C H A P T E R 1 6

THE FOURIER SERIES

Do not worry about your difficulties in mathematics, I assure you that
mine are greater.
—Albert Einstein

Historical Profiles
Jean Baptiste Joseph Fourier (1768–1830), a French mathematician, first presented
the series and transform that bear his name. Fourier’s results were not enthusiastically
received by the scientific world. He could not even get his work published as a paper.
Born in Auxerre, France, Fourier was orphaned at age 8. He attended a local
military college run by Benedictine monks, where he demonstrated great proficiency in
mathematics. Like most of his contemporaries, Fourier was swept into the politics of
the French Revolution. He played an important role in Napoleon’s expeditions to Egypt
in the later 1790s. Due to his political involvement, he narrowly escaped death twice.

Alexander Graham Bell (1847–1922) inventor of the telephone, was a Scottish-

American scientist.
Bell was born in Edinburgh, Scotland, a son of Alexander Melville Bell, a
well-known speech teacher. Alexander the younger also became a speech teacher
after graduating from the University of Edinburgh and the University of London. In
1866 he became interested in transmitting speech electrically. After his older brother
died of tuberculosis, his father decided to move to Canada. Alexander was asked to
come to Boston to work at the School for the Deaf. There he met Thomas A. Watson,
who became his assistant in his electromagnetic transmitter experiment. On March
10, 1876, Alexander sent the famous first telephone message: “Watson, come here I
want you.” The bel, the logarithmic unit introduced in Chapter 14, is named in his honor.

707
▲
▲

| | e-Text Main Menu | Textbook Table of Contents | Problem Solving Workbook Contents
708 PART 3 Advanced Circuit Analyses

16.1 INTRODUCTION
We have spent a considerable amount of time on the analysis of circuits
with sinusoidal sources. This chapter is concerned with a means of an-
alyzing circuits with periodic, nonsinusoidal excitations. The notion of
periodic functions was introduced in Chapter 9; it was mentioned there
that the sinusoid is the most simple and useful periodic function. This
chapter introduces the Fourier series, a technique for expressing a peri-
odic function in terms of sinusoids. Once the source function is expressed
in terms of sinusoids, we can apply the phasor method to analyze circuits.
The Fourier series is named after Jean Baptiste Joseph Fourier
(1768–1830). In 1822, Fourier’s genius came up with the insight that
any practical periodic function can be represented as a sum of sinusoids.
Such a representation, along with the superposition theorem, allows us
to find the response of circuits to arbitrary periodic inputs using phasor
techniques.
We begin with the trigonometric Fourier series. Later we consider
the exponential Fourier series. We then apply Fourier series in circuit
analysis. Finally, practical applications of Fourier series in spectrum
analyzers and filters are demonstrated.

16.2 TRIGONOMETRIC FOURIER SERIES

While studying heat flow, Fourier discovered that a nonsinusoidal periodic
function can be expressed as an infinite sum of sinusoidal functions.
Recall that a periodic function is one that repeats every T seconds. In
other words, a periodic function f (t) satisfies

f (t) = f (t + nT ) (16.1)

where n is an integer and T is the period of the function.

According to the Fourier theorem, any practical periodic function
of frequency ω0 can be expressed as an infinite sum of sine or cosine
functions that are integral multiples of ω0 . Thus, f (t) can be expressed
as
f (t) = a0 + a1 cos ω0 t + b1 sin ω0 t + a2 cos 2ω0 t
(16.2)
+ b2 sin 2ω0 t + a3 cos 3ω0 t + b3 sin 3ω0 t + · · ·
or

∞
f (t) = a0 + (an cos nω0 t + bn sin nω0 t) (16.3)


n=1
dc
 
ac

where ω0 = 2π/T is called the fundamental frequency in radians per

second. The sinusoid sin nω0 t or cos nω0 t is called the nth harmonic
of f (t); it is an odd harmonic if n is odd and an even harmonic if n is
The harmonic frequency ωn is an integral multiple even. Equation 16.3 is called the trigonometric Fourier series of f (t).
of the fundamental frequency ω0 , i.e., ωn = nω0 . The constants an and bn are the Fourier coefficients. The coefficient a0
is the dc component or the average value of f (t). (Recall that sinusoids
▲
▲

| | e-Text Main Menu | Textbook Table of Contents | Problem Solving Workbook Contents
 CHAPTER 16 The Fourier Series 709

have zero average values.) The coefficients an and bn (for n = 0) are the
amplitudes of the sinusoids in the ac component. Thus,

The Fourier series of a periodic function f (t) is a representation that resolves

f (t) into a dc component and an ac component comprising an
infinite series of harmonic sinusoids.

A function that can be represented by a Fourier series as in Eq. (16.3)

must meet certain requirements, because the infinite series in Eq. (16.3)
may or may not converge. These conditions on f (t) to yield a convergent
Fourier series are as follows:
1. f (t) is single-valued everywhere.
2. f (t) has a finite number of finite discontinuities in any one
period.
3. f (t) has a finite number of maxima and minima in any one
period.
 t0 +T
4. The integral |f (t)| dt < ∞ for any t0 .
t0
These conditions are called Dirichlet conditions. Although they are not Historical note: Although Fourier published his
necessary conditions, they are sufficient conditions for a Fourier series to theorem in 1822, it was P. G. L. Dirichlet (1805–
exist. 1859) who later supplied an acceptable proof of
A major task in Fourier series is the determination of the Fourier the theorem.
coefficients a0 , an , and bn . The process of determining the coefficients is
called Fourier analysis. The following trigonometric integrals are very
helpful in Fourier analysis. For any integers m and n,
A software package like Mathcad or Maple can
be used to evaluate the Fourier coefficients.
 T
sin nω0 t dt = 0 (16.4a)
0
 T
cos nω0 t dt = 0 (16.4b)
0
 T
sin nω0 t cos mω0 t dt = 0 (16.4c)
0
 T
sin nω0 t sin mω0 t dt = 0, (m = n) (16.4d)
0
 T
cos nω0 t cos mω0 t dt = 0, (m = n) (16.4e)
0
 T
T
sin2 nω0 t dt = (16.4f)
0 2
 T
T
cos2 nω0 t dt = (16.4g)
0 2

Let us use these identities to evaluate the Fourier coefficients.

▲
▲

| | e-Text Main Menu | Textbook Table of Contents | Problem Solving Workbook Contents
710 PART 3 Advanced Circuit Analyses

We begin by finding a0 . We integrate both sides of Eq. (16.3) over

one period and obtain
 T  T  ∞

f (t) dt = a0 + (an cos nω0 t + bn sin nω0 t) dt
0 0 n=1
 T ∞ 
 T
= a0 dt + an cos nω0 t dt (16.5)
0 n=1 0
 T

+ bn sin nω0 t dt dt
0
Invoking the identities of Eqs. (16.4a) and (16.4b), the two integrals in-
volving the ac terms vanish. Hence,
 T  T
f (t) dt = a0 dt = a0 T
0 0
or
 T
1
a0 = f (t) dt (16.6)
T 0

showing that a0 is the average value of f (t).

To evaluate an , we multiply both sides of Eq. (16.3) by cos mω0 t
and integrate over one period:
 T
f (t) cos mω0 t dt
0
  
T 
∞
= a0 + (an cos nω0 t + bn sin nω0 t) cos mω0 t dt
0 n=1
 T ∞ 
 T
= a0 cos mω0 t dt + an cos nω0 t cos mω0 t dt
0 n=1 0
 T

+ bn sin nω0 t cos mω0 t dt dt (16.7)

0

The integral containing a0 is zero in view of Eq. (16.4b), while the

integral containing bn vanishes according to Eq. (16.4c). The integral
containing an will be zero except when m = n, in which case it is T /2,
according to Eqs. (16.4e) and (16.4g). Thus,
 T
T
f (t) cos mω0 t dt = an , for m = n
0 2
or

2 T
an = f (t) cos nω0 t dt (16.8)
T 0

In a similar vein, we obtain bn by multiplying both sides of Eq.

(16.3) by sin mω0 t and integrating over the period. The result is
 T
2
bn = f (t) sin nω0 t dt (16.9)
T 0
▲
▲

| | e-Text Main Menu | Textbook Table of Contents | Problem Solving Workbook Contents
 CHAPTER 16 The Fourier Series 711

Be aware that since f (t) is periodic, it may be more convenient to carry

the integrations above from −T /2 to T /2 or generally from t0 to t0 + T
instead of 0 to T . The result will be the same.
An alternative form of Eq. (16.3) is the amplitude-phase form


∞
f (t) = a0 + An cos(nω0 t + φn ) (16.10)
n=1

We can use Eqs. (9.11) and (9.12) to relate Eq. (16.3) to Eq. (16.10), or
we can apply the trigonometric identity
cos(α + β) = cos α cos β − sin α sin β (16.11)

to the ac terms in Eq. (16.10) so that


∞ 
∞
a0 + An cos(nω0 t + φn ) = a0 + (An cos φn ) cos nω0 t
n=1 n=1 (16.12)
− (An sin φn ) sin nω0 t
Equating the coefficients of the series expansions in Eqs. (16.3) and
(16.12) shows that
an = An cos φn , bn = −An sin φn (16.13a)

or
 bn
An = an2 + bn2 , φn = − tan−1 (16.13b)
an

To avoid any confusion in determining φn , it may be better to relate the

terms in complex form as
An φn = an − j bn (16.14)

The convenience of this relationship will become evident in Section 16.6.

The plot of the amplitude An of the harmonics versus nω0 is called the
amplitude spectrum of f (t); the plot of the phase φn versus nω0 is the
phase spectrum of f (t). Both the amplitude and phase spectra form
the frequency spectrum of f (t). The frequency spectrum is also known as the
line spectrum in view of the discrete frequency
components.
The frequency spectrum of a signal consists of the plots of the amplitudes
and phases of the harmonics versus frequency.

Thus, the Fourier analysis is also a mathematical tool for finding the
spectrum of a periodic signal. Section 16.6 will elaborate more on the
spectrum of a signal.
To evaluate the Fourier coefficients a0 , an , and bn , we often need
to apply the following integrals:

1
cos at dt = sin at (16.15a)
a

1
sin at dt = − cos at (16.15b)
a
▲
▲

| | e-Text Main Menu | Textbook Table of Contents | Problem Solving Workbook Contents
712 PART 3 Advanced Circuit Analyses


1 1
t cos at dt =
2
cos at + t sin at (16.15c)
a a

1 1
t sin at dt = 2 sin at − t cos at (16.15d)
a a
It is also useful to know the values of the cosine, sine, and exponential
functions for integral multiples of π . These are given in Table 16.1, where
n is an integer.

TABLE 16.1 Values of cosine, sine, and

exponential functions for integral
multiples of π .

Function Value

cos 2nπ 1
sin 2nπ 0
cos nπ (−1)n
sin nπ 0

nπ (−1)n/2 , n = even
cos
2 0, n = odd

nπ (−1)(n−1)/2 , n = odd
sin
2 0, n = even
ej 2nπ 1
j nπ
e (−1)n

(−1)n/2 , n = even
ej nπ/2
j (−1) (n−1)/2
, n = odd

E X A M P L E 1 6 . 1
f(t) Determine the Fourier series of the waveform shown in Fig. 16.1. Obtain
the amplitude and phase spectra.
1
Solution:
The Fourier series is given by Eq. (16.3), namely,
–2 –1 0 1 2 3 t
∞
f (t) = a0 + (an cos nω0 t + bn sin nω0 t) (16.1.1)
Figure 16.1 For Example 16.1; a square wave.
n=1

Our goal is to obtain the Fourier coefficients a0 , an , and bn using Eqs.

(16.6), (16.8), and (16.9). First, we describe the waveform as

1, 0 < t < 1
f (t) = (16.1.2)
0, 1 < t < 2
and f (t) = f (t + T ). Since T = 2, ω0 = 2π/T = π . Thus,
  1  2
1
1 T 1 1  1
a0 = f (t) dt = 1 dt + 0 dt = t  = (16.1.3)
T 0 2 0 1 2 0 2
▲
▲

| | e-Text Main Menu | Textbook Table of Contents | Problem Solving Workbook Contents
 CHAPTER 16 The Fourier Series 713

Using Eq. (16.8) along with Eq. (16.15a),

 T
2
an = f (t) cos nω0 t dt
T 0
 1  2

2
= 1 cos nπ t dt + 0 cos nπ t dt (16.1.4)
2 0 1
1
1  1
= sin nπ t  = sin nπ = 0
nπ 0 nπ

From Eq. (16.9) with the aid of Eq. (16.15b),

 T
2
bn = f (t) sin nω0 t dt
T 0
 1  2

2
= 1 sin nπ t dt + 0 sin nπ t dt
2 0 1
1
1 
=− cos nπ t  (16.1.5)
nπ 0
1
=− (cos nπ − 1), cos nπ = (−1)n
nπ

 2
1 , n = odd
= [1 − (−1)n ] = nπ
nπ 
0, n = even

Substituting the Fourier coefficients in Eqs. (16.1.3) to (16.1.5) into Eq.

(16.1.1) gives the Fourier series as

1 2 2 2
f (t) = + sin π t + sin 3π t + sin 5π t + · · · (16.1.6)
2 π 3π 5π
Since f (t) contains only the dc component and the sine terms with the
fundamental component and odd harmonics, it may be written as

1 2
∞
1
f (t) = + sin nπ t, n = 2k − 1 (16.1.7)
2 π k=1 n

By summing the terms one by one as demonstrated in Fig. 16.2, Summing the Fourier terms by hand calculation
we notice how superposition of the terms can evolve into the original may be tedious. A computer is helpful to com-
square. As more and more Fourier components are added, the sum gets pute the terms and plot the sum like those shown
closer and closer to the square wave. However, it is not possible in in Fig. 16.2.
practice to sum the series in Eq. (16.1.6) or (16.1.7) to infinity. Only a
partial sum (n = 1, 2, 3, . . . , N, where N is finite) is possible. If we plot
the partial sum (or truncated series) over one period for a large N as in
Fig. 16.3, we notice that the partial sum oscillates above and below the
actual value of f (t). At the neighborhood of the points of discontinuity
(x = 0, 1, 2, . . .), there is overshoot and damped oscillation. In fact, an
overshoot of about 9 percent of the peak value is always present, regardless Historical note: Named after the mathematical
of the number of terms used to approximate f (t). This is called the Gibbs physicist Josiah Willard Gibbs, who first ob-
phenomenon. served it in 1899.
▲
▲

| | e-Text Main Menu | Textbook Table of Contents | Problem Solving Workbook Contents
714 PART 3 Advanced Circuit Analyses

f (t)
1
2
1

dc component t

0 1 2 t
t

Figure 16.3 Truncating the Fourier series at

Fundamental ac component N = 11; Gibbs phenomenon.
(a)

Finally, let us obtain the amplitude and phase spectra for the signal
in Fig. 16.1. Since an = 0,

t   2 , n = odd
An = an + bn = |bn | = nπ
2 2 (16.1.8)

0, n = even
Sum of first two ac components
and

bn −90◦ , n = odd
φn = − tan−1 = (16.1.9)
an 0, n = even
t The plots of An and φn for different values of nω0 = nπ provide the
amplitude and phase spectra in Fig. 16.4. Notice that the amplitudes of
the harmonics decay very fast with frequency.
Sum of first three ac components

An 2
p

0.5
t
2
3p
2
5p
Sum of first four ac components

0 p 2p 3p 4p 5p 6p v
(a)

t f

p 2p 3p 4p 5p 6p
0°
v
Sum of first five ac components
(b)
–90°
Figure 16.2 Evolution of a
(b)
square wave from its Fourier
components.
Figure 16.4 For Example 16.1: (a) ampli-
tude and (b) phase spectrum of the function
shown in Fig. 16.1.
▲
▲

| | e-Text Main Menu | Textbook Table of Contents | Problem Solving Workbook Contents
746 PART 3 Advanced Circuit Analyses

† 16.8 APPLICATIONS
We demonstrated in Section 16.4 that the Fourier series expansion per-
mits the application of the phasor techniques used in ac analysis to cir-
cuits containing nonsinusoidal periodic excitations. The Fourier series
has many other practical applications, particularly in communications
and signal processing. Typical applications include spectrum analysis,
filtering, rectification, and harmonic distortion. We will consider two of
these: spectrum analyzers and filters.
TABLE 16.4 Frequency ranges of
typical signals. 16.8.1 Spectrum Analyzers
Signal Frequency Range The Fourier series provides the spectrum of a signal. As we have seen, the
spectrum consists of the amplitudes and phases of the harmonics versus
Audible sounds 20 Hz to 15 kHz frequency. By providing the spectrum of a signal f (t), the Fourier series
AM radio 540–1600 kHz helps us identify the pertinent features of the signal. It demonstrates
Short-wave radio 3–36 MHz which frequencies are playing an important role in the shape of the output
Video signals dc to 4.2 MHz and which ones are not. For example, audible sounds have significant
(U.S. standards) components in the frequency range of 20 Hz to 15 kHz, while visible
VHF television, 54–216 MHz light signals range from 105 GHz to 106 GHz. Table 16.4 presents some
FM radio other signals and the frequency ranges of their components. A periodic
UHF television 470–806 MHz function is said to be band-limited if its amplitude spectrum contains only
Cellular telephone 824–891.5 MHz a finite number of coefficients An or cn . In this case, the Fourier series
Microwaves 2.4–300 GHz becomes
Visible light 105 –106 GHz 
N 
N
X-rays 108 –109 GHz f (t) = cn ej nω0 t = a0 + An cos(nω0 t + φn ) (16.75)
n=−N n=1

This shows that we need only 2N +1 terms (namely, a0 , A1 , A2 , . . . , AN ,

φ1 , φ2 , . . . , φN ) to completely specify f (t) if ω0 is known. This leads to
the sampling theorem: a band-limited periodic function whose Fourier
series contains N harmonics is uniquely specified by its values at 2N + 1
instants in one period.
A spectrum analyzer is an instrument that displays the amplitude of
the components of a signal versus frequency. In other words, it shows the
various frequency components (spectral lines) that indicate the amount
of energy at each frequency. It is unlike an oscilloscope, which displays
the entire signal (all components) versus time. An oscilloscope shows the
signal in the time domain, while the spectrum analyzer shows the signal
in the frequency domain. There is perhaps no instrument more useful to a
circuit analyst than the spectrum analyzer. An analyzer can conduct noise
and spurious signal analysis, phase checks, electromagnetic interference
and filter examinations, vibration measurements, radar measurements,
and more. Spectrum analyzers are commercially available in various
sizes and shapes. Figure 16.42 displays a typical one.

16.8.2 Filters
Filters are an important component of electronics and communications
systems. Chapter 14 presented a full discussion on passive and active fil-
ters. Here, we investigate how to design filters to select the fundamental
component (or any desired harmonic) of the input signal and reject other
harmonics. This filtering process cannot be accomplished without the
▲
▲

| | e-Text Main Menu | Textbook Table of Contents | Problem Solving Workbook Contents
 CHAPTER 16 The Fourier Series 747

Figure 16.42 A typical spectrum analyzer.

(Courtesy of Hewlett-Packer.)

Fourier series expansion of the input signal. For the purpose of illustra-
tion, we will consider two cases, a lowpass filter and a bandpass filter. In
Example 16.6, we already looked at a highpass RL filter.
The output of a lowpass filter depends on the input signal, the trans-
fer function H (ω) of the filter, and the corner or half-power frequency
ωc . We recall that ωc = 1/RC for an RC passive filter. As shown in
Fig. 16.43(a), the lowpass filter passes the dc and low-frequency com-
ponents, while blocking the high-frequency components. By making ωc
sufficiently large (ωc  ω0 , e.g., making C small), a large number of the

|H |

1
1
2
0 v0 2v0 3v0 v 0 v0 2v0 3v0 v

0 vc v

(a)

A Lowpass
A dc
filter
2

vc << v0

(b)

Figure 16.43 (a) Input and output spectra of a lowpass filter, (b) the lowpass filter passes
only the dc component when ωc  ω0 .
▲
▲

| | e-Text Main Menu | Textbook Table of Contents | Problem Solving Workbook Contents
748 PART 3 Advanced Circuit Analyses

harmonics can be passed. On the other hand, by making ωc sufficiently

small (ωc  ω0 ), we can block out all the ac components and pass only
dc, as shown typically in Fig. 16.43(b). (See Fig. 16.2(a) for the Fourier
series expansion of the square wave.)
Similarly, the output of a bandpass filter depends on the input signal,
the transfer function of the filter H (ω), its bandwidth B, and its center
In this section, we have used ωc for the center frequency ωc . As illustrated in Fig. 16.44(a), the filter passes all the
frequency of the bandpass filter instead of ω0 as harmonics of the input signal within a band of frequencies (ω1 < ω < ω2 )
in Chapter 14, to avoid confusing ω0 with the centered around ωc . We have assumed that ω0 , 2ω0 , and 3ω0 are within
fundamental frequency of the input signal. that band. If the filter is made highly selective (B  ω0 ) and ωc = ω0 ,
where ω0 is the fundamental frequency of the input signal, the filter passes
only the fundamental component (n = 1) of the input and blocks out all
higher harmonics. As shown in Fig. 16.44(b), with a square wave as
input, we obtain a sine wave of the same frequency as the output. (Again,
refer to Fig. 16.2(a).)

|H|

1
1
0 v0 2v0 3v0 v 2 0 v0 2v0 3v0 v

0 v1 vc v2 v

(a)

Bandpass
filter

T
vc = v0
B << v0 T

(b)

Figure 16.44 (a) Input and output spectra of a bandpass filter, (b) the bandpass filter
passes only the fundamental component when B  ω0 .

E X A M P L E 1 6 . 1 4
If the sawtooth waveform in Fig. 16.45(a) is applied to an ideal lowpass
filter with the transfer function shown in Fig. 16.45(b), determine the
output.
Solution:
The input signal in Fig. 16.45(a) is the same as the signal in Fig. 16.9.
From Practice Prob. 16.2, we know that the Fourier series expansion is
1 1 1 1
x(t) = − sin ω0 t − sin 2ω0 t − sin 3ω0 t − · · ·
2 π 2π 3π
▲
▲

| | e-Text Main Menu | Textbook Table of Contents | Problem Solving Workbook Contents
 CHAPTER 16 The Fourier Series 749

where the period is T = 1 s and the fundamental frequency is ω0 = 2π

rad/s. Since the corner frequency of the filter is ωc = 10 rad/s, only the
dc component and harmonics with nω0 < 10 will be passed. For n = 2,
nω0 = 4π = 12.566 rad/s, which is higher than 10 rad/s, meaning that
second and higher harmonics will be rejected. Thus, only the dc and
fundamental components will be passed. Hence the output of the filter is
1 1
y(t) = − sin 2π t
2 π

x(t) |H |

1 1

–1 0 1 2 3 t 0 10 v
(a) (b)

Figure 16.45 For Example 16.14.

PRACTICE PROBLEM 16.14

Rework Example 16.14 if the lowpass filter is replaced by the ideal band- |H |
pass filter shown in Fig. 16.46. 1
1 1 1
Answer: y(t) = − sin 3ω0 t − sin 4ω0 t − sin 5ω0 t.
3π 4π 5π 0 15 35 v

Figure 16.46 For Practice Prob. 16.14.

16.9 SUMMARY
1. A periodic function is one that repeats itself every T seconds; that
is, f (t ± nT ) = f (t), n = 1, 2, 3, . . . .
2. Any nonsinusoidal periodic function f (t) that we encounter in
electrical engineering can be expressed in terms of sinusoids using
Fourier series:
 ∞
f (t) = a0 + (an cos nω0 t + bn sin nω0 t)


n=1
dc
 
ac

where ω0 = 2π/T is the fundamental frequency. The Fourier series

resolves the function into the dc component a0 and an ac compo-
nent containing infinitely many harmonically related sinusoids. The
▲
▲

| | e-Text Main Menu | Textbook Table of Contents | Problem Solving Workbook Contents
 C H A P T E R 1 7
FOURIER TRANSFORM
No human investigation can claim to be scientific if it doesn’t pass the
test of mathematical proof.
—Leonardo da Vinci

Enhancing Your Career

Career in Communications Systems Communications
systems apply the principles of circuit analysis. A com-
munication system is designed to convey information from
a source (the transmitter) to a destination (the receiver) via
a channel (the propagation medium). Communications en-
gineers design systems for transmitting and receiving infor-
mation. The information can be in the form of voice, data,
or video.
We live in the information age—news, weather,
sports, shopping, financial, business inventory, and other
sources make information available to us almost instantly
via communications systems. Some obvious examples of
communications systems are the telephone network, mobile
cellular telephones, radio, cable TV, satellite TV, fax, and
radar. Mobile radio, used by police and fire departments,
aircraft, and various businesses is another example.
The field of communications is perhaps the fastest
growing area in electrical engineering. The merging of
the communications field with computer technology in re-
cent years has led to digital data communications networks
such as local area networks, metropolitan area networks,
and broadband integrated services digital networks. For ex-
ample, the Internet (the “information superhighway”) al-
lows educators, business people, and others to send elec-
tronic mail from their computers worldwide, log onto remote
databases, and transfer files. The Internet has hit the world
like a tidal wave and is drastically changing the way people
do business, communicate, and get information. This trend
will continue.
A communications systems engineer designs sys-
tems that provide high-quality information services. The
systems include hardware for generating, transmitting, and
receiving information signals. Communications engineers Cordless phone. Source: M. Nemzow, Fast Ethernet Implemen-
are employed in numerous communications industries and tation and Migration Solutions [New York: McGraw-Hill, 1997],
p. 176.
places where communications systems are routinely used.
More and more government agencies, academic depart-
ments, and businesses are demanding faster and more
accurate transmission of information. To meet these needs,
communications engineers are in high demand. Therefore,
the future is in communications and every electrical
engineer must prepare accordingly.

759
▲
▲

| | e-Text Main Menu | Textbook Table of Contents | Problem Solving Workbook Contents
760 PART 3 Advanced Circuit Analyses

17.1 INTRODUCTION
Fourier series enable us to represent a periodic function as a sum of sinu-
soids and to obtain the frequency spectrum from the series. The Fourier
transform allows us to extend the concept of a frequency spectrum to non-
periodic functions. The transform assumes that a nonperiodic function is
a periodic function with an infinite period. Thus, the Fourier transform
is an integral representation of a nonperiodic function that is analogous
to a Fourier series representation of a periodic function.
The Fourier transform is an integral transform like the Laplace
transform. It transforms a function in the time domain into the frequency
domain. The Fourier transform is very useful in communications systems
and digital signal processing, in situations where the Laplace transform
does not apply. While the Laplace transform can only handle circuits
with inputs for t > 0 with initial conditions, the Fourier transform can
handle circuits with inputs for t < 0 as well as those for t > 0.
We begin by using a Fourier series as a stepping stone in defining the
Fourier transform. Then we develop some of the properties of the Fourier
transform. Next, we apply the Fourier transform in analyzing circuits. We
discuss Parseval’s theorem, compare the Laplace and Fourier transforms,
and see how the Fourier transform is applied in amplitude modulation
and sampling.

17.2 DEFINITION OF THE FOURIER TRANSFORM

We saw in the previous chapter that a nonsinusoidal periodic function can
be represented by a Fourier series, provided that it satisfies the Dirichlet
conditions. What happens if a function is not periodic? Unfortunately,
there are many important nonperiodic functions—such as a unit step or
an exponential function—that we cannot represent by a Fourier series.
As we shall see, the Fourier transform allows a transformation from the
time to the frequency domain, even if the function is not periodic.
p(t)
Suppose we want to find the Fourier transform of a nonperiodic
function p(t), shown in Fig. 17.1(a). We consider a periodic function f (t)
A whose shape over one period is the same as p(t), as shown in Fig. 17.1(b).
If we let the period T → ∞, only a single pulse of width τ [the desired
nonperiodic function in Fig. 17.1(a)] remains, because the adjacent pulses
0 t t have been moved to infinity. Thus, the function f (t) is no longer periodic.
(a) In other words, f (t) = p(t) as T → ∞. It is interesting to consider the
spectrum of f (t) for A = 10 and τ = 0.2 (see Section 16.6). Figure 17.2
f(t) shows the effect of increasing T on the spectrum. First, we notice that
A the general shape of the spectrum remains the same, and the frequency at
which the envelope first becomes zero remains the same. However, the
amplitude of the spectrum and the spacing between adjacent components
−T 0 t T t
both decrease, while the number of harmonics increases. Thus, over a
range of frequencies, the sum of the amplitudes of the harmonics remains
(b) almost constant. Since the total “strength” or energy of the components
within a band must remain unchanged, the amplitudes of the harmonics
Figure 17.1 (a) A nonperiodic function,
(b) increasing T to infinity makes f (t) must decrease as T increases. Since f = 1/T , as T increases, f or ω
become the nonperiodic function in (a). decreases, so that the discrete spectrum ultimately becomes continuous.
▲
▲

| | e-Text Main Menu | Textbook Table of Contents | Problem Solving Workbook Contents
 CHAPTER 17 Fourier Transform 761

t = 0.2

−5 0 5 Hz
T=1
1
t = 0.2

−5 0 5 Hz
T=2

t = 0.2

0.4

−5 0 5 Hz
T=5

Figure 17.2 Effect of increasing T on the spectrum of the periodic

pulse trains in Fig. 17.1(b).
(Source: L. Balmer, Signals and Systems: An Introduction
[London: Prentice-Hall, 1991], p. 229.)

To further understand this connection between a nonperiodic func-

tion and its periodic counterpart, consider the exponential form of a
Fourier series in Eq. (16.58), namely,

∞
f (t) = cn ej nω0 t (17.1)
n=−∞

where
 T /2
1
cn = f (t)e−j nω0 t dt (17.2)
T −T /2

The fundamental frequency is

2π
ω0 = (17.3)
T
and the spacing between adjacent harmonics is

2π
ω = (n + 1)ω0 − nω0 = ω0 = (17.4)
T
▲
▲

| | e-Text Main Menu | Textbook Table of Contents | Problem Solving Workbook Contents
762 PART 3 Advanced Circuit Analyses

Substituting Eq. (17.2) into Eq. (17.1) gives

∞   
1 T /2 −j nω0 t
f (t) = f (t)e dt ej nω0 t
n=−∞ T −T /2

∞   
ω T /2 −j nω0 t
= f (t)e dt ej nω0 t (17.5)
n=−∞ 2π −T /2
∞  T /2 
1

= f (t)e−j nω0 t dt ωej nω0 t
2π n=−∞ −T /2
If we let T → ∞, the summation becomes integration, the incremen-
tal spacing ω becomes the differential separation dω, and the discrete
harmonic frequency nω0 becomes a continuous frequency ω. Thus, as
T → ∞,

∞  ∞
⇒
n=−∞ −∞
(17.6)
ω ⇒ dω
nω0 ⇒ ω
so that Eq. (17.5) becomes
 ∞  ∞ 
1
f (t) = f (t)e−j ωt dt ej ωt dω (17.7)
2π −∞ −∞
The term in the brackets is known as the Fourier transform of f (t) and
Some authors use F( jω) instead of F(ω) to rep- is represented by F (ω). Thus
resent the Fourier transform.
 ∞
F (ω) = F[f (t)] = f (t)e−j ωt dt (17.8)
−∞

where F is the Fourier transform operator. It is evident from Eq. (17.8)

that:

The Fourier transform is an integral transformation of f (t) from the

time domain to the frequency domain.

In general, F (ω) is a complex function; its magnitude is called the

amplitude spectrum, while its phase is called the phase spectrum. Thus
F (ω) is the spectrum.
Equation (17.7) can be written in terms of F (ω), and we obtain the
inverse Fourier transform as
 ∞
1
f (t) = F −1 [F (ω)] = F (ω)ej ωt dω (17.9)
2π −∞

The function f (t) and its transform F (ω) form the Fourier transform
pairs:
f (t) ⇐⇒ F (ω) (17.10)

since one can be derived from the other.

▲
▲

| | e-Text Main Menu | Textbook Table of Contents | Problem Solving Workbook Contents
 CHAPTER 17 Fourier Transform 763

The Fourier transform F (ω) exists when the Fourier integral in Eq.
(17.8) converges. A sufficient but not necessary condition that f (t) has
a Fourier transform is that it be completely integrable in the sense that
 ∞
|f (t)| dt < ∞ (17.11)
−∞

For example, the Fourier transform of the unit ramp function tu(t) does
not exist, because the function does not satisfy the condition above.
To avoid the complex algebra that explicitly appears in the Fourier
transform, it is sometimes expedient to temporarily replace j ω with s and
then replace s with j ω at the end.

E X A M P L E 1 7 . 1
Find the Fourier transform of the following functions: (a) δ(t − t0 ),
(b) ej ω0 t , (c) cos ω0 t.
Solution:
(a) For the impulse function,
 ∞
F (ω) = F[δ(t − t0 )] = δ(t − t0 )e−j ωt dt = e−j ωt0 (17.1.1)
−∞

where the sifting property of the impulse function in Eq. (7.32) has been
applied. For the special case t0 = 0, we obtain
F[δ(t)] = 1 (17.1.2)

This shows that the magnitude of the spectrum of the impulse function
is constant; that is, all frequencies are equally represented in the impulse
function.
(b) We can find the Fourier transform of ej ω0 t in two ways. If we let
F (ω) = δ(ω − ω0 )
then we can find f (t) using Eq. (17.9), writing
 ∞
1
f (t) = δ(ω − ω0 )ej ωt dω
2π −∞
Using the sifting property of the impulse function gives
1 j ω0 t
f (t) = e
2π
Since F (ω) and f (t) constitute a Fourier transform pair, so too must
2π δ(ω − ω0 ) and ej ω0 t ,
F[ej ω0 t ] = 2π δ(ω − ω0 ) (17.1.3)

Alternatively, from Eq. (17.1.2),

δ(t) = F −1 [1]
Using the inverse Fourier transform formula in Eq. (17.9),
 ∞
1
δ(t) = F −1 [1] = 1ej ωt dω
2π −∞
▲
▲

| | e-Text Main Menu | Textbook Table of Contents | Problem Solving Workbook Contents
764 PART 3 Advanced Circuit Analyses

or
 ∞
ej ωt dω = 2π δ(t) (17.1.4)
−∞

Interchanging variables t and ω results in

 ∞
ej ωt dt = 2π δ(ω) (17.1.5)
−∞

Using this result, the Fourier transform of the given function is

 ∞  ∞
F[ej ω0 t ] = ej ω0 t e−j ωt dt = ej (ω0 −ω) dt = 2π δ(ω0 − ω)
−∞ −∞

Since the impulse function is an even function, with δ(ω0 − ω) = δ(ω −

ω0 ),
F[ej ω0 t ] = 2π δ(ω − ω0 ) (17.1.6)

By simply changing the sign of ω0 , we readily obtain

F[e−j ω0 t ] = 2π δ(ω + ω0 ) (17.1.7)

Also, by setting ω0 = 0,
F[1] = 2π δ(ω) (17.1.8)

(c) By using the result in Eqs. (17.1.6) and (17.1.7), we get

 j ω0 t 
e + e−j ω0 t
F[cos ω0 t] = F
2
1 1 (17.1.9)
= F[ej ω0 t ] + F[e−j ω0 t ]
2 2
= π δ(ω − ω0 ) + π δ(ω + ω0 )
The Fourier transform of the cosine signal is shown in Fig. 17.3.

f (t) F(v)

1 p p

0 t −v0 0 v0 v

Figure 17.3 Fourier transform of f (t) = cos ω0 t.

PRACTICE PROBLEM 17.1

Determine the Fourier transforms of the following functions: (a) gate
function g(t) = u(t − 1) − u(t − 2), (b) 4δ(t + 2), (c) sin ω0 t.
Answer: (a) (e−j ω − e−j 2ω )/j ω, (b) 4ej 2ω ,
(c) j π [δ(ω + ω0 ) − π δ(ω − ω0 )].
▲
▲

| | e-Text Main Menu | Textbook Table of Contents | Problem Solving Workbook Contents
 CHAPTER 17 Fourier Transform 765

E X A M P L E 1 7 . 2
Derive the Fourier transform of a single rectangular pulse of width τ and f (t)
height A, shown in Fig. 17.4. A
Solution:
 
τ/2
−j ωt A −j ωt τ/2
F (ω) = Ae dt = − e 
−τ/2 jω −τ/2 −t 0 t t
2 2
 j ωτ/2 −j ωτ/2 
2A e −e
= Figure 17.4 A rectangular
ω 2j pulse; for Example 17.2.
sin ωτ/2 ωτ
= Aτ = Aτ sinc
ωτ/2 2
|F(v)|
If we make A = 10 and τ = 2 as in Fig. 16.27 (like in Section 16.6), 20
then
F (ω) = 20 sinc ω
whose amplitude spectrum is shown in Fig. 17.5. Comparing Fig. 17.4
with the frequency spectrum of the rectangular pulses in Fig. 16.28, we
notice that the spectrum in Fig. 16.28 is discrete and its envelope has the −3p −2p −p 0 p 2p 3p v
same shape as the Fourier transform of a single rectangular pulse.
Figure 17.5 Amplitude spectrum
of the rectangular pulse in Fig. 17.4;
for Example 17.2.

PRACTICE PROBLEM 17.2

Obtain the Fourier transform of the function in Fig. 17.6. f (t)
2(cos ω − 1) 1
Answer: .
jω
1
−1 0 t

−1

Figure 17.6 For Practice Prob. 17.2.

E X A M P L E 1 7 . 3
Obtain the Fourier transform of the “switched-on” exponential function
shown in Fig. 17.7.
Solution:
From Fig. 17.7,
 −at
e , t >0
f (t) = e−at u(t) =
0, t <0
▲
▲

| | e-Text Main Menu | Textbook Table of Contents | Problem Solving Workbook Contents
766 PART 3 Advanced Circuit Analyses

f(t)
Hence,
1  ∞  ∞  ∞
F (ω) = f (t)e−j ωt dt = e−at e−j ωt dt = e−(a+j ω)t dt
−∞ 0 0
e−at 
−1 −(a+j ω)t ∞ 1
= e  =
a + jω 0 a + jω

0 t

Figure 17.7 For Example 17.3.

PRACTICE PROBLEM 17.3

f(t) Determine the Fourier transform of the “switched-off” exponential func-
1 tion in Fig. 17.8.
1
Answer: .
a − jω
e at

0 t

Figure 17.8 For Practice Prob. 17.3.

17.3 PROPERTIES OF THE FOURIER TRANSFORM

We now develop some properties of the Fourier transform that are useful
in finding the transforms of complicated functions from the transforms
of simple functions. For each property, we will first state and derive it,
and then illustrate it with some examples.

Linearity
If F1 (ω) and F2 (ω) are the Fourier transforms of f1 (t) and f2 (t), respec-
tively, then

F[a1 f1 (t) + a2 f2 (t)] = a1 F1 (ω) + a2 F2 (ω) (17.12)

where a1 and a2 are constants. This property simply states that the Fourier
transform of a linear combination of functions is the same as the linear
combination of the transforms of the individual functions. The proof of
the linearity property in Eq. (17.12) is straightforward. By definition,
 ∞
F[a1 f1 (t) + a2 f2 (t)] = [a1 f1 (t) + a2 f2 (t)]e−j ωt dt
−∞
 ∞  ∞
−j ωt
= a1 f1 (t)e dt + a2 f2 (t)e−j ωt dt
−∞ −∞
= a1 F1 (ω) + a2 F2 (ω)
(17.13)
▲
▲

| | e-Text Main Menu | Textbook Table of Contents | Problem Solving Workbook Contents
 CHAPTER 17 Fourier Transform 767

For example, sin ω0 t = 1

2j
(ej ω0 t − e−j ω0 t ). Using the linearity
property,
1
F [sin ω0 t] = [F(ej ω0 t ) − F(e−j ω0 t )]
2j
(17.14)
π
= [δ(ω − ω0 ) − δ(ω + ω0 )]
j
Time Scaling
If F (ω) = F[f (t)], then

1 ω

F[f (at)] = F (17.15)

|a| a

where a is a constant. Equation (17.15) shows that time expansion

(|a| > 1) corresponds to frequency compression, or conversely, time
compression (|a| < 1) implies frequency expansion. The proof of the
time-scaling property proceeds as follows.
 ∞
F[f (at)] = f (at)e−j ωt dt (17.16)
−∞

If we let x = at, so that dx = a dt, then

 ∞
dx 1 ω

F[f (at)] = f (x)e−j ωx/a = F (17.17)

−∞ a a a
For example, for the rectangular pulse p(t) in Example 17.2,
ωτ
F[p(t)] = Aτ sinc (17.18a)
2
Using Eq. (17.15),
Aτ ωτ
F[p(2t)] = sinc (17.18b)
2 4
It may be helpful to plot p(t) and p(2t) and their Fourier transforms.
Since

A, − τ < t < τ
p(t) = 2 2 (17.19a)
0, otherwise

then replacing every t with 2t gives

 
A, − τ < 2t < τ A, −
τ
<t <
τ
p(2t) = 2 2 = 4 4 (17.19b)
0, otherwise 0, otherwise
showing that p(2t) is time compressed, as shown in Fig. 17.9(b). To plot
both Fourier transforms in Eq. (17.18), we recall that the sinc function
has zeros when its argument is nπ, where n is an integer. Hence, for the
transform of p(t) in Eq. (17.18a), ωτ/2 = 2πf τ/2 = nπ → f = n/τ ,
and for the transform of p(2t) in Eq. (17.18b), ωτ/4 = 2πf τ/4 =
nπ → f = 2n/τ . The plots of the Fourier transforms are shown in Fig.
17.9, which shows that time compression corresponds with frequency
▲
▲

| | e-Text Main Menu | Textbook Table of Contents | Problem Solving Workbook Contents
768 PART 3 Advanced Circuit Analyses

expansion. We should expect this intuitively, because when the signal is

squashed in time, we expect it to change more rapidly, thereby causing
higher-frequency components to exist.

F[ p(t)]
At

p(t)

−t 0 t t
−3 −2 −1 0 1 2 3 f
2 2 t t t t t t

(a)

p(2t)
F[ p(2t)]
A At
2

−t 0 t t −t −t 0 t t f
4 4 2 2
(b)

Figure 17.9 The effect of time scaling: (a) transform of the pulse, (b) time compres-
sion of the pulse causes frequency expansion.

Time Shifting
If F (ω) = F[f (t)], then

F[f (t − t0 )] = e−j ωt0 F (ω) (17.20)

that is, a delay in the time domain corresponds to a phase shift in the
frequency domain. To derive the time shifting property, we note that
 ∞
F[f (t − t0 )] = f (t − t0 )e−j ωt dt (17.21)
−∞

If we let x = t − t0 so that dx = dt and t = x + t0 , then

 ∞
F[f (t − t0 )] = f (x)e−j ω(x+t0 ) dx
−∞
 ∞
(17.22)
−j ωt0 −j ωx −j ωt0
=e f (x)e dx = e F (ω)
−∞

Similarly, F[f (t + t0 )] = ej ωt0 F (ω).

For example, from Example 17.3,
1
F[e−at u(t)] = (17.23)
a + jω
▲
▲

| | e-Text Main Menu | Textbook Table of Contents | Problem Solving Workbook Contents
782 PART 3 Advanced Circuit Analyses

17.5 PARSEVAL’S THEOREM

Parseval’s theorem demonstrates one practical use of the Fourier trans-
form. It relates the energy carried by a signal to the Fourier transform
of the signal. If p(t) is the power associated with the signal, the energy
carried by the signal is
 ∞
W = p(t) dt (17.57)
−∞

In order to be able compare the energy content of current and voltage

signals, it is convenient to use a 1-0 resistor as the base for energy cal-
culation. For a 1-0 resistor, p(t) = v 2 (t) = i 2 (t) = f 2 (t), where f (t)
stands for either voltage or current. The energy delivered to the 1-0 resis-
tor is
 ∞
W10 = f 2 (t) dt (17.58)
−∞

Parseval’s theorem states that this same energy can be calculated in the
frequency domain as
 ∞  ∞
1
W10 = f (t) dt =
2
|F (ω)|2 dω (17.59)
−∞ 2π −∞

Parseval’s theorem states that the total energy delivered to a 1-0 resistor equals
the total area under the square of f (t) or 1/2π times the total area under the
square of the magnitude of the Fourier transform of f (t).

Parseval’s theorem relates energy associated with a signal to its Fourier

transform. It provides the physical significance of F (ω), namely, that
|F (ω)|2 is a measure of the energy density (in joules per hertz) corre-
In fact, |F(ω)|2 is sometimes known as the energy sponding to f (t).
spectral density of signal f (t). To derive Eq. (17.59), we begin with Eq. (17.58) and substitute Eq.
(17.9) for one of the f (t)’s. We obtain
 ∞  ∞   ∞ 
1
W10 = f (t) dt =
2
f (t) j ωt
F (ω)e dω dt (17.60)
−∞ −∞ 2π −∞
The function f (t) can be moved inside the integral within the brackets,
since the integral does not involve time:
 ∞ ∞
1
W10 = f (t)F (ω)ej ωt dω dt (17.61)
2π −∞ −∞
Reversing the order of integration,
 ∞  ∞ 
1 −j (−ω)t
W10 = F (ω) f (t)e dt dω
2π −∞ −∞
 ∞  ∞ (17.62)
1 1
= F (ω)F (−ω) dω = F (ω)F ∗ (ω) dω
2π −∞ 2π −∞
But if z = x + jy, zz∗ = (x + jy)(x − jy) = x 2 + y 2 = |z|2 . Hence,
▲
▲

| | e-Text Main Menu | Textbook Table of Contents | Problem Solving Workbook Contents
 CHAPTER 17 Fourier Transform 783

 ∞  ∞
1
W10 = f (t) dt =
2
|F (ω)|2 dω (17.63)
−∞ 2π −∞

as expected. Equation (17.63) indicates that the energy carried by a signal

can be found by integrating either the square of f (t) in the time domain
or 1/2π times the square of F (ω) in the frequency domain.
Since |F (ω)|2 is an even function, we may integrate from 0 to ∞
and double the result, that is,
 ∞ 
1 ∞
W10 = f (t) dt =
2
|F (ω)|2 dω (17.64)
−∞ π 0
We may also calculate the energy in any frequency band ω1 < ω < ω2
as

1 ω2
W10 = |F (ω)|2 dω (17.65)
π ω1
Notice that Parseval’s theorem as stated here applies to nonperiodic
functions. Parseval’s theorem for periodic functions was presented in
Sections 16.5 and 16.6. As evident in Eq. (17.63), Parseval’s theorem
shows that the energy associated with a nonperiodic signal is spread over
the entire frequency spectrum, whereas the energy of the periodic signal
is concentrated at the frequencies of its harmonic components.

E X A M P L E 1 7 . 9
The voltage across a 10-0 resistor is v(t) = 5e−3t u(t) V. Find the total
energy dissipated in the resistor.
Solution:
We can find the energy using either f (t) = v(t) or F (ω) = V (ω). In the
time domain,
 ∞  ∞
W100 = 10 f (t) dt = 10
2
25e−6t dt
−∞ 0

−6t ∞
e  250
= 250 = = 41.67 J
−6 0 6
In the frequency domain,
5
F (ω) = V (ω) =
3 + jω
so that
25
|F (ω)|2 = F (ω)F ∗ (ω) =
9 + ω2
Hence, the energy dissipated is
 
10 ∞ 10 ∞ 25
W100 = |F (ω)|2 dω = dω
2π −∞ π 0 9 + ω2
 ∞  
250 1 −1 ω 
 250 1 π
250
= tan = = = 41.67 J
π 3 3 0 π 3 2 6
▲
▲

| | e-Text Main Menu | Textbook Table of Contents | Problem Solving Workbook Contents
784 PART 3 Advanced Circuit Analyses

PRACTICE PROBLEM 17.9

(a) Calculate the total energy absorbed by a 1-0 resistor with i(t) =
10e−2|t| A in the time domain. (b) Repeat (a) in the frequency domain.
Answer: (a) 50 J, (b) 50 J.

E X A M P L E 1 7 . 1 0
Calculate the fraction of the total energy dissipated by a 1-0 resistor in
the frequency band 0 < ω < 10 rad/s when the voltage across it is
v(t) = e−2t u(t).
Solution:
Given that f (t) = v(t) = e−2t u(t), then
1 1
F (ω) = ⇒ |F (ω)|2 =
2 + jω 4 + ω2
The total energy dissipated by the resistor is
 
1 ∞ 1 ∞ dω
W10 = |F (ω)|2 dω =
π 0 π 0 4 + ω2
 ∞   
1 1 −1 ω  1 1 π
= tan  = = 0.25 J
π 2 2 0 π 2 2
The energy in the frequencies 0 < ω < 10 is
   10 
1 10 1 10 dω 1 1 1 ω
W = |F (ω)|2 dω = = tan 
π 0 π 0 4 + ω2 π 2 2 0
 
1 −1 1 78.69◦
= tan 5 = π = 0.218 J
2π 2π 180◦
Its percentage of the total energy is
W 0.218
= = 87.4 %
W10 0.25

PRACTICE PROBLEM 17.10

A 2-0 resistor has i(t) = e−t u(t). What percentage of the total energy
is in the frequency band −4 < ω < 4 rad/s?
Answer: 84.4 percent.

17.6 COMPARING THE FOURIER AND LAPLACE

TRANSFORMS
It is worthwhile to take some moments to compare the Laplace and Fourier
transforms. The following similarities and differences should be noted:
1. The Laplace transform defined in Chapter 14 is one-sided in
that the integral is over 0 < t < ∞, making it only useful for
positive-time functions, f (t), t > 0. The Fourier transform is
applicable to functions defined for all time.
▲
▲

| | e-Text Main Menu | Textbook Table of Contents | Problem Solving Workbook Contents
 CHAPTER 17 Fourier Transform 785

2. For a function f (t) that is nonzero for positive time only (i.e.,
∞
f (t) = 0, t < 0) and |f (t)| dt < ∞, the two transforms
0
are related by In other words, if all the poles of F(s) lie in the

F (ω) = F (s)s=j ω (17.66) left-hand side of the s plane, then one can obtain
the Fourier transform F(ω) from the correspond-
This equation also shows that the Fourier transform can be re- ing Laplace transform F(s) by merely replacing s
garded as a special case of the Laplace transform with s = j ω. by jω. Note that this is not the case, for example,
Recall that s = σ + j ω. Therefore, Eq. (17.66) shows that the for u(t) or cos atu(t).
Laplace transform is related to the entire s plane, whereas the
Fourier transform is restricted to the j ω axis. See Fig. 15.1.
3. The Laplace transform is applicable to a wider range of func-
tions than the Fourier transform. For example, the function
tu(t) has a Laplace transform but no Fourier transform. But
Fourier transforms exist for signals that are not physically
realizable and have no Laplace transforms.
4. The Laplace transform is better suited for the analysis of tran-
sient problems involving initial conditions, since it permits the
inclusion of the initial conditions, whereas the Fourier trans-
form does not. The Fourier transform is especially useful for
problems in the steady state.
5. The Fourier transform provides greater insight into the fre-
quency characteristics of signals than does the Laplace trans-
form.
Some of the similarities and differences can be observed by comparing
Tables 15.1 and 15.2 with Tables 17.1 and 17.2.
† 17.7 APPLICATIONS
Besides its usefulness for circuit analysis, the Fourier transform is used
extensively in a variety of fields such as optics, spectroscopy, acoustics,
computer science, and electrical engineering. In electrical engineering,
it is applied in communications systems and signal processing, where
frequency response and frequency spectra are vital. Here we consider
two simple applications: amplitude modulation (AM) and sampling.

17.7.1 Amplitude Modulation

Electromagnetic radiation or transmission of information through space
has become an indispensable part of a modern technological society.
However, transmission through space is only efficient and economical
at radio frequencies (above 20 kHz). To transmit intelligent signals—
such as for speech and music—contained in the low-frequency range
of 50 Hz to 20 kHz is expensive; it requires a huge amount of power
and large antennas. A common method of transmitting low-frequency
audio information is to transmit a high-frequency signal, called a carrier,
which is controlled in some way to correspond to the audio information.
Three characteristics (amplitude, frequency, or phase) of a carrier can
be controlled so as to allow it to carry the intelligent signal, called the
modulating signal. Here we will only consider the control of the carrier’s
amplitude. This is known as amplitude modulation.
▲
▲

| | e-Text Main Menu | Textbook Table of Contents | Problem Solving Workbook Contents