Documenti di Didattica
Documenti di Professioni
Documenti di Cultura
Arthur L. Schoenstadt
Department of Applied Mathematics
Naval Postgraduate School
Code MA/Zh
Monterey, California 93943
March 9, 2011
c 1992 - Professor Arthur L. Schoenstadt
1
Contents
1 Infinite Sequences, Infinite Series and Improper Integrals 1
1.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2 Functions and Sequences . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.3 Limits . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.4 The Order Notation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.5 Innite Series . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.6 Convergence Tests . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.7 Error Estimates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.8 Sequences of Functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
2 Fourier Series 11
2.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
2.2 Derivation of the Fourier Series Coecients . . . . . . . . . . . . . . . . . . 12
2.3 Odd and Even Functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
2.4 Convergence Properties of Fourier Series . . . . . . . . . . . . . . . . . . . . 29
2.5 Interpretation of the Fourier Coecients . . . . . . . . . . . . . . . . . . . . 31
2.6 The Complex Form of the Fourier Series . . . . . . . . . . . . . . . . . . . . 37
2.7 Fourier Series and Ordinary Dierential Equations . . . . . . . . . . . . . . . 42
2.8 Fourier Series and Digital Data Transmission . . . . . . . . . . . . . . . . . . 45
i
4.7 Frequency Domain Analysis of the Circular Drum . . . . . . . . . . . . . . . 89
4.8 Time Domain Analysis of the Circular Membrane . . . . . . . . . . . . . . . 90
ii
List of Figures
1
1 Plot of f (x) = 1x along with the partial sums S1 , S5 , S10 , S20 . . . . . . . . 8
2 Graph of f (x) and the N th partial sums for N = 2, 5, 10, 20 . . . . . . . . . . 15
3 Graph of f (x) and the N th partial sums for N = 2, 5, 10, 20 . . . . . . . . . . 16
4 Graph of f (x) and the N th partial sums for N = 2, 5, 10, 20 . . . . . . . . . . 19
5 Graph of f (x) for problem 4a . . . . . . . . . . . . . . . . . . . . . . . . . . 21
6 Graph of f (x) for problem 4b . . . . . . . . . . . . . . . . . . . . . . . . . . 22
7 Graph of f (x) for problem 4c . . . . . . . . . . . . . . . . . . . . . . . . . . 23
8 Graph of f (x) for problem 1a of 2.3 . . . . . . . . . . . . . . . . . . . . . . . 24
9 Graph of f (x) for problem 1b of 2.3 . . . . . . . . . . . . . . . . . . . . . . . 25
10 Graph of f (x) for problem 1c of 2.3 . . . . . . . . . . . . . . . . . . . . . . . 26
11 Graph of f (x) for problem 1d of 2.3 . . . . . . . . . . . . . . . . . . . . . . . 28
12 Graph of spectrum for problem 1a of 2.5 . . . . . . . . . . . . . . . . . . . . 32
13 Graph of spectrum for problem 1b of 2.5 . . . . . . . . . . . . . . . . . . . . 33
14 Graph of spectrum for problem 1c of 2.5 . . . . . . . . . . . . . . . . . . . . 34
15 Graph of spectrum for problem 1d of 2.5 . . . . . . . . . . . . . . . . . . . . 35
16 Graph of spectrum for problem 1a of 2.6 . . . . . . . . . . . . . . . . . . . . 40
17 Graph of spectrum for problem 1b of 2.6 . . . . . . . . . . . . . . . . . . . . 40
18 Graph of S10 for problem 1 of 2.7 . . . . . . . . . . . . . . . . . . . . . . . . 44
19 Graph of u(x, t) for problem 1 of 3.6 for t = 1, 2, 3, 4 . . . . . . . . . . . . . . 57
20 The rst three modes for problem 1 of 3.11 . . . . . . . . . . . . . . . . . . . 79
21 The rst three frequencies for problem 1 of 3.11 . . . . . . . . . . . . . . . . 79
22 Graph of u(x, 0) for problem 1a of 13.3 . . . . . . . . . . . . . . . . . . . . . 82
23 Graph of u(x, 1) for problem 1a of 13.3 . . . . . . . . . . . . . . . . . . . . . 83
24 Graph of u(x, 2.5) for problem 1a of 13.3 . . . . . . . . . . . . . . . . . . . . 84
25 Graph of u(x, 4) for problem 1a of 13.3 . . . . . . . . . . . . . . . . . . . . . 84
26 Graph of u(x, 0) for problem 1b of 13.3 . . . . . . . . . . . . . . . . . . . . . 85
27 Graph of u(x, 1) for problem 1b of 13.3 . . . . . . . . . . . . . . . . . . . . . 85
28 Graph of u(x, 2.5) for problem 1b of 13.3 . . . . . . . . . . . . . . . . . . . . 86
29 Graph of u(x, 4) for problem 1b of 13.3 . . . . . . . . . . . . . . . . . . . . . 86
30 Graph of u(x, 0) for problem 1c of 13.3 . . . . . . . . . . . . . . . . . . . . . 87
31 Graph of u(x, 1) for problem 1c of 13.3 . . . . . . . . . . . . . . . . . . . . . 87
32 Graph of u(x, 2.5) for problem 1c of 13.3 . . . . . . . . . . . . . . . . . . . . 88
33 Graph of u(x, 4) for problem 1c of 13.3 . . . . . . . . . . . . . . . . . . . . . 88
34 Graph of h(t) for problem 1a of rst set of Chapter 5 for = 1 . . . . . . . . 112
35 Graph of H(f ) for problem 1a of rst set of Chapter 5. The left plot for = 1
and the right for = 0.05 . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112
36 Graph of h(t) for problem 1b of rst set of Chapter 5 . . . . . . . . . . . . . 113
37 Graph of (H(f )) (on the left) and (H(f )) (on the right) for problem 1b of
rst set of Chapter 5 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113
38 Graph of |H(f )| (on the left) and (f ) (on the right) for problem 1b of rst
set of Chapter 5 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114
39 Graph of h(t) for problem 1c of rst set of Chapter 5 . . . . . . . . . . . . . 115
iii
40 Graph of (H(f )) (on the left) and (H(f )) (on the right) for problem 1c of
rst set of Chapter 5 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116
41 Graph of |H(f )| (on the left) and (f ) (on the right) for problem 1c of rst
set of Chapter 5 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116
42 Graph of h(t) for problem 1d of rst set of Chapter 5 . . . . . . . . . . . . . 117
43 Graph of |H(f )| (on the left) and (f ) (on the right) for problem 1d of rst
set of Chapter 5 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 118
44 Graph of h(t) for problem 1e of rst set of Chapter 5 . . . . . . . . . . . . . 119
45 Graph of (H(f )) (on the left) and (H(f )) (on the right) for problem 1e of
rst set of Chapter 5 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 120
46 Graph of |H(f )| (on the left) and (f ) (on the right) for problem 1e of rst
set of Chapter 5 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 120
47 Graph of h(t) for problem 1f of rst set of Chapter 5, using A = 2, = 1 . . 121
48 Graph of H(f ) for problem 1e of rst set of Chapter 5, using A = 2, = 1 . 121
49 Graph of H(f ) for problem 1e of rst set of Chapter 5, using A = 2, = .05 122
50 Graph of h(t) for problem 1g of rst set of Chapter 5 . . . . . . . . . . . . . 123
51 Graph of (H(f )) (on the left) and (H(f )) (on the right) for problem 1g of
rst set of Chapter 5 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 124
52 Graph of |H(f )| (on the left) and (f ) (on the right) for problem 1g of rst
set of Chapter 5 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 124
53 Graph of h(t) with A = = 1 for problem 1h of rst set of Chapter 5 . . . 125
54 Graph of (H(f )) with A = = 1 for problem 1h of rst set of Chapter 5 . 126
55 Graph of |H(f )| (on the left) and (f ) (on the right) with A = = 1 for
problem 1h of rst set of Chapter 5 . . . . . . . . . . . . . . . . . . . . . . . 126
56 Graph of h(t) for problem 1i of rst set of Chapter 5 . . . . . . . . . . . . . 127
57 Graph of (H(f )) for problem 1i of rst set of Chapter 5 . . . . . . . . . . 127
58 Graph of |H(f )| (on the left) and (f ) (on the right) for problem 1i of rst
set of Chapter 5 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128
59 Graph of H(f ) for problem 2a of rst set of Chapter 5 . . . . . . . . . . . . 129
60 Graph of |H(f )| (on the left) and h(t) (on the right) for problem 2a of rst
set of Chapter 5 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 130
61 Graph of H(f ) for problem 2b of rst set of Chapter 5 . . . . . . . . . . . . 131
62 Graph of h(t) for problem 2b of rst set of Chapter 5 . . . . . . . . . . . . . 131
63 Graph of h(t) for problem 1a of second set of Chapter 5 . . . . . . . . . . . 133
64 Graph of g(t) for problem 1a of second set of Chapter 5 . . . . . . . . . . . 133
65 Graph of Re(H(f )) (on the left) and Im(H(f )) (on the right) for problem 1a
of second set of Chapter 5 . . . . . . . . . . . . . . . . . . . . . . . . . . . . 134
66 Graph of |H(f )| (on the left) and (f ) (on the right) for problem 1a of second
set of Chapter 5 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 134
67 Graph of h(t) for problem 1b of second set of Chapter 5 . . . . . . . . . . . 135
68 Graph of g(t) for problem 1b of second set of Chapter 5 . . . . . . . . . . . 135
69 Graph of Re(H(f )) (on the left) and Im(H(f )) (on the right) for problem 1b
of second set of Chapter 5 . . . . . . . . . . . . . . . . . . . . . . . . . . . . 136
iv
70 Graph of |H(f )| (on the left) and (f ) (on the right) for problem 1b of second
set of Chapter 5 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 136
71 Graph of h(t) for problem 1c of second set of Chapter 5 . . . . . . . . . . . 137
72 Graph of g(t) for problem 1c of second set of Chapter 5 . . . . . . . . . . . 137
73 Graph of Im(H(f )) (on the left) and |H(f )| (on the right) for problem 1c of
second set of Chapter 5 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 138
74 Graph of (f ) for problem 1c of second set of Chapter 5 . . . . . . . . . . . 138
75 Graph of h(t) for problem 1d of second set of Chapter 5 . . . . . . . . . . . 139
76 Graph of g(t 3) for problem 1d of second set of Chapter 5 . . . . . . . . . 139
77 Graph of Re(H(f )) (on the left) and Im(H(f )) (on the right) for problem 1d
of second set of Chapter 5 . . . . . . . . . . . . . . . . . . . . . . . . . . . . 140
78 Graph of |H(f )| (on the left) and (f ) (on the right) for problem 1d of second
set of Chapter 5 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 140
79 Graph of h(t) for problem 1e of second set of Chapter 5 . . . . . . . . . . . 141
80 Graph of g(t) for problem 1e of second set of Chapter 5 . . . . . . . . . . . 141
81 Graph of Re(H(f )) for problem 1e of second set of Chapter 5 . . . . . . . . 142
82 Graph of h(t) for problem 1f of second set of Chapter 5 . . . . . . . . . . . 143
83 Graph of Re(H(f )) for problem 1f of second set of Chapter 5 . . . . . . . . 144
84 Graph of h(t) for problem 1g of second set of Chapter 5 . . . . . . . . . . . 145
85 Graph of Re(H(f )) (on the left) and Im(H(f )) (on the right) for problem 1g
of second set of Chapter 5 . . . . . . . . . . . . . . . . . . . . . . . . . . . . 145
86 Graph of |H(f )| (on the left) and (f ) (on the right) for problem 1g of second
set of Chapter 5 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 146
87 Graph of h1 (t) (on the left) and h2 (t) (on the right) for problem 1g of second
set of Chapter 5 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 146
88 Graph of H(f ) for problem 2a of second set of Chapter 5 . . . . . . . . . . 147
89 Graph of G(f ) (on the left) and g(t) (on the right) for problem 2a of second
set of Chapter 5 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 147
90 Graph of h(t) for problem 2a of second set of Chapter 5 . . . . . . . . . . . 148
91 Graph of G(f ) (on the left) and h(t) (on the right) for problem 2b of second
set of Chapter 5 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 149
92 Graph of H(f ) for problem 2c of second set of Chapter 5 . . . . . . . . . . . 150
93 Graph of G(f ) (on the left) and h(t) (on the right) for problem 2c of second
set of Chapter 5 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151
94 Graph of h(t) (on the left) and g(t) (on the right) for problem 1a of Chapter 6.2154
95 Graph of h(t u) for problem 1a of Chapter 6.2 . . . . . . . . . . . . . . . . 154
96 Graph of g and h for problem 1a of Chapter 6.2 when t < 0 . . . . . . . . . 155
97 Graph of g and h for problem 1a of Chapter 6.2 when 0 < t < 2 . . . . . . . 155
98 Graph of g and h for problem 1a of Chapter 6.2 when 2 < t . . . . . . . . . 155
99 Graph of the convolution for problem 1a of Chapter 6.2 . . . . . . . . . . . 155
100 Graph of h(t) for problem 1b of Chapter 6.2 . . . . . . . . . . . . . . . . . . 157
101 Graph of h(t u) for problem 1b of Chapter 6.2 . . . . . . . . . . . . . . . . 157
102 Graph of h and h for problem 1b of Chapter 6.2 when t < 4 . . . . . . . . 158
103 Graph of h and h for problem 1b of Chapter 6.2 when 4 t 0 . . . . . . 158
v
104 Graph of g and h for problem 1b of Chapter 6.2 when 0 < t < 4 . . . . . . . 158
105 Graph of g and h for problem 1b of Chapter 6.2 when 4 < t . . . . . . . . . 158
106 Graph of the convolution for problem 1b of Chapter 6.2 . . . . . . . . . . . 159
107 Graph of h(t) (on the left) and g(t) (on the right) for problem 1c of Chapter 6.2160
108 Graph of h(t u) for problem 1c of Chapter 6.2 . . . . . . . . . . . . . . . . 160
109 Graph of g(u)h(t u) for problem 1c of Chapter 6.2 . . . . . . . . . . . . . . 161
110 Graph of the convolution for problem 1c of Chapter 6.2 . . . . . . . . . . . 161
111 Graph of h(t) for problem 1a of Chapter 6.10 . . . . . . . . . . . . . . . . . 165
112 Graph of |H(f )| for problem 1c of Chapter 6.10 . . . . . . . . . . . . . . . . 167
113 Graph of (f ) for problem 1c of Chapter 6.10 . . . . . . . . . . . . . . . . . 167
114 Graph of Vout (t) for problem 2a of Chapter 6.10 . . . . . . . . . . . . . . . . 168
115 Graph of |H(f )| for problem 2c of Chapter 6.10 . . . . . . . . . . . . . . . . 170
116 Graph of (f ) for problem 2c of Chapter 6.10 . . . . . . . . . . . . . . . . . 170
vi
1 Infinite Sequences, Infinite Series and Improper In-
tegrals
1.1 Introduction
1.2 Functions and Sequences
1.3 Limits
1.4 The Order Notation
1.5 Infinite Series
1.6 Convergence Tests
1.7 Error Estimates
1
1.8 Sequences of Functions
PROBLEMS
1. For each of the following sequences, determine if the sequence converges or diverges. If
the sequence converges, determine the limit
2n+1 (n + 1)2 sin(n)
a. an = n+2 b. an = 2 c. an =
3 5n + 2n + 1 n+1
cos(n) en n sin(n)
g. an = h. a n = i. an =
n2 + 1 n! n+1
2. Determine the order (big Oh) of the following sequences
n3 + 2n2 + 1000 cos(n)
a. an = 7 6
b. an = 2
n + 600n + n n +1
[ ]
n n 1
c. an = 2 2 sin((n + ))
n 1 n +1 2
3 n 2
10n e + n
d. an = 2
cos(n2 )
(2n + 1)
3. Consider the innite series
(n + 1)2 2n
n=0 (2n)!
a. Compute, explicitely, the partial sums S3 and S6
b. Write the equivalent series obtained by replacing n by k 2, i.e. by shifting the index.
4. Determine whether each of the following innite series diverges or converges:
n2 + 1 n2 cos(n)
a. en b. 3
c. 3 2
n=0 n=0 (n + 1) n=0 (n + 1)
n en 1
d. e. cos(n) f.
n=0 n + 3 n=0 n! n=2 n ln(n)
5. Determine an (approximate) upper bound to the error when each of the following innite
series is approximated by a twenty-term partial sum (S20 ).
2n + 1 1 (2n + 1)2
a. 4
b. 5
c.
n=0 3n + n + 1 n=1 n n=1 n4
6. Consider the series:
xn
n=0
a. plot the partial sums S1 (x), S5 (x), S10 (x), and S20 (x) for 2 < x < 2.
b. What can you conclude about the convergence of the partial sums in this interval?
2
c. What, if anything, dierent can you conclude about the convergence of these partial
1 1
sums in the interval < x < .
2 2
3
1. a. ( )n
2n+1 2 2 2
an = n+2
= 0 since <1
3 9 3 3
1. b.
2 1
(n + 1)2 n2 + 2n + 1 1+ + 1+0+0 1
an = 2
= 2
= n
2
n2
1 =
5n + 2n + 1 5n + 2n + 1 5+ n
+ n2
5+0+0 5
1. c.
sin n
an = Note | sin n| 1
n+1
1
Thus |an | 0.
n+1
Thus an 0.
1. d.
an = cos n does not approach a limit
Thus the sequence diverges.
1. e.
2(n + 1)2 + en 2n2 2
an = 2
2
= since en 0
3n + 5n + 10 3n 3
1. f.
( ) ( )
n n n
an = cos cos
n+1 2 2
( )
n
Since cos = 0, 1, depending on n the sequence diverges.
2
1. g.
cos (n)
an =
n2 + 1
1
Thus |an | 0 and an 0.
n2 +1
1. h.
en en
an = =
n! n(n 1) 2 1
Note n! = n(n 1) 2 1 3| {z
3} 2 1 = 3n2 2
| {z }
n terms n2 terms
n n 2 ( )n2
e e e e
Thus an = = 0 since e < 3.
n! 23 n2 2 3
4
1. i.
n sin(n)
an = = 0 since sin(n) = 0
n+1
Thus an 0.
2. a.
n3 + 2n2 + 1000
an =
n7 + 600n6 + n
n3 1 ( )
For n large an behaves like 7 = 4 . Thus an = O n14 .
n n
Note that for n 10 the numerator is n3 + 2n3 + n3 = 4n3 and the
4n3 1
denominator n7 . Thus for n 10 we have an 7 = 4 4 .
n n
2. b.
cos(n)
an =
n2 + 1
( )
1 1 1
Since | cos(n)| = 1, we have |an | = . Thus a n = O .
n2 + 1 n2 n2
2. c. [ ]
n n
an = 2 sin((n + 1/2))
n 1
2 n +1
Note sin((n + 1/2)) = sin(n) cos(/2) + cos(n) sin(/2) = cos(n)
| {z } | {z }
=0 =1
n3 + n (n3 n) 2n
The rst factor in brackets is = 4 .
( ) n 1 n 1
4
2n 1
So an = 4 cos(n) = O 3 since | cos(n)| = 1.
n 1 n
2. d.
10n3 en + n2
an = cos(n2 )
(2n + 1)2
Note cos(n2 ) = 1 depending on n even or odd. Also x3 ex 0 by
LHospital rule, so the rst term in numerator is negligible compared to the second term
there.
n2 1
So |an | 2
= . Therefore an = O(1).
(2n) 4
5
3.
(n + 1)2 2n 4 2 9 4 16 8 25 16 36 32
= 1+ + + + + + ...
n=0 (2n)! 2! 4! 6! 8! 10!
3
(n + 1)2 2n 4 2 9 4 16 8
a. S3 = = 1+ + + = 6.678
n=0 (2n)! 2! 4! 6!
6
(n + 1)2 2n 4 2 9 4 16 8 25 16 36 32 49 64
S6 = = 1+ + + + + + = 6.688
n=0 (2n)! 2! 4! 6! 8! 10! 12!
b. Let k = n + 2 (or n = k 2)
(n + 1)2 2n (k 1)2 2k2 1 (k 1)2 2k
= =
n=0 (2n)! k2=0 (2k 4)! 4 k=2 (2k 4)!
(
)n
4. a. en = e1 converges as a geometric series with r = e1 < 1.
n=0 n=0
1 e
The sum is en = =
n=0 1e 1 e1
2
n +1
b. 3
. Note that for large n an is of the order of 1/n.
n=0 (n + 1)
1
Now use the comparison test, comparing our series to .
n=0 n
Since the latter diverges, our series diverges.
2
n cos n n2
c. 3 2
. Note that for large n, |an | behaves like n6
.
n=0 (n + 1)
1
Now use the comparison test, comparing our series to .
n4 n=0
Since the latter converges (p test with p = 4), our series converges.
n
d. . Note that for large n, an behaves like 1 = 0.
n=0 n + 3
Since limn an = 0 the series diverges.
6
en
4. e. cos(n). See problem 1.h.
n!
n=0
Let us estimate n!, by takeing n! = 1 2 3 n > 1 2 3| {z
3} . Therefore
n2times
( )n2
en e2 e e2 n2
|an | = = r
n! 2 3 2
e e2 n2 e2
where |r| = < 1. Thus r = 2
rn converges as a geometric series.
3 n=0 2 2r n=0
Thus our series coverges by comparison test.
1 1
f. . Note that an > 0. Taking f (x) = we can use the integral test.
n=2 n ln(n) x ln(x)
dx du
= , where we used u = ln(x).
2 x ln(x) ln(2) u
Thus the anti derivative is ln(ln(x)) which tend to innity at the upper limit.
Therefore the series diverges.
2n + 1 2n + 1
5. a. , an = 4
3n4 +n+1 3n + n + 1
n=0 ( )
2n 1 1
For large n, |an | 4 3 . Thus an = O 3 .
3n n n
2n + 1
E20 = S S20 =
n=21 3n4+n+1
2n + 1 1 1
|E20 |
n=21
4
3n + n + 1 n=21 n
3 2 202
1
Note the above is for N = 20, p = 3. Thus
(p 1)N p1
1
|E20 | = .00125
800
( )
1 1 1
b. 5
, an = 5 = O 5 .
n=1 n n n
So
1 dx 1
|EN | = 5
=
n=N +1 n N x5 4N 4
1
|E20 | = 1.56 106
4 20 4
7
(2n + 1)2 (2n + 1)2 4
c. , an = 2 . So
n=1 n4 n4 n
4 4
|EN | 2
n=N +1 n N
4
|E20 | = .2
20
6. Consider
xn = 1 + x + x2 + x3 +
n=0
1x
N +1
N
, x = 1
SN = xn = 1 + x + x2 + x3 + + xN = 1x
N + 1,
n=0 x=1
a. S1 = 1 + x
1 x6
S5 = 1 + x + x2 + x3 + x4 + x5 =
1x
1 x11
S10 = 1 + x + x + x + + x
2 3 10
=
1x
1 x21
S20 = 1 + x + x2 + x3 + + x20 =
1x
In Table 1 we list the sum for various values of 1 < x < 1. The graphs of f (x) = 1
1x
along
with the partial sums S1 , S5 , S10 , S20 is given in Figure 1.
4
y
3
0
-2 -1 0 1 2
-1 x
-2
S1
S5
S10
S20
1
Figure 1: Plot of f (x) = 1x
along with the partial sums S1 , S5 , S10 , S20
8
x S S1 S5 S10 S20
-1.00 0.50 0.00 0.00 1.00 1.00
-0.96 0.51 0.04 0.11 0.84 0.73
-0.92 0.52 0.08 0.21 0.73 0.61
-0.88 0.53 0.12 0.28 0.66 0.57
-0.84 0.54 0.16 0.35 0.62 0.56
-0.80 0.56 0.20 0.41 0.60 0.56
-0.76 0.57 0.24 0.46 0.60 0.57
-0.72 0.58 0.28 0.50 0.60 0.58
-0.68 0.60 0.32 0.54 0.60 0.60
-0.64 0.61 0.36 0.57 0.61 0.61
-0.60 0.63 0.40 0.60 0.63 0.63
-0.56 0.64 0.44 0.62 0.64 0.64
-0.52 0.66 0.48 0.64 0.66 0.66
-0.48 0.68 0.52 0.67 0.68 0.68
-0.44 0.69 0.56 0.69 0.69 0.69
-0.40 0.71 0.60 0.71 0.71 0.71
-0.36 0.74 0.64 0.73 0.74 0.74
-0.32 0.76 0.68 0.76 0.76 0.76
-0.28 0.78 0.72 0.78 0.78 0.78
-0.24 0.81 0.76 0.81 0.81 0.81
-0.20 0.83 0.80 0.83 0.83 0.83
-0.16 0.86 0.84 0.86 0.86 0.86
-0.12 0.89 0.88 0.89 0.89 0.89
-0.08 0.93 0.92 0.93 0.93 0.93
-0.04 0.96 0.96 0.96 0.96 0.96
0.00 1.00 1.00 1.00 1.00 1.00
0.04 1.04 1.04 1.04 1.04 1.04
0.08 1.09 1.08 1.09 1.09 1.09
0.12 1.14 1.12 1.14 1.14 1.14
0.16 1.19 1.16 1.19 1.19 1.19
0.20 1.25 1.20 1.25 1.25 1.25
0.24 1.32 1.24 1.32 1.32 1.32
0.28 1.39 1.28 1.39 1.39 1.39
0.32 1.47 1.32 1.47 1.47 1.47
0.36 1.56 1.36 1.56 1.56 1.56
0.40 1.67 1.40 1.66 1.67 1.67
0.44 1.79 1.44 1.77 1.79 1.79
0.48 1.92 1.48 1.90 1.92 1.92
0.52 2.08 1.52 2.04 2.08 2.08
0.56 2.27 1.56 2.20 2.27 2.27
0.60 2.50 1.60 2.38 2.49 2.50
0.64 2.78 1.64 2.59 2.76 2.78
0.68 3.13 1.68 2.82 3.08 3.12
0.72 3.57 1.72 3.07 3.48 3.57
0.76 4.17 1.76 9 3.36 3.96 4.15
0.80 5.00 1.80 3.69 4.57 4.95
0.84 6.25 1.84 4.05 5.33 6.09
0.88 8.33 1.88 4.46 6.29 7.76
0.92 12.50 1.92 4.92 7.50 10.33
0.96 25.00 1.96 5.43 9.04 14.39
1.00 2.00 6.00 11.00 21.00
b. The series only converges for 1 < x < 1.
The series diverges for 2 < x 1 or for 1 x < 2.
1 1
c. Series converges uniformly to a continuous function for < x < .
2 2
10
2 Fourier Series
2.1 Introduction
11
2.2 Derivation of the Fourier Series Coefficients
PROBLEMS
2. For each of the following functions, nd the Fourier coecients, the Fourier series, and
sketch the partial sums S2 (x), S5 (x), and S10 (x):
{
0 , 1 < x < 0
a. f (x) =
1 , 0x<1
f (x + 2) = f (x)
{
3 + x , 3 < x < 0
b. f (x) = f (x + 6) = f (x)
3x , 0x<3
0 , 2 < x < 0
c. f (x) = x , 0x<1
2x , 1x<2
f (x + 4) = f (x)
d. f (x) = 1 cos(x), 1 x 1
12
where the formula for a0 is no longer the n = 0 special case of the formula for an .
b. Show that the alternative Fourier Series representation
{ ( ) ( )}
a0 2nt 2nt
f (t) = + an cos + bn sin
2 n=1 T0 T0
(note the independent variable here is t), where L has been replaced by T0 /2, leads to
( )
2 T0 /2 2nt
an = f (x) cos dt , n0
T0 T0 /2 T0
T0 /2 ( )
2 2nt
bn = f (x) sin dt , n>0 .
T0 T0 /2 T0
Note that here the formula for a0 is again the n = 0 special case of the formula for an .
4. In each of the following, nd each point in L < x < L where f (x) has a discontinuity.
Find the left and right-hand limits of f (x) and f (x) at each point of discontinuity and at
the end points of the interval. Without computing the Fourier coecients, indicate to what
values the series should converge at these points.
x
2
, 1x3
a. f (x) = 0 , 2 x < 1
2x , 3 x < 2
{
3 , /2 x
b. f (x) =
2x 2 , x < /2
x2 , 2 x < 0
c. f (x) = 0 , 0x<1
4(x 1) , 1 x 2
13
1. { ( ) ( )}
a0 nx nx
f (x) = + an cos + bn sin
2 n=1 L L
( )
kx
Let k denote a xed integer. Multiply both sides by sin and integrate
L
( )
( )
( ) L
( )
L kx
L kx nx kx
L f (x) sin dx = a0 sin dx + n=1 an cos sin dx
L
L L
L L L
| {z } | {z }
=0 =0 for all n,k
( )
L ( )
nx kx
+ bn sin sin dx
L L L
|
{z }
=0 for all n=k, =L for n=k
or ( )
L kx
f (x) sin dx = Lbk
L L
( )
1L kx
bk = f (x) sin dx
L L L
Replacing k by n on both sides
( )
1L nx
bn = f (x) sin dx
L L L
{
0 , 1 < x < 0
2. a.f (x) =
1 , 0x<1
f (x + 2) = f (x)
L=1
a0
f (x) = + {an cos (nx) + bn sin (nx)}
2 n=1
1 1
an = f (x) cos(nx)dx
1 1
1 1
bn = f (x) sin(nx)dx
1 1
1
1 1 a0 1
a0 = f (x)dx = dx = x = 1 =
1 0 0 2 2
1 1 1 1
an = f (x) cos(nx)dx = cos(nx)dx = sin(nx) = 0
1 0 n 0
14
1.2 1.2
1 1
0.8 0.8
0.6 0.6
0.4 0.4
0.2 0.2
0 0
0.2 0.2
1 0.5 0 0.5 1 1 0.5 0 0.5 1
1.2 1.2
1 1
0.8 0.8
0.6 0.6
0.4 0.4
0.2 0.2
0 0
0.2 0.2
1 0.5 0 0.5 1 1 0.5 0 0.5 1
1 1 1 1 1 cos(n)
bn = f (x) sin(nx)dx = sin(nx)dx = cos(nx) =
1 0 n 0 n
Thus
1 1 cos(n)
f (x) = + sin(nx)
2 n=1 n
or
1 2) 2) 2)
f (x) = + sin(x) + sin(3x) + sin(5x) +
2 3 5
{
3 + x , 3 < x < 0
2. b. f (x) =
3x , 0x<3
f (x + 6) = f (x)
L=3
{ ( ) ( )}
a0 nx nx
f (x) = + an cos + bn sin
2 n=1 3 3
where
1 3 nx
an = f (x) cos( )dx
3 3 3
1 3 nx
bn = f (x) sin( )dx
3 3 3
[ 0 ]
1 3 1 3
a0 = f (x)dx = (3 + x)dx + (3 x)dx
3 3 3 3 0
15
[ ] [ ( )]
1 (3 + x)2 0 (3 x)2 3 1 9 9
a0 = = = 3
3 2 3 2 0 3 2 2
a0 3
= .
2 2
[ 0 3 ]
1 nx nx
an = (3 + x) cos( )dx + (3 x) cos( )dx
3 3 3 0 3
( )2
1
3 nx 0 3 nx 0
an = (3 + x) sin( ) + cos( )
3 n | {z3 } 3 n 3 3
=0
3 ( )2 3
3 nx 3 nx
(3 x) sin( ) cos( )
n | {z3 } 0 n 3 0
=0
[ ]2
1 3 6
an = (1 cos(n) cos(n) + 1) = (1 cos(n))
3 n n2 2
3 3
2.5 2.5
2 2
1.5 1.5
1 1
0.5 0.5
0 0
4 2 0 2 4 4 2 0 2 4
3 3
2.5 2.5
2 2
1.5 1.5
1 1
0.5 0.5
0 0
4 2 0 2 4 4 2 0 2 4
[ 0 3 ]
1 nx nx
bn = (3 + x) sin( )dx + (3 x) sin( )dx
3 3 3 0 3
16
( )2
1
3 nx 0 3 nx 0
bn = (3 + x) cos( ) + sin( )
3 n 3 3 n | {z3 } 3
=0
3 ( )2 3
3 nx 3 nx
(3 x) cos( ) cos( )
n 3 0 n | {z 3 } 0
=0
[ ( )]
1 3 3
bn = 3 cos(0) 3 cos(0) = 0
3 n n
Thus ( )
3 6 nx
f (x) = + 2 2
(1 cos(n)) cos
2 n=1 n 3
( ) ( ) ( )
3 12 x 12 3x 12 5x
f (x) = + 2 cos + 2 cos + 2
cos +
2 3 9 3 25 3
0 , 2 < x < 0
c. f (x) = x , 0x<1
2x , 1x<2
f (x + 4) = f (x)
L=2
{
( ) ( )}
a0 nx nx
f (x) = + an cos + bn sin
2 n=1 2 2
where
1 2 nx
an =
f (x) cos( )dx
2 2 2
1 2 nx
bn = f (x) sin( )dx
2 2 2
[ 1 ]
1 2 1 2
a0 = f (x)dx = xdx + (2 x)dx
2 2 2 0 1
[ ] [ ]
1 x2 1 (2 x)2 2 1 1 1 1
a0 = = + =
2 2 0 2 1 2 2 2 2
a0 1
=
2 4
[ 1 2 ]
1 nx nx
an = x cos( )dx + (2 x) cos( )dx
2 0 2 1 2
17
[( ( )2 )
1 2 nx 2 nx 1
an = x sin( )+ cos( )
2 n 2 n 2 0
( ( )2 ) ]
2 nx 2 nx 2
+ (2 x) sin( ) cos( )
n 2 n 2 1
[ ( )2 ( )
1 2 n 2 n
an = sin( ) + cos( )1
2 n 2 n 2
( )2 ( )]
2 n 2 n
sin( ) cos(n) cos(
n 2 n 2
Note that the terms with sin( n
2
) cancel out.
[ ]
2 n
an = 2 2 cos( ) cos(n) 1
(n) 2
and [ 2 ]
1 1 nx nx
bn = x sin( )dx + (2 x) sin( )dx
2 0 2 1 2
[( ( )2 )
1 2 nx 2 nx 1
bn = x cos( )+ sin( )
2 n 2 n 2 0
( ( )2 ) ]
2 nx 2 nx 2
(2 x) cos( ) sin( )
n 2 n 2 1
[ ( )2 ( )2 ( )]
1 2 n 2 n 2 n 2 n
bn = cos( ) + sin( ) + cos( ) + sin
2 n 2 n 2 n 2 n 2
( )2
2 n
bn = sin( )
n 2
Thus
{( )2 ( ) ( )2 ( )}
1 2 2 cos( n
2
) cos(n) 1 nx 2 n nx
f (x) = + cos + sin( ) sin
4 n=1 n 2 2 n 2 2
( )2 ( ) ( )2 ( ) ( )2 ( ) ( )2 ( )
1 2 x 2 2x 2 3x 2 6x
f (x) = + sin 2 cos sin 2 cos +
4 | {z 2 } | 2 {z 2 } | 3 {z 2 } | 6 {z 2 }
n=1 n=2 n=3 n=6
18
1.2 1.2
1 1
0.8 0.8
0.6 0.6
0.4 0.4
0.2 0.2
0 0
0.2 0.2
2 1 0 1 2 2 1 0 1 2
1.2 1.2
1 1
0.8 0.8
0.6 0.6
0.4 0.4
0.2 0.2
0 0
0.2 0.2
2 1 0 1 2 2 1 0 1 2
d. f (x) = 1 cos(x), 1 x 1
L=1
so
a0
f (x) = + {an cos(nx) + bn sin(nx)}
2 n=1
a0 = 2, a1 = 1, an = 0, for n 2
bn = 0 for all n
If you dont see this, it will emerge by brute force (recalling the orthogonality integrals.)
19
3. a. Show that the alternative Fourier Series representation
{
( ) ( )}
nx nx
f (x) = a0 + an cos + bn sin (2.2.1)
n=1 L L
Since the only change is in the rst term before the summation, we will show how a0 is
computed. Integrate both sides of (2.2.1)
L L
L L ( ) ( )
nx nx
f (x)dx = a0 dx + an cos dx +bn sin dx
L L n=1
| L {z
L } |
L
{z
L }
=0 =0
so L
f (x)dx = 2La0
L
and therefore
1 L
a0 = f (x)dx
2L L
(note the independent variable here is t), where L has been replaced by T0 /2, leads to
( )
2 T0 /2 2nt
an = f (x) cos dt , n0
T0 T0 /2 T0
( )
2 T0 /2 2nt
bn = f (x) sin dt , n>0 .
T0 T0 /2 T0
Note that (2.2.2) is identical to the Fourier series in the text, except that x is replaced here
by t and L is here T0 /2.
20
Thus
1L nt 1 T0 /2 nt
an = f (t) cos( )dt = f (t) cos( )dt
L L L T0 /2 T0 /2 T0 /2
2 T0 /2 2nt
= f (t) cos( )dt
T0 T0 /2 T0
and
1L nt 1 T0 /2 nt
bn = f (t) sin( )dx = f (t) sin( )dt
L L L T0 /2 T0 /2 T0 /2
2 T0 /2 2nt
= f (t) sin( )dt
T0 T0 /2 T0
4. In each of the following, nd each point in L < x < L where f (x) has a discontinuity.
Find the left and right-hand limits of f (x) and f (x) at each point of discontinuity and at
the end points of the interval. Without computing the Fourier coecients, indicate to what
values the series should converge at these points.
x
2
, 1x3
a. f (x) = 0 , 2 x < 1
2x , 3 x < 2
0
-3 -2 -1 0 1 2 3
-2 x
-4
-6
21
Clearly f (x) is discontinuous at x = 2 and x = 1, and the periodic extension will also
be discontinuos at x = 3.
Since f (2 ) = 4 and f (2+ ) = 0, the series converges to (4 + 0)/2 = 2 at x = 2.
Since f (1 ) = 0 and f (1+ ) = 1, the series converges to 21 at x = 1.
Since f (3 ) = 9 and f (3+ ) = 6, the series converges to 32 at x = 3.
{
3 , /2 x
b. f (x) =
2x 2 , x < /2
2
x
-3 -2 -1 0 1 2 3
0
-2
-4
-6
-8
22
x2 , 2 x < 0
c. f (x) = 0 , 0x<1
4(x 1) , 1 x 2
0
-2 -1 0 1 2
x
Clearly f (x) is continuous and so is its periodic extension. So the series converges to
f (x) everywhere.
23
2.3 Odd and Even Functions
PROBLEMS
1. Find the Fourier series for the following functions
{
1 , 2 x 0
a. f (x) =
1 , 0<x<2
f (x + 4) = f (x)
f (x) is clearly odd, and L = 2. Thus
( )
nx
f (x) = bn sin
n=1 2
where ( ) ( )
2 2 nx 2 nx 2 2
bn = 1 sin dx = cos = [1 cos(n)]
2 0 2 n 2 0 n
or ( )
2 nx
f (x) = [1 cos(n)] sin
n=1 n 2
1.5 1.5
1 1
0.5 0.5
0 0
0.5 0.5
1 1
1.5 1.5
2 1 0 1 2 2 1 0 1 2
1.5 1.5
1 1
0.5 0.5
0 0
0.5 0.5
1 1
1.5 1.5
2 1 0 1 2 2 1 0 1 2
24
b. f (x) = |x| , f (x + 2) = f (x)
f (x) is clearly even, L = 1. Thus
a0
f (x) = + an cos (nx)
2 n=1
2
an = [cos(n) 1] for n = 0
n2 2
and 1
1
a0 = 2 xdx = x2 = 1
0 0
a0 1
=
2 2
or
1 2
f (x) = + [cos(n) 1] cos (nx)
2 n=1 n2 2
1 1
0.8 0.8
0.6 0.6
0.4 0.4
0.2 0.2
0 0
1 0.5 0 0.5 1 1 0.5 0 0.5 1
1 1
0.8 0.8
0.6 0.6
0.4 0.4
0.2 0.2
0 0
1 0.5 0 0.5 1 1 0.5 0 0.5 1
25
c. f (x) = | sin(x)|
Pick L = /2. Since f (x) is even
a0
f (x) = + an cos (nx)
2 n=1
where
2 /2 4 /2
an = | sin(x)| cos (2nx) dx = sin(x) cos (2nx) dx
/2 0 0
or [ ]
4 cos(1 2n)x cos(1 + 2n)x /2
an = for n = 0
2(1 2n) 2(1 + 2n) 0
[ ]
4 cos(1 2n)/2 cos(1 + 2n)/2 1 1
an = + + for n = 0
2(1 2n) 2(1 + 2n) 2(1 2n) 2(1 + 2n)
Note that cos(1 2n)/2 = cos(/2) cos(n) + sin(/2) sin(n), also cos(1 + 2n)/2 = 0.
| {z } | {z }
=0 =0
Thus [ ]
4 1 1 4
an = + = for n = 0
2 1 2n 1 + 2n (1 4n2 )
1 1.4
1.2
0.8
1
0.6 0.8
0.4 0.6
0.4
0.2
0.2
0 0
2 1 0 1 2 2 1 0 1 2
1 1
0.8 0.8
0.6 0.6
0.4 0.4
0.2 0.2
0 0
2 1 0 1 2 2 1 0 1 2
and /2
4 /2 4 4
a0 = sin xdx = cos x =
0 0
a0 2
=
2
26
or
1 4
f (x) = + cos(2nx)
2 n=1 (1 4n2 )
2 x ,
2 x 1
d. f (x) = x , 1 x 1
2x , 1x 2
f (x + 4) = f (x)
L = 2.
Since f (x) is odd,
( )
n
f (x) = bn sin x
n=1 2
where
( )
2 2 n
bn = f (x) sin x
2 0 2
1 ( )
2 ( )
n n
= x sin x + (2 x) sin x
0 2 1 2
{ ( ) ( )2 ( )} 1
2 n 2 n
= x cos x + sin x
n 2 n 2 0
{ ( ) ( )2 ( )} 2
2 n 2 n
+ (2 x) cos x sin x
n 2 n 2 1
( ) ( )2 ( ) ( ) ( )2 ( )
2 n 2 n 2 n 2 n
= cos + sin + cos + sin
n 2 n 2 n 2 n 2
( )2 ( )
2 n
= 2 n
sin 2
Thus ( )2 ( ) ( )
2 n n
f (x) = 2 sin sin x
n=1 n 2 2
or ( ) ( ) ( )
8 8 3 8 5
f (x) = 2 sin x 2 sin x + 2
sin x
2 9 2 25 2
27
1 1
0.5 0.5
0 0
0.5 0.5
1 1
2 1 0 1 2 2 1 0 1 2
1 1
0.5 0.5
0 0
0.5 0.5
1 1
2 1 0 1 2 2 1 0 1 2
28
2.4 Convergence Properties of Fourier Series
PROBLEMS
1. For each of the following Fourier series, determine whether the series will converge uni-
formly, converge only in mean square, or diverge:
( )
1 1 nx
a. + sin
3 n=1 n3 3 3
( )
1 nx
b. 1 + cos(n) cos
n=1 n 2
{ }
n cos(n) 1
c. 2 + 2
cos(nx) 2 2 sin(nx)
n=1 n +1 n
n
d. sin(nx)
n=1 (n + 1)
2. For each convergent Fourier series in problem 1 above, determine the highest derivative
of the periodic extension of f (x) that should be continuous.
f (x) = |x|, 1 x 1
(found earlier). Dierentiate this series once term by term, and compare your answer to the
actual series of f (x).
29
( )
1 1
1. a. an = 0, bn = = O , p=3>1
n3 2 n3
The series converges uniformly.
( )
1 1
1. b. an = cos(n) = O , p = 1 bn = 0
n n
The series converges in the mean but not uniformly.
( ) ( )
n cos(n) 1 1 1
1. c. an = = O , p = 1 bn = = O
n2 + 1 n n2 2 n2 ( )
The series converges in the mean but not uniformly (because the dominant term is O n1 ).
n
1. d. an = 0, bn = = O (1)
n+1
The series does not converge. ( )
1 1
2. a. an = 0, bn = 3 2 = O 3 , p = 3 > 1
n n
Since p = 3, both f (x) and f (x) should be continuous but not f (x)
( )
1 1
2. b. an = cos(n) = O , p = 1 bn = 0
n n
f (x) should not be continuous.
( ) ( )
n cos(n) 1 1 1
2. c. an = 2
= O , p = 1 bn = 2 2 = O 2
n +1 n n n
f (x) should not be continuous.
n
2. d. an = 0, bn = = O (1)
n+1
The series does not converge and so we cant talk about the limit function.
Note that g(x) is the derivative of |x| and also that it is odd and so we expect a Fourier sine
series.
30
2.5 Interpretation of the Fourier Coefficients
PROBLEMS
1. Plot the amplitude and phase as a function of frequency for each of the Fourier series
found for the problems in the rst section of this chapter.
31
1. a.
1 1 cos(n)
f (x) = + sin(nx)
2 n=1 n
f (x) = A0 + An cos (nx n )
n=1
where ( )
bn
1
An = a2n + b2n , n = tan
an
Note that n = 0 by denition if An = 0. Now
1 cos(n) |1 cos(n)| 1
an = 0, bn = , An = , A0 =
n n 2
{
0, n = 0, 2, 4, . . .
n =
/2, n = 1, 3, 5, . . .
An n
/2
.64
.5
.21
.13
n n
32
1. b. ( )
3 6 nx
f (x) = + 2 2
(1 cos(n)) cos
2 n=1 n 3
f (x) = A0 + An cos (nx n )
n=1
where ( )
bn
1
An = a2n + b2n , n = tan
an
Now
a0 3 6
= , an = 2 2 (1 cos(n)), bn = 0,
2 2 n
12
, n = 1, 3, 5, . . .
3 n22
A0 = , An =
2
0, n = 2, 4, 6, . . .
n = 0
An n
/2
1.5
1.22
.14
.05
n n
33
1. c.
{( )2 ( ) ( )2 ( )}
1 2 2 cos( n
2
) cos(n) 1 nx 2 n nx
f (x) = + cos + sin( ) sin
4 n=1 n 2 2 n 2 2
Now ( )2 ( )2
a0 1 2 2 cos( n
2
) cos(n) 1 2 n
= , an = , bn = sin( ),
2 4 n 2 n 2
( )2
1 2 (2 cos( n ) cos(n) 1)2 n
A0 = , An = 2
+ sin2 ( )
4 n 4 2
( )
2 sin( n
1 2
)
n = tan
2 cos( 2 ) cos(n) 1
n
1 = , 2 = , 3 = , 4 = 0
2 2
For n 5 these values repeat.
An
.5
/2
.41
.20
n
.04
.02
n
/2
34
1. d.
f (x) = 1 cos(x)
a0 = 1, a1 = 1, all others are zero
A0 = 1, A1 = 1, An = 0, n 2
1 =
An
n
1
35
2. ( ( ) ( ))
a0 n n
f (x) = + an cos x + bn sin x
2 n=1 L L
So [ ( ( ) ( ))]2
1 L 1 L a0
n n
[f (x)]2 dx = + an cos x + bn sin x dx
2L L 2L L 2 n=1 L L
[ ( ( ) ( ))
1 L a20 a0 n n
= +2 an cos x + bn sin x
2L L 4 2 n=1 L L
(
( ) ( )) ( ( ) ( ))]
n n m m
+ an cos x + bn sin x am cos x + bm sin x dx
n=1 m=1 L L L L
( ) ( )
1 L a20 1
L
n L
n
= dx + a0 an cos x dx +bn sin x dx
2L L 4 2L n=1 | L {z
L } |
L
{z
L }
=0 =0
L ( ) L( ) ( ) ( )
1 a n a m
n m n m
+ cos x cos x dx +a b
n m cos x sin x dx
2L n=1 m=1 |
L L {z
L } |
L L {z
L }
=0, m=n =0, for all m,n
( ) ( L ) ( ) ( )
L m n n m
+am bn cos x sin x dx +bn bm sin x sin x dx
|
L L {z L } |
L L {z L }
=0, for all m,n =0, m=n
(
L ) ( )
1 a2 1 2 L n n
= 2L 0 + an cos2 x dx +b2n sin2 x dx
2L 4 2L n=1 | L {z
L } |
L
{z
L }
=L =L
a20 1
( ) a2 1 ( )
= + La2n + Lb2n = 0 + a2n + b2n
4 2L n=1 4 2 n=1
1
= A20 + A2
2 n=1 n
36
2.6 The Complex Form of the Fourier Series
PROBLEMS
1. Find the complex Fourier series for each of the following functions:
a. f (x) = x, 3 x 3 , f (x + 6) = f (x)
{
0 , 1 < x < 0
b. f (x) =
1 , 0<x<1
f (x + 2) = f (x)
2. Plot the complex amplitude spectrum for each of the series found in problem 1 above.
3. Show that if we use T0 for the period of a signal, rather than 2L, the formula for the
complex Fourier series coecients reduces to
1 T0 /2
cn = f (x)e2inx/T0 dx
T0 T0 /2
4. Using the complex form of the Fourier series, prove the following form of Parsevals
theorem
1 L
|f (x)| dx =
2
|cn |2
2L L n=
(Hint: Show
[
][
]
|f (x)| = f (x)f (x) =
2
cn e inx/L
cm eimx/L
n= m=
[ ]
= cn cm ei(nm)x/L ,
n= m=
then integrate.)
37
1. a.
f (x) = cn ei(nx/L)
n=
where L = 3 and
1 L 1 3 i(nx/3)
cn = f (x)ei(nx/L) dx = xe
2L L 6 3
[ ( )2 ]
1 3 3 3
cn = xei(nx/3) e i(nx/3)
6 in in 3
[ ( )2 ( )2 ]
1 9 in 3 in 9 in 3
cn = e e e + e in
, n = 0
6 in in in in
Note that eix = cos x + i sin x ein = cos(n) i sin(n) = cos(n). Also
| {z }
=0
3
1
c0 = xdx = 0
6 3
So [ ]
1 18 3i 1
cn = cos(n) = cos(n), note = i
6 in n i
Thus
3i
f (x) = cos(n)ei(nx/3)
n= n
3i 3
|cn | = | cos(n)| = | |, n = 0
n n
|c0 | = 0
To compute the argument of cn , we note that
3i
cn = , for n = 1, 3, 5, . . .
n
and
3i
, for n = 2, 4, 6, . . .
cn =
n
Notice that since n changes sign, we have
2,
n = 1, 3, 5, . . .
arg(cn ) =
2
, n = 1, 3, 5, . . .
and
2
, n = 2, 4, 6, . . .
arg(cn ) =
2 , n = 2, 4, 6, . . .
38
1. b.
f (x) = cn ei(nx/L)
n=
where L = 1 and
1 1 i(nx) 1 1 i(nx)
cn = f (x)e dx = e
2 1 2 0
1 i(nx) 1 1 [ in ]
cn = e = e 1 , n = 0
2in 0 2in
Note that ein = cos(n) i sin(n) = cos(n). Also
| {z }
=0
1 1 1
c0 = dx =
2 0 2
So
i
cn = (cos(n) 1)
2n
Thus
1 i
f (x) = + (cos(n) 1) ei(nx)
2 n=,n=0 2n
{
i | n
1
|, n odd
|cn | = | (cos(n) 1) | =
2n 0, n even
1
|c0 | =
2
Argument of cn is 2 where n = 1, 3, 5, . . . and is 0 for n = 2, 4, 6, . . .
39
2. a.
arg(cn)
/2
|cn|
.95
.48 n
.32
.24
n
/2
2. b.
arg(cn)
/2
|cn|
.50 n
.32
.11
n
/2
40
3.
If f (x) = cn ei(nx/L) then
n=
1 L
cn = f (x)ei(nx/L) dx
2L L
T0
Now let T0 = 2L, L = . Thus
2
1 T0 /2
cn = f (x)einx/(T0 /2) dx
T0 T0 /2
T0 /2
1
cn = f (x)e2inx/T0 dx
T0 T0 /2
4.
L L L
1 1 1
|f (x)| dx =
2
f (x)f (x) dx = cn cm ei(nm)x/L dx
2L L 2L L 2L L n= m=
L L
1
i(nm)x/L 1
= cn c e dx = cn c dx
2L n= m= m | L {z }
2L n= n | L{z }
=0 for n=m =2L
Thus
1 L
|f (x)|2 dx = |cn |2
2L L n=
41
2.7 Fourier Series and Ordinary Differential Equations
PROBLEM
42
1. A plot of the periodic extension of f (x) shows that f (x) = |x|, 1 x 1.
Therefore
1 2(cos(n) 1)
f (x) = + cos(nx)
2 n=1 n2 2
So, if
1 2(cos(n) 1)
y + 2y + y = + cos(nx)
2 n=1 n2 2
then y(x) = y0 (x) + yn (x) where
n=1
1 1
y0 + 2y0 + y0 = y0 =
2 2
2(cos(n) 1)
yn + 2yn + yn = cos(nx)
n2 2
Thus
yn (x) = n cos(nx) + n sin(nx)
and substitution yields (after comparing like terms)
( ) 2(cos(n) 1)
n2 2 + 1 n + 2nn =
n2 2
( )
2nn + n2 2 + 1 n = 0
Solving the system of two equations for the two unknowns, we have
( )
2(cos(n) 1) (1 n2 2 ) 1
n = = O
2
n 2
(1 n2 2 ) + 4n2 2
2
n4
( )
2(cos(n) 1) 2n 1
n = =O 5
2
n 2
(1 n ) + 4n
2 2 2 2 2 n
So
1
2(cos(n) 1) {( ) }
y(x) = + [ ] 1 n 2 2
cos(nx) + 2n sin(nx)
2 n=1 n2 2 (1 n2 2 )2 + 4n2 2
43
y(x)
.5
1 1 x
44
2.8 Fourier Series and Digital Data Transmission
45
3 The One-Dimensional Wave Equation
3.1 Introduction
46
3.2 The One-Dimensional Wave Equation
PROBLEMS
1. Show that if a uniform, thin, tightly stretched elastic string is acted upon by no forces
other than internal tension and an external air resistance proportional to the vertical velocity,
then Newtons second law leads to a partial dierential equation of the form:
2u u 2u
+ d = ,
t2 t x2
where d is some positive constant of proportionality.
2. Show that if a uniform, thin, tightly stretched elastic string is acted upon by no forces
other than internal tension and an external spring-like restoring force proportional to the
vertical displacement, then Newtons second law leads to a partial dierential equation of
the form:
2u 2u
= s u ,
t2 x2
where s is some positive constant of proportionality.
47
1. If we add air resistance proportional to the vertical velocity then the net vertical force
acting on the small segment of the string becomes
{ }
u u 1 u
(x + x, t) (x, t) + g(x + x, t)x K x
x x 2 t
Therefore
2u 2u u
(x, t) = (x, t) + g(x, t) K d
t2 x2 t
2. If instead of air resistance we have external spring-like restoring force then the rst
u
equation in problem 1 will have Kux instead of K x and thus the nal equation
t
becomes
2u 2u
2 (x, t) = 2 (x, t) + g(x, t) Ks u(x, t)
t x
48
3.3 Boundary Conditions
PROBLEMS
1. Having physically correct algebraic signs in boundary conditions can be critical. Show,
both mathematically and physically, that the following boundary conditions:
a. u(L, t) ux (L, t) = 0
b. u(0, t) + ux (0, t) = 0
are not physically realistic.
2. Show that if a uniform, thin, tightly stretched elastic string is attached at its right-
hand boundary to a slip-ring on a pole that is not frictionless, but in which the friction
is proportional to the vertical velocity along the pole, then the boundary condition at that
point becomes
u u
+ d =0 ,
x t
where d is some positive constant of proportionality.
49
1. a.-b. The balance of forces, as discussed in the notes, leads to
u
s u(L, t) + (L, t) = 0
x
u
s u(0, t) (0, t) = 0
x
Physically, the spring constant s and the string tension cannot be negative.
2. The vertical force acting on a weightless slip ring due to string tension is, as discussed
u
in the notes, (x, t). The vertical force acting on the ring due to pole friction is opposite
x
u
to the direction of its motion: d (x, t). The sum of forces acting on the ring must be
t
zero, therefore
u u
(x, t) d (x, t) = 0
x t
50
3.4 Initial Conditions
51
3.5 Introduction to the Solution of the Wave Equation
PROBLEMS
1. Briey describe, in a few sentences, a physical model for each of the following boundary
value problems:
a.
utt = 4uxx
u(0, t) = u{x (3, t) = 0
2x , 0 < x < 1
u(x, 0) =
0 ,1 x < 3
ut (x, 0) = 0
b.
utt = uxx
u(0, t) = u(1, t) = 0
u(x, 0) = 0
ut (x, 0) = 1
c.
utt = 9uxx
ux (0, t) = ux (2, t) = 0
u(x, 0) = x
ut (x, 0) = 0
d.
utt = uxx
u(0, t) = 0
u(3, t) + 2u
{ x (3, t) = 0
2x , 0 < x < 1
u(x, 0) =
0 ,1 x < 3
ut (x, 0) = 10
52
1. a. A vibrating string of length L = 3 and c = 2. The initial displacement is linear
from 0 to 1 and at rest from 1 to 3. The initial velocity is zero. The left end is xed and the
right end is free.
53
3.6 The Fixed End Condition String
PROBLEMS
1. Solve:
utt = uxx
u(0, t) = u(3,
t) = 0
2x , 0 < x < 1/2
u(x, 0) = 2 2x , 1/2 x < 1
0 ,1 x < 3
ut (x, 0) = 0
Sketch the ten-term partial sum of your computed solution at
t = 0, 1, 2, 4 .
2. Solve:
utt = uxx
u(0, t) = u(, t) = 0
0 , 0 < x < /4
(4x )/ , /4 x /2
u(x, 0) =
(3 4x)/ , /2 x 3/4
0 , 3/4 < x <
ut (x, 0) = 0
3. Solve:
utt = uxx
u(0, t) = u(, t) = 0
u(x, 0) = x( x), 0 < x <
ut (x, 0) = 0
4. Solve:
utt = uxx
u(0, t) = u(3, t) = 0
u(x, 0) = 0
ut (x, 0) = x
5. Solve:
utt = 9uxx
u(0, t) = u(, t) = 0
u(x, 0) = sin(x)
ut (x, 0) = 1
54
6. Solve:
utt = 4uxx
u(0, t) = u(5,
{ t) = 0
x , 0 < x < 5/2
ut (x, 0) =
5 x , 5/2 x < 5
u(x, 0) = 0
7. The dissipation of heat in a very large solid slab of thickness L whose faces are held at
a xed reference temperature of 0o is described by the partial dierential equation:
ut = kuxx
u(0, t) = u(L, t) = 0
u(x, 0) = f (x)
where: ( )
2L nx
bn = f (x) sin dx .
L 0 L
55
1.
utt = uxx
u(0, t) = u(3,
t) = 0
2x , 0 < x < 1/2
u(x, 0) = 2 2x , 1/2 x < 1
0 ,1 x < 3
ut (x, 0) = 0
Separation of variables leads to
{ ( ) ( ) ( ) ( ) ( )2 ( )
2 3 n 3 2 n 3 n
= cos +2 sin 2 sin
3 n 6 n 6 n 3
( ) ( ) ( )2 ( )}
3 n 3 n
+ cos +2 sin
n 6 n 6
( ) { ( ) ( )}
2 3 2 n n
= 2 2 sin sin
3 {n ( ) 6 ( )} 3 ( )
12 n n 1
= 2
2 sin sin O 2
(n) 6 3 n
56
Thus { ( ) ( )} ( ) ( )
12 n n n n
u(x, t) = 2
2 sin sin cos t sin x
n=1 (n) 6 3 3 3
The plot of u(x, t) for various values of t is given in Figure 19
1 1
t=0 t=1
0.5 0.5
0 0
0.5 0.5
1 1
1 0 1 2 3 1 0 1 2 3
1 1
0 0
0.5 0.5
1 1
1 0 1 2 3 1 0 1 2 3
2.
utt = uxx
u(0, t) = u(, t) = 0
0 , 0 < x < /4
(4x )/ , /4 x /2
u(x, 0) =
(3 4x)/ , /2 x 3/4
0 , 3/4 < x <
ut (x, 0) = 0
Separation of variables yields:
u(x, t) = {An cos (nt) + Bn sin (nt)} sin (nx)
n=1
Initial conditions
2
u(x, 0) = An sin (nx) = f (x) An = f (x) sin (nx) dx
n=1 0
ut (x, 0) = nBn sin (nx) = 0 Bn = 0
n=1
57
Thus { }
2 /2 4 3/4
4 3
An = (x ) sin (nx) dx + ( x) sin (nx) dx
/4 4 /2 4
{[ ]
8 x 4 1 /2
An = cos (nx) + 2 sin (nx)
[ n n ]
3/4 }
/4
3
x 1
+ 4 cos (nx) 2 sin (nx)
n n /2
{ ( ) ( ) ( )
8 n 1 n 1 n
= 2 cos + 2 sin 2 sin
4n ) 2
( (n ) 2 n( )} 4
n 1 3n 1 n
+ cos 2 sin + 2 sin
4n 2 n 4 n 2
{ ( ) ( ) ( )}
8 2 n 1 n 1 3n
= 2 2
sin 2 sin 2 sin
n 2 n 4 n 4
8 { ( ) ( ) ( )}
2 n 1 n 1 3n
u(x, t) = 2
sin sin sin cos (nt) sin (nx)
n=1 n2 2 n2 4 n2 4
58
3.
utt = uxx
u(0, t) = u(, t) = 0
u(x, 0) = x( x), 0 < x <
ut (x, 0) = 0
Separation of variables yields:
u(x, t) = {An cos (nt) + Bn sin (nt)} sin (nx)
n=1
Initial conditions
2
u(x, 0) = An sin (nx) = (x x2 ) An = (x x2 ) sin (nx) dx
n=1 0
ut (x, 0) = nBn sin (nx) = 0 Bn = 0
n=1
Thus
2
An = (x x2 ) sin (nx) dx
0
{ }
2 x x2
2x 2
An = cos (nx) + sin (nx) cos (nx)
n n2 n3 0
{ }
2 2
= 0 + 0 3 (cos (n) 1)
n
4
= (1 cos (n))
n3
4 1 cos (n)
u(x, t) = cos (nt) sin (nx)
n=1 n3
59
4.
utt = uxx
u(0, t) = u(3, t) = 0
u(x, 0) = 0
ut (x, 0) = x
Separation of variables leads to
60
5.
utt = 9uxx
u(0, t) = u(, t) = 0
u(x, 0) = sin(x)
ut (x, 0) = 1
Separation yields
u(x, t) = {An cos (3nt) + Bn sin (3nt)} sin (nx)
n=1
Initial conditions
u(x, 0) = An sin (nx) = sin x A1 = 1, An = 0, n > 1
n=1
1 2
ut (x, 0) = 3nBn sin (nx) = 1 Bn = 1 sin (nx) dx
n=1 3n 0
Computing Bn , we have
2(1 cos(n))
Bn =
3n2
Thus
2 1 cos(n)
u(x, t) = cos(3t) sin(x) + sin (3nt) sin (nx)
3 n=1 n2
61
6.
utt = 4uxx
u(0, t) = u(5,
{ t) = 0
x , 0 < x < 5/2
ut (x, 0) =
5 x , 5/2 x < 5
u(x, 0) = 0
Separation yields
{
( ) ( )} ( )
2n 2n n
u(x, t) = An cos t + Bn sin t sin x
n=1 5 5 5
Initial conditions ( )
n
u(x, 0) = An sin x = 0 An = 0
n=1 5
( ) ( ) {
2n n x, 0 < x < 5/2
ut (x, 0) = Bn sin x =
n=1 5 5 5 x, 5/2 < x < 5
So { ( ) ( ) }
5 2 5/2 n 5
n
Bn = x sin x dx + (5 x) sin x dx
2n 5 0 5 5/2 5
Computing Bn , we have { ( )2 ( )}
1 5 n
Bn = 2 sin
n n 2
Thus ( ) ( ) ( )
50 n 2n n
u(x, t) = 3
sin sin t sin x
n=1 (n) 2 5 5
62
7.
ut = kuxx
u(0, t) = u(L, t) = 0
u(x, 0) = f (x)
Separation of variables, as in the wave equation case, yields:
T (t)
X(x)
= =
kT (t) X(x)
and
u(0, t) = X(0)T (t) = 0, X(0) = 0
u(L, t) = X(L)T (t) = 0, X(L) = 0
Thus
T (t) + kT (t) = 0 X (x) + X(x) = 0
X(0) = 0
X(L) = 0
( )2
n
The problem for X(x) was already solved and the eigenvalues are: n = , n = 1, 2, . . .
( ) L
n
and the eigenfunctions are: Xn (x) = sin x , n = 1, 2, . . . So now
L
( )2
n
T (t) = 0 T (t) = ek(n/L)
2t
T (t) + k
L
So the linearly independent solutions for u(x, t) are
( )
k(n/L)2 t n
un (x, t) = e sin x , n = 1, 2, . . .
L
So to form the general solution, take the linear combination
( )
k(n/L)2 t n
u(x, t) = bn e sin x
n=1 L
and so ( )
2L n
bn = f (x) sin x dx
L 0 L
63
3.7 The Free End Conditions Problem
PROBLEMS
1. Solve:
utt = 25uxx
ux (0, t) =
ux (1, t) = 0
0 , 0 < x < 1/4
u(x, 0) = x 1/4 , 1/4 < x < 3/4
1/2 , 3/4 < x < 1
ut (x, 0) = 0
Interpret the solution physically.
2. Solve:
utt = uxx
ux (0, t) =
ux (2, t) = 0
2x , 0 < x < 1/2
u(x, 0) = 2 2x , 1/2 x 3/2
2x 4 , 3/2 < x < 2
ut (x, 0) = 1
Interpret the solution physically.
64
1.
utt = 25uxx
ux (0, t) =
ux (1, t) = 0
0 , 0 < x < 1/4
u(x, 0) = x 1/4 , 1/4 < x < 3/4
1/2 , 3/4 < x < 1
ut (x, 0) = 0
Separation of variables leads to
Eigenvalues are n = (n)2 and eigenfunctions are Xn (x) = cos (nx) , n = 0, 1, 2, . . . The
solution of the equation for T (t) is
Initial conditions
0
, 0 < x < 1/4
1
u(x, 0) = A0 + An cos (nx) = x 1/4 , 1/4 < x < 3/4
2
n=1 1/2 , 3/4 < x < 1
Computing A0 , we have 1 1
A0 = 2 u(x, 0)dx =
0 2
and ( ) ( )
1 2(cos 3n
cos n
)
4 4
An = 2 u(x, 0) cos (nx) dx =
0 n2 2
Thus ( ) ( )
1 2(cos
3n
4
cos n
4
)
u(x, t) = + cos (5nt) cos (nx)
4 n=1 n2 2
65
2.
utt = uxx
ux (0, t) =
ux (2, t) = 0
2x , 0 < x < 1/2
u(x, 0) = 2 2x , 1/2 x 3/2
2x 4 , 3/2 < x < 2
ut (x, 0) = 1
Separation of variables leads to
Initial conditions
( )
2x
, 0 < x < 1/2
1 n
u(x, 0) = A0 + An cos x = 2 2x , 1/2 x 3/2
2 2
n=1 2x 4 , 3/2 < x < 2
( )
1 n n
ut (x, 0) = B0 + Bn cos x = 1
2 n=1 2 2
Computing A0 , B0 , we have
A0 = 0, B0 = 2
and ( ) ( )
8 + 16 cos n
4
16 cos 3n
4
+ 8 cos (n)
An =
n2 2
and ( )
2 2 2 n
Bn = cos x dx = 0
n 2 0 2
Thus
( ) ( )
8 + 16 cos
n
16 cos 3n
+ 8 cos (n) ( ) ( )
4 4 n n
u(x, t) = t + cos t cos x
n=1 n2 2 2 2
66
3.8 The Mixed End Conditions Problem
PROBLEMS
1. Solve:
utt = uxx
u(0, t) = 0
ux (2, t) = {0
x , 0<x1
u(x, 0) =
1 , 1<x<2
ut (x, 0) = 0
2. Solve:
utt = 4uxx
ux (0, t) = 0
u(1, t) = 0{
1 , 0 < x 1/2
u(x, 0) =
2 2x , 1/2 < x < 1
ut (x, 0) = 0
3. Solve:
utt = 9uxx
ux (0, t) = u(2, t) = 0
u(x, 0) = 0
ut (x, 0) = (2 x), 0 < x < 2
provided:
a. f (x) is odd,
b. f (x) is periodic of period 4L, and
c. f (x + L) = f (L x) , 0 < x < L
67
1.
utt = uxx
u(0, t) = 0
ux (2, t) = {0
x , 0<x1
u(x, 0) =
1 , 1<x<2
ut (x, 0) = 0
Separation of variables leads to
Initial conditions
[( ) ] {
1 x , 0<x1
u(x, 0) = An sin n x =
n=1 2 2 1 , 1<x<2
Computing An , we have
[( ) ] [( ) ]
2 1 1 2 2 1
An = x sin n x dx + sin n x dx
2 0 2 2 2 1 2 2
[ ]
2n1
16 sin 4
An =
(2n 1)2 2
Thus [ ]
16 sin 2n1
[( ) ] [( ) ]
4 1 1
u(x, t) = (2n 1)2 2
cos n t sin n x
n=1 2 2 2 2
68
2.
utt = 4uxx
ux (0, t) = 0
u(1, t) = 0{
1 , 0 < x 1/2
u(x, 0) =
2 2x , 1/2 < x < 1
ut (x, 0) = 0
Separation of variables leads to
Initial conditions
[( ) ] {
1 1 , 0 < x 1/2
u(x, 0) = An cos n x =
n=1 2 2 2x , 1/2 < x < 1
(
) [( ) ]
1 1
ut (x, 0) = n Bn cos n x = 0 Bn = 0
n=1 2 2
By Sturm-Liouville (p(x) = w(x) = 1, q(x) = 0), we have
1 1 [( ) ]
f (x)Xn (x)dx 1
An = 0
1 = 2 f (x) cos n x dx
2
0 Xn (x)dx 0 2
{ [( ) 1 ] [( ) ] }
1/2 1 1
An = 2 cos n x dx + (2 2x) cos n x dx
0 2 1/2 2
{ [( ) ]
2 1 1/2
An = 2 sin n x
(2n 1) 2 0
( )2
[( ) ] [( )
1 ]
2 1 2 1
+ (2 2x) sin n x 2 cos n x
(2n 1) 2 (2n 1) 2 1/2
( )2 [( ) ]
2 1
An = 4 cos n
(2n 1) 2 2
69
Thus
( )2 [( ) ] [( ) ]
2 1 1
u(x, t) = 2 cos n cos [(2n 1) t] cos n x
n=1 (2n 1) 2 2 2
70
3.
utt = 9uxx
ux (0, t) = u(2, t) = 0
u(x, 0) = 0
ut (x, 0) = (2 x), 0 < x < 2
Separation of variables leads to
Initial conditions
[( ) ]
1
u(x, 0) = An cos n x = 0 An = 0
n=1 2 2
{ (
) [( ) ]}
1 1
ut (x, 0) = 3 n Bn cos n x = (2 x)
n=1 2 2 2 2
Thus [ ] [( ) ]
4 2 2 1
Bn = (2 x) cos n x dx
3 (2n 1) 2 0 2 2
[ ] ( )2 [ ]
4
4 (2n 1) 4 (2n 1) 2
Bn = (2 x) sin x cos x
3 (2n 1) (2n 1) 4 (2n 1) 4 0
( )3
1 4
Bn =
3 (2n 1)
Thus ( )3
[ ( ) ] [( ) ]
1 4 1 1
u(x, t) = sin 3 n t cos n x
n=1 3 (2n 1) 2 2 2 2
71
4. Given an odd function f (x) with a period of 4L, we know that the Fourier series
should have only sine functions, i.e.
( )
nx
f (x) = bn sin
n=1 2L
Now we would like to use the fact that f (x + L) = f (L x), 0 < x < L, i.e. symmetry
about x = L. This requires that the sine functions satisty the same thing.
( ) ( )
n n
sin (x + L) = sin (L x)
2L 2L
( ) ( ) ( ) ( ) ( ) ( ) ( ) ( )
n n n n n n n n
sin x cos + sin cos x = sin cos x sin x cos
2L 2 2
| {z } 2L | {z 2 } 2L 2L 2
=0 =0
72
3.9 Generalizations on the Method of Separation of Variables
73
3.10 Sturm-Liouville Theory
PROBLEMS
1. For each of the following problems, determine if the given equation is in Sturm-Liouville
form, and, if so, identify the values of the appropriate function p(x), q(x), w(x), and the
values of i and i :
b. [(x2 1)u ] + 3u = 0
u(0) = 0
u(1/2) = 0
c. y + xy = 0
y(0) = 0
y(3) + 2y (3) = 0
d. y + xy + y = 0
y(0) = 0
y(1) = 0
2. Following similar steps to those used in class, show that the eigenfunctions of the singular
Sturm-Liouville problem:
corresponding to dierent eigenvalues are orthogonal with respect to the weighting func-
tion w(x).
74
1. In all of these, the key is to identify values (if any) such that the given dierential
equation is a special case of the one in the notes.
a.
((x + 1)y ) + (x + 1)y y = 0
y(1) = 0
y(2) = 0
This is in Sturm-Liouville form, since p(x) = x + 1, w(x) = x + 1, q(x) = 1 and the interval
is a = 1, b = 2. The coecients in the boundary conditions are: 1 = 2 = 1, 1 = 2 = 0.
b.
[(x2 1) u ] + 3u = 0
u(0) = 0
u(1/2) = 0
This is almost in Sturm-Liouville form, since p(x) = (x2 1), w(x) = 3, q(x) = 0 but
p(x) < 0 for the interval from a = 0 to b = 1/2. This violates the condition on p(x). If we
multiply the equation by 1 and incorporating the sign with (i.e., let = ) then it is
S-L. The coecients in the boundary conditions are: 1 = 2 = 1, 1 = 2 = 0.
c.
y + xy = 0
y(0) = 0
y(3) + 2y (3) = 0
This is in Sturm-Liouville form, since p(x) = 1, w(x) = x, q(x) = 0. The interval is from a =
0 to b = 3. The coecients in the boundary conditions are: 1 = 1, 2 = 1, 1 = 0, 2 = 2.
d.
y + xy + y = 0
y(0) = 0
y(1) = 0
This is not in Sturm-Liouville form, since the equation has a separate y term which does
not occur in the S-L form.
75
2. Suppose y1 (x) and y2 (x) are solutions of
[p(x)y ] +w(x)y + q(x)y = 0 , a < x < b
y(a), y (a) nite ,
2 y(b) + 2 y (b) = 0
where p (x), w(x), and q(x) are continuous, and
p(x) > 0 , a < x b ,
p(a) = 0
w(x) > 0 , a < x < b
q(x) 0 , a x b
corresponding to dierent eigenvalues, i.e.
[p(x)y1 ] + 1 w(x)y1 + q(x)y1 = 0
[p(x)y2 ] + 2 w(x)y2 + q(x)y2 = 0
Multiply the rst equation by y1 (x) and the second equation by y1 (x) and subtract. Then
y1 [py1 ] y1 [py2 ] + (1 2 )wy1 y2 = 0
or
d
[y2 p(x)y1 y1 p(x)y2 ] + (1 2 )wy1 y2 = 0
dx
Integrate b
b
[y2 p(x)y1 y1 p(x)y2 ] + (1 2 ) wy1 y2 dx = 0
a a
or
b
p(b) [y1 (b)y2 (b) y2 (b)y1 (b)] p(a) [y1 (a)y2 (a) y2 (a)y1 (a)] + (1 2 ) wy1 y2 dx = 0
| {z } a
=0
Thus b
p(b) [y1 (b)y2 (b) y2 (b)y1 (b)] + (1 2 ) wy1 y2 dx = 0
a
But now if both y1 (x) and y2 (x) satisfy the B.C. at x = b
2 y2 (b) + 2 y2 (b) = 0
2 y1 (b) + 2 y1 (b) = 0
and 2 and 2 not both zero implies that the homogeneous system can have a solution only if
the determinant is zero, i.e. y1 (b)y2 (b)y2 (b)y1 (b) = 0. Thus p(b) [y1 (b)y2 (b) y2 (b)y1 (b)] = 0
and so b
(1 2 ) wy1 y2 dx = 0
a
Since 1 = 2 we must have b
wy1 y2 dx = 0
a
76
3.11 The Frequency Domain Interpretation of the Wave Equation
PROBLEM
1. Find the three lowest natural frequencies, and sketch the associated modes, for the
equation:
utt = c2 uxx
u(0, t) = ux (L, t) = 0
Plot, in the frequency domain, the natural frequencies of this string.
77
1. Separation of variables yields
Therefore ( ) 2
1
n+
n = , n = 0, 1, 2, . . .
2
L
and the eigenfunctions
( )
1
n+
Xn (x) = sin x , n = 0, 1, 2, . . .
2
L
78
( )
3
X1 (x) = sinx
2L
( )
5
X2 (x) = sin x
2L
The modes are given in Figure 20 and the frequencies in Figure 21
1
0.5
0.5
1
2 1.5 1 0.5 0 0.5 1 1.5 2
0.5
0.5
1
2 1.5 1 0.5 0 0.5 1 1.5 2
0.5
0.5
1
2 1.5 1 0.5 0 0.5 1 1.5 2
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
1 0.5 0 0.5 1 1.5 2 2.5
79
3.12 The DAlembert Solution of the Wave Equation
80
3.13 The Effect of Boundary Conditions
PROBLEM
1. Sketch the DAlembert solutions at t = 0, 1, 2.5 and 4 to:
a. utt = uxx
u(0, t) = u(3, t) = 0
2x , 0 < x < 1/2
1 , 1/2 < x < 3/2
u(x, 0) =
4 2x , 3/2 < x < 2
0 , 2<x<3
ut (x, 0) = 0
b. utt = uxx
u(0, t) = ux (3, t) = 0
2x , 0 < x < 1/2
1 , 1/2 < x < 3/2
u(x, 0) =
4 2x , 3/2 < x < 2
0 , 2<x<3
ut (x, 0) = 0
c. utt = 4uxx
ux (0, t) = u(1, t) = 0
{
1 , 0 < x < 1/2
u(x, 0) =
2 2x , 1/2 < x < 1
ut (x, 0) = 0
81
1. a. The initial displacement is given by:
2x , 0 < x < 1/2
1 , 1/2 < x < 3/2
u(x, 0) =
4 2x , 3/2 < x < 2
0 , 2<x<3
The solution is
u(x, t) = F (x + t) + F (x t)
where F (x) is odd and periodic with period 2L = 6. The solution at various time is given
in the following 4 gures:
0.5
0
0 L
x
-0.5
-1
u(x, t) = F (x + t) + F (x t)
where the function F (x) is odd around 0 and even around L of period 4L = 12. The solution
at various time is given in the following 4 gures:
82
0.4
0.2
0
0 L
x
-0.2
-0.4
where the function F (x) is even around x = 0 and odd around x = 1 of period 4L = 4. The
solution at various time is given in the following 4 gures:
83
1
0.5
0
0 L
x
-0.5
-1
0.4
0.2
0
0 L
x
-0.2
-0.4
84
1
0.5
0
0 L
x
-0.5
-1
0.4
0.2
0
0 L
x
-0.2
-0.4
85
1
0.5
x
0 L
0
-0.5
-1
0.5
0
0 L
x
-0.5
-1
86
1
0.8
0.6
0.4
0.2
0
0 L
x
x
0 L
0
-0.2
-0.4
-0.6
-0.8
-1
87
1
0.5
0
0 L
x
-0.5
-1
0.8
0.6
0.4
0.2
0
0 L
x
88
4 The Two-Dimensional Wave Equation
4.1 Introduction
4.2 The Rigid Edge Problem
4.3 Frequency Domain Analysis
4.4 Time Domain Analysis
4.5 The Wave Equation in Circular Regions
4.6 Symmetric Vibrations of the Circular Drum
4.7 Frequency Domain Analysis of the Circular Drum
89
4.8 Time Domain Analysis of the Circular Membrane
PROBLEMS
1. It can be shown that the small free vertical vibrations of a uniform beam (e.g. a bridge
girder) are governed by the fourth order partial dierential equation:
2u 4
2 u
+ c =0
t2 x4
where c2 is a constant involving the elasticity, moment of inertia, density and cross sectional
area of the beam. If the beam if freely supported at both ends, e.g. sitting on a piling, the
boundary conditions for this problem become:
u(0, t) = u(L, t) = 0
uxx (0, t) = uxx (L, t) = 0
Show that separation of variables works in this problem, and, in case the beam is initially
at rest, i.e.
ut (x, 0) = 0
produces a general solution of the form:
( ) ( )
n2 2 ct nx
An cos 2
sin
n=1 L L
90
4. Find the eigenvalues and the form of the eigenfunctions for:
41 = 22 ; 61 = 23 ; 62 = 43 ; 72 = 14
6. Show that in the square membrane, certain natural frequencies may have four independent
modes (shapes) associated with them.
91
1.
2u 4
2 u
+ c =0
t2 x4
Since the PDE and BC are linear and homogeneous, assume
u(x, t) = F (x)G(t)
Thus uxx (0, t) = F (0)G(t) = 0, F (0) = 0. Similarly for the other conditions.
There is no a-priori reason to assume a particular value for k, so consider all three cases. If
k > 0, we write k = 4 4 (just to avoid radicals) then try F (x) = erx to get
r4 + 4 4 = 0
F (x) = 2 2 {C1 sinh x cos x C2 sinh x sin x + C3 cosh x cos x C4 cosh x sin x}
F (0) = C2 = 0
F (0) = C3 = 0
Thus at x = L
F (L) = C1 cosh L sin L + C4 sinh L cos L = 0
F (L) = 2 2 {C1 sinh L cos L C4 cosh L sin L} = 0
The determinant of the coecients is equal
{ }
2 2 cosh2 L sin2 L sinh2 L cos2 L = 0
92
Thus by Cramers rule, the only solution is C1 = C4 = 0 which is the trivial solution.
Now try k = 0. Thus F = 0 and F (x) = C1 + C2 x + C3 x2 + C4 x3
F (0) = C1 = 0
F (0) = C3 = 0
F (L) = C2 L + C4 L3 = 0
F (L) = 6C4 L = 0, C4 = 0
Substituting in the previous equation F (L) = C2 L + C4 L3 = 0 we have C2 = 0 and again
this is the trivial solution.
So k < 0, so we take k = 4 and we nd that the roots are now
r = , i
C1 sinh L = 0, C1 = 0
Therefore
C3 sin L = 0, sin L = 0
Therefore
n
n =
L
and ( )
n
Fn (x) = sin x
L
Now recall that k = 4 , so
( )4
G(t) + n
kc2 G = 0, G(t) c2 G = 0
L
93
Thus ( ) ( )
n2 2 n2 2
G(t) = c1 cos ct + c2 sin ct
L2 L2
( )
n2 2
G(0) = 0, c c2 = 0, c2 = 0
L2
( )
n2 2
Gn (t) = cos ct
L2
So the solution becomes
( ) ( )
n2 2 n
un (x, t) = cos 2
ct sin x
L L
and ( ) ( )
n2 2 n
u(x, t) = An cos 2
ct sin x
n=1 L L
94
2.
utt = uxx + uyy
u(0, y, t) = u(1, y, t) = 0
u(x, 0, t) = u(x, 1, t) = 0
u(x, y, 0) = .01xy(1 x)(1 y)
ut (x, y, 0) = 0
Let u(x, y, t) = X(x)Y (y)T (t), then
T(t) + T (t) = 0 X + X = 0 Y + ( )Y = 0
X(0) = 0 Y (0) = 0
X(1) = 0 Y (1) = 0
So {
}
u(x, y, t) = Amn cos( mn t) + Bmn sin( mn t) sin(nx) sin(my)
n=1 m=1
where mn = m2 + n2 , so
u(x, y, 0) = Amn sin(nx) sin(my) = .01xy(1 x)(1 y)
n=1 m=1
ut (x, y, 0) = mn Bmn sin(nx) sin(my) = 0, Bmn = 0
n=1 m=1
Thus
4 1 1
Amn = (.01xy(1 x)(1 y) sin(nx) sin(my)dydx
11 0 0
1 1
Amn = .04 x(1 x) sin(nx)dx y(1 y) sin(my)dy
0 0
But
[ ]
1 x(1 x) 1 2x 2 1
x(1 x) sin(nx)dx = cos(nx) + sin(nx) cos(nx)
0 n (n)2 (n)3 0
1 cos(n)
which is 2 .
(n)3
Similarly
1 1 cos(m)
y(1 y) sin(my)dy = 2
0 (m)3
Therefore
1 cos(m) 1 cos(n)
Amn = .16
(m)3 (n)3
95
and
1 cos(m) 1 cos(n)
u(x, y, t) = .16 cos( mn t) sin(nx) sin(my)
n=1 m=1 (m)3 (n)3
where
mn = m2 + n2
96
3.
utt = 16 (uxx + uyy )
u(0, y, t) = u(3, y, t) = 0
u(x, 0, t) = u(x, 2, t) = 0( )
u(x, y, 0) = y(2 y) sin 2x
3
ut (x, y, 0) = 0
Let u(x, y, t) = X(x)Y (y)T (t), then
97
( )3 3 ( ) ( )
8 2 2 n
Amn = (1 cos(m)) sin x sin x dx
6 m |0 3 {z 3 }
3
= , n = 2, otherwise = 0
2
Thus ( )3
2
Am 2 = 2 (1 cos(m)), Am n = 0, n = 2
m
So ( )3 ( ) ( )
2 2 m
u(x, y, t) = 2 (1 cos(m)) cos(4m 2 t) sin x sin y
m=1 m 3 2
where
( )2 ( )2
2 m
m 2 = +
3 2
98
4.
utt = 9 (uxx + uyy )
u(0, y, t) = u(4, y, t) = 0
uy (x, 0, t) = u(x, 1, t) = 0
Let u(x, y, t) = X(x)Y (y)T (t), then
T(t) + 9T (t) = 0 X + X = 0 Y + ( )Y = 0
X(0) = 0 Y (0) = 0
X(4) = 0 Y (1) = 0
with
( )2 ( )2 ( )2 ( )2
n (2m + 1) n (2m + 1)
n = , ( )m = , mn = +
4 2 4 2
Let mn = mn , then
( )2 ( )2
n 2m + 1
mn = +
4 2
So
Tmn = Amn cos(3mn t) + Bmn sin(3mn t)
Thus the frequencies are given by
( )2 ( )2
3 3 n 2m + 1
fmn = mn = +
2 2 4 2
m= 0 1 2
n = 1 0.839 2.281 3.769
n = 2 1.061 2.372 5.303
n = 3 1.352 2.516
n = 4 1.677
The four lowest frequencies are: 0.839, 1.061, 1.352, 1.677. Notice that in this case they all
come from m = 0.
99
5. If L = 2 and W = 1, the eigenvalues are,
( )2 ( )2 ( )2
n m n 2
nm = + = + m2 = n + (2m)2
L W 2 2
The eigenfunctions are ( )
nx
sin(my) sin
2
Thus two dierent eigenfunctions correspond to the same eigenvalue if n21 + (2m1
)2 = n22 +
(2m2 )2 . For example, m = 1, n = 4 and m = 2, n = 2 give the same eigenvalue 2 20. Also,
by trial and error, we nd: m = 1, n = 6 and m = 3, n = 2 give the same eigenvalue 2 40,
m = 2, n = 6 and m = 3, n = 4 give the same eigenvalue 2 52, and m = 2, n = 7 and
m = 4, n = 1 give the same eigenvalue 2 65.
( ) ( )
n 2 m 2
6. For a square membrane L = W and the eigenvalues are mn = + =
( ) ( ) L L
L 2 my nx
n + m2 . The eigenfunctions are sin sin . Thus dierent eigenfunctions
L L
have the same eigenvalues only when dierent m, n have the same m2 + n2 . The following
4 eigenfunctions have the same eigenvalue:
m = 1, n = 8; m = 4, n = 7; m = 8, n = 1; m = 7, n = 4; m2 + n2 = 65
m = 2, n = 9; m = 6, n = 7; m = 9, n = 2; m = 7, n = 6; m2 + n2 = 85
m = 2, n = 11; m = 5, n = 10; m = 11, n = 2; m = 10, n = 5; m2 + n2 = 125
m = 3, n = 11; m = 7, n = 9; m = 11, n = 3; m = 9, n = 7; m2 + n2 = 130
100
PROBLEMS
1. Show that separation of variables (u(r, , t) = R(r)()T (t)), applied to the wave equation
in a circular region of radius A,
{ [ ] }
2u 1 u 1 2u
2
= c2 r + 2 2
t r r r r
u(A, , t) = 0
u
u(0, , t) , (0, , t) nite
r
u(r, , t) = u(r, + 2, t)
leads to
T + c2 T = 0 r [rR ] + (r2 )R = 0 + = 0
R(0) , R (0) nite () = ( + 2)
R(A) = 0
u(4, t) = 0
u
u(0, t) , (0, t) nite
r
u(r, 0) = f (r)
ut (r, 0) = 0
4. Solve the following problems. (Non-zero coecients may be left in terms of denite
integrals of known functions.)
a. [ ]
1 u
utt = r
r r r
u(2, t) = 0
u
u(0, t) , (0, t) nite
r
u(r, 0) = sin(r)
ut (r, 0) = 0
101
b. [ ]
4 u
utt = r
r r r
u(1, t) = 0
u
u(0, t) , (0, t) nite
r
u(r, 0) = 1 r2
ut (r, 0) = 0
c. [ ]
21 u
utt =c r
r r r
u(2, t) = 0
u
u(0, t) , (0, t) nite
r
u(r, 0) = 0
ut (r, 0) = 1
5. Solve the following problem. (Non-zero coecients may be left in terms of denite
integrals of known functions.) Physically interpret the boundary conditions, and relate this
to the properties of the solution:
[ ]
21 u
utt =c r
r r r
ur (L, t) = 0
u
u(0, t) , (0, t) nite
r
u(r, 0) = f (r)
ut (r, 0) = 0
102
1. { [ ] }
2u 1 u 1 2u
= c2 r + 2 2
t2 r r r r
u(A, , t) = 0
u
u(0, , t) , (0, , t) nite
r
u(r, , t) = u(r, + 2, t)
Let u(r, , t) = R(r)()T (t), thus
[ ]
1 1
RT = c2 T (rR ) + 2 RT
r r
Divide by c2 RT ,
T 1 1
2
= (rR ) + 2
=
|c{zT} |rR {z r }
function of t only function of r,
Thus
T + c2 T = 0
and
1 1
(rR ) + 2 =
rR r
or, multiply by r2 and separate the variables:
r
(rR ) + r2 = =
|R {z } | {z}
function of r only function of only
Thus
+ = 0, r(rR ) + (r2 )R = 0
The condition u(A, , t) = 0 implies R(A) = 0 and u(0, , t) nite, implies R(0) nite,
u
similarly (0, , t) nite implies R (0) is nite. The periodicity u(r, , t) = u(r, + 2, t)
r
implies
(0) = (2)
Thus
T + c2 T = 0 r(rR ) + (r2 )R = 0 + = 0
R(0), R (0) nite (0) = (2)
R(A) = 0
103
2. Mathematically, the PDE is dened only for 0 2 (due to polar coordinates
system.
Physically, (r, ) and (r, + 2) are the same point and must have the same displacement.
Thus u(r, , t) = u(r, + 2, t) .
104
3. [ ]
6 u
utt = r
r r r
u(4, t) = 0
u
u(0, t) , (0, t) nite
r
u(r, 0) = f (r)
ut (r, 0) = 0
Separation yields
T + 6c2 T = 0 r(rR ) + (r2 )R = 0
R(0), R (0) nite
R(4) = 0
105
4. a. [ ]
1 u
utt = r
r r r
u(2, t) = 0
u
u(0, t) , (0, t) nite
r
u(r, 0) = sin(r)
ut (r, 0) = 0
Separation yields
T + T = 0 (rR ) + rR = 0
R(0), R (0) nite
R(2) = 0
d2 R dR
x2 2
+x + x2 R = 0
dx dx
which is Bessels equation of order zero. The general solution is
C1 J0 (2) = 0
Thus
J0 (2) = 0, 2n = n , n = 1, 2, . . .
where J0 (n ) = 0.
2.4048 5.5201
1 = , 2 = ,...
2 2
and
Tn (t) = An cos(n t) + Bn sin(n t)
The general solution is then
u(r, t) = {An cos(n t) + Bn sin(n t)} J0 (n r)
n=1
106
ut (r, 0) = n Bn J0 (n r) = 0, Bn = 0
n=1
2
0 r sin(r)J0 (n r)dr
An = 2 2
0 r [J0 (n r)] dr
So the solution is
u(r, t) = An cos(n t)J0 (n r)
n=1
4. b. [ ]
4 u
utt = r
r r r
u(1, t) = 0
u
u(0, t) , (0, t) nite
r
u(r, 0) = 1 r2
ut (r, 0) = 0
Proceeding as in the previous case, the general solution is
u(r, t) = {An cos(2n t) + Bn sin(2n t)} J0 (n r)
n=1
ut (r, 0) = n Bn J0 (n r) = 0, Bn = 0
n=1
1
0 r (1 r2 ) J0 (n r)dr
An = 1 2
0 r [J0 (n r)] dr
So the solution is
u(r, t) = An cos(2n t)J0 (n r)
n=1
107
4. c. [ ]
21
u
utt =c r
r r r
u(2, t) = 0
u
u(0, t) , (0, t) nite
r
u(r, 0) = 0
ut (r, 0) = 1
Proceeding as in the previous case, 4a, the general solution is
u(r, t) = {An cos(n ct) + Bn sin(n ct)} J0 (n r)
n=1
ut (r, 0) = n cBn J0 (n r) = 1
n=1
2
1 0 rJ0 (n r)dr
Bn = 2
n c 0 r [J0 (n r)]2 dr
So the solution is
u(r, t) = Bn sin(n ct)J0 (n r)
n=1
108
5. [ ]
21 u
utt =c r
r r r
ur (L, t) = 0
u
u(0, t) , (0, t) nite
r
u(r, 0) = f (r)
ut (r, 0) = 0
Separation of variables yields
T + c2 T = 0 (rR ) + rR = 0
R(0), R (0) nite
R (L) = 0
The condition R(0), R (0) nite implies C2 = 0 and the condition R (L) = 0 gives
J0 (L) = 0
Note that the eigenfunctions are still Bessel functions of order zero, due to the ODE. Only
the eigenvalues are dierent. The initial conditions
u(r, 0) = A0 + An J0 (n r) = f (r)
n=1
ut (r, t) = B0 + n cBn J0 (n r) = 0, Bn = 0
n=1
and L L
rf (r)J0 (n r)dr
0 0 rf (r)dr
An = L 2 , A0 = L
0 r [J0 (n r)] dr 0 rdr
109
5 Introduction to the Fourier Transform
5.1 Periodic and Aperiodic Functions
5.2 Representation of Aperiodic Functions
5.3 The Fourier Transform and Inverse Transform
5.4 Examples of Fourier Transforms and Their Graphical Repre-
sentation
5.5 Special Computational Cases of the Fourier Transform
5.6 Relations Between the Transform and Inverse Transform
5.7 General Properties of the Fourier Transform - Linearity, Shift-
ing and Scaling
5.8 The Fourier Transform of Derivatives and Integrals
5.9 The Fourier Transform of the Impulse Function and Its Im-
plications
110
5.10 Further Extensions of the Fourier Transform
PROBLEMS
1. Compute, from the denition, and using the properties of even and odd functions
where appropriate, the Fourier transform of each of the following functions. In each case,
plot h(t) and the Amplitude spectrum and phase angle graphs.
2. Find, directly from the denition, the inverse of the following Fourier transforms, and
plot h(t) and the amplitude and phase graphs:
{
(1 f 2 )2 , 1 < f < 1
a. H(f ) =
0 , otherwise
111
1. a. h(t) = e|t| , < t < , > 0. Note that h(t) is even (see Figure 34).
0.8
0.6
0.4
0.2
0
-4 -2 0 2 4
t
Figure 34: Graph of h(t) for problem 1a of rst set of Chapter 5 for = 1
Thus
2jf t
H(f ) = h(t)e dt = 2 et cos(2f t)dt
0
{ }
t cos(2f t) + 2f sin(2f t)
H(f ) = 2e
2 + (2f )2 0
2
H(f ) =
2 + (2f )2
2
For = 1, H(f ) = , real and positive, (f ) = 0. The plot of H(f ) is
1 + (2f )2
given in Figure 35 on the left. If we decrease to 0.05, then the plot of H(f ) is now on the
right of the same gure. Note the vertical scale, does this suggest something?
2 40
1.5 30
1 20
0.5 10
0
-1 -0.5 0 0.5 1 -1 -0.5 0 0.5 1
f f
Figure 35: Graph of H(f ) for problem 1a of rst set of Chapter 5. The left plot for = 1
and the right for = 0.05
112
{
1 , 0t1
1. b. h(t) = Note that h(t) is neither even nor odd, see Figure 36.
0 , otherwise
Thus
0.8
0.6
0.4
0.2
0
-2 -1 0 1 2
0.6
0.8
0.4
0.6
0.2 f
-2 -1 0 1 2
0.4 0
0.2 -0.2
-0.4
0
-2 -1 0 1 2
-0.6
f
-0.2
Figure 37: Graph of (H(f )) (on the left) and (H(f )) (on the right) for problem 1b of
rst set of Chapter 5
113
1
3
0.8 2
1
0.6
0
-3 -2 -1 0 1 2 3
0.4
-1
0.2 -2
-3
0
-2 -1 0 1 2
f
Figure 38: Graph of |H(f )| (on the left) and (f ) (on the right) for problem 1b of rst set
of Chapter 5
114
{
tet , 0 < t <
1. c. h(t) =
0 , otherwise
Note that h(t) is neither even nor odd, see Figure 36. Thus
0.35
0.3
0.25
0.2
0.15
0.1
0.05
0
-4 -2 0 2 4
2jf t
H(f ) = h(t)e dt = tet e2jf t dt
0
{ }
te(1+2jf )t e(1+2jf )t
H(f ) =
1 + 2jf (1 + 2jf )2 0
1
H(f ) =
(1 + 2jf )2
Or, in terms of real and imaginary parts
1 4 2 f 2 4f
H(f ) = 2 2 2
j
(1 + 4 f ) (1 + 4 2 f 2 )2
Thus
1 4 2 f 2
(H(f )) =
(1 + 4 2 f 2 )2
and
4f
(H(f )) =
(1 + 4 2 f 2 )2
and
1
|H(f )| =
1 + 4 2 f 2
( )
1
Note that |H(f )| = O
f2
115
1
0.6
0.8 0.4
0.6 0.2
0
0.4 -2 -1 0 1 2
f
-0.2
0.2
-0.4
0
-2 -1 0 1 2
-0.6
f
Figure 40: Graph of (H(f )) (on the left) and (H(f )) (on the right) for problem 1c of
rst set of Chapter 5
1 3
2
0.8
1
0.6
0
-3 -2 -1 0 1 2 3
0.4 f
-1
0.2
-2
0 -3
-2 -1 0 1 2
f
Figure 41: Graph of |H(f )| (on the left) and (f ) (on the right) for problem 1c of rst set
of Chapter 5
116
{
(1 t2 ) , 1 < t < 1
d. h(t) =
0 , otherwise
0.8
0.6
0.4
0.2
0
-4 -2 0 2 4
1
H(f ) = h(t)e2jf t dt = 2 (1 t2 ) cos(2f t)dt, since h(t) is even
0
{ }
(1 t2 ) sin(2f t) 2t cos(2f t) 2 sin(2f t) 1
H(f ) = 2 +
2f (2f )2 (2f )3 0
{ }
4 cos(2f ) 4 sin(2f )
H(f ) = +
(2f )2 (2f )3
Or
1
H(f ) = {sin(2f ) 2f cos(2f )}
2 3 f 3
Note that H(f ) is real and even!! Thus
1
|H(f )| = {| sin(2f ) 2f cos(2f )|}
2 3 |f |3
( )
1
Note that |H(f )| = O
f2
117
3
1.2
2.5
1
2
0.8
1.5
0.6
0.4 1
0.2 0.5
0 0
-2 -1 0 1 2 -1 -0.5 0 0.5 1
f
Figure 43: Graph of |H(f )| (on the left) and (f ) (on the right) for problem 1d of rst set
of Chapter 5
118
{
(1 t)2 , 1 < t < 1
e. h(t) =
0 , otherwise
0
-4 -2 0 2 4
1
H(f ) = h(t)e2jf t dt = (1 t)2 e2jf t dt
1
{ }
(1 t)2 e2jf t 2(1 t)e2jf t 2e2jf t 1
H(f ) = +
2jf (2jf )2 (2jf )3 1
{ }
4e2jf 4e2jf 2 [ 2jf 2jf
]
H(f ) = + e e
2jf (2jf )2 (2f )3
Or
4(2jf 1)e2jf 4j
H(f ) = 2
+ sin(2f )
(2jf ) (2jf )3
(1 2jf ) 1
H(f ) = 2
[cos(2f ) + j sin(2f )] sin(2f )
(f ) 2(f )3
2f cos(2f ) + (4 2 f 2 1) sin(2f ) sin(2f ) 2f cos(2f )
H(f ) = 3
+j
2(f ) (f )2
( )
1
Note that H(f ) = O
f
119
2.5
1.5
2
1
1.5
0.5
1
0
-2 -1 0 1 2
0.5 f
-0.5
0
-2 -1 0 1 2 -1
f
-0.5
-1.5
Figure 45: Graph of (H(f )) (on the left) and (H(f )) (on the right) for problem 1e of
rst set of Chapter 5
7 3
6
2
5
1
4
0
-3 -2 -1 0 1 2 3
3
-1
2
-2
1
-3
0
-2 -1 0 1 2
Figure 46: Graph of |H(f )| (on the left) and (f ) (on the right) for problem 1e of rst set
of Chapter 5
120
f. h(t) = Ae|t| cos(2t) , <t< , >0 .
0
-3 -2 -1 0 1 2 3
t
-1
-2
Figure 47: Graph of h(t) for problem 1f of rst set of Chapter 5, using A = 2, = 1
{ }
t (2f + 2) sin((2f + 2)t) cos((2f + 2)t)
+ 2Ae
2 [2 + (2f + 2)2 ] 0
A A
H(f ) = + 2
2 + (2(f 1)) 2 + (2(f + 1))2
Note that H(f ) is real and positive. Note the change of scale as we change the value of
from = 1 in Figure 48 to a value of = 0.05 in Figure 49.
1.5
0.5
0
-3 -2 -1 0 1 2 3
f
Figure 48: Graph of H(f ) for problem 1e of rst set of Chapter 5, using A = 2, = 1
121
40
30
20
10
0
-3 -2 -1 0 1 2 3
f
Figure 49: Graph of H(f ) for problem 1e of rst set of Chapter 5, using A = 2, = .05
122
(1 + t) , 1 < t < 0
1 , 0t1
g. h(t) = Since h(t) is neither even nor odd
(2 t) , 1<t<2
0 , otherwise
0.8
0.6
0.4
0.2
0
-4 -2 0 2 4
0 1 2
H(f ) = h(t)e2jf t dt = (1 + t)e2jf t dt + e2jf t dt + (2 t)e2jf t dt
1 0 1
{ }
(1 + t)e2jf t e2jf t 0
H(f ) =
2jf (2jf )2 1
{ } { }
e2jf t 1 (2 t)e2jf t e2jf t 2
+ + +
2jf 0 2jf (2jf )2 1
123
and
sin(4f ) 2 sin(2f )
| {z }
{ }
2 sin(2f ) cos(2f ) cos(2f ) 1
(f ) = arctan = arctan
1 cos(4f )
sin(2f )
| {z }
2 sin2 (2f )
Or { }
sin(f )
(f ) = arctan
cos(f )
In the following gures we plot the real and imaginary parts of H(f ), the absolute value of
H(f ) and (f ).
2
0.8
1.5 0.4
-2 -1 0 1 2
1 0
0.5 -0.4
-0.8
0
-2 -1 0 1 2
Figure 51: Graph of (H(f )) (on the left) and (H(f )) (on the right) for problem 1g of
rst set of Chapter 5
2
1.5
1
1.5
0.5
1 0
-1 -0.5 0 0.5 1
-0.5
0.5
-1
-1.5
0
-2 -1 0 1 2
Figure 52: Graph of |H(f )| (on the left) and (f ) (on the right) for problem 1g of rst set
of Chapter 5
124
h. h(t) = Ate|t| , < t < , > 0
Note that h(t) is odd, see Figure 53. Thus
0.3
0.2
0.1 t
-10 -5 0 5 10
0
-0.1
-0.2
-0.3
Figure 53: Graph of h(t) with A = = 1 for problem 1h of rst set of Chapter 5
2jf t
H(f ) = h(t)e dt = 2jA tet sin(2f t)dt
0
{
tet [ sin(2f t) 2f cos(2f t)]
H(f ) = 2jA
2 + (2f )2 0
}
et [(2 (2f )2 ) sin(2f t) + 2(2f ) cos(2f t)]
[2 + (2f )2 ]2 0
[ ] ( )
2(2f ) 8Af 1
H(f ) = 2jA = 2 j = O
[2 + (2f )2 ]2 [ + (2f )2 ] 2
f3
Note that H(f ) is purely imaginary. The |H(f )| is given by
8A|f |
|H(f )| =
[2 + (2f )2 ]2
125
1
0.5
-1 -0.5 0 0.5 1
0
-0.5
-1
Figure 54: Graph of (H(f )) with A = = 1 for problem 1h of rst set of Chapter 5
1.5
1.2
1
1
0.5
0.8
0
0.6 -4 -2 0 2 4
-0.5
0.4
-1
0.2
-1.5
0
-1 -0.5 0 0.5 1
Figure 55: Graph of |H(f )| (on the left) and (f ) (on the right) with A = = 1 for problem
1h of rst set of Chapter 5
126
{
t , 1 < t < 1
i. h(t) =
0 , otherwise
Note that h(t) is odd, see Figure 56. Thus
0.5
-2 -1 0 1 2
0
-0.5
-1
{ }
sin(2f ) cos(2f )
= 2j
(2f )2 2f
( )
sin(2f ) 2f cos(2f ) 1
= 2j 2
= O
(2f ) f
Note that H(f ) is purely imaginary. The |H(f )| is given by
0.8
0.4
-2 -1 0 1 2
0
-0.4
-0.8
| sin(2f ) 2f cos(2f )|
|H(f )| =
2(f )2
127
1.5
0.8
0.6
0.5
0
0.4 -2 -1 0 1 2
-0.5
0.2
-1
-1.5
0
-2 -1 0 1 2
Figure 58: Graph of |H(f )| (on the left) and (f ) (on the right) for problem 1i of rst set
of Chapter 5
128
{
(1 f 2 )2 , 1 < f < 1
2. a. H(f ) = H(f ) is even, see Figure 59. Thus
0 , otherwise
0.8
0.6
0.4
0.2
0
-2 -1 0 1 2
1
h(t) = H(f )e2jf t df = (1 f 2 )2 e2jf t df
1
1
h(t) = 2 (1 2f 2 + f 4 ) cos(2f t)df, Since H(f ) is even
0
{
1 2f 2 + f 4 4f + 4f 3 4 + 12f 2
= 2 sin(2f t) + cos(2f t) sin(2f t)
2t (2t)2 (2t)3
}
24f 24 f =1
cos(2f t) + sin(2f t)
(2f )4 (2t)5 f =0
[ ]
8 24 24
= 2 3
sin(2t) 4
cos(2t) + sin(2t)
(2t) (2t) (2t)5
16 {[ ] }
= (2t) 3
+ 3 sin(2t) 3(2t) cos(2t)
(2t)5
Note that |H(f )| = H(f ).
129
1
1
0.8
0.8
0.6 0.6
0.4 0.4
0.2 0.2
0
0 -2 -1 0 1 2
-2 -1 0 1 2
t
Figure 60: Graph of |H(f )| (on the left) and h(t) (on the right) for problem 2a of rst set
of Chapter 5
130
2. b. H(f ) = |f | e2|f | , < f < .
H(f ) is real and even, see Figure 62. Thus
0.16
0.12
0.08
0.04
0
-4 -2 0 2 4
f
2jf t
h(t) = H(f )e df = 2 H(f ) cos(2f t)df
0
h(t) = 2 f e2f cos(2f t)df
0
s2 2
= 2 , Using Laplace Transform tables
(s2 + 2 )2
4 (2t)2
= 2
(4 + (2t)2 )2
0.5
0.4
0.3
0.2
0.1
0
-2 -1 0 1 2
t
131
PROBLEMS
1. Compute the Fourier transform of each of the following functions, using tables, shifting
and scaling, etc. when appropriate. In each case, plot h(t) and the Amplitude spectrum and
phase angle graphs.
{
2 , 1 < t < 5
a. h(t) =
0 , otherwise
t , 0t2
b. h(t) = 4t , 2<t4
0 , otherwise
( )
t
c. h(t) = sin
3
{
2 , 3<t<
d. h(t) =
0 , otherwise
1
e. h(t) =
4 + t2
sin2 (3t)
f. h(t) =
6t2
{
et , 0 < t < 2
g. h(t) =
0 , otherwise
2. Find, using tables, shifting and/or scaling, etc., the inverse of each of the following Fourier
transforms, and plot h(t) and the amplitude and phase graphs:
1
a. H(f ) = , <f <
1 + f2
b. H(f ) = e3jf e2|f | , <f <
{
2 , 3 < f < 3
c. H(f ) =
0 , otherwise
132
{
2 , 1 < t < 5
1. a. h(t) =
0 , otherwise
The function h(t) is given in Figure 63. The closest function for which we have a Fourier
transform is the function
{
1 , T0 /2 < t < T0 /2
g(t) =
0 , otherwise
given in Figure 64.
1.5
0.5
0
-2 0 2 4 6
0.8
0.6
0.4
0.2
0
-4 -2 0 2 4
133
The real and imaginary parts of H(f ) are given in Figure 65, the |H(f )| and (f ) are given
in Figure 66.
12
6
10
4
8
2
f
6 -10 -5 0 5 10
0
4
-2
2 -4
0 -6
-10 -5 0 5 10
f
Figure 65: Graph of Re(H(f )) (on the left) and Im(H(f )) (on the right) for problem 1a of
second set of Chapter 5
12
2
10
1
8
6 0
-0.4 -0.2 0 0.2 0.4
f
4
-1
-2
0
-10 -5 0 5 10
f
Figure 66: Graph of |H(f )| (on the left) and (f ) (on the right) for problem 1a of second
set of Chapter 5
134
t , 0t2
1. b. h(t) = 4 t , 2 < t 4
0 , otherwise
The function h(t) is given in Figure 67. The closest function for which we have a Fourier
transform is the function
1 + t/T0 , T0 < t 0
g(t) =
1 t/T0 , 0 t < T0
0 , otherwise
1.5
0.5
0
-1 0 1 2 3 4 5
1.5
0.5
0
-3 -2 -1 0 1 2 3
So
H(f ) = F [2g(t 2)] = 2F [g(t 2)] , by linearity
135
{ }
H(f ) = 2 e2j(2)f G(f ) , by shifting
( )2
1 sin(2f )
H(f ) = 2e4jf
2 f
or ( )2
4jf sin(2f )
H(f ) = e
f
The real and imaginary parts of H(f ) are given in Figure 69, the |H(f )| and (f ) are given
in Figure 70.
4
3
3
2
2 1
-1 -0.5 0 0.5 1
1 0
-1
0
-1 -0.5 0 0.5 1
-2
-1
-3
-2
Figure 69: Graph of Re(H(f )) (on the left) and Im(H(f )) (on the right) for problem 1b of
second set of Chapter 5
4
3
2
3
-1 -0.5 0 0.5 1
2 0
-1
1
-2
-3
0
-1 -0.5 0 0.5 1
f
Figure 70: Graph of |H(f )| (on the left) and (f ) (on the right) for problem 1b of second
set of Chapter 5
136
( )
t
1. c. h(t) = sin
3
The function h(t) is given in Figure 71. The closest function for which we have a Fourier
transform is the function g(t) = A sin(2f0 t) given in Figure 72.
0.5
-20 -10 0 10 20
0
-0.5
-1
0.5
0
-1 -0.5 0 0.5 1
-0.5
-1
In this case
1 1
h(t) = g(t), for A = 1, and 2f0 = f0 =
3 6
So { ( ) ( )}
1 1 1
H(f ) = F [g(t)] = j
f+ f
2 6 6
Note that H(f ) is purely imaginary, so we plot Im(H(f )) and |H(f )| in Figure 73. The
graph of (f ) is given in Figure 74.
137
|H(f)|
.5
Im(H(f))
.5
1/6
f
1/6
f
1/6 1/6
Figure 73: Graph of Im(H(f )) (on the left) and |H(f )| (on the right) for problem 1c of
second set of Chapter 5
pi/2
f
1/6
pi/2
138
{
2 , 3<t<
1. d. h(t) =
0 , otherwise
The function h(t) is given in Figure 75. The closest function for which we have a Fourier
transform is the function g(t) = sgn(t) given in Figure 76.
1.5
0.5
0
-1 0 1 2 3 4 5
0.5
0
-1 0 1 2 3 4 5
-0.5
-1
139
2 2
f
-2 -1 0 1 2
y 1 y 0
0 -2
-2 -1 0 1 2
f
-1 -4
-2 -6
Figure 77: Graph of Re(H(f )) (on the left) and Im(H(f )) (on the right) for problem 1d of
second set of Chapter 5
5
3
4 2
1
3 f
-0.4 -0.2 0 0.2 0.4
0
2
-1
1 -2
-3
0
-2 -1 0 1 2
f
Figure 78: Graph of |H(f )| (on the left) and (f ) (on the right) for problem 1d of second
set of Chapter 5
140
1
1. e. h(t) =
4 + t2
The closest transform pair to h(t), given in Figure 79, is
2
G(f ) =
2 + 4 2 f 2
0.25
0.2
0.15
0.1
0.05
-10 -5 0 5 10
t
0.16
0.14
0.12
0.1
0.08
0.06
0.04
0.02
-10 -5 0 5 10
t
Thus we have to use the fact that if X(f ) is the Fourier transform of x(t) then x(f ) is
2
the Fourier transform of X(t). So for g(t) = 2 (see Figure 80) we have
+ 4 2 t2
G(f ) = e|f | = e|f | .
But
2(4)
h(t) = 2 2 2
= g(t), with = 4
2 (4) + 4 t 2
Thus [ ]
H(f ) = F g(t) = G(f )
2 2
141
So
4|f |
H(f ) = e
2
Note that the transform is a real function, so we have plotted H(f ) which is the same as
Re(H(f )) and |H(f )| in Figure 81.
1.2
0.8
0.4
0
-1 -0.5 0 0.5 1
f
142
sin2 (3t)
1. f. h(t) =
6t2
The closest transform pair to h(t), given in Figure 82, is
sin2 (f0 t)
g(t) =
f0 2 t 2
whose transform is given in Figure 68 and can be described as
( )
|f |
G(f ) = 1 (1 uf0 (|f |))
f0
1.4
1.2
0.8
0.6
0.4
0.2
0
-3 -2 -1 0 1 2 3
t
But now
sin2 (3t) sin2 (f0 t) 3
h(t) = = , with f0 =
6t2 2 2 t 2 f0
So ( ) { ( )
|f | ( )
1 |f |
, |f | < 3
H(f ) = 1 1 u3/ (|f |) = 2 3
2 3 0, otherwise
Note that H(f ) is real, so the plot of Re(H(f )) and |H(f )| are identical and given in Figure
83.
143
1.2
0.8
0.4
0
-3 -2 -1 0 1 2 3
144
{
et , 0 < t < 2
1. g. h(t) =
0 , otherwise
The plot of h(t) is given in Figure 84.
0.8
0.6
0.4
0.2
0
-3 -2 -1 0 1 2 3
e(1+2jf )t 2 1 e2(1+2jf )
H(f ) = =
1 + 2jf 0 1 + 2jf
The plots of Re(H(f )) and Im(H(f )) are given in Figure 85. The plots of |H(f )| and
(H(f )) are given in Figure 86.
0.8
0.15
0.1
0.6
0.05
0
0.4 -2 -1 0 1 2
-0.05
0.2 -0.1
-0.15
0
-2 -1 0 1 2
Figure 85: Graph of Re(H(f )) (on the left) and Im(H(f )) (on the right) for problem 1g of
second set of Chapter 5
Another way to solve the problem is to view h(t) as the product of h1 (t) and h2 (t) shown
in Figure 87. The functions are
h1 (t) = et , t > 0
145
1.5
0.8
0.6
0.5
f
-1.5 -1 -0.5 0 0.5 1 1.5
0
0.4
-0.5
0.2
-1
-1.5
-2 -1 0 1 2
Figure 86: Graph of |H(f )| (on the left) and (f ) (on the right) for problem 1g of second
set of Chapter 5
and {
1, 0<t<2
h2 (t) =
0, otherwise
The transforms of these functions are
1 sin(2f )
H1 (f ) = , H2 (f ) = e2jf
1 + 2jf f
So
H(f ) = F[h1 (t)h2 (t)] = H1 (f ) H2 (f ) = H1 (u)H2 (f u)du
or 1 sin(2(f u))
H(f ) = e2j(f u) du
1 + 2ju (f u)
Now you see that the previous method is easier!!!
1 1
0.8 0.8
0.6 0.6
0.4 0.4
0.2 0.2
0 0
-2 0 2 4 -2 0 2 4
Figure 87: Graph of h1 (t) (on the left) and h2 (t) (on the right) for problem 1g of second set
of Chapter 5
146
1
2. a. H(f ) = , <f <
1 + f2
The graph of H(f ) is given in Figure 88. The closest pair is
2
G(f ) = , g(t) = e|t| .
2 + 4 2 f 2
With = 2 we have
H(f ) = G(f )
See Figure 89 for the graphs of G(f ) and g(t) and Figure 90 for h(t).
0.8
0.6
0.4
0.2
-4 -2 0 2 4
f
1
0.3
0.8
0.25
0.2 0.6
0.15
0.4
0.1
0.2
0.05
0
-4 -2 0 2 4 -2 -1 0 1 2
t
f
Figure 89: Graph of G(f ) (on the left) and g(t) (on the right) for problem 2a of second set
of Chapter 5
Thus
h(t) = F 1 [H(f )] = F 1 [G(f )] = e2|t|
147
3
2.5
1.5
0.5
0
-4 -2 0 2 4
t
148
2. b. H(f ) = e3jf e2|f | , <f <
0.8
0.6
0.4
0.2
0
-2 -1 0 1 2
f
0.8
0.6
0.4
0.2
4 2 0 2 4
Figure 91: Graph of G(f ) (on the left) and h(t) (on the right) for problem 2b of second set
of Chapter 5
149
{
2 , 3 < f < 3
2. c. H(f ) =
0 , otherwise
Note that H(f ), given in Figure 92, is real and even and we can nd h(t) directly or from
the tables.
1.5
0.5
0
-4 -2 0 2 4
3 3
2jf t 2jf t
h(t) = H(f )e df = 2e df = 2 2 cos(2f t)df
3 0
So
sin(2f t) f =3 sin(6t)
h(t) = 4 = 2
2f f =0 t
{
1/(2f0 ), f0 < f < f0
The other way is to note that H(f ) = 12G(f ) where G(f ) =
0, otherwise
is given in Figure 93 with f0 = 3. Thus
So
sin(2(3)t) sin(6t)
h(t) = 12 = 2
2(3)t t
as before.
150
12
0.16
10
0.12 8
0.08
4
2
0.04
0
-4 -2 0 2 4
-2 t
0
-4 -2 0 2 4
Figure 93: Graph of G(f ) (on the left) and h(t) (on the right) for problem 2c of second set
of Chapter 5
151
6 Applications of the Fourier Transform
6.1 Introduction
152
6.2 Convolution and Fourier Transforms
PROBLEMS
1. Compute, using the denition, the convolution ( h(t) g(t) ) in the following cases. Then,
in each case, compute the Fourier transform of the convolution and verify the result agrees
with the convolution theorem:
{
2 , 0<t<2
a. h(t) =
0 , otherwise
{
et , 0<t
g(t) =
0 , otherwise
{
2 , 2 < t < 2
b. h(t) = g(t) =
0 , otherwise
153
{
2 , 0<t<2
1. a. h(t) =
0 , otherwise
{
et , 0<t
g(t) =
0 , otherwise
The graphs of h(t) and g(t) are given in Figure 94 and the graph of h(t u) is given in
Figure 95.
2 1
0.8
1.5
0.6
0.4
0.5
0.2
0 0
-2 -1 0 1 2 3 4 -2 -1 0 1 2 3 4
Figure 94: Graph of h(t) (on the left) and g(t) (on the right) for problem 1a of Chapter 6.2
h(tu)
u
t2 t
Three possible cases are described below. Case I (Figure 96) when t < 0 and there is no
overlap, Case II (Figure 97) when 0 < t < 2 and there is an overlap for 0 < u < t, Case III
(Figure 98) when 2 < t and there is an overlap for t 2 < u < t. The convolution
0 case I
( )
t
gh = g(u)h(t u)du = 2eu du = 2 1 et case II
0 ( )
t
2eu du = 2 e(t2) et case III
t2
154
t<0
u
t
Figure 96: Graph of g and h for problem 1a of Chapter 6.2 when t < 0
0<t<2
u
t
Figure 97: Graph of g and h for problem 1a of Chapter 6.2 when 0 < t < 2
2<t
u
t
Figure 98: Graph of g and h for problem 1a of Chapter 6.2 when 2 < t
1.6
1.2
0.8
0.4
0
-2 0 2 4 6
155
By denition
F[g h] = g he2jf t dt
2 ( ) ( )
F[g h] = 2 1 et e2jf t dt + 2 e(t2) et e2jf t dt
0 2
{ }
(1+2jf )t (1+2jf )t
e2jf t 2 e(1+2jf )t 2 e +2
e
F[g h] = 2 + 2e
2
2jf 0 1 + 2jf 0 1 + 2jf 2 1 + 2jf 2
{ }
1 e4jf e2(1+2jf ) 1 e2(1+2jf ) e2(1+2jf )
F[g h] = 2 + + e2
2jf 2jf 1 + 2jf 1 + 2jf 1 + 2jf 1 + 2jf
{ }
1 e4jf 1 e4jf )
F[g h] = 2 +
2jf 2jf 1 + 2jf 1 + 2jf
{ }
( ) 1 1
4jf
F[g h] = 2 1 e
2jf 1 + 2jf
1 e4jf
F[g h] = 2
2jf (1 + 2jf )
By convolution theorem
2 1 e4jf
H(f ) = 2e2jf t dt = 2
0 2jf
1
G(f ) =
1 + 2jf
therefore
1 e4jf
F[g h] = G(f )H(f ) = 2
2jf (1 + 2jf )
156
{
2 , 2 < t < 2
1. b. h(t) = g(t) =
0 , otherwise
The graphs of h(t) and g(t) are the same and given in Figure 100 and the graph of h(tu)
is given in Figure 101.
1.5
0.5
0
-4 -2 0 2 4
u
t2 t+2
Four possible cases are described below. Case I (Figure 102) when t < 4 and there is no
overlap, Case II (Figure 103) when 4 t 0 and there is an overlap for 2 < u < t + 2,
Case III (Figure 104) when 0 < t < 4 and there is an overlap for t 2 < u < 2, case IV
(Figure 105) when 4 < t and there is no overlap again. The convolution
0 case I
t+2
4du = 4(t + 4) case II
gh = g(u)h(t u)du = 2
2
4du = 4(2 (t 2)) = 4(4 t)) case III
t2
0 case IV
157
t < 4
u
t+2 2 2
Figure 102: Graph of h and h for problem 1b of Chapter 6.2 when t < 4
4 < t < 0
u
2 t+2 2
0<t<4
u
2 2 t+2
Figure 104: Graph of g and h for problem 1b of Chapter 6.2 when 0 < t < 4
4<t
u
2 2 t+2
Figure 105: Graph of g and h for problem 1b of Chapter 6.2 when 4 < t
158
By denition
F[g h] = g he2jf t dt
0 4
F[g h] = 4(4 + t)e2jf t dt + 4(4 t)e2jf t dt
4 0
sin2 (4f )
F[g h] = 4
(f )2
By convolution theorem
sin(4f )
H(f ) = G(f ) = 2
f
therefore
sin2 (4f )
F[g h] = G(f )H(f ) = 4
(f )2
16
12
0
-6 -4 -2 0 2 4 6
159
1. c. h(t) = e|t| , <t<
g(t) = cos(2t) , < t <
The graphs of h(t) and g(t) are given in Figure 107 and the graph of h(t u) is given in
Figure 108.
1 1
0.8
0.5
0.6
0
-2 -1 0 1 2
0.4
-0.5
0.2
0 -1
-4 -2 0 2 4
Figure 107: Graph of h(t) (on the left) and g(t) (on the right) for problem 1c of Chapter 6.2
u
t
By denition
gh = g(u)h(t u)du
160
u
0.04
0.02
0
-2 -1 0 1 2
-0.02
-0.04
By denition
F[g h] = g he2jf t dt
[ ]
2
F[g h] = F cos(2t)
1 + 4 2
1
F[g h] = [(f 1) + (f + 1)]
1 + 4 2
By convolution theorem
1
H(f ) = 2
1 + 4 2 f 2
1
G(f ) = [(f 1) + (f + 1)]
2
therefore
1
F[g h] = G(f )H(f ) = [(f 1) + (f + 1)]
1 + 4 2
161
Note that this last step uses the fact that (x)(x x0 ) (x0 )(x x0 ).
162
6.3 Linear, Shift-Invariant Systems
6.4 Determining a Systems Impulse Response and Transfer Func-
tion
6.5 Applications of Convolution - Signal Processing and Filters
6.6 Applications of Convolution - Amplitude Modulation and Fre-
quency Division Multiplexing
6.7 The DAlembert Solution Revisited
6.8 Dispersive Waves
6.9 Correlation
163
6.10 Summary
PROBLEMS
dI
L + RI = E(t)
dt
Vout = RI
a. Directly determine, e.g. using the Laplace transform, the impulse response of this
system. Sketch this response.
b. (1) Find the Transfer function of this system by computing the Fourier transform of
the impulse response determined in part a. above.
b. (2) Show that the alternative method of nding the Transfer function, i.e. as the
response of a system to the forcing function e2jf0 t , produces the same result as in part (1),
c. Sketch the amplitude and phase spectra of the Transfer function computed in part b.
2. Repeat problem 1 for the same circuit, except with the output taken as the voltage across
the inductor, i.e.
dI
L + RI = E(t)
dt
Vout = L dI
dt
164
1. a.
dI
L + RI = E(t)
dt
Vout = RI
Laplace transform: Let us denote L[I(t)] = I(s),
then
+ RI(s)
sLI(s) =1
Thus
1 R R 1
I(s) = Vout (s) = =
sL + R sL + R L s + (R/L)
So {
R Rt/L
L
e , t>0
Vout (t) =
0, t<0
This Vout is plotted in Figure 111.
R/L
t
L/R
165
1. b. (1)
Transfer function
R Rt/L 2jf t R 1
H(f ) = h(t)e2jf t dt = e e dt =
L 0 L R/L + 2jf
R
H(f ) =
R + 2jf L
dI
L + RI = e2jf0 t
dt
Using undetermined coecients, we substitute I(t) = A(f0 )e2jf0 t in the equation and we
have
(2jf0 LA + RA)e2jf0 t = e2jf0 t
or
1
A(f0 ) =
R + 2jf0 L
Therefore
1
I(t) = e2jf0 t
R + 2jf0 L
and
R
Vout = RI = e2jf0 t = H(f0 )e2jf0 t
R + 2jf0 L
Therefore
R
H(f0 ) =
R + 2jf0 L
or
R
H(f ) =
R + 2jf L
1. c.
We rewrite H(f ) as
R2 2f RL
H(f ) = 2 2
j 2
R + (2f L) R + (2f L)2
Therefore
R
|H(f )| =
R2 + (2f L)2
The plots of |H(f )| and (f ) are given in Figures 112 and 113.
166
|H(f)|
f
R/L R/L
1.5
0.5
0
-2 -1 0 1 2
-0.5
-1
-1.5
167
2. a.
dI
L + RI = E(t)
dt
Vout = L dI
dt
Impulse response:
dI
L + RI = (t)
dt
Vout = L dI
dt
Laplace transform
+ RI(s)
sLI(s) =1
Vout (s) = sLI(s)
or
1 sL
I(s) = Vout (s) =
sL + R sL + R
Now we can rewrite
R R 1
Vout (s) = 1 =1
sL + R L s + (L/R)
so
R Rt/L
Vout (t) = (t) e
L
This function is given in Figure 114.
h(t)
L/R
t
R/L
168
2. b. (1)
[ ]
R Rt/L
H(f ) = F[h(t)] = F (t) e
L
R 2jf L
H(f ) = 1 =
R + 2jf L R + 2jf L
2. b. (2)
dI
L + RI = e2jf0 t
dt
As before
1
I(t) = e2jf0 t
R + 2jf0 L
dI 2jf0 L
Vout = L = e2jf0 t
dt R + 2jf0 L
| {z }
H(f0 )
So
2jf L
H(f ) =
R + 2jf L
same as before.
2. c.
We rewrite H(f ) as
(2f L)2 2f LR
H(f ) = 2 2
+j 2
R + (2f L) R + (2f L)2
Therefore
2L|f |
|H(f )| =
R2 + (2f L)2
and
R
tan (f ) =
2f L
The |H(f )| and (f ) are given in Figures 115 and 116, respectively.
169
1
0.8
0.6
0.4
0.2
-2 -1 0 1 2
1.5
0.5
0
-2 -1 0 1 2
-0.5
-1
-1.5
170
7 Appendix A - Bessels Equation
7.1 Bessels Equation
7.2 Properties of Bessel Functions
PROBLEMS
1. Using the recurrence formulas, and a table of values for J0 (x) and J1 (x), nd
a. J1 (x) in terms of J0 (x) and J1 (x)
b. J2 (2.0)
c. J3 (1.0)
171
1. a. Using the recurrence relation N JN (x) + xJN (x) = xJN 1 (x) with N = 1, we have
Therefore
xJ0 (x) J1 (x)
J1 (x) =
x
2N
1. b. Using the recurrence relation JN +1 = JN (x) JN 1 (x) with N = 1, we have
x
2
J2 = J1 (x) J0 (x)
x
From CRC tables (or Maple, Matlab), J0 (2.0) = .2239, J1 (2.0) = .5767 therefore
2
J2 (2.0) = .0(.5767) .2239 = .3528
2
1. c. Using the recurrence relation N JN (x) + xJN (x) = xJN 1 (x) with N = 3 we have
But
2N
JN +1 = JN (x) JN 1 (x)
x
with N = 2 yields
4
J3 (x) = J2 (x) J1 (x)
x
Combining these two, we get
( )
4
3 J2 (x) J1 (x) + xJ3 (x) = xJ2 (x)
x
So [( ) ]
1 12
J3 (x) = x J2 (x) + 3J1 (x)
x x
[( )( ) ]
1 12 2
J3 (x) = x J1 (x) J0 (x) + 3J1 (x)
x x x
Thus
J3 (1.0) = [11 (2J1 (1.0) J0 (1.0)) + 3J1 (1.0)]
J3 (1.0) = 19J1 (1.0) + 11J0 (1.0) = 19(.4401) + 11(.7652) = .0553
172
2. a. The general solution of x2 y + xy + ( 2 x2 n2 ) y = 0 is
C1 Jn (x) + C2 Yn (x)
C1 J0 (2x) + C2 Y0 (2x)
C1 J2 (3x) + C2 Y2 (3x)
1
2. c. The equation 4x2 y + 4xy + (x2 1)y = 0 matches the general case with n =
2
1
and = , therefore the solution is
2
( ) ( )
1 1
C1 J1/2 x + C2 Y1/2 x
2 2
173
7.3 Variants of Bessels Equation
PROBLEMS
Use Bessel Functions (the big ugly equation) to nd the general solution to each ODE
below.
1. ( )
x2 y + 3 xy + 3 + 4 x4 y = 0
2. ( ) ( )
x2 y + x + 2 x2 y + 4 + 9 x2 + x + x2 y = 0
3. ( )
x2 y 5 xy + 9 + 4 x2 y = 0
4. ( ) ( )
x2 y + x 2 x3 y + 1/4 + x2 2 x2 + x4 y = 0
174
Answers
1.
y(x) = A J1 (x2 )/x + B Y1 (x2 )/x
2.
y(x) = ex (A J2 (3/x) + B Y2 (3/x))
3.
y(x) = x3 (A J0 (2x) + B Y0 (2x))
4. ( )
2 /2
y(x) = ex A J1/2 (1/x) + B Y1/2 (1/x)
175