0 T/2 T 3T/2 2T

–1

2
Here, o =
T
Fourier series of y(t) is given by,
T /2
1
Dn =  y (t ) e  jnot dt
T T /2

T /2
1          jnot
=    t  2     t  2   e dt
T T /2     

1  jn 2   jn
2  
= e T 2  e T 2 
T  

2 j   n  
Dn = sin 
T   T  

d
we have, y(t) = x(t )
dt
Let, Cn be the Fourier series of x(t). Using time differentiation property, we get
Dn = jn0Cn
1
Cn = D
jn0 n

 T  2 j  n   1  n 
=    sin      sin   
 jn 2   T  T   n  T 

 sin( n / T )    n 
=  sinc    
T ( n / T )  T T 

Example 3.23

The waveform shown in figure (a) is applied to the circuit shown in figure (b). Determine the rms value of
third harmonic of current in the circuit.

GATE MASTERS PUBLICATION

 Continuous-Time Fourier Series 139

v(t)

25 25  0.5 H i (t)

–  t
0 2

–25
v (t)
(a) (b)

Solution 3.23
First of all lets find Ck of v(t),
2
1  jo kt
Ck =
2  v(t ) e dt
0

Since, T0 = 2, o = 1
2
1  jkt
 Ck =
2  v(t ) e dt
0

 2 
1  jkt
=   25 e dt   25 e jkt dt 
2  0  

1  25( e jk  1)  25( e j 2 k  e j k )

=
2(  jk )  

25 25   jk 
= ( e jk   1)  (1  e j k ) = (e  1) 
2(  jk )   (  jk ) 

25
Ck = [1  ( 1)k ]
j k

 50
 , k is odd
So, Ck =  j k
 0, k is even

 
So, v(t) =  Ck exp ( jo kt )   Ck exp ( jkt )
k   k  

For third harmonic k = 3 and –3

So, 3rd harmonic component in v(t) is
= C3 exp(3jt) + C–3 exp(–3jt)
50 3 jt 50 3 jt
= e  e
j 3 j 3

100
3rd harmonic = sin 3t
3

GATE MASTERS PUBLICATION

 140 Signals & Systems

100
sin 3t
v (t )
So, 3rd harmonic component of i(t) =  3
R  jL 25  j (3) (0.5)

= 0.423sin(3t  3.4)

0.423
So, rms value is  0.299 A .
2

3.7 Summary
Property Section Periodic signal Fourier series coefficients
x (t ) Periodic with period T and ak

y (t ) Fundamental freqeuncy 0 = 2/T bk
Linearity 3.5.1 A x( t )  B y (t ) Aak  Bbk
 jk0t0  ak
Time shifting 3.5.2 x(t  t0 ) ak e e jk(2  / T )t0
Frequency shifting e jM0t x(t )  e jM(2 / T )t x(t ) ak  M
Conjugation 3.5.6 x(t ) a k
Time reversal 3.5.3 x( t ) ak
Time scaling 3.5.4 x( t ),   0 (periodic with period T/  ) Tak bk

Periodic convolution  x( ) y (t   ) d  a l bk  l
T l  
2
Multiplication 3.5.5 x( t ) y( t ) jk 0 a k  jk ak
T
dx(t )  1   1 
Differentiation   ak    ak
dt  jk 0   jk 2  / T 
 a k  a *k

t
(finite valued and  Re{ak }  Re{a k }

Integration  x(t ) dt periodic only if a0 = 0) Im{a k }   Im{a  k }


 ak  ak
  a k   a k
Conjugate symmetry for real signals 3.5.6 x( t ) real a k real and even
Real and Even signals 3.5.6 x( t ) real and even a k purely imaginary and odd
Real and Odd signals 3.5.6 x(t ) real and o dd Re{a k }
 x e( t )  ev {x(t )} [ x( t ) real]
Even-Odd decomposition  j Im{ak }
 x0 (t ) = od{x( t )} [ x( t ) real]
Parseval's relation for periodic signals

1 2 2
T T
x( t ) dt   ak
k  

Table-3.1

GATE MASTERS PUBLICATION

 Continuous-Time Fourier Series 141

Time domain signal x(t) Fourier series coefficient C k

Real Conjugate symmetric

Imaginary Conjugate anti-symmetric
Even Even
Odd Odd
Real and Even Real and Even
Real and Odd Imaginary and Odd
Imaginary and Even Imaginary and Even
Imaginary and Odd Real and Odd

Table-3.2

Study Note
In the Table 3.2 the arrow is both sided that is if x(t) is real then Ck is conjugate symmetric and if Ck is real then x(t) is
conjugate symmetric and if Ck is real then x(t) conjugate symmetric. This apply to all in Table 3.2.



GATE MASTERS PUBLICATION

 142 Signals & Systems

fundamental period T = /6 and Fourier series

coefficients ak, it is found that,
S
x(t )   y(t )  x(t )

1 Subjective Practice Problems For what values of k is it guaranteed that ak = 0?

 t, 0  t 1
1. A continuous-time periodic signal x(t) is real valued and 5. Let, x(t) = 
2  t , 1  t  2
has a fundamental period T = 8. The non zero Fourier
be a periodic signal with fundamental period T = 2 and
series coefficients for x(t) are
Fourier coefficients ak:
a1 = a–1 = 2, a3  a3  4 j (a) Determine the value of a0.
Express x(t) in the form (b) Determine the Fourier series representation of
 dx(t)/dt.
x(t) =  Ak cos(k   k ) (c) Use the result of part (b) and the differentiation
k 0
property of the continuous-time Fourier series to
help determine the Fourier series coefficients of x(t).
2. Consider three continuous-time periodic signals whose
Fourier series representations are as follows: 6. Consider the following three continuous-time signals
100 k 2 with a fundamental period of T = 1/2.
1 jk
50
t
x1(t) =   2  e x(t) = cos(4t)
k 0 y(t) = sin(4t)
100 2 z(t) = x(t) y(t)
jk t
x2(t) =  cos(k ) e 50 (a) Determine the Fourier series coefficient of x(t).
k  100 (b) Determine the Fourier series coefficients of y(t).
2
(c) Use the results of parts (a) and (b), along with the
100
 k   jk 50 t multiplication property of the continuous-time
x3(t) =  j sin  e
k  100  2  Fourier series, to determine the Fourier series
Use Fourier series properties to help answer the following coefficients of z(t) = x(t) y(t).
questions: (d) Determine the Fourier series coefficients of z(t)
(a) Which of the three signals is/are real valued? through direct expansion of z(t) in trigonometric
(b) Which of the three signals is/are even? form, and compare your result with that of part (c).

3. Suppose we are given the following information about a 7. Let x(t) be a periodic signals whose Fourier series
signal x[n]: coefficients are:
(i) x[n] is a real and even signal.  2, k0
(ii) x[n] has period N = 10 and Fourier coefficient ak. 
ak =   1  k
(iii) a11 = 5  j   , otherwise
 2
9
1 2 Use Fourier series properties to answer the following
(iv)
10
 x[ n]  50
n0 questions:
Show that, x[n] = A cos(Bn + C), and specify numerical (a) Is x(t) real?
values for the constants A, B and C. (b) Is x(t) even?
(c) Is dx(t)/dt even?
4. Consider a continuous-time ideal low-pass filter S whose
frequency response is 8. Suppose we are given the following information about a
continuous-time periodic signal with period 3 and
1,   100
H( j) =  Fourier coefficients ak:
0,   100
1. ak = ak + 2 2. ak = a–k
When the input to this filter is a signal x(t) with
0.5 2
3. 0.5 x(t ) dt  1 4. 1 x(t ) dt  2

GATE MASTERS PUBLICATION

 Continuous-Time Fourier Series 143

Determine x(t). (a) constant coefficients.

(b) only cosine functions.
9. If the Fourier series coefficients of a signal are periodic,
(c) only sine functions.
then the signal must be
(d) orthogonal functions.
(a) continuous time, periodic
(b) discrete time, periodic 14. What is the magnitude of the exponential Fourier series
(c) continuous time, non-periodic coefficient of the fundamental term of the periodic signal
(d) discrete time, non-periodic x(t) = 3 + 2 cos(3t) + sin(3t)

10. If a periodic function f( t) of period T satisfies 3 5

(a) (b)
2 2
 T
f (t )   f  t   , then in its Fourier series expansion, 1
 2 (c) (d) 1
2
(a) the constant term will be zero.
(b) there will be no cosine terms. 
(c) there will be no sine terms.
(d) there will be no even harmonics.

11. A periodic signal which can be expanded in Fourier series

(a) is a power signal.
(b) is an energy signal.
(c) is neither a power signal or an energy signal.
(d) can be either a power signal or an energy signal
depending on the nature.

12. The Fourier series of a real, even periodic signal will

contain only
(a) cosine terms (b) sine terms
(c) even terms (d) odd harmonic

13. The Fourier series representations are based on using

GATE MASTERS PUBLICATION