8

FOurier Analysis of
ACCircuits
INTRODUCTION TO FOURIER THEOREM

discontinuous functions by
Furier theorem provides a method for representing cosine and sine terns and are
igonometric series. Fourier series are series of
represent general periodic signals. The ideas and techniques of Fourier
sed to non-periodic signals. The signal representation for such
cies can be extended to
signals given by Fourier integrals and Fourier transforms.
IO-periodic series and Fourier transforms and has
several
analysis deals with Fourier
Fourier
Mathematics, Science and Engineering, particularly in the area of
eplications in
ammunications and signal processing.
represented as the sum of a finite or infinite
number
Any periodic signal can be responses of linear systems to non-sinusoidal
sinusoidal functions, the the superposition integral. The Fourier
determined by applying
cCitationscan be
provides the ways of solving the above problem.
talysis
TRIGONOMETRIC SERIES
are
sum of sinusoids whose frequencies
represented by
Periodic signals can bemultiples fundamental frequency. The Fourier series
of
Cs or integer will be of theform.
of a periodic signal 2ox+ ...
tepresentation + a, cos 2 ot + ... +b, sin N + b, sin
series is called a
o+ ajcos or constants. Such a
... are real coeficients of the series.
Nhere a ..., b,, b
called the
Onometric series aand a, and b, are
write this series
Using the summation, we may
sinnor)
(a, cos nOr + b,
x)= ay +2
& Network Theory

Let us assume x(0 is aperiodic function of period 2u/o and is integrable over
trigonometric series.
a period 2nlex Assume x() can be represented by a
x()= t a, cos no1 +b, sin no) (8.1)
We axsune that this series converges and has x(0) as its sum. From the fune:.
A) we can determine the coefficients a, and b,:
8.2.1 Determination of Constant Term a,
to
Integrating on both sides of Eq. 8.1 from |

We get x()dt = 4+ (a, cosnot +b, sin nor |ad (8.2)

The term by term integration gives

J 0d= a| d+4,| cosno1 dt +b,| sin not dt (83)

n=l

27
The first term on the right equals -do.
All the other integrals on the right are zero.
27
2
Jo x(t)dt =
27

An alternate form of the

evaluation integral with the variable ot and the
corresponding period of 2r radians is
ay = x()d(01) (8.4)
27 Jo
8.2.2 Determination of the Coefficients a, of the Cosine Terms
We multiply Eq. 8.1 by cos m ot, where m is any fixed positive
integer ald
integrate from to .

JX)coS mOr dt
-j 4,+ la, cosn 0r +b, sin
ntor) cos Smor
m
dt (8.5)
 Fouler Anaysis of AC Circts B3
term by term, the
right hand side
becones,
cOs MOtdt +

(8.6)

Thefirst third integrals are

and
zero. The second integral a when
becomes
Therefore, a=

x() cos n ot d(o) (8.7)

Determination of the Coefficients b, of the Sine Terms

8.2-3
integer and
1'e multiply Eq. (8.I) by sin mo, where mis the fixed positive
integrate to

= 4 +(4, cos n0r +b, sin non) sin mot dt

(8.8)
)Sin mOt dt

the right side becomes

Integrating term by term,
motdt +b,sin n otsinmot dt
cos nor sin
4o-sin mot dt+ n=1
(8.9)
when n = m.
the last integral becomes
are zero and
The first two integrals
dt
r()sinnot

(8.10)

b,= x0)sinnord(a) coefficients of x().

Or Fourier
calledtthe
and
a,andb,
The coefficients ao,
 N e t k They
SERIES
TRIGONOMETRIC FOURIER
COMPACT
&3 can be combincd
as
The sine and cosine tems of
same frequency single sine or
sine with a phase angle c, cos(n0r+ O,)
c0s n + ,sin nor
C, sin (not + ) (8.11)
4 cOs nO +b, sin not
where

8, = tan and , =
form
Fourier series in Eq. 8.1 can be expressed in the compact
x() = te, cos(n01 +0,) (8.12)
where co = a, is the average value of x() and c, , are computed from a, and b
8.3.1 Existance of the Fourier Series-Dirichlet Conditions
Any periodic function x() =x([+ I) can be expressed by a Fourier series provided
the following conditions are satisfied.
1. The function x() is absolutely integrable over one period.

where T is the period.

2. The function x() has only a finite number of positive and
in one period.
negative maxima
3. The function x() has a finite number of
The above conditions are known as discontinuities in one period.
Dirichlet conditions and hence, possess a
convergent Fourier series.

Exarmple 8.1 Find the Fourier series for the waveform

shown in Fig. 8.1.

20

21
4T
ot
Fig. 8.1
Solution The waveform equation is given by
 Fourier Analyt of AC Crcuts 85
20
2
averagevalue offthe waveformis
The

1 2r 20
=10
21t o r

(2r 20
ot cos not d(ot)

20 Or 1 2
2Tn -sin n0t +
n cos n0r
20
-(cos n2T-cos 0]
=0 for all integer values of n

2r 20 |Ot sin not d(o)

20
a20-0
-cosnOt +sin n0r TUn

average term, the series is

Ilsing these sine term coefficients and the
20 20
20 sin 20t-sin 301
v(t) = 10-sin Ot 2 3t

SYMMETRY-FUNCTIONS OF ANY
84 CONSIDERATION OF
PERIOD 2T-SYMMETRY CONSIDERATION

far had period 2I. In several applications, periodic

Ihe functions considered so
periods.
lunctions will generally have otherhas a Fourier series, then this series is
period 2T
afunction x() of n (8.13)
or tb, sin or
T
x() = 4g

coefficients of x() given by

ih the Fourier
-7

for
dor) n=1,2..
7
8.6 Network Theory

1,2, ...

function shown in Fig. 82

Example 8.2 Find the Fourier series of the
<-I
0 if -21t <01
-I< OI<I (8.14)
I<O1 < 21

-2 -1 211

Fig. 8.2

Solution From the Eg. (8.14) the Fourier coefficients are

k
1 2r kd (or) =
a= = v()d(o) 4T -1 2
(8.15)

a, =
2
-d(on)

L2k Sin nT
nT 2 (8.16)

Thus, a, =0ifn is even

2k -2k
and ifn =1, 5,9, ...,a, = if n=3, 7, 11, ...
We also find that b, =0 for n = 1, 2, ... Hence, the result is

k 2k 37 1
v)
=cos0-cos:
32 -ottcOS (8.17)
5

Even and Odd Functions Afunction x() is even if

such a function is symmetric with respect to x(-)=x(). The waveform of
y-axis. A function x(t) is odd 1
x(-)=-x(). The function cos not is even and sin not is odd. The even and odd
functions are shown in Fig. 8.3(a) and (b).
 Fourier Analysis of AC Circuits 8.7
x(0)

(a) Even function

(b) Odd function
Fig. 8.3
Thesum
oftwo or more even functions is an
even function and with the addition
constant, the even nature of the function
remains, The sum of two or more
adfunctions is an odd function, but the addition ofa constant removes the odd
fthe function.
patureoft The product of two odd functions is an even
function.
TheFourier series of an even function of period 27Tis a Fourier cosine series.

x() = 4, ta, cos or (8.18)

n=l
With coefficients

2 f x() cos d(o)

a Fourier sine series.
Fourier series of an odd function of period 2T is
The
(8.19)
x()= (Ot)

With coefficients
x()sin T
half-wave symmetry if x() =
have
function x(t) is said to waveform has half-wave symmetry,
Aperiodic the period. If the series will contain both
Tis This
T+ 12) where present in the
series.
odd or even. In any
case, an
only odd harmonics are is also half-wave
unless the function any waveform with
Sine and cosine terms ... for
to zero forn = 2, 4. 6
Bn are equal triangular
Symmetry. series for the
trigonometricFourier 8.4.
Example 8.3 the
Findwaveform shownin Fig.
even
 &&Netwok Theory

10 V

0 2r 31

Fig. 8.4
Solution The waveform shown in Fig. 8.4 is an even function.
Since. v() = v)
By observation, the average value of the wave
d =5
The waveform has half-wave symmetry

V) =
The equation for the waveform

V) = 10+ for -t< Ot <0

10
= 10 for 0 < Ot < I
(8.20)
Since, even waveforms have only cosine terms
b,,= 0 for all n

-
cos not d(o) +| cos nor d(ot)-cos not d(o)
T

5cos n0r + sin nor 5 cos nor + -sinnor

10
-(cos 0-cos(-nT) - cos nTt + cos 0}

20
Tn
(l-cos nt)
a, =0 for n= 2, 4, 6, ...
a, =
40
r'2Or n= 1, 3, 5, ...
 Fourler Awyis of AC CHts # 89
Fourier series is

cOs 04 40 40
cos 30t4
s cos Sot4
(3n) (5n)
pample
8.4 Find the
trigonomnetric
shown in Fig. 8.5. Fourier series for the waveform
20

37

-20

Fig. 8.5
colution By inspection, the average value of the waveform a, = 0. The
saveformis odd andI contains only sine terms. The expression of the
waveform
v)=20 0t for-I< 0t <I (8.21)

20
ot sin not d(or)

-40
sin nOt -cOS n0t CoS nt (8.22)
nt

The Fourier series is

40 Sin3or
v() = Sin 0t sin 2ot ASn401+... (8.23)

EXPONENTIAL FORM OF FOURIER SERIES

The Fourier series

coS not+b, sin no) (8.24)

x() = 4, t (a,
Can nbe
We knowwritten
form.
that in the complex
ea = cos na +jsin nON (8.25)
(8.26)
en cOs not- Sin nor
&o Netwok They

By adting the liqs 8.25S and 8.26.

We get cOs nOx
get
Subtracting Eq. 8.26 rom Eo. 8.25 and division by 2j, we
sin nox =

te
a, cos nOt + b, sin not=

-4,-j,)waM , +jb,)e -jnor

Consider

;(a, - jb,) =G,

and L(a, +jb,) =S
The Eq. 8.24 becomes

x(0) =, +c,eo +ce hay

n=l
(8.27)

27T
xt)(cos nor- jsin nor)d(ot)

C-n= ;4, +j0)

x(0)(cos not +jsin no)d(o)
1
= 2r0)e d(or)
By combining the two formulas and writing c, = C- we get

x() = (8.28)

where CH= |x()em" d(or) for n =0, + 1, + 2, ...

2 -
This is called complex form of Fourier series of x().
The c, are called the complex Fourier coefficients of x(t).
 Fourier Andlyis ofACCircut B.1
p m p l8
e. s Findthe complex
in Fig. 8.6. Fourier series for the square wave shown

10V

2r 37 4n

-10V

Fig. 8.6
Solution The expression of the waveform
v(t) =-10 for - <ON<0
= 10 for 0<a<n
average value of the wave isSzero.
The

10
2r

10
-j2Tn

j10 (8.29)
n7

For n even eint = land c, =0

-020)
For n odd ent = -l and c, =

The Fourier series is

20 r . . (830)
jeje
j20e r - j e20j o 31
x() = ...+

PHASE SPECTRA
8.6 LINE SPECTRA AND
complex quantity and can be
exponential formis a
Fourier r coefficient G, Oofthe (8.31)
Tepresented by
(Gl
G,=Re lcl+jIm
8.12 Network Theory

The real part of c, is (8.32)

Re le))cosnor

of c, is
The imaginary part x()sin nor
(8.33)

odd function of n
2r
Im [c,] is an
even function of n, whereas series is given by
Re l lis an spectrum of the Fourier
Theretore, the amplitude (8.34)
I)2
lc,l= (Re-lG,]+Im²le,
by
While the phase spectrum is given
Im(G,1 (8.35)
., = tan Relc,)J

THE FREQUENCY SPECTRUM

8.7

Consider the Fourier series

nor)
x() = a, +(a, cos nOr +b, sin
n=l

a, = Cn COS a
Let (8.36)
and
b, = C, Sin a,,
nO1)
Therefore, x(t) = at2(c, cos , coS not t c, sin a,, sin

= 4 t c , cos(n01 -a,) (8.37)

n=l

From Eq. 8.36, we have

b,
tan a, =

.,, = tan
(8.38)

Also, we have
(8.39)

The magnitude of c, is plotted against no, and the graph obtained is called the
frequency spectrum of the given waveform x(). The amplitudes decrease rapidly
for waves with rapidly convergent series. Waves with discontinuities such as
 Fourler Antyt of AC Creti &

square wave have spectra with slowly decreasing amplitudes si

and
h havestrong high harmonics.

FOURIERINTEGRALS AND FOURIER TRANSFORMS

powertul
series ar tools in dealing various problems involving periodic
InseveralI applications many practical problems involve non-periodic

NTNVIS.
7he
method of Fourier series can be extended to such non-periodic
vponentialform of Fourier series is given for periodic signal belowfor
The

i()) = (8.40)

T
1
X,=
2

T
(8.41)
2

thefundamental frequency
of the periodic signal x().
fo
shere is non-periodicsignal
Ifwedefine
(8.42)
o
x()=i)forltk 2
=0 otherwise
Ax(0)

(b)
To
(a)
Fig. 8-7
(8.43)

2 (8.44)

ToX, s
Ne can define the envelope
&4Network Theory

(845)
where
Therefore, the coefficient X, becomes

X,=X(0)
To
X,=
T
-X(n0) (8.46)
Substituting Eq. 8.46 in Eq. 8.45, we get

i() =

i() =
n=-00

27t
Substituting ,
T
() = 270 X(n0,)e nay

i() = X(n0, )e

X() = (8.47)
2T
In the above complex sum, let the frequency ,
approach, zero, and the index
n approach, infinity, such that the product n0, approaches a continues
variable o and the discrete sum is replaced by continuous integral. frequency
Therefore, i() ’x() as a, approaches zero, then Eq. 8.45 becomes

X(0) = x)e dt (8.48)

and Eq. 847 becomes

x() = x(0)e do (8.49)

These expressions define a Fourier transform pair for the
signal x(t) and are
denoted by the notation
x(0) > X(o)
 Fourler Anlysis of AC Circuits B.5

(8.48) is refered to as Fourier transform of x() and dis denoted

X(0) =Fx0)
inegral Ba. 8.49 is referred to as the inverse Fourier transform and is
The by the symbol x() =FX(o).
transtormis essentially, a transformation from afunction of the time
kurier l h e
t to
le
frequency variable o. Not allfunctions can be transformed but
Dirichlet conditions in any finite interval and

sulticientconditions for the existence of Fourier transforms.

writing X(0) in terms of magnitude
ier transtorm may be represented by
n phaseas (8.50)
X(0) = LX(O)l e0)
B0) = X(0) versus frequency fare referred to as the
Dlots of IX(0)l and
and phase spectra of x(/) respectively.
nplitude
SPECTRUM
80 THE ENERGY
interval
signal if it has finite energy over the
an energy expressed in the
Thesignal is defined as zero. The energy of asignal can be
is
oo, oo) so that power
fequency domain. defined
normalised energy for a signal is
The (8.51)
E= r'od
representation for x(), that is
transform
Using the Fourier (8.52)
x() =
Xo)e"do

We can write the energy as

(8.53)
x0)x(-0)do
2
 &6 Network Theony

Therefore. the energy

E(@)X*(0) do
Sine, A-)N(o)

x@Y do (8.54)

E- [r'od=LX(of'do (8.55)
X(o)?
This is referred to as Parseval's theorem for Fourier transforms. Thus, I
has the units of energy density with frequency and is called as energy density
spoctrumn of x().
Go) =1X(oP
Integration of G(o) over all frequencies gives the total energy contained in a
signal. Similarly, integration of G(o) over a finite range of frequencies gives the
energy contained in a signal within the limits of integration.

8.10 FOURIER TRANSFORM OF POWER SIGNALS

The signal is defined as power signal if and only if, it has finite power and infinite
energy. The power signals are not absolutely integrable, power signals do not
satisfy the condition.

This class of signals are called power signals because the signal energy is
infinite over the entire interval., but the power is finite. Therefore, the power
T
1
P= lim Jrod <o (8.56)

Example 8.6 Determine the Fourier transformn ofa signal given in Fig. 8.8.
Solution From thc waveform shown in
Fig. 8.8.
x) = v for (8.57)
=0 otherwise
The Fourier transform of the expression
Fig. 8.8
 Fourler Analyss of AC Crcuits B./7

-
(8.58)
jo

joD -j0
X(0) = VI, 2-e
20-0
2

Sin
= VT, (8.59)
2
transform of the signal is illustrated in Fig. 8.9.
The Fourier

VTo

2n 0
10 10

Fig. 8.9

of the general impulse,

the Fourier transform
Example 8-7 Determine x(t)
x() = AS(t).
function for the
Oution The impulse
gVen equation in Fig. 8.10.
function is
Ihe transform of the above
Fig.8.10
dt (860)
X(o) =( AS)e M X(0)

A
X(0) = A
function
ctransform of impulse
hown in Fig. 8.11.
Fig. 8.1
&18 Network Theory

signal shoyn i
Example 8.8 Determine the Fourier transform of the
x()
Fie &/2.
Solution This waveform does not satisly
the condition.

(8.61)
Fig. 8.12
However, the transform of x() in Fig.
Fig. 8.12 can be
812 can be found, ifa is allowed to tend to zero. Therefore, the
represented by Fig. 8.13.
x()
1

Fig. 8.13
The transform of the signal in Fig. 8.13 can be determined

1 20
a- jo a+ jo a +o'
(8.62)

As a’ 0, X(0) is zero except when o=0.

2
where X(0) = lim
a’0 2a

Thus, the transform is an impulse whose strength may be obtained by integrating

X(0) over the frequency range.
2a
Therefore. x(0)do = do = 21 (8.63)

Th transform of the signal is shown in Fig. 8.14.

Fig. 8.14
 Fourier Analysis of ACCirets B19
pample
6.9
Determine the Fourier
shown in Fig. 8.15. Iransform of the sigum fncim
The function is defined as
csbtin 41
Sign () =Ifor r>0
=0for t=0 (8.64)
=-1 for t<0
1
Fouriertransform of the function
Jhe
Fig, 8.45

-f_e"da-lea (8.65)

X(0) = 12 (8.66)
jo jo jo
TheFouriertransform of signum function is shown in Fig. 8.16.

Xo) =2
jo

Fig. 8.16

Example 8.10 Determine the Fourier transform of unit step function

shown in Fig. 8.17.
x(0
Solution The function is defined as
u() = 1for 1>0
+1

= for t <0 (8.67)

This function is obtained by the sum
dtwo functions shown in Fig. 8.14 and
nFig. 8.16 and dividing by 2. The Fig. 8.17
Tnslorm is half the sum of transforms of
ndvidual time functions.
(8.68)
X(0) = 12+2nÖ
2 j0
o) jo
The Fourier transform is given below
1
(8.69)
+Ô O)
X(o) =
jO
 &30 Netwoek Theory

Fig. 8.18

8.1 FOURIER TRANSFORM OF PERIODIC SIGNALS

In the previous section, Fourier transform representation for non-periodic signals

have been discussed. Fourier transform of periodic signals have been discussed
in this section. In general, Fourier transform of periodic signals are not absolutely
integrable and have infinite discontinuties. Here, we obtain the Fourier transform
of a periodic signal by Fourier transforming its complex Fourier series term-by
term. The Fourier transform consists of atrain of impulses in the frequency domain.
The area of these impulses is directly proportional to the Fourier series coefficiens.
Consider a signal x() with Fourier transform X(0).
X(0) = 2I Ô(0- o); an impulse at ) = o. To determine the signal x() for
the Fourier transfom X(0), we use inverse Fourier
transform relation
x() = X(0)eo do (8.70)

2r8(0 - o, )edo (8.71)

x() = ej
Any periodic signal is represented by alinear
spaced in frequency. combination of impulses equally
X(0) = 2TtX,d 0-n0,) (8.72)
From the Fourier inverse
transform
x(t) = X(0)e do (8.73)

(8.74)

x(0) = x, m (8.75)
The above Eq. 8.75 gives the
exponential form of the Fourier series.
 Tourier Analyslt of AC Creuits 8.31
xample
8.11
Find the
train shownFourier 8, 19.
in Fig. transform of a perlodie unit impulse

-3T -2T -T 0 T 2T 3T 4T
Fig, 8.19
solution The periodic signal is
represented by
x() = a -nT) (8.76)
Byexpanding x() in a Fourier series

(8.77)

where

T-T|2

(8.78)

Each of the X, gives the same constant T

impulse train is
Therefore, Fourier series representation of the unit
(8.79)

on both sides, we get

If we take the Fourier transform

T
(8.80)
210
T

2r
where domain is transtormedintoan
impulse
impulse train inthe time an each
impulse is a.
Therefore, a unit
domain, where
an area of
rain in the frequency
8.22 Network Theory

8.12 PROPERTIES OF THE FOURIER TRANSFORM

Properties of the Fourier transform facilitate the transformation from the time
domain to the frequency domain and vice-versa.
8.12.1 Linearity
The Fourier transform satisfy linearity and principle of superposition. Consider
two signals xj() and x().

xl0) > X, (o)

then ajx) + azx) > aX(0) + a,X,(o) (8.81)
Proof: F [ax() + ayr()]
= a, F[x()] + a, Fx,(01
=aX(0) +aX,(0) (8.82)
Therefore, the linearity is proved.
8.12.2 Scaling
Consider Fourier transform of x(t) is X(0)
If x() > X(o)
then for real constant a

(8.83)
Proof: Assume a > 0. Then Fourier transform of x(at) is
F|xa)) =| x{at)e aràt d(8.84)
Let m = at
Then dm=a dt

dm
Therefore, Fuan)) =x(n)e (8.85)
If a<0, then

F\xa) = x(an)e io" dt (8.86)

Now -1= + at

dm
 orie Analysis of AC Cireuts &y

dm

(onbiningthese two
values, we get (8.87)
X(a)es lao)
(8.88)
S23 Symmetry
x() X(o)
hen
X() > 2nX(-0)

Proof X(0)e*"do (8.89)

Then replacing the dummy variable oby o

2 x(-t) =| X(0)edo' (8.90)

Now replace t by o
2rx(-0) = X(0)e do' (8.91)
Finally replace o by t, thus,
2nx(-) = X0e"dt =s{xo} (8.92)

X()2r X(0)
8.12.4 Convolution
two signals into the product of their
Fourier transform makes the convolution of convolution properties, one for the
of
Fourier transforms. There are two types
frequency domain.
lime domain and one for the
() Time convolution
x() X(0)
) X(0)

= (t)y((- rldte (O)

then )
(8.93)
=X(0) X{0)
(8.94)

Proof: Fly0)
8.24 Network Theory

(8.95)

Now let &a[- t then dk =dt and t=k+ T

Therefore, (8.96)

(8.97)

Y(0) = X(0) X,(0) (8.98)

(i) Frequency convolution
If x(0) X(0)
x) X(0)
1
then x) x() > 21
X(0) * X,(0) (8.99)
Proof: Consider the inverse transform of [X,(0) *X(0))/2r
We have

-
x,x,(o -)e"do. du (8.10)
Let m = )-uthen dm = do and @= m+ u
Thus,

=x() x ) (8.101)
Therefore, x)-) X(0) *X, (o) (8.102)
2
8.12.5 Frequency Differentiation and Integration
If x() X()
then H) x() dX(0)
do
 Consider
Fourier
transfom of x(0) ore Ayil of Ac
CRts S9
pitterentiating with X(0)=x)e d
respective o (8.103)
do
Fourier
transformof Hj) x() is X(0)
(8.104)
do
(H) x()’ dX (0)
do
Ingeneral
d"X(0) (8.105)
do'
Similarly the
If
frequency
x() > X(O)integration
(8.106)

then
-jt
(8.107)

X(0) = x()e dt (8.108)

x(0)do =x0d|"do
= x(0)d ea
--x0a
x) °x(0)d0 (8.109)
-jt

8.12.6 Time Shifting

x(t) > X(o)
x(1-to) e X(o)
then
implies shift in phase in
can say that shift in time domain
rom the above, we
the transform.
(8.110)
-t6)eMd
Proof: Fut-)]= fx(
=m + I,;dt =dmn
Let t-t, = m ort