11/16/99 (T.F.

Weiss)

Lecture #18: Continuous time periodic signals Outline:

• Fourier series of periodic functions

Motivation:
• Examples of Fourier series — periodic impulse train

• Representation of continuous time, periodic signals in the • Fourier transforms of periodic functions — relation to
frequency domain Fourier series

• Periodic signals occur frequently — motion of planets and • Conclusions

their satellites, vibration of oscillators, electric power dis-
tribution, beating of the heart, vibration of vocal chords,
etc.

1 2

Fourier series coefficients

Fourier series of a periodic function The coefficients of the Fourier series can be found as follows.
 
Periodic time function   ∞
1 T /2 1 T /2 

x(t) is a periodic time function with period T . x(t)e−j2πnt/T dt = X[k]ej2πkt/T  e−j2πnt/T dt,
T −T /2 T −T /2 k=−∞
 ∞   
1 T /2
1 T /2 j2π(k−n)t/T
x(t) x(t)e−j2πnt/T dt = X[k] e dt .
T −T /2 k=−∞
T −T /2
The integral can be evaluated as follows.

0 T t 1 T /2 j2π(k−n)t/T 1 if k = n,
e dt =
T −T /2 0 if k = n.
The set of exponential time functions are said to be an orthonor-
Such a periodic function can be expanded in an infinite series of
mal basis.
exponential time functions called the Fourier series,
∞

The coefficients are
x(t) = X[n]ej2πnt/T .

n=−∞ 1 T /2
X[n] = x(t)e−j2πnt/T dt.
T −T /2
3 4
 Examples of Fourier series of periodic time functions
Definition of line spectra, harmonics Periodic impulse train
The fundamental frequency fo = 1/T . The Fourier series co-
efficients plotted as a function of n or nfo is called a Fourier The periodic impulse train is an important periodic time function
spectrum. and we derive its Fourier series coefficients.
s(t)
X[n] ∞


1 1 1 1 1 1 1 s(t) = δ(t − nT ).
n=−∞
0T t
0 1 2 3 4 n ⇐ Harmonic #
0 fo 2fo 3fo 4fo f ⇐ Frequency The Fourier series coefficients are found as follows

1 T /2
s(t)e−j2πnt/T dt,
3rd harmonic
DC

2nd harmonic
Fundamental

4th harmonic
S[n] =
T −T /2
 ∞
1 T /2

= δ(t − nT )e−j2πnt/T dt,
T −T /2 n=−∞

1 T /2 1
= δ(t)e−j2πnt/T dt = .
T −T /2 T
5 6

Periodic impulse train, cont’d

The Fourier series coefficients are
1
. S[n] = Periodic impulse train, cont’d
T
The time function and spectrum are shown below. The Fourier series of the periodic impulse train is
∞

∞
1

s(t) S[n] s(t) = δ(t − nT ) = ej2πnt/T .
n=−∞ T n=−∞
1/T
1 1 1 1 1 1 1 It is not obvious that the two expressions are equal. To investi-
gate this, we define the partial sum of the Fourier series, sN (t),
0T t 0123 n
1
N
sN (t) = ej2πnt/T ,
To summarize, the periodic impulse train can be represented by T n=−N
its Fourier series, and investigate its behavior as N → ∞.
∞

∞
1

s(t) = δ(t − nT ) = ej2πnt/T .
n=−∞ T n=−∞

7 8
 Periodic impulse train, cont’d
Periodic impulse train, cont’d
The partial sum of the Fourier series is  
sN (t) 1 sin(2N + 1)πt/T
sN (t) = .
1
N
1
N
n 4 N =2 T sin πt/T
sN (t) = ej2πnt/T = ej2πt/T . T =1
T n=−N T n=−N 2 Note that this function is pe-
We can use the summation formula for a finite geometric series riodic with period T , and
−2 −1 1 2 t
(Lecture 10) to sum this series, −2 2N + 1
10 sN (nT ) = .

−N
N +1 N =5 T
1 e
j2πt/T − ej2πt/T 5
T =1 The first zero of sN (t) is at
sN (t) = ,
T 1 − ej2πt/T T
1 ej(2N +1)πt/T − e−j(2N +1)πt/T t=
2N + 1
.
sN (t) = ,
T ejπt/T − e−jπt/T
 20
N = 10 Thus, as N → ∞, each lobe
1 sin(2N + 1)πt/T T =1 gets larger and narrower. To
sN (t) = . 10
T sin πt/T determine if each lobe acts
as an impulse, we need to
find its area.
9 10

Periodic impulse train, cont’d Fourier transform of a periodic function

Fourier transform of a periodic impulse train
We have two expressions for a periodic impulse train,
The area of each period of sN (t) is simply
∞

∞
 T /2  T /2
1

1

N s(t) = ej2πnt/T .
δ(t − nT ) =
Area = sN (t) dt = ej2πnt/T dt n=−∞ T n=−∞
−T /2 −T /2 T n=−N
The Fourier transform of each expression is


N 
1 T /2 j2πnt/T ∞

∞ 
= e dt 1
n
S(f ) = e−j2πnT f =δ f− .
n=−N T −T /2 T n=−∞ T
n=−∞
The integral is zero except when n = 0 when it equals T so that Therefore, the Fourier transform of a periodic impulse train in
Area = 1. Thus, each lobe of sN (t) becomes: tall, height is time is a periodic impulse train in frequency.
sN (nT ) = (2N + 1)/T ; narrow, width is 2T /(2N + 1); and its
area is 1. Thus, the partial sum approaches an infinite impulse s(t) S(f )
train of unit area, F
1 1 1 1 1 1 1 ⇐⇒ 1/T
∞

∞


1
s(t) = δ(t − nT ) = ej2πnt/T . 0T t 1
n=−∞ T n=−∞ 0 f
T
11 12
Relation of Fourier series spectrum to Fourier transform of Fourier transform of an arbitrary periodic function
a periodic impulse train Representation of a periodic function
An arbitrary periodic function can be generated by convolving a
pulse, xT (t), that represents one period of the periodic function

s(t) S[n] S(f ) with a periodic impulse train, s(t) = n δ(t − nT ),
1/T
1 1 1 1 1 1 1 ∞

∞


1/T x(t) = xT (t) ∗ s(t) = xT (t) ∗ δ(t − nT ) = xT (t − nT ).
0T t 0123 n 1 n=−∞ n=−∞
0 f
T
xT (t) s(t) x(t)

Therefore, the Fourier transform of the periodic impulse train has ∗ 1 1 1 1 1 =

an impulse at the frequency of each Fourier series component
0 t −T 0 T t −T 0 T t
and the area of the impulse equals the Fourier series coefficient.

13 14

Fourier transform of a periodic function

The Fourier transform of the periodic function is
F
x(t) = xT (t) ∗ s(t) ⇐⇒ X (f ) = XT (f ) × S(f ).

 
∞  ∞XT Tn 
1
n 


δ f −
n
X (f ) = XT (f ) × δ f− = .
T n=−∞ T n=−∞ T T Fourier transform of a periodic function, cont’d
An important conclusion is that the Fourier transform of a pe-
xT (t) s(t) x(t) riodic function consists of impulses in frequency at multiples
of the fundamental frequency. Thus, periodic continuous time
∗ 1 1 1 1 1 = functions can be represented by a countably infinite number of
0 t −T 0 T t −T 0 T t complex exponentials.
XT (f ) S(f ) X (f )
1
× =
T
−1 1
f 0 f f
T T
15 16
Fourier series coefficients
The Fourier transform of the periodic function is Fourier series of a square wave — generation of the square


∞  wave

XT Tn n
X (f ) =  δ f − . We will find the Fourier series of a square wave by finding the
n=−∞ T T
Fourier transform of one period.
Recall that
 T /2
xT (t) x(t)
XT (f ) = xT (t)e−j2πf t dt 1 1
−T /2
Therefore, 0

 −T T −T 0 T t
 0 t
XT Tn 1 T /2 4 4
= xT (t)e−j2πnt/T dt = X[n].
T T −T /2
Therefore, for an arbitrary periodic continuous time function, the
x(t) = xT (t) ∗ s(t)
Fourier transform consists of impulses (located at the harmonic
frequencies) whose areas are the Fourier series coefficients.

17 18

Fourier series of a square wave — Fourier transform of

square wave
Fourier series of a square wave — Fourier transform of one
To obtain the Fourier transform of the square wave, we take the
period of the square wave
Fourier transform of one period of the square wave and multiply
it by the Fourier transform of the periodic impulse train.
F
xT (t) ⇐⇒ XT (f )  ∞ 
sin(πf T /2)
T 1
n
X (f ) = XT (f ) × S(f ) = × δ f− ,
  2 πf T /2 T n=−∞ T
T sin(πf T /2) ∞   
XT (f ) =
1 sin(nπ/2) n
xT (t) 2 πf T /2 = δ f− .
n=−∞ 2 nπ/2 T
1 T /2

XT (f )
−T T t X (f )
−2 2 f
4 4
T T
T /2 −9 −7 −5 −3 −1 1 3 5 7 9
0 f
T T T T T T T T T T
19 20
 Two-minute miniquiz solution

Problem 18-1 — Fourier series of the square wave

Two-minute miniquiz problem

a) Since the Fourier transform of the square wave is

Problem 18-1 — Fourier series of the square wave   
∞


1 sin(nπ/2) n
X (f ) = δ f− ,
n=−∞ 2 nπ/2 T
the Fourier series coefficients are
a) Determine the Fourier series coefficients of the square wave.  
1 sin(nπ/2)
X[n] =
2 nπ/2
b) From the Fourier series coefficients determine the average
value of the square wave. x(t)
 1
1 T /2 1/2
b) X[0] = x(t) dt = 1/2.
T −T /2
−T 0 T t

20-1 20-2

Fourier series of a square wave — synthesis of a square

Fourier series of a square wave — synthesis of a square
wave, cont’d
wave
To investigate the synthesis of the square wave, we consider the
We can synthesize the square wave by adding complex exponen-
partial sum of the Fourier series,
tials weighted by their Fourier series coefficients,
 
∞ ∞   1

N
sin(nπ/2)


1 sin(nπ/2) j2πnt/T xN (t) = + cos(2πnt/T ).
x(t) = X[n]ej2πnt/T = e . 2 nπ/2
n=−∞ n=−∞ 2 nπ/2 n=1

Note that the coefficients are even functions of n. Hence, we

N=0 N=1 N=3 N=5
can rewrite the series as follows
∞  
1
1 sin(nπ/2)
j2πnt/T 
x(t) = + e + e−j2πnt/T ,
2 n=1 2 nπ/2
∞  
1
sin(nπ/2) N=7 N=9 N=11 N=13
x(t) = + cos(2πnt/T ). 1
2 n=1 nπ/2
0.5
Note that all the even harmonics of x(t) are zero except for the
term for n = 0. −1 −0.5 0.5 1
t/T

21 22
Fourier series of a square wave — synthesis of a square
wave, cont’d
Demo illustrating the Gibbs’ phenomenon.
Fourier series of a square wave — Gibbs’ phenomenon
As harmonics are added to synthesize the square wave, the partial
sum of the Fourier series converges to the square wave every-
where except near the discontinuity where the partial sum takes
on the value of 1/2. There are oscillations on either side of
the discontinuity whose maximum over and undershoot approach
9% of the discontinuity independent of N . We can interpret the
Gibbs’ phenomenon by examining the partial sum of the Fourier
series as a filtering problem.

23 24

Gibbs’ phenomenon interpretation as an ideal LPF of a Gibbs’ phenomenon — step response of an ideal LPF
square wave The Gibbs’ phenomenon can be investigated further by examin-
Truncation of the Fourier series of a square wave can be in- ing the step response of an ideal lowpass.
terpreted in the frequency domain as passing the square wave
through an ideal lowpass filter that truncates the spectrum. That H(f )
can be interpreted in the time domain as the convolution of the x(t) xw (t)
H(f )
square wave with a sinc function. The response of the ideal low-
−W W f
pass filter to the discontinuity of the square wave gives rise to
x(t) h(t) xw (t)
the oscillations.
∗ =
X (f ) H(f ) XN (f )
× = t t t

f f f
Hence,
x(t) h(t) xN (t)  
 t
sin(2πW t)
∗ = xw (t) = u(t) ∗ h(t) = h(τ ) dτ where h(t) = 2W .
−∞ 2πW t
t t t
25 26
Gibbs’ phenomenon — step response of an ideal LPF, Gibbs’ phenomenon — step response of an ideal LPF,
cont’d cont’d
We evaluate the integral The sine integral function,
   t
 t sin y
sin(2πW τ ) Si(t) = dy,
xw (t) = 2W dτ, 0 y
−∞ 2πW τ
is plotted below.
by changing variables y = 2πW t which yields
 2πW t  
sin y Si(t)
xw (t) = dy. 1.5 π/2
−∞ y
1 sin t
This is closely related to a tabulated function called the sine 0.5
t
integral function Si(t) defined as t
−20 −10 10 20
 t -0.5
sin y -1
Si(t) = dy.
0 y −π/2 -1.5

27 28

Conclusions
Gibbs’ phenomenon — step response of an ideal LPF, Time Function Fourier Transform Fourier Series
X (f )
cont’d x(t)
We express the step with the truncated spectrum in terms of the
sine integral function as t f
1 1 x(t) X (f ) X[n]
xw (t) = + Si(2πW t),
2 π
which is shown plotted below for several values of W . t f n
1 1
W=1 W=2 F
0.5 0.5 • Aperiodic, continuous functions of time ⇐⇒ aperiodic, con-
tinuous functions of frequency.
1 1
0.5 W=3 0.5 W=4 F
• Periodic, continuous functions of time ⇐⇒ impulses whose
−3−2−1 1 2 3 −3−2−1 1 2 3 areas are the Fourier series coefficients and located at dis-
crete frequencies.

29 30
 Joseph Fourier, continued
Historical perspective
Jean Baptiste Joseph Fourier (1768-1830)
• Born of humble origins — his father was the town tailor, he
was orphaned at age 9, and raised by a neighbor.

• Went to military school and discovered mathematics at age

Joseph Fourier was born in Aux-
erre, France on March 21, 1768
and died in Paris on May 4,
1830. This is one of two por-
traits that has survived.
13.
• Taught mathematics, rhetoric, philosophy, and history in a
Benedictine school in his home town.
• Became active in local politics and social causes. His political
honesty got him in hot water — he was arrested first by the
Robespierre regime and later by the enemy post-Robespierre
regime. He was just barely saved from the guillotine twice.

31 32

Joseph Fourier, continued

Joseph Fourier, continued

• Taught at the Ecole Polytechnique where he developed an

excellent reputation as a lecturer and began publishing math-
ematical research. He taught with Lagrange and Monge —
two eminent mathematicians.
• When Napoleon Bonaparte was put in charge of a French ex-
pedition to Egypt, Fourier was chosen as a scientific member.
He held political and administrative positions in Egypt and
proved to be an able administrator. When the French occu-
pation of Egypt ended, Fourier returned to teaching mathe-
matics in France.
• In 1802, Napoleon appointed Fourier as the prefect of Isère, a
French department whose center was Grenoble. This was an
important administrative/political position somewhat akin to
the governor of a US state. Fourier always hoped to complete
his tasks and return to a scholarly life, but he continued to
take on important administrative posts throughout his life.
His mathematical and scientific studies were largely part-time
efforts.

33 34
Joseph Fourier, continued
• In 1807, Fourier submitted a paper on the use of trigonomet-
ric series to solve problems in heat conduction to the Institute
Joseph Fourier, continued
Fourier’s work and that of those that followed him have had a
profound effect on science and mathematics. Fourier transforms
are common tools in many different fields of science.
12

of France. It was reviewed by four famous French mathe- 50

10
100
maticians — Lagrange, Laplace, Lacroix, and Monge. Three 150 8

of the four referees voted to accept the paper, Lagrange was 200
6
250

opposed. Lagrange had been on one side of a raging con- 300 4

troversy on the representation of functions by trigonomet- 350

400
2

ric series. Lagrange did not believe that arbitrary functions 450 0

500
could be expanded in trigonometric series as Fourier’s paper 550
-2

claimed. Fourier’s original paper was never published, but in 100 200 300 400 500

1822 he published his work in a book The Analytical Theory Fourier Fourier transformed
of Heat in 1822. Although he did not originate trigonometric
series, nor did he determine the precise conditions for their
The image of Fourier (left) is 582 by 582 pixels. The logarithm
validity, he did use them to solve problems in heat conduction
of the magnitude of the Fourier transform of this image is plotted
thus illustrating their utility.
on the right.
35 36