Chapter 3
Q1. Determine the poles and zeros of H(z)=1+z-1+z-2.
1 z z2
H ( z ) 1 z 1 z 2
z2
1 j 3
Solving 1 z z 2 =0, gives us zeros at
2
Solving z 2 =0, two poles at z 0.
Q2. The causal system H(z)=1/(1-0.9z-1)2 is excited by a unit-step sequence. Use the residue theorem
to determine the transient and steady state responses of this system.
z z2
Y ( z)
z 1 ( z 0.9) 2
y (n) RES[ z n 1 Y ( z )]z 1 RES[ z n 1 Y ( z )]z 0.9
( z 1) z n 2 d ( z 0.9) 2 z n 2
2
2
( z 1)( z 0.9) z 1 dz ( z 1)( z 0.9) z 0.9
1 (n 2) z n 1 z n2
2
(1 0.9) ( z 1) ( z 1) 2 z 0.9
y ss (n) 100
(n 2)0.9n 1 0.9n 2
y tr (n)
(0.1) (0.1) 2
Q3. Determine the following z-transforms and give their ROC:
a) x(n) = 0.6|n|.
z 1 (0.6 0.61 )
z transform:
(1 0.6 z 1 )(1 0.61 z 1 )
1
ROC: 0.6 | z |
0.6
b) x(n) = 5n (u(n-1) – u(n-10)).
5 z 1 510 z 10
z transform
1 5 z 1
ROC: | z | 0
Q4. Determine the following inverse z-transforms:
c) X ( z ) 5 3z 2 z 1 , 0 | z | .
h(n)=5δ(n)+3δ(n+1)+ 2δ(n-1)
z2
d) X ( z ) 2 , 0.6 | z | 0.8.
z 1.4 z 0.48
z2 z2 z z
X ( z) C1 C2 ,
z 1.4 z 0.48 ( z 0.6)( z 0.8)
2
( z 0.6) ( z 0.8)
0.6 | z | 0.8.
h(n) C1 0.6n u (n) C2 0.8n u (n 1)
C1 3 C2 4
Q5.
Q2. The following digital filter is given
The filter coefficients are quantized to 5 bits in fixed point where the binary representation is sign
magnitude of the form, xxx.xx (b5 b4 b3 .b2 b1), with the most significant bit being the sign. The fixed
point numerical representation here is of the form b5 (sign bit), b4 x 21 + b3 x 20 + b2 x 2-1 + b1 x 2-2
a) Write the quantized frequency response function.
20 (20 21 ) z 1 (21 22 ) z 2
H ( z)
20 (20 21 22 ) z 1 (21 22 ) z 2
１ 1.5 z 1 0.75 z 2
１ 1.75 z 1 0.75 z 2
b) Plot the quantized frequency response function magnitude in dB and compare with the
unquantized frequency response given above.
Comparison between two frequency responses
Q6. The transfer function below is given
1 1 1 2 1 3
1 z z z
H ( z) 4 9 36
1 1 2 1 3
1 z z z
2 2
a) Give a cascade realization of the filter.
1 2 1 1 1 1 1 1
1 z 1 z 1 z 1 z 1 1 z 1
H z 4
9 3 3 1 4 1
1
1 z 1 z 1 z 1 1 z 1 z
1 2 1
2 2 H 2 ( z ) H3 ( z )
H1 ( z )
b) Give a parallel realization of the filter.
Using the long division, we have
1 2 2 1 17
z z
1 18 9 18
H ( z)
18 1 z 3 z 2 1 z 1 1
2 2
1 C1 C2 C3
H ( z)
18 1 z 1 z 1 1
1 1
1 z
2
C1 1 z 1 H ( z ) |z 1
2
3
C2 1 z 1 H ( z ) |z 1
10
27
1 5
C1 1 z 1 H ( z ) |z 1/2
2 54
2 10 5
1 3
Therefore, we have H ( z ) 271 54
18 1 z 1 z 1 1
1
1 z
Ha ( z ) Hb ( z ) 2
Hc ( z )
Chapter 4 and 5
Q1
600
1 =2 =0.3
4000
1200
2 =2 =0.6
4000
1 2
hd (n) sin c 1n n sin c 2 n
Therefore, by applying a window, we have
hFIR (n) hd (n) w(n)
a. for rectangular window, w(n)=1
so
1 2
hFIR (n) sin c 1n n sin c 2 n
For the 4th order filter, by taking n = -2,-1,0,1,2, we have
h 2 0.2449
h( 1) 0.0452
h(0) 0.7
h(1) 0.0452
h(2) 0.2449
b) for triangular window, L = 6, w(n)=[1/3, 2/3, 1, 2/3, 1/3]
For the 4th order filter, we have
h 2 0.0816
h( 1) -0.0301
h(0) 0.7
h(1) -0.0301
h(2) 0.0816
c) Frequency response of rectangular window
Frequency response of triangular window
d) The cutoff characteristics for the triangular window are much closer to the ideal band stop filter.
Chapter 4
Q23
0.4
1
e d 0.4sinc(0.4 n)
jn
hLP (n) 3 n 3
2 0.4
hLP(n)=[
-0.0624 0.0935 0.3027 0.4000 0.3027 0.0935 -0.0624] -3<n<3
a) bn= hLP(n-3); rectangular window
b) Triangular Window samples: w(n)=[¼, ½, ¾, 1, ¾, ½, ¼] -3<n<3
bn= hLP(n-3)*w(n-3); triangular window
hLP(n = [0.0156 0.0468 0.2270 0.4000 0.2270 0.0468 -0.0156]
c) Hamming window
w(n)=[0.0800 0.3100 0.7700 1.0000 0.7700 0.3100 0.0800] -3<n<3
bn= hLP(n-3)*w(n-3); Hamming window
h=[-0.0050 0.0290 0.2331 0.4000 0.2331 0.0290 -0.0050]
50
Magnitude Response (dB)
0
-50
-100
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Normalized frequency (Nyquist == 1)
0
Phase (degrees)
-100
-200
-300
-400
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Normalized frequency (Nyquist == 1)
0
Magnitude Response (dB)
-10
-20
-30
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Normalized frequency (Nyquist == 1)
0
Phase (degrees)
-200
-400
-600
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Normalized frequency (Nyquist == 1)
0
Magnitude Response (dB)
-20
-40
-60
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Normalized frequency (Nyquist == 1)
0
Phase (degrees)
-200
-400
-600
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Normalized frequency (Nyquist == 1)
Chapter 5
Q14
Use analog filter approximation theory to convert a first order high-pass circuit to a digital filter with: (a)
the impulse invariance method and (b) the bilinear transformation. Take RC=1.
s
H ( s) and hence h(t ) (t ) e t u (t )
s 1
T
a. h(n) T (n) Te nT u (n) and H ( z ) T
1 e T z 1
z 1 z 1
b. s , H (s ) 0.5 0.5 z 1
z 1 z 1
Q18
Determine the order and poles of a digital Butterworth filters
f s 8kHz
p 2 *1200 / 8000 0.3
st 0.45
p tan( p / 2) 0.5095, st tan( st / 2) 0.854
1
p
2M 1
G 0.794
1
st 1
.001
log(2.5*104)
Thus M 8
log(0.356)
16
0.854
999 . Thus c 0.5549radians.
c
j
The s domain poles can be written as si re i . There are a total of 16 poles with angular separation
between 2 poles 2 /16 . The first pole is at an angle /16 , second pole at 3 /16 and so on and the last
pole at 31 /16 .
The left half s plane poles s5 to s12 are chosen for stability.
s5 = -0.1082 - 0.5442i z5 = -0.4539 + 0.7140i
s6 =-0.3082 - 0.4613i z6 = -0.3596 + 0.4794i
s7 = -0.4613 - 0.3082i z7 = -0.3102 + 0.2764i
s8 = -0.5442 - 0.1082i z8 = -0.28880 + 0.09034i
s9 = -0.5442 + 0.1082i z9 = -0.28880 - 0.09034i
s10 = -0.4613 + 0.3082i z10 = -0.3102 - 0.2764i
s11 = -0.3082 + 0.4613i z11 = -0.3596 - 0.4794i
s12 = -0.1082 + 0.5442i z12 = -0.4539 - 0.7140i
