Professional Documents
Culture Documents
Lathi Ls SM Chapter 5
Lathi Ls SM Chapter 5
5.1-1. Given the fact that the time-domain signal is finite in duration, the region of con-
5-1
vergence should include the entire z-plane, except possibly z = 0 or z = ∞.
P∞ P7 P7 8
X(z) = n=−∞ x[n]z −n = n=0 (−1)n z −n = n=0 (−1/z)n = 1−(−1/z) 1−(−1/z) . Thus,
1 − z −8
X(z) = ; ROC |z| > 0.
1 + z −1
In this form, X(z) appears to have eight finite zeros and one finite pole. The eight
zeros are the eight roots of unity, or z = e2πk/8 for k = (0, 1, . . . , 7). The apparent
pole is at z = −1. However, there is also a zero z = −1 (k = 4) that cancels this
pole. Thus, there are actually no finite poles and only seven finite zeros, z = e2πk/8
for k = (0, 1, 2, 3, 5, 6, 7). MATLAB is used to plot the zeros in the complex plane; the
unit circle is also plotted for reference.
0.8
0.6
0.4
0.2
Im(z)
−0.2
−0.4
−0.6
−0.8
−1
−1 −0.5 0 0.5 1
Re(z)
5.1-2. (a)
∞
X
X [z] = z −n = z −m + z −(m+1) + z −(m+2) + · · ·
n=m
213
190
Therefore
X [z] = eγ/z
(k)
£ n−1 ¤
x[n] = z − (−2)n−1 u[n]
"∞ µ ¶ ∞ µ ¶n #
1 X 2 X
n
−2
X[z] = +
2 n=0 z n=0
z
· ¸
1 1 1
= +
2 1 − z2 1 + z2
· ¸
1 z z
= +
2 z−2 z+2
z2
=
z2 − 4
(l)
X∞ X∞ µ ¶n
1 1 ln α
X [z] = (ln α)n z −n =
n=0
n! n=0
n! z
∞
X 6
n(−3/2)−n = − = −0.24.
n=0
25
5.1-4.
5-2 (a)
Hence
1 z+1
u[n] − u[n − 2] ⇐⇒ 1 + =
z z
(b)
1
γ n−2 u[n − 2] = {γ n u[n] − δ[n] − γδ[n − 1]}
γ2
Hence · ¸
1 z γ 1
γ n−2 u[n − 2] ⇐⇒ − 1 − =
γ2 z − γ z z(z − γ)
(c)
1
x [n] = (2)n+1 u [n − 1] + (e)n−1 u [n] = 4(2)n−1 u [n − 1] + (e)n u [n]
e
216
191
Therefore
4 1 z
X [z] = +
z−2 ez−e
(d)
h πn i πn
x [n] = (2)−n cos u [n − 1] = (2)−n cos u [n] − δ [n]
3 3
Therefore
z(z − 0.25) 0.25(z − 1)
X [z] = −1= 2
z 2 − 0.5z + 0.25 z − 0.5z + 0.25
(e)
Therefore
γz
X [z] =
(z − γ)2
(f) Because n(n − 1)(n − 2) = 0 for n = 0, 1, and 2
n = 0, 1, or 2. Therefore
and · ¸
−3 3!(2)3 z 6z
X [z] = (2) =
(z − 2)4 (z − 2)4
(g)
1 2 3 4 5
X[z] = 0− + 2 − 3 + 4 − 5 + ···
z z z z z
X∞ µ ¶n X ∞ µ ¶2n−1
1 1
= 2n 2
− (2n + 1) 2
n=0
z n=0
z
X∞ µ ¶n ∞ µ ¶n ∞ µ ¶n
1 2X 1 1X 1
= 2 n − n −
n=0
z2 z n=0 z2 z n=0 z 2
2z 2 2 z2 z
X[z] = 2 2
− 2 2
− 2
(z − 1) z (z − 1) z −1
−z 3 + 2z 2 − z
=
(z 2 − 1)2
−z(z − 1)2
=
(z + 1)2 (z − 1)2
−z
=
(z + 1)2
217
192
(h)
∞
X
x[n] = kδ[n − 2k + 1]
k=0
∞ ∞
̰ !
X X X
−n
X[z] = x[n]z = kδ[n − 2k + 1] z −n
n=0 n=0 k=0
We obtain
∞
à ∞
!
X X
−n
X[z] = k δ[n − 2k + 1]z
k=0 n=0
X∞
= kz −(2k−1)
k=0
X∞ µ ¶k
1
= z k
z2
k=0
1/z 2
= z 2
[(1/z)2 − 1]
z3
= 2
(z 2 − 1)
5.1-5. (a)
5-3
X [z] z−4 2 1
= = −
z (z − 2)(z − 3) z−2 z−3
z z
X [z] = 2 −
z−2 z−3
x [n] = [2(2)n − (3)n ] u [n]
(b)
(c)
X [z] e−2 − 2 1 1
= −2
= −2
−
z (z − e )(z − 2) z−e z−2
z z
X [z] = −
z − e−2 z−2
218
193
£ −2n ¤
x [n] = e − 2n u [n]
(d)
(z − 1)2 z 2 − 2z + 1 1 2 1
X[z] = 3
= 3
= − 2+ 3
z z z z z
Hence
x[n] = δ[n − 1] − 2δ[n − 2] + δ[n − 3]
(e)
(f)
X [z] −5z + 22 3 k 4
= 2
= + +
z (z + 1)(z − 2) z + 1 z − 2 (z − 2)2
0 = 3 + k + 0 =⇒ k = −3
z z z
X [z] = 3 −3 +4
z+1 z−2 (z − 2)2
x [n] = [3(−1)n − 3(2)n + 2n(2)n ] u [n]
(g)
0 = 1 + k =⇒ k = −1
z z z
X [z] = − +2
z − 0.2 z − 0.8 (z − 0.8)2
· ¸
5
x [n] = (0.2)n − (0.8)n + n(0.8)n u [n]
2
219
194
(i)
2 = 1 + A =⇒ A = 1
Setting z = 1 on both sides yields
1.95 1+B
=1+ =⇒ B = −0.9
1.85 1.85
z(z − 0.9)
X [z] = 1 +
z2
+ 0.6z + 0.25
For the second fraction on right side, we use pair 12c with A = 1, B = −0.9,
a = 0.3, and |γ| = 0.5. This yields
√ −0.3 1.2
r= 10 β = cos−1 ( ) = 2.214 θ = tan−1 ( ) = 1.249
0.5 0.4
√
x [n] = δ [n] + 10(0.5)n cos(2.214n + 1.249)u [n]
(j)
0 = −2 + A =⇒ A = 2
Set z = 0 on both sides to obtain
46 B
=2+ =⇒ B = −4
25 25
z z(2z − 4)
X [z] = −2 + + 2
z − 1 z − 6z + 25
For the second fraction on the right-hand side, we use pair 12c with A = 2,
B = −4, a = −3, and |γ| = 5.
√
17 3 −1
r= β = cos−1 ( ) = 0.927 θ = tan−1 ( ) = −0.25
2 5 4
" √ #
17 n
x [n] = −2 + (5) cos(0.927n − 0.25) u [n]
2
(k)
0 = 1 + A =⇒ A = −1
220
195
Setting z = 0 on both sides yields
11.34 1 B
=− + =⇒ B = 6.83
−50 2 25
z z(−z + 6.83)
X [z] = + 2
z−2 z − 5z + 25
For the second fraction on right-hand side, use pair 12c with A = −1, B = 6.83,
a = −2.5, and |γ| = 5.
√ π −4.33 3π
r= 2 β = cos−1 (0.5) = θ = tan−1 ( )=−
3 −4.33 4
· ¸
√ π 3π
x [n] = (2)n + 2(5)n cos( n − ) u [n]
3 4
(l)
X [z] z(−2z 2 + 8z − 7) 1 k1 k2 2
= 3
= + + 2
+
z (z − 1)(z − 2) z − 1 z − 2 (z − 2) (z − 2)3
−2 = 1 + k1 =⇒ k1 = −3
z z z z
X [z] = z−1 − 3 z−2 − (z−2) 2 + 2 (z−2)3
£ ¤
x [n] = 1 − 3(2)n − n2 (2)n + 41 n(n − 1)(2)n u [n]
5.1-6.
5-4 (a) Long division of 2z 3 + 13z 2 + z by z 3 + 7z 2 + 2z + 1 yields
1 4
X [z] = 2 − + 2 + ···
z z
Therefore x [0] = 2, x [1] = −1, x [2] = 4.
(b) Long division of 2z 4 + 16z 3 + 17z 2 + 3z by z 3 + 7z 2 + 2z + 1 yields
1 4
X [z] = 2z + 2 − + 2 + ···
z z
Therefore x [−1] = 2, x [0] = 2, x [1] = −1, x [2] = 4.
5.1-7.
5-5
γz
X [z] =
z2 − 2γz + γ 2
γz γ ³ γ ´2 ³ γ ´3
= + 2 + 3 + ···
z 2 − 2γz + γ 2 z z z
221
196
Therefore x [0] = 0, x [1] = γ, x [2] = 2γ 2 , x [3] = 3γ 3 , · · ·, and
x [n] = nγ n u [n]
5.1-8.
5-6 (a) We can express
X[1] X[2]
X[z] = X[0] + + 2 + ···
z z
Let Xn [z] and Xd [z] be the numerator and the denominator polynomials of X[z]
with powers M and N , respectively. If M = N , then the long division of Xn with
Xd in power series of z −1 yields x[0] = a nonzero constant. If N − M = 1, the
term x[0] = 0, but x[1], x[2], · · · are generally nonzero. In general, if N −M = m,
then the long division show that all x[0], x[1], · · · , x[m − 1] are zero. Only the
terms from x[m] on are generally nonzero. The difference N − M indicates that
the first N − M samples of x[n] are zero.
(b) In this case the first four samples of x[n] are zero. Hence N − M = 4.
1 1 1
5.2-1. (a) X[z] = 1 +
5-7 z + z2 +···+ z (m−1)
. This sum can be found using the result in Sec.
B.7-4, as
(1/z)m−1 − 1 1 − z −m
X[z] = =
(1/z) − 1 1 − z −1
(b)
x [n] = u [n] − u [n − m]
z z 1 − z −m
X [z] = − z −m =
z−1 z−1 1 − z −1
5.2-2.
5-8
x [n] = δ [n − 1] + 2δ [n − 2] + 3δ [n − 3] + 4δ [n − 4] + 3δ [n − 5] + 2δ [n − 6] + δ [n − 7]
Therefore
1 2 3 4 3 2 1
X [z] = + 2+ 3+ 4+ 5+ 6+ 7
z z z z z z z
z 6 + 2z 5 + 3z 4 + 4z 3 + 3z 2 + 2z + 1
=
z7
Alternate Method:
Therefore
z 2z z 2z z
X [z] = − 5 + 9 − 5 +
(z − 1)2 z (z − 1)2 z (z − 1)2 z (z − 1) z 9 (z − 1)
z £ 9 ¤
= 9 2
z − 2z 4 + 1 − 2z 4 (z − 1) + (z − 1)
z (z − 1)
222
197
1 £ 8 ¤
= z − 2z 4 + 1
z 7 (z − 1)2
x [n] = n2 γ n u [n]
z
Repeated application of Eq. (5.21) to γ n u [n] ⇐⇒ z−γ yields
γz
nγ n u [n] ⇐⇒
(z − γ)2
γz(z + γ)
n2 γ n u [n] ⇐⇒
(z − γ)3
z(z+1)
Let γ = 1 n2 u [n] ⇐⇒ (z−1)3
(b) Consider
z
γ n u[n] ⇐⇒
z−γ
Application of multiplication property to this pair yields
µ ¶
n d z γz
nγ u[n] ⇐⇒ −z =
dz z − γ (z − γ)2
γz(z+γ)
(c) Application of Eq. (5.21) to n2 γ n u [n] ⇐⇒ (z−γ)3 (found in part a) yields
· ¸
d γz(z + γ) γz(z 2 + 4γz + γ 2 )
n3 γ n u [n] = −z 3
=
dz (z − γ) (z − γ)3
z(z 2 + 4z + 1)
n3 u [n] =
(z − 1)3
(d)
(e)
223
198
· ¸
e−2m 1
= (1 − m) + mz
z(z − e−2 )2 e2
(f)
5.2-4.
5-10 Pair2:
u [n] = δ [n] + δ [n − 1] + δ [n − 2] + δ [n − 3] + · · ·
1 1 1 1 z
u [n] ⇐⇒ 1 + + 2 + 3 + ··· = 1 =
z z z 1− z
z−1
Repeated application of Eq. (5.21) to pair 2 yields pair 3, 4, and 5. Application of Eq.
(5.20) to pair 2 yields pair 6, and application of time-delay (5.15b) to pair 6 yields
pair 7. Repeated application of Eq. (5.21) to pair 7 yields pair 8 and 9.
5.2-5.
5-11 There is a typographic error in the book problem: cos(πn/A) should be cos(πn/4).
Given (Pair 11b with |γ| = 1)
z sin β
sin βnu[n] ⇐⇒
z 2 − (2 cos β)z + 1
³ π´
cos βnu[n] = − sin βn − u[n]
2
and
π ³ π´
cos nu[n] = − sin βn − u[n]
4 hπ 2 i
= − sin (n − 2) u[n]
hπ4 i ³ π´ ³ π´
= − sin (n − 2) u[n − 2] + sin − δ[n] + sin − δ[n − 1]
4 2 4
hπ i 1
= − sin (n − 2) u[n − 2] + δ[n] + √ δ[n − 1]
4 2
From Pair 11b, and shift property
hπ i √
z/ 2 1
sin (n − 2) u[n − 2] ⇐⇒ √ =√ √
4 2 2
z (z − 2z + 1) 2
2z(z − 2z + 1)
224
199
Therefore
π −1 1 z(z − √12 )
cos nu[n] ⇐⇒ √ √ +1+ √ = √
4 2z(z 2 − 2z + 1) 2z (z 2 − 2z + 1)
5.2-6.
5-12 Application of time reversal to pair 6 yiels
1/z 1
β −n u[−n] ⇐⇒ = |z| < 1/β
1/z − β 1 − βz
Moreover
β −n u[−n − 1] = β −n u[−n] − δ[n]
Hence
1 βz
β −n u[−n − 1] ⇐⇒ −1= |z| < 1/β
1 − βz 1 − βz
Setting β = 1/γ, we obtain
−z
γ n u[−n − 1] = |z| < |γ|
z−γ
5.2-7.
5-13 (a)
X∞ ∞
(−1)n x[n] X x[n]
(−1)n x[n] ⇐⇒ = = X[−z]
n=0
zn n=0
(−z)n
(c) (i)
1 n 1 z
2n−1 u[n] = 2 u[n] ⇐⇒
2 2z−2
Application of (b) to this result yields
1 1 z
(−2)n−1 u[n] = − (−2)n u[n] ⇐⇒ −
2 2z+2
Hence
· ¸
£ n−1 n−1
¤ 1 z z z2
2 − (−2) u[n] ⇐⇒ + = 2
2 z−2 z+2 z −4
(ii) To find
γ n cos πnu[n] = (−γ)n u[n]
Hence
z
γ n cos πnu[n] ⇐⇒
z+γ
225
200
5.2-8.
5-14 (a)
n
X n
X
x[k] = x[k]u[n − m] = x[n] ∗ u[n]
k=0 k=0
(b) Let x[n] = δ[n], which yields X[z] = 1. Application of the result in part (a) yields
n
X z
δ[k] = u[n] ⇐⇒
z−1
k=0
2
z
5.2-9.
5-15 (a) In the time-domain, (z−0.75) 2 is a convolution of two causal, decaying exponen-
tials. Thus, either plot 1 or plot 13 is possible. Using the IVT, the initial value
z2
is limz→∞ (z−0.75) 2 = 1. Thus,
z2
Plot 13 corresponds to .
(z − 0.75)2
√
z 2 −0.9z/ 2
(b) In the time-domain, z 2 −0.9
√
2z+0.81
is a decaying sinusoid. Thus, either plot 8 or
2
√
z −0.9z/ 2
plot 9 is possible. Using the IVT, the initial value is limz→∞ z2 −0.9√
2z+0.81
= 1.
Thus, √
z 2 − 0.9z/ 2
Plot 9 corresponds to √ .
z 2 − 0.9 2z + 0.81
P4
(c) Note that k=0 z −2k = 1 + z −2 + z −4 + z −6 + z −8 . Thus, the time-domain signal
is δ[n] + δ[n − 2] + δ[n − 4] + δ[n − 6] + δ[n − 8] and
4
X
Plot 18 corresponds to z −2k .
k=0
z −5
(d) By inspection, 1−z −1 corresponds to a unit step that is shifted to the right by
five. Thus,
z −5
Plot 10 corresponds to .
1 − z −1
2
(e) Using synthetic division on z4z−1 yields (z −2 + z −6 + z −10 + . . .). In the time-
domain, the first non-zero term therefore occurs at n = 2. Thus,
z2
Plot 15 corresponds to .
z4 − 1
0.75z
(f) In the time-domain, (z−0.75) 2 is a convolution of two causal, decaying exponen-
tials. Thus, either plot 1 or plot 13 is possible. Using the IVT, the initial value
226
201
0.75z
is limz→∞ (z−0.75)2 = 0. Thus,
0.75z
Plot 1 corresponds to .
(z − 0.75)2
√
z 2 −z/ 2
(g) In the time-domain, 2
√
z − 2z+1
has the form of a sinusoid. Thus, either plot 3
√
z 2 −z/ 2
or plot 4 is possible. Using the IVT, the initial value is limz→∞ √
z 2 − 2z+1
= 1.
Thus, √
z 2 − z/ 2
Plot 4 corresponds to √ .
z 2 − 2z + 1
−1 −5 −6
(h) The apparent repeated root at z = 1 of z 5(1−z −5z +4z
−1 )2 suggests a signal that
grows linearly in the time-domain. Thus, plot 6 or plot 12 are possible. To dis-
tinguish between the two, first determine whether or not a root at z = 1 really
−1
−5z −5 +4z −6 z 5 −5z+4 z 5 −5z+4
exists. First, limz→1 z 5(1−z −1 )2 = limz→1 5(z 3 −z 2 )2 = limz→1 5(z 6 −2z 5 +z 4 ) =
z −1 −5z −5 +4z −6
0/0. This is indeterminant so use L’Hospital’s rule: limz→1 5(1−z −1 )2 =
d 5
dz (z −5z+4) 5z 4 −5
limz→1 d
(5z 6 −10z 5 +5z 4 ) = limz→1 30z5 −50z 4 +20z 3 = 0/0. This too is
dz
−1
z −5z −5 +4z −6
indeterminant so use L’Hospital’s rule again: limz→1 5(1−z −1 )2 =
d2
(z 5 −5z+4) 20z 3
dz 2
limz→1 d2
= limz→1 150z 4 −200z 3 +60z 2 = 20/10 = 2. Since
(5z 6 −10z 5 +5z 4 )
dz 2
−1 −5 −6
z −5z +4z
limz→1 5(1−z −1 )2 = 2 6= ∞, no root exists at z = 1 and
z −1 − 5z −5 + 4z −6
Plot 6 corresponds to .
5(1 − z −1 )2
z
(i) In the time-domain, z−1.1 is a growing exponential due to the root outside the
unit circle. Thus, plot 16 or plot 17 is possible. Using the IVT, the initial value
z
is limz→∞ z−1.1 = 1. Thus,
z
Plot 16 corresponds to .
z − 1.1
−1
(j) In the time-domain, (1−z−10.25z
)(1−0.75z −1 ) is the convolution of a decaying sinusoid
and a unit step. Thus, plot 7 or plot 11 are possible. Using the IVT, the initial
−1
value is limz→∞ (1−z−10.25z
)(1−0.75z −1 ) = 0. Thus,
0.25z −1
Plot 7 corresponds to .
(1 − z −1 )(1 − 0.75z −1 )
5.3-1.
5-16
y [n + 1] − γy [n] = x [n + 1]
P
X [z] =
z−1
y [n] ⇐⇒ Y [z] y [n + 1] ⇐⇒ zY [z] + M z
227
202
The z-transform of the system equation is
Pz
zY [z] + M z − γY [z] =
z−1
Pz
(z − γ)Y [z] = −M +
z−1
and
−M z Pz
Y [z] = +
z−γ (z − γ)(z − 1)
· ¸
Y [z] −M P −M P 1 1
= + = + −
z z−γ (z − γ)(z − 1) z−γ γ−1 z−γ z−1
· ¸
z P z z
Y [z] = −M + −
z−γ γ−1 z−γ z−1
· ¸
P (γ n − 1)
y [n] = −M γ n + u [n] r =γ−1
r
The loan balance is zero for n = N , that is, y [N ] = 0. Setting n = N in the above
equation we obtain · ¸
N P (γ N − 1)
y [N ] = −M γ + =0
r
This yields
rγ N rM
P = M=
γN −1 1 − (1 + r)−N
5.3-2. Because part (b) requires us to separate the response into zero-input and zero-state
5-17
components, we shall start with the delay operator form of the equations, as
To determine the initial condition y[−1], we set n = 0 in this equation and substitute
y[0] = 1 to obtain
The term (1 − e) on the right-hand side is due to the initial condition, and hence rep-
resents the zero-input component. the second term on the right-hand side represents
the zero-state component of the response. Thus
· ¸
Y [z] 1−e ez
= +
z z+2 (z − e−1 )(z + 2)
228
203
1−e 2e2 e
= + +
z + 2 (2e + 1)(z + 2) (2e + 1)(z − e−1 )
Hence
z 2e2 z e z
Y [z] = (1 − e) + +
z + 2 2e + 1 z + 2 2e + 1 z − e−1
The first term on the right-hand side is the zero-input component and the remaining
two terms represent the zero-state component. Thus
2e2 e
y[n] = (1 − e)(−2)n u[n] + (−2)n u[n] + e−n u[n]
| {z } 2e + 1 2e + 1
| {z }
zero-input comp. zero-state comp.
5.3-3.
5-18 (a) The system equation in delay form is
2y [n] − 3y [n − 1] + y [n − 2] = 4x [n] − 3x [n − 1]
Also
1 1
y [n] ⇐⇒ Y [z] y [n − 1] ⇐⇒ Y [z] y [n − 2] ⇐⇒ Y [z] + 1
z z2
z 1
x [n] ⇐⇒ X [z] = x [n − 1] ⇐⇒
z − 0.25 z − 0.25
The z-transform of the equation is
3 1 4z 3 4z − 3
2Y [z] − Y [z] + 2 Y [z] + 1 = − =
z z z − 0.25 z − 0.25 z − 0.25
or µ ¶
3 1 4z − 3 3z − 2.75
2 − + 2 Y [z] = −1 + =
z z z − 0.25 z − 0.25
and
Y [z] z(3z − 2.75)
=
z (2z 2 − 3z + 1)(z − 0.25)
z(3z − 2.75)
=
2(z − 0.5)(z − 1)(z − 0.25)
5/2 1/3 4/3
= + −
z − 1/2 z − 1 z − 0.25
· ¸
1 5 4
y [n] = + (0.5)n − (0.25)n u [n]
3 2 3
· ¸
1 5 −n 4 −n
= + (2) − (4) u [n]
3 2 3
229
204
(b) From part (a), we have
µ ¶
3 1 4z − 3
2 − + 2 Y [z] = −1 +
z z |{z} z − 0.25
I.C. term | {z }
input term
2z 2 − 3z + 1 4z − 3
Y [z] = −1 +
z2 z − 0.25
and
Y [z] −z z(4z − 3)
= +
z 2(z − 0.5)(z − 1) 2(z − 0.5)(z − 1)(z − 0.25)
| {z } | {z }
z−i z−s
0.5 1 2 4 1 4 1
= − + + −
z − 0.5 z − 1 z − 0.5 3 z − 1 3 z − 0.25
and
z z z 4 z 4 z
Y [z] = 0.5 − +2 + −
z − 0.5 z − 1 z − 0.5 3 z − 1 3 z − 0.25
and
· ¸ · ¸
1 4 4
y[n] = (0.5)n − 1 u[n] + 2(0.5)n + − (0.25)n u[n]
2 3 3
| {z } | {z }
yzi yzs
· ¸
n 4 n 1
= 2.5(0.5) − (0.5) u[n] + u[n]
3 3
| {z } | {z }
ytr yss
5-19 For initial conditions y [0], y [1], we require equation in advance form:
5.3-4.
2y [n + 2] − 3y [n + 1] + y [n] = 4x [n + 2] − 3x [n + 1]
Also
3 3 35
y [n] ⇐⇒ Y [z] y [n + 1] ⇐⇒ zY [z] − z y [n + 2] ⇐⇒ z 2 Y [z] − z 2 − z
2 2 4
z 0.25z
x [n] ⇐⇒ X [z] = x [n + 1] ⇐⇒ zX [z] − z =
z − 0.25 z − 0.25
and
1 z
x [n + 2] ⇐⇒ z 2 X [z] − z 2 − z =
4 16(z − 0.25)
The z-transform of the equation is
· ¸ · ¸
3 35 3 −z/2
2 z 2 Y [z] − z 2 − z − 3 zY [z] − z + Y [z] =
2 4 2 z − 0.25
or
z(3z 2 + 12.25z − 3.75)
(2z 2 − 3z + 1)Y [z] =
(z − 0.25)
230
205
and
Y [z] 3z 2 + 12.25z − 3.75 46/3 4/3 25/2
= = − −
z 2(z − 0.25)(z − 1)(z − 0.5) z − 1 z − 0.25 z − 0.5
46 z 4 z 25 z
Y [z] = − −
3 z − 1 3 z − 0.25 2 z − 0.5
· ¸
46 4 n 25 n
y [n] = − (0.25) − (0.5) u [n]
3 3 2
4y [n] + 4y [n − 1] + y [n − 2] = x [n − 1]
Also
1 1
y [n] ⇐⇒ Y [z] y [n − 1] ⇐⇒ Y [z] y [n − 2] ⇐⇒ Y [z] + 1
z z2
z 1
x [n] ⇐⇒ x [n − 1] ⇐⇒ (x [−1] = 0)
z−1 z−1
The z-transform of the system equation is
4 1 1
4Y [z] + Y [z] + 2 Y [z] + 1 = (1)
z z z−1
4z 2 + 4z + 1 2−z
2
Y [z] = (2)
z z−1
and
Y [z] z(2 − z) z(2 − z)
= =
z 4(z − 1)(z 2 + z + 0.25) 4(z − 1)(z + 0.5)2
· ¸
1 4/9 13/9 5/6
= − +
4 z − 1 z + 0.5 (z + 0.5)2
· ¸
1 4 z 13 z 5 z
Y [z] = − +
4 9z−1 9 z + 0.5 6 (z + 0.5)2
· ¸
1 13 5
y [n] = − (−0.5)n − n(−0.5)n u [n]
9 36 12
(b) To find the zero-input and the zero-state components, we observe that the only
term arising because of the initial conditions is 1 on the left-hand side of Eq. (1).
Hence, we can rewrite Eq. (2) with explicit zero-input and zero-state components
as
4z 2 + 4z + 1 1
Y [z] = −1 +
z2 z−1
Here the term −1 on the right-hand side represents the zero-input terms and
the second term on the right-hand side represents the zero-state component.
Rearranging the equation, we obtain
· ¸
Y [z] z 1
= −1 +
z 4(z + 0.5)2 z−1
231
206
−z z
= +
4(z + 0.5)2 4(z − 1)(z + 0.5)2
| {z } | {z }
zero-input zero-state
−1/4 1/8 1/9 1/9 1/12
= + 2
+ − +
z + 0.5 (z + 0.5) z − 1 z + 0.5 (z + 0.5)2
| {z } | {z }
zero-input zero-state
Therefore
(−1/4)z (1/8)z (1/9)z (1/9)z (1/12)z
Y [z] = + 2
+ − +
z + 0.5 (z + 0.5) z−1 z + 0.5 (z + 0.5)2
| {z } | {z }
zero-input zero-state
n
−1 n(−0.5)n 1 (−0.5)n n(−0.5)
y[n] = (−0.5)n − + − − u[n]
| 4 {z 4 } |9 9 {z 6 }
zero-input zero-state
(c) The terms which vanish as n → ∞ correspond to the transient component and
the terms which do not vanish correspond to the steady-state component. Hence
· ¸
13 5 1
y[n] = − (−0.5)n − n(−0.5)n u[n] + u[n]
36 12 |9 {z }
| {z }
transient steady-state
y [n] − 3y [n − 1] + 2y [n − 2] = x [n − 1]
Also
1 1 2
y [n] ⇐⇒ Y [z] y [n − 1] ⇐⇒ Y [z] + 2 y [n − 2] ⇐⇒ 2
Y [z] + + 3
z z z
1
x [n] ⇐⇒ X [z] x [n − 1] ⇐⇒ X [z]
z
z
X [z] =
z−3
The z-transform of the system equation is
· ¸ · ¸
1 1 2 1
Y [z] − 3 Y [z] + 2 + 2 2 Y [z] + + 3 =
z z z z−3
µ ¶
3 2 4 1 −3z + 12
1 − + 2 Y [z] = − + =
z z z z−3 z(z − 3)
232
207
9 z z 3 z
Y [z] = −6 +
2z−1 z−2 2z−3
· ¸
9 3
y [n] = − 6(2)n + (3)n u [n]
2 2
y [n] − 2y [n − 1] + 2y [n − 2] = x [n − 2]
1 1 1
y [n] ⇐⇒ Y [z] y [n − 1] ⇐⇒ Y [z] + 1 y [n − 2] ⇐⇒ Y [z] +
z z2 z
1 z
x [n − 2] ⇐⇒ X [z] and X [z] =
z2 z−1
The z-transform of the difference equation is
· ¸ · ¸
1 1 1 1
Y [z] − 2 Y [z] + 1 + 2 2 Y [z] + =
z z z z(z − 1)
(z 2 − 2z + 2) 2z 2 − 4z + 3
Y [z] =
z2 z(z − 1)
Y [z] 2z 2 − 4z + 3 1 z−1
= = +
z (z − 1)(z 2 − 2z + 2) z − 1 z 2 − 2z + 2
z z(z − 1)
Y [z] = +
z − 1 z 2 − 2z + 2
For the second fraction on the right-hand side, we use pair 12c with A = 1, B = −1,
a = −1, |γ|2 = 2. This yields r = 1, β = π4 , and θ = 0. Therefore
h √ π i
y [n] = 1 + ( 2)n cos( n) u [n]
4
5.3-8.
5-23 The equation in advance form is
y [n + 2] + 2y [n + 1] + 2y [n] = x [n + 1] + 2x [n]
z2 2z z(e + 2)
z 2 Y [z] − z + 2zY [z] + 2Y [z] = −z+ =
z−e z−e z−e
z(e + 2) z(z + 2)
(z 2 + 2z + 2)Y [z] = z+ =
z−e z−e
Therefore
Y [z] z+2 0.318 −0.318z − 0.502
= = +
z (z − e)(z 2 + 2z + 2) z−e z 2 + 2z + 2
233
208
z z(0.318z + 0.502)
Y [z] = 0.318 −
z−e z 2 + 2z + 2
For the second fraction on the right-hand side, we use pair 12c with A = 0.318,
B = 0.502, a = 1, |γ|2 = 2 and
−1 3π −0.184
r = 0.367 β = cos−1 ( √ ) = θ = tan−1 ( ) = −0.525
2 4 0.318
· ¸
n
√ n 3π
y [n] = 0.318(e) − 0.367( 2) cos( n − 0.525) u [n]
4
2z/3
5.3-9. In transform domain, H(z) = z −1 z−1/3
5-24 and Y (z) = z −1 −2z
z+2 . Since Y (z) = H(z)X(z),
z −1 −2z
we know X(z) = Y (z)/H(z) = −1 2z/3 z+2
. Thus, X(z) = −3 z−1/3
z+2 . Using tables, x[n] =
z
¡ ¢
z−1/3
¡ ¢
−3 (−2)n u[n] − 31 (−2)n−1 u[n − 1] = −3 −2(−2)n−1 u[n] − 31 (−2)n−1 u[n − 1] or
5.3-10. Taking the z-transform of b[m] = (1.01)b[m − 1] + p[m] and solving for B(z) yields
5-25
1
B(z) = P (z) 1−1.01z −1 . Thus, P (z) is required to solve for b[m]. One way to represent
P∞
Sally’s deposit schedule is p[m] = 100 (u[m] − k=0 δ[m − (12k + 11)]). Defined this
way, Sally deposits one hundred dollars on the first day of every month m except for
Decembers, (m = 12k + 11 for k = {0, 1, 2, . . .}).
³ P∞ P∞ ´
Taking the z-transform yields P (z) = 100 1−z1 −1 − m=−∞ k=0 δ[m − (12k + 11)]z −m =
³ P∞ P∞ ´
100 1−z1 −1 − k=0 m=−∞ δ[m − (12k + 11)]z −m =
³ P ´
∞
100 1−z1 −1 − k=0 z −(12k+11) . Substituting P (z) into the expres-
³ P∞ ´
sion for B(z) yields B(z) = 100 1−z1 −1 − k=0 z −(12k+11) 1−1.01z 1
−1 =
³ P −(12k+11)
´ ³ P −(12k+11)
´
1 ∞ z 101 −100 ∞ z
100 (1−1.01z−1 −1
)(1−z ) + k=0 1−1.01z −1 = 100 1−1.01z −1 + 1−z −1 + k=0 1−1.01z −1 .
The first two terms are easily inverted using a table of z-transform pairs, while the
last sum is inverted using tables and the shifting property.
à ∞
!
X
m m−(12k+11)
b[m] = 100 101(1.01) u[m] − 100u[m] − (1.01) u[m − (12k + 11)] .
k=0
5.3-11.
5-26 (a) Note, h1 [n] = (−1 + (0.5)n ) u[n] = −(1)n u[n] + (1/2)n u[n]. Thus, two real poles
are evident at z = 1 and z = 1/2. Since h[n] is not absolutely summable, the
system is not BIBO stable. Thought of another way, the pole on the unit-circle
makes the system marginally stable, at best. Marginally stable systems are not
BIBO stable.
(b) Notice, h2 [n] = ()n (u[n] − u[n − 10]) is a finite duration, causal signal. Thus,
h2 [n] has no poles (other than at zero). Since h2 [n] is absolutely summable, the
system is BIBO stable.
5.3-12.
5-27 (a) Let
n
X
y[n] = k
k=0
234
209
then
y[n] − y[n − 1] = n with y[0] = 0
Setting n = 0 in this equation and y[0] = 0, yields
and
Y [z] z 1 1
= 3
= 2
+
z (z − 1) (z − 1) (z − 1)3
Hence
z z
Y [z] = +
(z − 1)2 (z − 1)3
and
n(n − 1) n(n + 1)
y[n] = n + =
2 2
(b) Let
n
X
y[n] = k2
k=0
then
y[n] − y[n − 1] = n2 with y[0] = 0
Setting n = 0, we get
0 − y[−1] = 0 =⇒ y[−1] = 0
The z-transform of the difference equation is
z−1 z(z + 1)
Y [z] =
z (z − 1)3
Hence
Y [z] z(z + 1) 1 3 2
= = + +
z (z − 1)4 (z − 1)2 (z − 1)3 (z − 1)4
z 3z 2z
Y [z] = + +
(z − 1)2 (z − 1)3 (z − 1)4
· ¸
3n(n − 1) n(n − 1)(n − 2)
y[n] = n+ + u[n]
2 3
2n3 + 3n2 + n n(n + 1)(2n + 1)
= u[n] = u[n]
6 6
5.3-13.
5-28 Let
n
X
y[n] = k3
k=0
then
y[n] − y[n − 1] = n3 with y[0] = 0
235
210
Setting n = 0 in this equation and using y[0] = 0, we get
0 − y[−1] = 0 =⇒ y[−1] = 0
z−1 z(z 2 + 4z + 1)
Y [z] =
z (z − 1)4
Y [z] z(z 2 + 4z + 1) 1 7 12 6
= 5
= 2
+ 3
+ 4
+
z (z − 1) (z − 1) (z − 1) (z − 1) (z − 1)5
z z z z
Y [z] = +7 + 12 +6
(z − 1)2 (z − 1)3 (z − 1)4 (z − 1)5
7 1
y[n] = n + n(n + 1) + 2n(n − 1)(n − 2) + n(n − 1)(n − 2)(n − 3)
2 4
n4 + 2n3 + n2 n2 (n + 1)2
= =
4 4
5.3-14.
5-29 Let
n
X
y[n] = kak a 6= 1
k=0
then
y[n] − y[n − 1] = nak with y[0] = 0
Setting n = 0 and y[0] = 0 in this equation yields
0 − y[−1] = 0 =⇒ y[−1] = 0
5.3-15.
5-30 (i) Let
x[n] = nu[n]
Then
z
X[z] =
(z − 1)2
236
211
Use of the result in Prob. 5.2-8a
5-14a yields
n
X z2 z(z + 1) z
k ⇐⇒ = −
(z − 1)3 (z − 1)3 (z − 1)3
k=0
Hence
n
X n(n − 1) n(n + 1)
k = n2 − =
2 2
k=0
(ii) Let
x[n] = n2 u[n]
Then
z(z + 1)
X[z] =
(z − 1)3
Use of the result in Prob. 5.2-8a
5-14a yields
n
X z 2 (z + 1) z(z 2 + 4z + 1) 3z(z − 1) 4z
k 2 ⇐⇒ 4
= 4
− 4
−
(z − 1) (z − 1) (z − 1) (z − 1)4
k=0
Hence
n
X 3n(n − 1) 2n(n − 1)(n − 2) 2n3 + 3n2 + n n(n + 1)(2n + 1)
k 2 = n3 − − = =
2 3 6 6
k=0
5.3-16. Let
5-31
x[n] = n3 u[n]
Then
z(z 2 + 4z + 1)
X[z] =
(z − 1)4
Use of the result in Prob. 5.2-8a
5-14a yields
n
X z 2 (z 2 + 4z + 1) z 7z 12z 6z
k 3 ⇐⇒ 5
= 2
+ 3
+ 4
+
(z − 1) (z − 1) (z − 1) (z − 1) (z − 1)5
k=0
Hence
n
X 7 n(n − 1)(n − 2)(n − 3) n4 + 2n3 + n2 n2 (n + 1)2
k 3 = n+ n(n−1)+2n(n−1)(n−2)+ = =
2 4 4 4
k=0
5.3-17.
5-32 Let
x[n] = nan u[n]
Then
az
X[z] =
(z − a)2
Use of the result in Prob. 5.2-8a
5-14a yields
n
X · ¸
k az 2 a z z z
ka ⇐⇒ = − + a(a − 1)
(z − 1)(z − a)2 (a − 1)2 z − 1 z − a (z − a)2
k=0
237
212
Hence
n
X a a + an+1 [n(a − 1) − 1]
kak = [1 − a n
+ (a − 1)nan
] u[n] = a 6= 1
(a − 1)2 (a − 1)2
k=0
5-33
5.3-18. (a)
ez
x [n] = een u [n] X [z] =
z−e
ez 2
Y [z] = X [z] H [z] =
(z − e)(z + 0.2)(z − 0.8)
Therefore
Y [z] ez 1.32 0.186 1.13
= = − −
z (z − e)(z + 0.2)(z − 0.8) z − e z + 0.2 z − 0.8
z z z
Y [z] = 1.32 − 0.186 − 1.13
z−e z + 0.2 z − 0.8
y [n] = [1.32(e)n − 0.186(−0.2)n − 1.13(0.8)n ] u [n]
or
y[n] − 0.6y[n − 1] − 0.16y[n − 2] = x[n − 1]
5.3-19. (a)
5-34
z(2z + 3)
Y [z] = X [z] H [z] =
(z − 1)(z − 2)(z − 3)
Therefore
Y [z] 2z + 3 5/2 7 9/2
= = − +
z (z − 1)(z − 2)(z − 3) z−1 z−2 z−3
5 z z 9 z
Y [z] = −7 +
2z−1 z−2 2z−3
· ¸
5 9
y [n] = − 7(2) + (3)n u [n]
n
2 2
238
213
or
y[n] − 5y[n − 1] + 6y[n − 2] = 2x[n − 1] + 3x[n − 2]
5.3-20.
5-35 All cases use the same transfer function. From the given H[z] (after dividing the
numerator and the denominator by 6), we can write
µ ¶
5 1
z2 − z + Y [z] = (5z − 1)X[z]
6 6
6z(5z − 1) z(5z − 1)
Y [z] = X [z] H [z] = 1 =
(z − 4 )(6z
2
− 5z + 1) (z − 4 )(z − 31 )(z − 21 )
1
Therefore
Y [z] 5z − 1 12 48 36
= = − +
z (z − 41 )(z − 31 )(z − 21 ) z − 41 z − 31 z − 21
z z z
Y [z] = 12 − 48 + 36
z − 41 z − 31 z − 21
· ¸
1 n 1 n 1 n
y [n] = 12( ) − 48( ) + 36( ) u [n]
4 3 2
£ ¤
= 12 4−n − 4(3)−n + 3(2)−n u [n]
(b) Here the input is 4−(n−2) u [n − 2] which is identical to the input in part (a)
delayed by 2 units. Therefore the response will be the output in part (a) delayed
by 2 units (time-invariance property). Therefore
h i
y [n] = 12 4−(n−2) − 4(3)−(n−2) + 3(2)−(n−2) u [n − 2]
This input is 16 times the input in part (a). Therefore the response will be 16
times the output in part (a) (linearity property). Therefore
£ ¤
y [n] = 192 4−n − 4(3)−n + 3(2)−n u [n]
239
214
times the output in part (b). Therefore
3 h −(n−2) i
y [n] = 4 − 4(3)−(n−2) + 3(2)−(n−2) u [n − 2]
4
5.3-21.
5-36 (a)
z(2z − 1)
Y [z] = X [z] H [z] =
(z − 1)(z 2 − 1.6z + 0.8)
Y [z] 2z − 1 5 5(z − 1)
= 2
= − 2
z (z − 1)(z − 1.6z + 0.8) z − 1 z − 1.6z + 0.8
z z(z − 1)
Y [z] = 5 −5 2
z−1 z − 1.6z + 0.8
For the second fraction on the right-hand side, we use pair 12c with A = 1,
B = −1, a = −0.8, γ = √25 , |γ|2 = 0.8. Therefore
√
0.8 5 0.2
r = 1.118 β = cos−1 ( ) = 0.464 θ = tan−1 ( ) = 0.464
2 0.4
· µ ¶n ¸
2
y [n] = 5 − 5(1.118) √ cos(0.464n + 0.464) u [n]
5
· µ ¶n ¸
2
= 5 − 5.59 √ cos(0.464n + 0.464) u [n]
5
or
y[n] − 1.6y[n − 1] + 0.8y[n − 2] = 2x[n − 1] − x[n − 2]
z
5.3-22.
5-37 (a) z+2
4z 2 −3z
(b) 2z 2 −3z+1
z
(c) 4z 2 +4z+1
(d) We convert the equation to advanced operator form. This yields (E 2 + 2E +
z+2
2)y[n] = (E + 2)x[n]. Hence, H[z] = z2 +2z+2
5.3-23.
5-38 (a)
z 2 + 3z + 3 z 2 + 3z + 3
H [z] = =
z 2 + 3z + 2 (z + 1)(z + 2)
Therefore
H [z] z 2 + 3z + 3 3/2 1 1/2
= = − +
z z(z + 1)(z + 2) z z+1 z+2
240
215
3 z 1 z
H [z] = − +
2 z+1 2z+2
· ¸
3 1
h [n] = δ [n] − (−1)n + (−2)n u [n]
2 2
(b)
2z 2 − z z(2z − 1)
H [z] = =
z 2 + 2z + 1 (z + 1)2
Therefore
H [z] 2z − 1 2 3
= = −
z (z + 1)2 z + 1 (z + 1)2
z z
H [z] = 2( )−3
z+1 (z + 1)2
h [n] = [2(−1)n + 3n(−1)n ] u [n] = (2 + 3n)(−1)n u [n]
(c)
z 2 + 2z z(z + 2)
H [z] = = 2
z 2 − z + 0.5 z − z + 0.5
Therefore
H [z] z+2
= 2
z z − z + 0.5
We use pair 12c with A = 1, B = 2, a = −0.5, |γ|2 = 0.5, |γ| = √1 , and
2
√ π −2.5
r = 5.099 β = cos−1 (0.5 5) = θ = tan−1 ( ) = −1.373
4 0.5
µ ¶n
1 π
h [n] = 5.099 √ cos( − 1.373)u [n]
2 4
5-39
5.3-24. (a)
H [z] 1 −1 1
= = +
z (z + 0.2)(z − 0.8) z + 0.2 z − 0.8
z z
H [z] = − +
z + 0.2 z − 0.8
h [n] = [−(−0.2)n + (0.8)n ] u [n]
(b)
241
216
(c)
2
h [n] = −1.25δ [n] + 2.795( √ )n cos(0.464n − 1.107)u [n]
5
z 1 z−1
5.3-25.
5-40 (a) Noting that H(z) = z −3 z−1 , H −1 (z) = H(z) = z −2 = z 3 − z 2 . Thus,
(b) Since h−1 [n] is absolutely summable, the system inverse is stable. However,
h−1 [n] 6= 0 for n < 0 so the system is not causal.
(c) For systems that have time as the independent variable, it is only possible to
realize causal systems. Shifting h−1 [n] by three makes it causal and therefore
realizable. That is, implement h−1 [n] = h−1 [n − 3] = δ[n] − δ[n − 1], as shown
causal
in S5-40c
S5.3-25c.
. Within a delay factor, this implementation functions as the system
inverse.
1+
5.4-1. For convenience, let γ =
5-41 √
8
and rewrite h[n] = (γ n + (γ ∗ )n ) u[n].
242
217
5.4-2.
5-42 (a)
µ ¶µ ¶
3z 2 − 1.8z 3z(z − 0.6) 3z z − 0.6
H [z] = 2
= =
z − z + 0.16 (z − 0.2)(z − 0.8) z − 0.2 z − 0.8
Parallel form: To realize parallel form, we could expand H [z] or H [z] /z into
partial fractions. In our case:
1.2z − 0.48
H [z] = 3 +
(z − 0.2)(z − 0.8)
0.4 0.8
H [z] = 3 + +
(z − 0.2) (z − 0.8)
Alternatively we could expand H [z] /z into partial fractions as:
and
z z
H [z] = 2 +
z − 0.2 z − 0.8
The realizations are shown in Figure S5.4-2.
S5-42.
(b) Refer to Figure S5-42
S5.4-2.
.
5.4-3.
5-43 (a)
5z + 2.2 5z + 2.2
H [z] = = 2
(z + 0.2)(z + 0.8) z + z + 0.16
µ ¶µ ¶
1 5z + 2.2 2 3
= = +
z + 0.2 z + 0.8 z + 0.2 z + 0.8
3.8z − 1.1
H [z] =
z 3 − 0.8z 2 + 0.37z − 0.05
For a cascade form, we express H [z] as:
µ ¶µ ¶
1 3.8z − 1.1
H [z] =
z − 0.2 z 2 − 0.6z + 0.25
5.4-5. (a) Note: the complex conjugate poles must be realized together as a second order
5-45
factor
243
218
Cascade form: µ ¶µ ¶
z 1.6z − 1.8
H [z] =
z − 0.2 z 2 + z + 0.5
Parallel form:
−2z 2z 2 + 4z
H [z] = + 2
z − 0.2 z + z + 0.5
All the realizations are shown in Figure S5.4-5.
S5-45.
(b) Refer to Figure S5.4-5.
S5-45.
5.4-6.
5-46 (a) Cascade form: µ ¶µ ¶
z 2z 2 + 1.3z + 0.96
H [z] =
z + 0.5 z 2 − 0.8z + 0.16
Parallel form:
z z 2z
H [z] = + +
z + 0.5 z − 0.4 (z − 0.4)2
(b) Refer to Figure S5.4-6.
S5-46.
244
219
AT , B T , C T , DT¡, denote
¢ transposed realization of A, B, C, and D.
A B
The realization C D denote a direct
³ form´ canonic realization by cascading (s +
T
a)/(s + c) with (s + b)/(s + d), and AC D B
represents a canonic realization using
the cascade of the transpose of (s + a)/(s + c) with direct form of (s + b)/(s + d), etc
···
We can count at least the following realization.
1 Direct canonic
1 Transpose of the above
Cascade forms: The following 16 forms
³ ´ ³ ´ ³ ´ ³ ´
A B A B B A B A
C D , D C , D C , C D ,
³ T ´ ³ T ´ ³ T ´ ³ T ´
A B A B B A B A
C D , D C , D C , C D ,
³ ´ ³ ´ ³ ´ ³ ´
A BT A BT B AT B AT
C D , D C , D C , C D ,
³ T T´ ³ T T´ ³ T T´ ³ T T´
A B A B B A B A
C D , D C , D C , C D ,
251
221
and µ ¶
£ ¤ 0.4 sin Ω
6 H ejΩ = − tan−1
1 − 0.4 cos Ω
(c)
3z 2 − 1.8z
H [z] =
z2− z + 0.16
£ ¤ 3e2jΩ − 1.8ejΩ (3 cos 2Ω − 1.8 cos Ω) + j(3 sin 2Ω − 1.8 sin Ω)
and H ejΩ = =
e2jΩ − ejΩ + 0.16 (cos 2Ω − cos Ω + 0.16) + j(sin 2Ω − sin Ω)
· ¸· ¸
¯ £ jΩ ¤¯2 3e2jΩ − 1.8ejΩ 3e−2jΩ − 1.8e−jΩ
¯H e ¯ =
e2jΩ − ejΩ + 0.16 e−2jΩ − e−jΩ + 0.16
12.24 − 10.8 cos Ω
=
2.0256 − 2.32 cos Ω + 0.32 cos 2Ω
· ¸1/2
¯ £ jΩ ¤¯ 12.24 − 10.8 cos Ω
Therefore ¯H e ¯ =
2.0256 − 2.32 cos Ω + 0.32 cos 2Ω
and
µ ¶ µ ¶
£ ¤ 3 sin 2Ω − 1.8 sin Ω sin 2Ω − sin Ω
6 H ejΩ = tan−1 − tan −1
3 cos 2Ω − 1.8 cos Ω cos 2Ω − cos Ω + 0.16
5.5-2. (a)
5-51
0.5 2 2 0.5 1
H [z] = 1 + + 2+ 3+ 4 + 5
z z z z z
£ ¤
H ejΩ = 1 + 0.5e−jΩ + 2e−2jΩ + 2e−j3Ω + 0.5e−j4Ω + e−j5Ω
£ ¤
= e−j2.5Ω ej2.5Ω + 0.5ej1.5Ω + 2ej0.5Ω + 0.5e−j1.5Ω + e−j2.5Ω
· ¸
−j2.5Ω Ω 1 3Ω 5Ω
= 2e 2 cos + cos + cos
2 2 2 2
¯ ¯
¯ £ jΩ ¤¯ ¯ Ω 3Ω 5Ω ¯¯
Therefore ¯ H e ¯ ¯
= ¯4 cos + cos + 2 cos
2 2 2 ¯
£ jΩ ¤
and 6 H e = −2.5Ω
5.5-3.
5-52 The advance operator form of this equation is
1£ 4 ¤
E 4 y[n] = E + E3 + E2 + E + 1
5
252
222
and · ¸
1 z4 + z3 + z2 + z + 1
H[z] =
5 z4
· ¸
£ ¤ 1 ej4Ω + ej3Ω + ej2Ω + ejΩ + 1
H ejΩ =
5 ej4Ω
1 −j2Ω £ j2Ω ¤
= e e + ejΩ + 1 + e−jΩ + e−j2Ω
5
1 −j2Ω
= e [1 + 2 cos Ω + 2 cos 2Ω]
5
(i)
£ ¤ ejΩ 1 1
H ejΩ = = =
ejΩ + 0.9 1 + 0.9e−jΩ 1 + 0.9 cos Ω − j0.9 sin Ω
· ¸
¯ £ jΩ ¤¯ 1 £ ¤ −0.9 sin Ω
¯H e ¯ = √ , 6 H ejΩ = − tan−1
1.81 + 1.8 cos Ω 1 + 0.9 cos Ω
(ii)
£ ¤ ejΩ 1 1
H ejΩ = = =
ejΩ
− 0.9 1 − 0.9e −jΩ 1 − 0.9 cos Ω + j0.9 sin Ω
· ¸
¯ £ jΩ ¤¯ 1 £ jΩ ¤ 0.9 sin Ω
¯H e ¯ = √ , H e
6 = − tan −1
1.81 − 1.8 cos Ω 1 − 0.9 cos Ω
Filter(i) has a zero at the origin and a pole at −0.9. Because the only pole is
near Ω = π(z = −1), this is a highpass filter, as verified from the frequency
response shown in Figure S5-53a.
S5.5-4a.
Filter(ii) has a zero at the origin and a pole at 0.9. Because the only pole is near
Ω = 0(z = 1), this is a lowpass filter, as verified from the frequency response
shown in Figure S5.5-4b.
S5-53b.
253
223
£ ¤
6 H ej0.5 = 0.2784 − 0.9037 = −0.6253 rad
Therefore
π
y [n] = 2.86 cos(0.5n − − 0.6253) = 2.86 cos(0.5n − 1.6725)
3
5.5-6.
5-55
Therefore
N
X z £ ¤ z
Y [z] = ci + H ejΩ
i=1
z − γi z − ejΩ
and " #
N
X
n
£ jΩ ¤ jΩn
y [n] = ci γi +H e e u [n]
i=1
The sum on the right-hand side consists of N characteristic modes the system. For
an asymptotically stable system |γi | < 1 (i = 1, 2, · · · , N ) and the sum on the right-
hand side vanishes as n → £∞. This
¤ sum is therefore the transient component of the
response. The last terms H ejΩ ejΩn , which does not vanish as n → ∞, is the steady-
state component of the response yss [n]:
£ ¤
yss [n] = H ejΩ ejΩn
255
225
5.5-7.
5-56 (a) Ω = 0.8π is in the fundamental range. Hence it is the apparent frequency.
(b) 1.2π = 2π − 0.8π. Hence
³ π ´ ³ π ´
sin(1.2πn + θ) = cos 1.2πn − + θ = cos 0.8π + − θ = − sin(0.8π − θ)
2 2
cos(6.9n + θ) = cos(0.6168n + θ)
1 106
5.5-9. (a) fh =
5-58 2T = 50×2 = 10 kHz.
256
226
1
(b) fs ≥ 2fh = 100 kHz, = 10 µs.
T ≤ fs
P∞ k
5.5-10.
5-59 P Taking the z-transform ofPy[n] = k=0 (0.5) x[n − k] yields Y (z) =
∞ k −k ∞ k
k=0 (0.5) X(z)z = X(z) k=0 (0.5/z) . For |z| > 1/2, this becomes Y (z) =
1 Y (z) 1
X(z) 1−0.5z −1 . Thus, the system function is H(z) = X(z) = 1−0.5z −1 .
¯ ¯
(a) Using H(z) Ω ¯ Ω ¯
¯ ¯ and letting z = e , the magnitude response is H(e ) =
¯ 1 ¯ 1 1
¯ 1−0.5e −Ω ¯ = √ 2 2
= √ 2 2
.
(1−0.5 cos(Ω)) +(−0.5 sin(Ω)) 1−cos(Ω)+0.25(cos (Ω)+sin (Ω))
Thus,
¯ ¯ 1
¯H(eΩ )¯ = p .
5/4 − cos(Ω)
¯ ¯
MATLAB is used to plot ¯H(eΩ )¯.
>> Omega = linspace(-pi,pi,501);
>> Hmag = 1./sqrt(5/4-cos(Omega));
>> plot(Omega,Hmag,’k’); axis([-pi pi 0 2.5]); grid;
>> xlabel(’\Omega’); ylabel(’|H(e^{j\Omega})|’);
>> set(gca,’xtick’,[-pi:pi/4:pi],’xticklabel’,[’-p ’;...
’ ’;’ ’;’ ’;’ 0 ’;’ ’;’ ’;’ ’;’ p ’],...
’fontname’,’symbol’);
2.5
1.5
|H(e )|
jΩ
0.5
0
−π 0 π
Ω
¯ ¯
S5.5-10a: ¯H(eΩ )¯ = √
Figure S5-59a: 1
.
5/4−cos(Ω)
1
(b) Using H(z) and letting z = eΩ , the phase response is 6 H(eΩ ) = 6
1−0.5e−Ω =
−6 (1 − 0.5 cos(Ω) − 0.5 sin(Ω)). Thus,
µ ¶
−0.5 sin(Ω)
6 H(eΩ ) = − arctan .
1 − 0.5 cos(Ω)
257
227
(for Ω = 0). As Ω increases, the angle due to both poles increase. Hence, 6 H[ejΩ ]
increases in negative direction until it reaches the value −2π at Ω = π. For case (b),
similar behavior is observed. Note that angle −2π is the same as 0.
5.6-2.
5-61 The two systems are very similar and have identical steady-state characteristics. There
is an important difference, however, between the two systems. The system y[n] −
y[n − 1] = x[n] − x[n − 1] is first-order and can support an initial condition; the
system y[n] = x[n] is zero-order and cannot support an initial condition. If the initial
condition of the first system is non-zero, the output of the two systems can be quite
different.
5.6-3. (a) From the magnitude response plot, it is clear that this is a lowpass filter. Low
5-62
frequencies near Ω = 0 are attenuated, and high frequencies near Ω = ±π are
passed with unity gain.
(b) From the magnitude and phase response plots, H(eπ/2 ) = √1 e3π/2 . Thus, the
2
output to x1 [n] = 2 sin( π2 n + π4 ) is
√ π √ π
y1 [n] = 2 sin( n + π) = − 2 sin( n).
2 2
(c) Notice, H(e7π/4 ) = H(e−π/4 ). From the magnitude and phase response plots,
H(e−π/4 ) ≈ 0.071e2.43 . Thus, the output to x2 [n] = cos( 7π
4 n) is
7π
y2 [n] = 0.071 cos( n + 2.43).
4
259
229
See solution for 5.M-1.
Figure S5.6-4
S5-63
ω 1.176 0.187
Hence B= = = Hz
2π 2πT T
5.6-7.
5-66
1 1
T ≤ = = 25µs
2Bh 40000
Select T = 25µs Frequency 5000 Hz gives
260
230
5.6-9. (a)
5-68
Use of the property in Eq. (5.20) with γ = e±jπ , we obtain H2 [z] = H1 [ej±π z].
Hence
H2 [ejΩ ] = H1 [ej(Ω±π ]
(b) Figure S5.6-9
S5-68 shows the frequency response of an ideal lowpass filter with cutoff
frequency Ωc . Figure S5.6-9
S5-68 also shows the same frequency response shifted by
π (with 2π-periodicity). It is clear that the shifted response corresponds to an
ideal high-pass filter with cutoff frequency π − Ωc .
Figure S5.6-9
5.6-10. (a) Let H(z) represent the original filter, either highpass or lowpass. The transformed
5-69
filter has system function HT (z) = H(z)|z=−z = H(−z). The basic character
of the transformed filter can be assessed by its magnitude response, |HT (eΩ )| =
|H(−eΩ )| = |H(eπ eΩ )| = |H(e(Ω+π) )|. That is, the magnitude response of a
transformed filter is just the magnitude response of the original filter shifted in
frequency by π. If the magnitude response of a digital LPF is shifted in frequency
by π, it becomes a highpass filter. Similarly, if the magnitude response of a digital
HPF is shifted in frequency by π, it becomes a lowpass filter. Put another way,
an original passband centered at Ω = 0 (LPF) is shifted by the transformation
to a passband centered at Ω = π (HPF), and vice-versa.
P∞ −n
(b) P
If H(z) = n=−∞ h[n]z
P∞ is the original filter, HT (z) = H(−z) =
∞ −n n
n=−∞ h[n](−z) = n=−∞ (−1) h[n]z −n . Thus, the transformed filter has
impulse response
hT [n] = (−1)n h[n].
Put another way, the same transformation is accomplished by simply negating
the values of h[n] for every odd integer n.
5-70 The bilinear transformation states s = 2(1−z −1 ) 1+T s/2
5.6-11. T (1+z −1 ) . Rearranging yields z = 1−T s/2 .
1+ωT /2
Thus, s = ω maps to z = 1−ωT /2 .
¯ ¯
¯ 1+ωT /2 ¯ |1+ωT /2|
(a) The magnitude of this transformation is |z| = ¯ 1−ωT /2 ¯ = |1−ωT /2| = 1. Since
only the unit circle has |z| = 1, the bilinear transform maps s = ω onto the unit
circle in the z-plane.
1+ωT /2
(b) Using the solution to 5.6-11a,
5-70a, we know that z = 1−ωT /2 describes the unit circle
1+ωT /2
in the complex plane. This allows us to write z = eΩ = 1−ωT /2 . Thus, the
262
232
bilinear transformation maps (−∞ ≤ ω ≤ ∞) to (−π/2 ≤ Ωπ/2) in a monotonic,
1+ωT /2
although non-linear, manner according to Ω = 6 z = 6 1−ωT /2 = arctan(ωT /2) −
arctan(−ωT /2) = 2 arctan(ωT /2).
5.7-1. For the system in Eq. (3.15a), the transfer function H(s) is given by
5-71
1
H(s) =
s+c
and
h(t) = e−ct u(t)
Using impulse invariance criterion, we obtain the equivalent digital filter transfer func-
1
tion from Table 5.3 corresponding to H(s) = s+c as
Tz Tz
H[z] = ' assuming T → 0
z − e−ct z − (1 − cT )
b = βz
H[z]
z+α
T −1
Substituting β = 1+cT and α = 1+cT , we obtain
Tz
b =
H[z] 1+cT
1
z − 1+cT
1
As T → 0, 1+cT ' 1 − cT . Hence
b ' T (1 − cT )z '
H[z]
Tz
z − (1 − cT ) z − (1 − cT )
7s + 20 7s + 20 1 5/2
Ha (s) = = = +
2(s2 + 7s + 10) 2(s + 2)(s + 5) s+2 s+5
First we select T :
7
Ha (0) = 1 and for s À 5 =⇒ Ha (s) ≈
2s
7
and |Ha (jω)| ' ωÀ5
ω
(b) We shall choose the filter bandwidth to be that frequency where
7 π
|Ha (jω0 )| is 1% of |Ha (0)|. Hence = 0.01 and ω0 = 350, T =
2ω0 350
263
233
pair 11b, we see that a transfer function
z sin ω0 T
H[z] =
z 2 − 2z cos ω0 T + 1
has an impulse response (or zero-input response) of the form sin Ω0 n (Ω0 = ω0 T ).
The period of the sinusoid is T0 = 2π/Ω0 , and there are 10 samples in each cycle.
Therefore, the sampling interval T = T0 /10 = π/5Ω0 , and Ω0 T = π/5. hence,
¡ ¢
z sin π5
H[z] = ¡ ¢
z 2 − 2 cos π5 z + 1
0.5878z
=
z 2 − 1.618z + 1
This is one possible solution. By varying the phase in the impulse response, we
could obtain variations of this transfer function.
(b) Another approach is to consider an analog system with transfer function Ha (s)
such that its impulse response ( or zero-input response) is of the form sin ω0 t [or
cos(ω0 t + θ) for any value of θ]. From Table 6.1, pair 8b we find
ω0
Ha (s) =
s2 + ω02
Now using Table 12.1, we find the corresponding digital filter using impulse in-
variance method as
T z sin ω0 T
H[z] = 2
z − 2z cos ω0 T + 1
Earlier we found that ω0 T = π/5. Because ω0 = 2π(10, 000) = 20, 000π, the
period T0 = 10−4 . There are 10 samples in each period. Hence the sampling
interval T = 10−5 , and
0.5878z
H[z] = 10−5
z 2 − 1.618z + 1
This is identical to the answer in (a) except for an amplitude scaling by 10−5 .
Because, we did not specify any amplitude requirement on the oscillator, different
answers will differ by a constant multiplier. This is a marginally stable system
and will oscillate without input with the response of the form
266
236
Hence,
T T T π
|H[ejωT ]| = ¯¯q ¯= p
¯ = |ω| ≤
¯ (cos ωT − 1)2 + sin2 ωT ¯ | 2(1 − cos ωT )| 2| sin ωT
2 |
T
¯ ¯
z Tz 1
= H[z] or H[z] = (z − 1)
z−1 (z − 1)2 T
(b) For an integrator
1
H(s) =
s
The unit ramp response r(t) is given by
µ ¶
−1 1 1 1
r(t) = L 2
= t2 u(t)
s s 2
Now we design H[z] such that its response to nT u[n] is 21 n2 T 2 u[n], that is
½ ¾
1 2 2
Z n T u[n] = H[z]Z{nT u[n]}
2
or
T 2 z(z + 1) Tz
= H[z]
2(z − 1)3 (z − 1)2
Hence µ ¶
T z+1
H[z] =
2 z−1
268
238
5.7-8.
5-78
X ki X ki z
For Ha (s) = , H[z] = T
i
s − λi i
z − eλi T
If λi = αi + jβi , then eλi T = eαi T ejβi T . When λi is in LHP, αi < 0 and |eλi T | =
eαi T < 1. Hence if λi is in LHP, the corresponding pole of H[z] is within the unit
circle. Clearly if Ha (s) is stable, the corresponding H[z] is also stable.
−1
5-79 (a) The ω-axis is given by s = ω. Rewriting the transformation s = 1−z
5.7-9. T as
1 1
z = 1−sT and substituting s = ω yields z = 1−ωT . Thus, we need to show that
1
z = 1−ωT describes a circle centered at (1/2, 0) with a radius of 1/2.
For a complex variable z, the equation |z − 1/2|2 = (1/2)2 describes a circle
1
centered at (1/2, 0) with a radius of 1/2. The transformation rule z = 1−ωT is
substituted into this expression.
|z − 1/2|2 = (z ∗
³ − 1/2)(z ´−³ 1/2) ´
1 1 1 1
= 1−ωT − 2 1+ωT − 2
−1
= 1+ω12 T 2 + 2(1+ωT ) + −1
2(1−ωT ) +
1
4
−(1−ωT )−(1+ωT )
.
= 1+ω12 T 2 + 2(1+ωT )(1−ωT ) + 1
4
= 1+ω12 T 2 + −2 1
2(1+ω 2 T 2 ) + 4
2
= 1/4 = (1/2)
1
Since the equation is satisfied, the transformation rule z = 1−ωT maps the ω-axis
to a circle centered at (1/2, 0) with a radius of 1/2.
Notice that different values of ω can map to the same value of z (aliasing), which
makes an inverse transformation very problematic.
1−z −1 1
(b) First, rewrite the transformation s = T as z = 1−sT Next, notice
|z|2 = zz ∗
1 1
= 1−sT 1−s∗ T
1
= 1−sT −s∗ T +ss∗ T 2 .
1
= 1−(σ+ω)T −(σ−ω)T 2 2
+(σ +ω )T 2
1
= 1−2σT +(σ 2 +ω 2 )T 2
For σ < 0, the denominator 1 − 2σT + (σ 2 + ω 2 )T 2 > 1 and |z|2 < 1. That is,
the left-half plane of s (σ < 0) is guaranteed to map to the interior of the unit
circle in the z-plane.
5.9-1. (a)
5-80
z
x1 [n] ⇐⇒ |z| > 0.8
z − 0.8
−z
x2 [n] ⇐⇒ |z| < 2
z−2
269
239
Hence
z z
X [z] = − 0.8 < |z| < 2
z − 0.8 z − 2
−1.2z
= 0.8 < |z| < 2
z 2 − 2.8z + 1.6
(b)
z
X1 [z] = |z| > 2
z−2
z
X2 [z] = |z| < 3
z−3
Hence
z z
X [z] = + 2 < |z| < 3
z−2 z−3
z(2z − 5)
= 2 < |z| < 3
z 2 − 5z + 6
(c)
z
X1 [z] = |z| > 0.8
z − 0.8
−z
X2 [z] = |z| < 0.9
z − 0.9
z z
Hence X [z] = −
z − 0.8 z − 0.9
−z
= 2
0.8 < |z| < 0.9
10(z − 1.7z + 0.72)
(d)
µ ¶
−z 3z
[(0.8)n + 3(0.4)n ] u [−(n + 1)] ⇐⇒ − |z| < 0.4
z − 0.8 z − 0.4
−4z(z − 0.7)
= |z| < 0.4
(z − 0.4)(z − 0.8)
(e)
z 3z
[(0.8)n + 3(0.4)n ] u [n] ⇐⇒ + |z| > 0.8
z − 0.8 z − 0.4
4z(z − 0.7)
= |z| > 0.8
(z − 0.4)(z − 0.8)
(f)
The region of convergence for (0.8)n u [n] is |z| > 0.8. The region of convergence
for (0.4)n u [−(n + 1)] is |z| < 0.4. The common region does not exist. Hence the
z-transform for this function does not exist.
270
240
(g)
(h)
(b) The region of convergence is e−2 < |z| < 2. In this case the 1st term is causal
and the second is anticausal.
(c) The region of convergence is |z| < e−2 . Both terms are anticausal in this case.
1 1/2 1 −2 2
5.9-3. X(z) =
5-82 (2z+1)(z+1)(z+ 21 )
= (z+1/2)2 (z+1) = (z+1/2) 2 + (z+1/2) + (z+1) =
−z/2 z z
¡ ¢
−2z −1 (z+1/2) 2 − 2z
−1
+
(z+1/2) 2z −1 (z+1) . Since |z| < 21 , the time-domain signal
is
n−1 n−1
left-sided. Thus, x[n] = 2(n−1)(−1/2) u[−(n−1)−1]+2(−1/2) u[−(n−1)−1]−
2(−1)n−1 u[−(n − 1) − 1] = −4(n − 1)(−1/2)n u[−n] − 4(−1/2)n u[−n] + 2(−1)n u[−n].
Simplifying yields
271
241
>> plot(real(zz),imag(zz),’ko’,real(zp),imag(zp),’kx’);
>> xlabel(’Re(z)’); ylabel(’Im(z)’);
>> axis([-1.5 1.5 -1.5 1.5]); axis equal; grid;
1.5
0.5
Im(z)
0
−0.5
−1
−1.5
−1.5 −1 −0.5 0 0.5 1 1.5 2
Re(z)
z(z− 21 )
Figure S5.9-4a:
S5-83a: Pole-zero plot for H(z) = (z 3 − 27
.
8 )
There are two possible regions of convergence, both of which exclude the three
system poles: |z| < 3/2 or |z| > 3/2.
(b) The poles and zeros of H −1 (z) are just the zeros and poles, respectively, of H(z).
Thus, the three zeros of H(z) satisfy z 3 = 278 , or z = 3/2e
2πk/3
for k = (0, 1, 2).
There are two finite poles at z = 0 and z = 1/2 as well as a pole at infinity.
MATLAB is used to create the corresponding pole-zero plot.
>> k = [0:2]; zz = 3/2*exp(j*2*pi*k/3); zp = [0,1/2];
>> plot(real(zz),imag(zz),’ko’,real(zp),imag(zp),’kx’);
>> xlabel(’Re(z)’); ylabel(’Im(z)’);
>> axis([-1.5 1.5 -1.5 1.5]); axis equal; grid;
1.5
0.5
Im(z)
−0.5
−1
−1.5
−1.5 −1 −0.5 0 0.5 1 1.5 2
Re(z)
(z 3 − 27
8 )
S5-83b: Pole-zero plot for H −1 (z) =
Figure S5.9-4b: .
z(z− 21 )
There are two possible regions of convergence, both of which exclude the three
system poles: 0 < |z| < 1/2 or 1/2 < |z| < ∞.
272
242
5-84 It is known that x[n] has a mode (1/2)n and that x1 [n] = (1/3)n x[n] is absolutely
5.9-5.
summable. For this to be true, the mode at (1/2)n must be right-sided. That is,
(1/2)n (1/3)n u[n] is absolutely summable but (1/2)n (1/3)n u[−n] is not. It is also
known that x2 [n] = (1/4)n x[n] is not absolutely summable. For this to be true, there
must be a pole somewhere in the annulus 3 < |z| < 4 that corresponds to a left-sided
signal; such a mode when multiplied by (1/3)n is still absolutely summable but when
multiplied by (1/4)n is not absolutely summable. Thus,
4 n 1 n
y1 [n] = (−1/2) u[n] + (3/4) u[n].
5 5
(b) The idea of frequency response is used to determine the output in response to the
everlasting exponential input x2 [n] = (3/4)n . That is, H(z = 3/4) = 3/4−1/2
3/4+1/2 =
1/4
5/4 = 51 . Thus,
1 n
y2 [n] = (3/4) .
5
5-87 The given signal is x[n] = (−1)n u[n − n0 ] + αn u[−n]. If |α| = 2, the z-transform
5.9-8.
z −z
X(z) = (−1)n0 z −n0 z+1 + z−α has region of convergence 1 < |z| < |α| = 2, as desired.
Thus, the necessary constraint is
|α| = 2.
There is no constraint on the integer n0 , other than it be finite.
P∞ P∞
5.9-9. (a) X1 (z) =
5-88 n=−∞ x1 [n]z −n = n=−∞ ((−)
−n
u[−n] + δ[−n]) z −n =
P0 ³ ´ n P∞ P0 ¡ ¢ n
n=−∞
−1
z + n=−∞ δ[−n]z −n = n=−∞ z + 1. For |z| < 1, this be-
comes
0 − /z −
X1 (z) = 1 + =1+ ; ROC |z| < 1.
1 − /z z−
273
243
P∞ −n
P∞ n
(b) X2 (z) = n=−∞ x2 [n]z = n=−∞ () cos(n + 1)u[n] =
P∞ ¡ ¢ P∞ ³ ´n
n (n+1) −jmath(n+1) −n e
n=0 0.5 e + e z = n=0 0.5e z +
P∞ ³ − ´n
− e
n=0 0.5e z . For |z| > 1, this becomes
0.5e 0.5e−
X2 (z) = + ; ROC |z| > 1.
1 − e z −1 1 − e− z −1
P∞ P∞
(c) X3 (z) = x3 [n]z −n = n
n=−∞ (0.5(e − e
−n
)u[−n + 1]) z −n =
P1 ©¡n=−∞
¢
e n
¡ ¢
1 n
ª −1
0.5 z − ez . For |z| < e and |z| < e , this becomes X3 (z) =
nn=−∞ 2 o
0−(e/z) 0−(ez)−2
2 1−e/z − 1−(ez)−1 . Thus,
½ ¾
z −2 e−2 e2
X3 (z) = −1
+ ; ROC |z| < e−1 .
2 1 − (ez) 1 − ez −1
P∞ P∞ ³P ´
0
(d) X4 (z) = n=−∞ x4 [n]z −n = n=−∞
n
− 2k] z −n =
k=−∞ (2) δ[n
P0 P∞ ¡ 2 ¢n P0 ¡ 2 ¢2k P0 ¡ ¢k
k=−∞ n=−∞ z δ[n − 2k] = k=−∞ z = k=−∞ −4
z2 . For |z 2 | <
4, this becomes
0 − (−4z −2 ) 4z −2
X4 (z) = −2
= ; ROC |z| < 2.
1 − (−4z ) 1 + 4z −2
274
244
³ ´ ³ ´ ³ ´
6/5 −1/5
z+1
(b) H2 (z) = z −3 (z−2)(z+1/2) = z −3 z−2 + z+1/2 = z −4 6z/5 −z/5
z−2 + z+1/2 . Since
h2 [n] is stable, the region of convergence for H2 (z) must be 1/2 < |z| < 2. Using
this ROC and z-transform tables, the inverse transform is
−6 n−4 −1
h2 [n] = (2) u[−n + 3] + (−1/2)n−4 u[n − 4].
5 5
P∞ −n
P∞ P∞ P∞ P∞
5-91 H(z) =
5.9-12. n=−∞ h[n]z = δ[n − N k]z −n = δ[n −
−n
P∞ −N k P∞ ¡ n=−∞
¢
1 k
k=0 k=0 n=−∞
N k]z = k=0 z = k=0 zN
. For |z| > 1, this becomes
1
H(z) = ; ROC |z| > 1.
1 − zN
Notice, H(z) has N poles that correspond to the N roots of unity. That is, the poles
of H(z) satisfy z = e2πk/N for k = (0, 1, . . . , N − 1).
5.9-13.
5-92 For causal signals, the region of convergence may be ignored. We shall consider it only
for noncausal inputs
(a)
z2
Y [z] = X [z] H [z] =
(z − e)(z + 0.2)(z − 0.8)
(b)
−z
X [z] = |z| < 2
z−2
z
H [z] = |z| > 0.8
(z + 0.2)(z − 0.8)
−z 2
Y [z] = 0.8 < |z| < 2
(z + 0.2)(z − 0.8)(z − 2)
and
Y [z] −z 1/11 2/3 0.758
= = + −
z (z + 0.2)(z − 0.8)(z − 2) z + 0.2 z − 0.8 z − 2
1 z 2 z z
Therefore Y [z] = + − 0.758 0.8 < |z| < 2
11 z + 0.2 3 z − 0.8 z−2
· ¸
1 2
and y [n] = (−0.2)n + (0.8)n u [n] + 0.758(2)n u [−(n + 1)]
11 3
(c) The input in this case is the sum of the inputs in parts a and b hence the response
will be the sum of the responses in part a and b.
275
245
5.9-14.
5-93
z
X1 [z] = |z| > 2
z−2
−z
X2 [z] = |z| < 1
z−1
There is no region of convergence common to X1 [z] and X2 [z]
z
H [z] =
(z + 0.2)(z − 0.8)
The region of convergence of H [z] is |z| > 0.8 (assuming a causal system). We should
find the response to x1 [n] and x2 [n] separately.
z2
Y1 [z] = |z| > 2
(z − 2)(z + 0.2)(z − 0.8)
5.9-15.
5-94
−z
X [z] = |z| < e2
z − e2
and
z
H [z] = |z| > 0.8
(z + 0.2)(z − 0.8)
No common region of convergence for X [z] and H [z] exists. Hence
y [n] = ∞
¡ ¢
5.M-1. Taking the z-transform of 4y[n + 2] − y[n] = x[n + 2] + x[n] yields Y (z) 4z 2 − 1 =
276
246
¡ ¢ Y (z) z 2 +1 1+z −2
X(z) z 2 + 1 . Thus, the system function is H(z) = X(z) = 4z 2 −1 = 0.25 1−z −2 /4 .
0.8
0.6
0.4
0.2
Im(z)
−0.2
−0.4
−0.6
−0.8
−1
−1 −0.5 0 0.5 1
Re(z)
z 2 +1
Figure S5.M-1a: Pole-zero diagram for H(z) = 4z 2 −1 .
¯
z 2 +1 ¯
(b) H(eΩ ) = 4z 2 −1 ¯ . MATLAB is used to plot the magnitude response
z=eΩ
|H(eΩ )|.
>> Omega = linspace(-pi,pi,501);
>> z = exp(j*Omega); H = (z.^2+1)./(4*z.^2-1);
>> plot(Omega,abs(H),’k’); axis([-pi pi 0 1]); grid;
>> xlabel(’\Omega’); ylabel(’|H(e^{j\Omega})|’);
>> set(gca,’xtick’,[-pi:pi/4:pi],’xticklabel’,[’-p ’;...
’ ’;’ ’;’ ’;’ 0 ’;’ ’;’ ’;’ ’;’ p ’],...
’fontname’,’symbol’);
(c) The pole-zero plot of Figure S5.M-1a and the magnitude response plot of Figure
S5.M-1b confirm that this is a band-stop system.
(d) Yes, the system is asymptotically stable. Referring to Figure S5.M-1a, all the
system poles are within the unit circle.
(e) Yes, the system is real. Since the system is expressed as a constant-coefficient
linear difference equation with real coefficients, the impulse response h[n] and
system are both real.
(f) For an input of the form x[n] = cos(Ωn), the greatest possible amplitude of the
output corresponds to the greatest gain shown in the magnitude response plot of
Figure S5.M-1b. Thus, 2/3 is the greatest output amplitude given an input of
x[n] = cos(Ωn). This output amplitude occurs when Ω = kπ, for any integer k.
277
247
1
0.9
0.8
0.7
0.6
|H(e )|
jΩ
0.5
0.4
0.3
0.2
0.1
0
−π 0 π
Ω
Figure S5.M-1b: Magnitude response plot for 4y[n + 2] − y[n] = x[n + 2] + x[n].
Σ
Figure S5.M-1g: TDFII implementation of y[n] − 0.25y[n − 2] = 0.25x[n] + 0.25x[n − 2].
Y (z) −2
1+z
(g) Inverting H(z) = X(z) = 0.25 1−z −2 /4 provides y[n] − 0.25y[n − 2] = 0.25x[n] +
278
248
1
0.8
0.6
0.4
0.2
Im(z)
0
−0.2
−0.4
−0.6
−0.8
−1
−1 −0.5 0 0.5 1
Re(z)
z 2 −
Figure S5.M-2a: Pole-zero diagram for H(z) = z−0.9e3π/4
.
>> set(gca,’xtick’,[-2*pi:pi/4:2*pi],’xticklabel’,[’-2p’;
’ ’;’ ’;’ ’;’-p ’;’ ’;’ ’;’ ’;’ 0 ’;...
’ ’;’ ’;’ ’;’ p ’;’ ’;’ ’;’ ’;’ 2p’],’fontname’,’symbol’)
20
18
16
14
12
|H(ejΩ)|
10
0
−2π −π 0 π 2π
Ω
z 2 −
Figure S5.M-2b: Magnitude response plot for H(z) = z−0.9e3π/4
.
From Figure S5.M-2b, the system appears to be a type of bandpass filter. This
particular filter tends to pass only positive frequency inputs near Ω = 3π/4; the
corresponding negative frequencies near Ω = −3π/4 are significantly attenuated.
Since the magnitude response is not an even function of Ω, the output of the
filter will be complex-valued.
z 4 −1
5.M-3. (a) H(z) = 2(z 2 +0.81) has four finite zeros and two finite poles. The zeros are the four
roots of unity, and the poles are at z = 0.9eπ/4 and z = 0.9e−3π/4 . MATLAB
is used to compute the pole-zero plot.
>> zz = roots([1 0 0 0 -1]); zp = roots([1 0 j*0.81]);
>> theta = linspace(0,2*pi,201);
>> plot(real(zz),imag(zz),’ko’,real(zp),imag(zp),’kx’,...
cos(theta),sin(theta),’k’);
279
249
>> xlabel(’Re(z)’); ylabel(’Im(z)’);
>> axis([-1.1 1.1 -1.1 1.1]); axis equal; grid;
0.8
0.6
0.4
0.2
Im(z)
0
−0.2
−0.4
−0.6
−0.8
−1
−1 −0.5 0 0.5 1
Re(z)
z 4 −1
Figure S5.M-3a: Pole-zero diagram for H(z) = 2(z 2 +0.81) .
4
|H(ejΩ)|
0
−π 0 π
Ω
z 4 −1
Figure S5.M-3b: Magnitude response plot for H(z) = 2(z 2 +0.81) .
(c) Since H(z) is an improper rational function and has poles at infinity, the system
is non-causal. By dividing the numerator by the denominator (in such a way as
to yield a right-sided impulse response), an improper rational function H(z) will
yield an impulse response h[n] that is not zero for all negative n. Right-sided
280
250
functions with poles at infinity are sometimes called “not-quite causal” since they
are not causal only by a finite shift.
(d) Adding two poles at zero (a = b = 0) corresponds to a simple right-shift
¯ by two
¯
Ω ¯ H(eΩ ) ¯
and does not affect the magnitude response. That is, |Hcausal (e )| = ¯ (eΩ )2 ¯ =
|H(eΩ )|
= |H(eΩ )|. By adding two poles at zero to H(z), Hcausal (z) becomes
|(eΩ )2 |
a proper rational function with only finite poles, and the corresponding impulse
response function becomes causal.
Y (z) 4 −4
= z0.5z −0.5 0.5−0.5z
(e) Noting that Hcausal (z) = X(z) 4 +0.81z 2 = 1+0.81z −2 , the corresponding
difference equation is y[n] + 0.81y[n − 2] = 0.5x[n] − 0.5x[n − 4]. Figure S5.M-3e
shows the corresponding TDFII implementation.
Σ
Σ
Figure S5.M-3e: TDFII implementation of y[n] + 0.81y[n − 2] = 0.5x[n] − 0.5x[n − 4].
5.M-4. (a) From the block diagram, the corresponding difference equation is written directly.
¯ ¯ 1
¯H(eΩ )¯ = p .
5/4 − cos(2Ω)
¯ ¯
MATLAB is used to plot ¯H(eΩ )¯.
>> Omega = linspace(-pi,pi,501);
>> Hmag = 1./sqrt(5/4-cos(2*Omega));
281
251
>> plot(Omega,Hmag,’k’); axis([-pi pi 0 2.5]); grid;
>> xlabel(’\Omega’); ylabel(’|H(e^{j\Omega})|’);
>> set(gca,’xtick’,[-pi:pi/4:pi],’xticklabel’,[’-p ’;...
’ ’;’ ’;’ ’;’ 0 ’;’ ’;’ ’;’ ’;’ p ’],...
’fontname’,’symbol’);
2.5
1.5
|H(e )|
jΩ
0.5
0
−π 0 π
Ω
¯ ¯
Figure S5.M-4b: ¯H(eΩ )¯ = √ 1
.
5/4−cos(2Ω)
Standard filter types do not provide a good description of this filter. The system
appears most like a bandstop filter, but its stopband attenuation is quite poor.
The system boosts the gain of low and high frequencies more than it attenuates
the middle frequencies.
1 1/2 1/2
(c) Inverting H(z) = 1−0.5z −2 = √
1−z −1 / 2
+ √
1+z −1 / 2
yields
³ √ √ ´
h[n] = 0.5 (1/ 2)n + (−1/ 2)n u[n].
5.M-5. Label the first summation block output as v[n]. Thus, v[n] = x[n − 1] + 41 y[n] or
V (z) = z −1 X(z) + 41 Y (z). The output of the second summation block is y[n + 1] =
−1
x[n]+v[n−1]¡or zY (z) = X(z)+z ¢ V (z). ¡Combining ¢ the two equations
¡ ¢yields zY (z) =
X(z) + z −1
z X(z) + 4 Y (z) or Y (z) z − 41 z −1 = X(z) 1 + z −2 . Multiplying
−1 1
¡ ¢ ¡ ¢ Y (z)
both sides by z −1 yields Y (z) 1 − 41 z −2 = X(z) z −1 + z −3 . Thus, H(z) = X(z) =
z −1 +z −3 1+z 2
−4 5/2 5/2 −4 5z/2 5z/2
= z(z−1/2)(z+1/2)
1− 41 z −2
= z + z−1/2 + z+1/2 = z + z −1 z−1/2 + z −1 z+1/2 . Taking
the inverse transform yields
5 5
h[n] = −4δ[n − 1] + (1/2)n−1 u[n − 1] + (−1/2)n−1 u[n − 1].
2 2
282
252
>> Hmag = 2*abs(cos(Omega));
>> plot(Omega,Hmag,’k’); axis([-pi pi 0 2.5]); grid;
>> xlabel(’\Omega’); ylabel(’|H(e^{j\Omega})|’);
>> set(gca,’xtick’,[-pi:pi/4:pi],’xticklabel’,[’-p ’;...
’ ’;’ ’;’ ’;’ 0 ’;’ ’;’ ’;’ ’;’ p ’],...
’fontname’,’symbol’);
2.5
1.5
|H(e )|
jΩ
0.5
0
−π 0 π
Ω
¯ ¯
Figure S5.M-6: ¯H(eΩ )¯ = 2| cos(Ω)|.
Using Figure S5.M-6, this system is best described by a bandstop filter with gain.
That is, low and high frequencies have a boosted gain of two and middle frequencies
near Ω = π/2 are attenuated.
5.M-7. (a) >> Omega = linspace(-pi,pi,501); z = exp(j*Omega);
>> H = cos(1./z);
>> plot(Omega,abs(H),’k’); axis([-pi pi 0 2]); grid;
>> xlabel(’\Omega’); ylabel(’|H(e^{j\Omega})|’);
>> set(gca,’xtick’,[-pi:pi/4:pi],’xticklabel’,[’-p ’;...
’ ’;’ ’;’ ’;’ 0 ’;’ ’;’ ’;’ ’;’ p ’],...
’fontname’,’symbol’);
1.8
1.6
1.4
1.2
|H(e )|
jΩ
0.8
0.6
0.4
0.2
0
−π 0 π
Ω
283
253
From Figure S5.M-7a, it appears that this system behaves somewhat like a band-
pass filter. Frequencies near Ω = π/2 are boosted while frequencies near Ω = kπ
are somewhat attenuated.
P∞ 2 k
(b) A Maclaurin series expansion of cos(x) is k=0 (−x )
(2k)! . Substituting n = z
−1
P ∞ −2 k
yields H(z) = cos(z −1 ) = k=0 (−z )
(2k)! . Inverting yields
∞
X (−1)k 1 1 1
h[n] = δ[n − 2k] = δ[n] − δ[n − 2] + δ[n − 4] − δ[n − 6] + . . . .
(2k)! 2! 4! 6!
k=0
0.5
h[n]
−0.5
0 1 2 3 4 5 6 7 8 9 10
n
(c) Using Figure S5.M-7b, it is clear that the impulse response quickly decays to zero.
Thus, only the first few terms from h[n] are needed for a good approximation.
Note that h[n] ≈ δ[n] − δ[n − 2]/2 + δ[n − 4]/24. An FIR difference equation is
found by letting h[n] = y[n] and δ[n] = x[n]. That is,
This is a fourth-order FIR filter with only three non-zero coefficients. The mag-
nitude response is easily computed using MATLAB.
>> Omega = linspace(-pi,pi,501); z = exp(j*Omega);
>> H = 1 - z.^(-2)/2 + z.^(-4)/24;
>> plot(Omega,abs(H),’k’); axis([-pi pi 0 2]); grid;
>> xlabel(’\Omega’); ylabel(’|H(e^{j\Omega})|’);
>> set(gca,’xtick’,[-pi:pi/4:pi],’xticklabel’,[’-p ’;...
’ ’;’ ’;’ ’;’ 0 ’;’ ’;’ ’;’ ’;’ p ’],...
’fontname’,’symbol’);
Visually, Figure S5.M-7c is indistinguishable from Figure S5.M-7b. Thus, the
FIR filter appears to closely approximate the original system.
284
254
2
1.8
1.6
1.4
1.2
|H(e )|
jΩ
1
0.8
0.6
0.4
0.2
0
−π 0 π
Ω
5.M-8. Factored form is used to plot roots, and standard transfer function form is used to
compute magnitude response plots.
(a) Order-8 Butterworth LPF with Ωc = π/3.
>> Omega_c = pi/3; Omega = linspace(-pi,pi,1001);
>> [z,p,k] = butter(8,Omega_c/pi);
>> subplot(121),plot(real(p),imag(p),’kx’,...
real(z),imag(z),’ko’,cos(Omega),sin(Omega),’k’);
>> axis([-1.1 1.1 -1.1 1.1]); axis equal;
>> xlabel(’Re(z)’); ylabel(’Im(z)’);
>> [B,A] = butter(8,Omega_c/pi);
>> HLP = polyval(B,exp(j*Omega))./polyval(A,exp(j*Omega));
>> subplot(122),plot(Omega,20*log10(abs(HLP)),’k’);
>> axis([-pi pi -40 2]); grid;
>> xlabel(’\Omega’); ylabel(’|H_{LP}(e^{j\Omega})|’);
>> set(gca,’xtick’,[-pi:pi/3:pi],’xticklabel’,[’-p ’;...
’ ’;’ ’;’ 0 ’;’ ’;’ ’;’ p ’],...
’fontname’,’symbol’);
(b) Order-8 Butterworth HPF with Ωc = π/3.
>> Omega_c = pi/3; Omega = linspace(-pi,pi,1001);
>> [z,p,k] = butter(8,Omega_c/pi,’high’);
>> subplot(121),plot(real(p),imag(p),’kx’,...
real(z),imag(z),’ko’,cos(Omega),sin(Omega),’k’);
>> axis([-1.1 1.1 -1.1 1.1]); axis equal;
>> xlabel(’Re(z)’); ylabel(’Im(z)’);
>> [B,A] = butter(8,Omega_c/pi,’high’);
>> HHP = polyval(B,exp(j*Omega))./polyval(A,exp(j*Omega));
>> subplot(122),plot(Omega,20*log10(abs(HHP)),’k’);
>> axis([-pi pi -40 2]); grid;
>> xlabel(’\Omega’); ylabel(’|H_{HP}(e^{j\Omega})|’);
>> set(gca,’xtick’,[-pi:pi/3:pi],’xticklabel’,[’-p ’;...
’ ’;’ ’;’ 0 ’;’ ’;’ ’;’ p ’],...
’fontname’,’symbol’);
285
255
2
0
1.5
−5
1
−10
0.5
−15
|HLP(ejΩ)|
Im(z)
0
−20
−0.5
−25
−1
−30
−1.5
−35
−2
−40
−1 −0.5 0 0.5 1 −π 0 π
Re(z) Ω
2
0
1.5
−5
1
−10
0.5
−15
|HHP(ejΩ)|
Im(z)
0
−20
−0.5
−25
−1
−30
−1.5
−35
−2
−40
−1 −0.5 0 0.5 1 −π 0 π
Re(z) Ω
(c) Order-8 Butterworth BPF with passband between 5π/24 and 11π/24. Notice
that the command butter requires the parameter N = 3 to be used to obtain a
(2N = 8)-order bandpass filter.
>> Omega_c = [5*pi/24,11*pi/24]; Omega = linspace(-pi,pi,1001);
>> [z,p,k] = butter(4,Omega_c/pi);
>> subplot(121),plot(real(p),imag(p),’kx’,...
real(z),imag(z),’ko’,cos(Omega),sin(Omega),’k’);
>> axis([-1.1 1.1 -1.1 1.1]); axis equal;
>> xlabel(’Re(z)’); ylabel(’Im(z)’);
>> [B,A] = butter(4,Omega_c/pi);
>> HBP = polyval(B,exp(j*Omega))./polyval(A,exp(j*Omega));
>> subplot(122),plot(Omega,20*log10(abs(HBP)),’k’);
>> axis([-pi pi -40 2]); grid;
>> xlabel(’\Omega’); ylabel(’|H_{BP}(e^{j\Omega})|’);
>> set(gca,’xtick’,[-pi:pi/3:pi],’xticklabel’,[’-p ’;...
’ ’;’ ’;’ 0 ’;’ ’;’ ’;’ p ’],...
’fontname’,’symbol’);
286
256
2
0
1.5
−5
1
−10
0.5
−15
|HBP(ejΩ)|
Im(z)
0
−20
−0.5
−25
−1
−30
−1.5
−35
−2
−40
−1 −0.5 0 0.5 1 −π 0 π
Re(z) Ω
Figure S5.M-8c: Order-8 Butterworth BPF with passband between 5π/24 and 11π/24.
(d) Order-8 Butterworth BSF with stopband between 5π/24 and 11π/24. Notice
that the command butter requires the parameter N = 4 to be used to obtain a
(2N = 8)-order bandstop filter.
>> Omega_c = [5*pi/24,11*pi/24]; Omega = linspace(-pi,pi,1001);
>> [z,p,k] = butter(4,Omega_c/pi,’stop’);
>> subplot(121),plot(real(p),imag(p),’kx’,...
real(z),imag(z),’ko’,cos(Omega),sin(Omega),’k’);
>> axis([-1.1 1.1 -1.1 1.1]); axis equal;
>> xlabel(’Re(z)’); ylabel(’Im(z)’);
>> [B,A] = butter(4,Omega_c/pi,’stop’);
>> HBS = polyval(B,exp(j*Omega))./polyval(A,exp(j*Omega));
>> subplot(122),plot(Omega,20*log10(abs(HBS)),’k’);
>> axis([-pi pi -40 2]); grid;
>> xlabel(’\Omega’); ylabel(’|H_{BS}(e^{j\Omega})|’);
>> set(gca,’xtick’,[-pi:pi/3:pi],’xticklabel’,[’-p ’;...
’ ’;’ ’;’ 0 ’;’ ’;’ ’;’ p ’],...
’fontname’,’symbol’);
2
0
1.5
−5
1
−10
0.5
−15
|HBS(ejΩ)|
Im(z)
0
−20
−0.5
−25
−1
−30
−1.5
−35
−2
−40
−1 −0.5 0 0.5 1 −π 0 π
Re(z) Ω
Figure S5.M-8d: Order-8 Butterworth BSF with stopband between 5π/24 and 11π/24.
287
257
5.M-9. Factored form is used to plot roots, and standard transfer function form is used to
compute magnitude response plots.
(a) Order-8 Chebyshev Type I LPF with Ωc = π/3.
>> Omega_c = pi/3; Omega = linspace(-pi,pi,1001);
>> [z,p,k] = cheby1(8,3,Omega_c/pi);
>> subplot(121),plot(real(p),imag(p),’kx’,...
real(z),imag(z),’ko’,cos(Omega),sin(Omega),’k’);
>> axis([-1.1 1.1 -1.1 1.1]); axis equal;
>> xlabel(’Re(z)’); ylabel(’Im(z)’);
>> [B,A] = cheby1(8,3,Omega_c/pi);
>> HLP = polyval(B,exp(j*Omega))./polyval(A,exp(j*Omega));
>> subplot(122),plot(Omega,20*log10(abs(HLP)),’k’);
>> axis([-pi pi -40 2]); grid;
>> xlabel(’\Omega’); ylabel(’|H_{LP}(e^{j\Omega})|’);
>> set(gca,’xtick’,[-pi:pi/3:pi],’xticklabel’,[’-p ’;...
’ ’;’ ’;’ 0 ’;’ ’;’ ’;’ p ’],...
’fontname’,’symbol’);
2
0
1.5
−5
1
−10
0.5
−15
|HLP(ejΩ)|
Im(z)
0
−20
−0.5
−25
−1
−30
−1.5
−35
−2
−40
−1 −0.5 0 0.5 1 −π 0 π
Re(z) Ω
288
258
2
0
1.5
−5
1
−10
0.5
−15
|HHP(ejΩ)|
Im(z)
0
−20
−0.5
−25
−1
−30
−1.5
−35
−2
−40
−1 −0.5 0 0.5 1 −π 0 π
Re(z) Ω
(c) Order-8 Chebyshev Type I BPF with passband between 5π/24 and 11π/24. No-
tice that the command cheby1 requires the parameter N = 4 to be used to obtain
a (2N = 8)-order bandpass filter.
>> Omega_c = [5*pi/24,11*pi/24]; Omega = linspace(-pi,pi,1001);
>> [z,p,k] = cheby1(4,3,Omega_c/pi);
>> subplot(121),plot(real(p),imag(p),’kx’,...
real(z),imag(z),’ko’,cos(Omega),sin(Omega),’k’);
>> axis([-1.1 1.1 -1.1 1.1]); axis equal;
>> xlabel(’Re(z)’); ylabel(’Im(z)’);
>> [B,A] = cheby1(4,3,Omega_c/pi);
>> HBP = polyval(B,exp(j*Omega))./polyval(A,exp(j*Omega));
>> subplot(122),plot(Omega,20*log10(abs(HBP)),’k’);
>> axis([-pi pi -40 2]); grid;
>> xlabel(’\Omega’); ylabel(’|H_{BP}(e^{j\Omega})|’);
>> set(gca,’xtick’,[-pi:pi/3:pi],’xticklabel’,[’-p ’;...
’ ’;’ ’;’ 0 ’;’ ’;’ ’;’ p ’],...
’fontname’,’symbol’);
(d) Order-8 Chebyshev Type I BSF with stopband between 5π/24 and 11π/24. No-
tice that the command cheby1 requires the parameter N = 4 to be used to obtain
a (2N = 8)-order bandstop filter.
>> Omega_c = [5*pi/24,11*pi/24]; Omega = linspace(-pi,pi,1001);
>> [z,p,k] = cheby1(4,3,Omega_c/pi,’stop’);
>> subplot(121),plot(real(p),imag(p),’kx’,...
real(z),imag(z),’ko’,cos(Omega),sin(Omega),’k’);
>> axis([-1.1 1.1 -1.1 1.1]); axis equal;
>> xlabel(’Re(z)’); ylabel(’Im(z)’);
>> [B,A] = cheby1(4,3,Omega_c/pi,’stop’);
>> HBS = polyval(B,exp(j*Omega))./polyval(A,exp(j*Omega));
>> subplot(122),plot(Omega,20*log10(abs(HBS)),’k’);
>> axis([-pi pi -40 2]); grid;
>> xlabel(’\Omega’); ylabel(’|H_{BS}(e^{j\Omega})|’);
>> set(gca,’xtick’,[-pi:pi/3:pi],’xticklabel’,[’-p ’;...
’ ’;’ ’;’ 0 ’;’ ’;’ ’;’ p ’],...
289
259
2
0
1.5
−5
1
−10
0.5
−15
|HBP(ejΩ)|
Im(z)
0
−20
−0.5
−25
−1
−30
−1.5
−35
−2
−40
−1 −0.5 0 0.5 1 −π 0 π
Re(z) Ω
Figure S5.M-9c: Order-8 Chebyshev Type I BPF with passband between 5π/24 and
11π/24.
’fontname’,’symbol’);
2
0
1.5
−5
1
−10
0.5
−15
|HBS(ejΩ)|
Im(z)
0
−20
−0.5
−25
−1
−30
−1.5
−35
−2
−40
−1 −0.5 0 0.5 1 −π 0 π
Re(z) Ω
Figure S5.M-9d: Order-8 Chebyshev Type I BSF with stopband between 5π/24 and
11π/24.
290
260
>> legend(’R_p = 0.1’,’R_p = 1.0’,’R_p = 3.0’,0);
>> set(gca,’xtick’,[-pi:pi/3:pi],’xticklabel’,[’-p ’;...
’ ’;’ ’;’ 0 ’;’ ’;’ ’;’ p ’],...
’fontname’,’symbol’);
0 Rp = 0.1
Rp = 1.0
Rp = 3.0
−5
−10
−15
|HLP(ejΩ)|
−20
−25
−30
−35
−40
−π 0 π
Ω
Thus, reducing the allowable passband ripple Rp tends to broaden the transition
bands of the filter.
5.M-10. Factored form is used to plot roots, and standard transfer function form is used to
compute magnitude response plots.
291
261
2
0
1.5
−5
1
−10
0.5
−15
|HLP(ejΩ)|
Im(z)
0
−20
−0.5
−25
−1
−30
−1.5
−35
−2
−40
−1 −0.5 0 0.5 1 −π 0 π
Re(z) Ω
2
0
1.5
−5
1
−10
0.5
−15
|HHP(ejΩ)|
Im(z)
0
−20
−0.5
−25
−1
−30
−1.5
−35
−2
−40
−1 −0.5 0 0.5 1 −π 0 π
Re(z) Ω
(c) Order-8 Chebyshev Type II BPF with passband between 5π/24 and 11π/24.
Notice that the command cheby2 requires the parameter N = 4 to be used to
obtain a (2N = 8)-order bandpass filter.
>> Omega_c = [5*pi/24,11*pi/24]; Omega = linspace(-pi,pi,1001);
>> [z,p,k] = cheby2(4,20,Omega_c/pi);
>> subplot(121),plot(real(p),imag(p),’kx’,...
real(z),imag(z),’ko’,cos(Omega),sin(Omega),’k’);
>> axis([-1.1 1.1 -1.1 1.1]); axis equal;
>> xlabel(’Re(z)’); ylabel(’Im(z)’);
>> [B,A] = cheby2(4,20,Omega_c/pi);
292
262
>> HBP = polyval(B,exp(j*Omega))./polyval(A,exp(j*Omega));
>> subplot(122),plot(Omega,20*log10(abs(HBP)),’k’);
>> axis([-pi pi -40 2]); grid;
>> xlabel(’\Omega’); ylabel(’|H_{BP}(e^{j\Omega})|’);
>> set(gca,’xtick’,[-pi:pi/3:pi],’xticklabel’,[’-p ’;...
’ ’;’ ’;’ 0 ’;’ ’;’ ’;’ p ’],...
’fontname’,’symbol’);
2
0
1.5
−5
1
−10
0.5
−15
|HBP(ejΩ)|
Im(z)
0
−20
−0.5
−25
−1
−30
−1.5
−35
−2
−40
−1 −0.5 0 0.5 1 −π 0 π
Re(z) Ω
Figure S5.M-10c: Order-8 Chebyshev Type II BPF with passband between 5π/24 and
11π/24.
(d) Order-8 Chebyshev Type II BSF with stopband between 5π/24 and 11π/24.
Notice that the command cheby2 requires the parameter N = 4 to be used to
obtain a (2N = 8)-order bandstop filter.
>> Omega_c = [5*pi/24,11*pi/24]; Omega = linspace(-pi,pi,1001);
>> [z,p,k] = cheby2(4,20,Omega_c/pi,’stop’);
>> subplot(121),plot(real(p),imag(p),’kx’,...
real(z),imag(z),’ko’,cos(Omega),sin(Omega),’k’);
>> axis([-1.1 1.1 -1.1 1.1]); axis equal;
>> xlabel(’Re(z)’); ylabel(’Im(z)’);
>> [B,A] = cheby2(4,20,Omega_c/pi,’stop’);
>> HBS = polyval(B,exp(j*Omega))./polyval(A,exp(j*Omega));
>> subplot(122),plot(Omega,20*log10(abs(HBS)),’k’);
>> axis([-pi pi -40 2]); grid;
>> xlabel(’\Omega’); ylabel(’|H_{BS}(e^{j\Omega})|’);
>> set(gca,’xtick’,[-pi:pi/3:pi],’xticklabel’,[’-p ’;...
’ ’;’ ’;’ 0 ’;’ ’;’ ’;’ p ’],...
’fontname’,’symbol’);
To demonstrate the effect of increasing Rs , consider magnitude response plots
for Chebyshev Type II LPFs with Rs = {10, 20, 30}.
>> Omega_c = pi/3; Omega = linspace(-pi,pi,1001);
>> [B,A] = cheby2(8,10,Omega_c/pi);
>> HLP1 = polyval(B,exp(j*Omega))./polyval(A,exp(j*Omega));
>> [B,A] = cheby2(8,20,Omega_c/pi);
>> HLP2 = polyval(B,exp(j*Omega))./polyval(A,exp(j*Omega));
>> [B,A] = cheby2(8,30,Omega_c/pi);
293
263
2
0
1.5
−5
1
−10
0.5
−15
|HBS(ejΩ)|
Im(z)
0
−20
−0.5
−25
−1
−30
−1.5
−35
−2
−40
−1 −0.5 0 0.5 1 −π 0 π
Re(z) Ω
Figure S5.M-10d: Order-8 Chebyshev Type II BSF with stopband between 5π/24 and
11π/24.
0 Rs = 10
Rs = 20
Rs = 30
−5
−10
−15
|HLP(ejΩ)|
−20
−25
−30
−35
−40
−π 0 π
Ω
294
264
>> [z,p,k] = ellip(8,3,20,Omega_c/pi);
>> subplot(121),plot(real(p),imag(p),’kx’,...
real(z),imag(z),’ko’,cos(Omega),sin(Omega),’k’);
>> axis([-1.1 1.1 -1.1 1.1]); axis equal;
>> xlabel(’Re(z)’); ylabel(’Im(z)’);
>> [B,A] = ellip(8,3,20,Omega_c/pi);
>> HLP = polyval(B,exp(j*Omega))./polyval(A,exp(j*Omega));
>> subplot(122),plot(Omega,20*log10(abs(HLP)),’k’);
>> axis([-pi pi -40 2]); grid;
>> xlabel(’\Omega’); ylabel(’|H_{LP}(e^{j\Omega})|’);
>> set(gca,’xtick’,[-pi:pi/3:pi],’xticklabel’,[’-p ’;...
’ ’;’ ’;’ 0 ’;’ ’;’ ’;’ p ’],...
’fontname’,’symbol’);
2
0
1.5
−5
1
−10
0.5
−15
|HLP(ejΩ)|
Im(z)
0
−20
−0.5
−25
−1
−30
−1.5
−35
−2
−40
−1 −0.5 0 0.5 1 −π 0 π
Re(z) Ω
295
265
2
0
1.5
−5
1
−10
0.5
−15
|HHP(ejΩ)|
Im(z)
0
−20
−0.5
−25
−1
−30
−1.5
−35
−2
−40
−1 −0.5 0 0.5 1 −π 0 π
Re(z) Ω
2
0
1.5
−5
1
−10
0.5
−15
|HBP(ejΩ)|
Im(z)
0
−20
−0.5
−25
−1
−30
−1.5
−35
−2
−40
−1 −0.5 0 0.5 1 −π 0 π
Re(z) Ω
Figure S5.M-11c: Order-8 Elliptic BPF with passband between 5π/24 and 11π/24.
(d) Order-8 Elliptic BSF with stopband between 5π/24 and 11π/24. Notice that the
command ellip requires the parameter N = 4 to be used to obtain a (2N = 8)-
order bandstop filter.
>> Omega_c = [5*pi/24,11*pi/24]; Omega = linspace(-pi,pi,1001);
296
266
>> [z,p,k] = ellip(4,3,20,Omega_c/pi,’stop’);
>> subplot(121),plot(real(p),imag(p),’kx’,...
real(z),imag(z),’ko’,cos(Omega),sin(Omega),’k’);
>> axis([-1.1 1.1 -1.1 1.1]); axis equal;
>> xlabel(’Re(z)’); ylabel(’Im(z)’);
>> [B,A] = ellip(4,3,20,Omega_c/pi,’stop’);
>> HBS = polyval(B,exp(j*Omega))./polyval(A,exp(j*Omega));
>> subplot(122),plot(Omega,20*log10(abs(HBS)),’k’);
>> axis([-pi pi -40 2]); grid;
>> xlabel(’\Omega’); ylabel(’|H_{BS}(e^{j\Omega})|’);
>> set(gca,’xtick’,[-pi:pi/3:pi],’xticklabel’,[’-p ’;...
’ ’;’ ’;’ 0 ’;’ ’;’ ’;’ p ’],...
’fontname’,’symbol’);
2
0
1.5
−5
1
−10
0.5
−15
|HBS(ejΩ)|
Im(z)
0
−20
−0.5
−25
−1
−30
−1.5
−35
−2
−40
−1 −0.5 0 0.5 1 −π 0 π
Re(z) Ω
Figure S5.M-11d: Order-8 Elliptic BSF with stopband between 5π/24 and 11π/24.
297
267