of ECE
KL University, Vaddeswaram,
Dept. of ECE,
Signal Processing (B. Tech all branches) 13-ES205
Lesson-17: IIR-Digital Filters
A digital filter is a Linear Time Invariant System. The input
sequence is modified according to the characteristics of the
system and gives some output. Hence the system is acting
some kind of filtering operation. In time domain the input and
outputs are related by y[n] h[n] x[n] , where x[n] , y[n] and
sequence and impulse response of the system respectively. In frequency domain, they are related as
Y [ z] H [ z] X [ z] , where H [ z ]
Y [ z]
is referred to as System Function or Transfer Function or
X [ z]
Frequency Response of the system. Hence any filter can be characterized by either impulse response
h[n] 0, for n 0 .
2. Classification of Digital Filters: Digital filters are categorized into two types.
(a) IIR Filter: Infinite duration Impulse Response. Recursive type. The present output sample
depends on the present input sample, past input samples and output samples.
(b) FIR Filter: Finite duration Impulse Response. Non-recursive type. The present output sample
depends on the present input sample and past input samples.
Frequency Selective Filters: A filter is one which rejects un-wanted frequencies from the input signal
and allow the desired frequencies to obtain the required shape of the output signal. The filters are
categorized into four types as described below:
1, for 0 || c
0, for c ||
0, for 0 || c
1, for c ||
0, for 0 || c
1
0, for c2 ||
1, for 0 || c
1
1, for c2 ||
Practical Filters
3. IIR Digital Filters: The input output relation of a IIR digital filter is described by a difference
equation:
N
k 1
k 0
H [ z]
k
bk z
k 0
N
1 ak z k
k 10
The design of an IIR filter for the given specifications is determining filter co-efficients ak and bk of the
filter.
2
4 Analog Lowpass Filter Design: The most general form of analog filter transfer function is described
M
bk s k
N (s)
by H [ s ]
, where H [ s] is L.T. of h(t ) , ak and bk are filter coefficients, N is the
k 0
D( s) 1 N a s k
k
k 1
order of the filter, N > M. For stable filter, the poles of H [ s] must lie in the left half of s-plane.
3
5. Analog Lowpass Butterworth Filter: All pole filter. The magnitude function of the Butterworth
LPF
| H ( j) |
2N
, N = 1, 2, . .
| H ( j) |2
1
, N = 1, 2, . . .
1 () 2 N
1
1
For stable filter the poles must lie on the left half of s-plane. Therefore the stable files are derived from
the following equation:
sk e jk ,
where k
2
(2k 1)
,
2N
k 1, 2,. . . , N .
Denominator of H ( s)
(s 1)
(s 2 2s 1)
(s 1)(s 2 s 1)
100.1 s 1
log
100.1 p 1
N
log s
p
log 2
1
2 1
1
N
log s
p
log
N
log s
p
Ex1: Determine the order of LPF if it has pass band attenuation of - 3dB at 500 Hz and stop
band attenuation of - 40dB at 1000 Hz.
Ans: The pass band attenuation is -3 dB.
The stop band attenuation is 40 dB.
The pass band cut off frequency in radian is p = 500 2 1000 rad / sec
5
The stop band cut off frequency in radian is s 1000 2 2000 rad / sec
100.1 p 1 log
2
0.9976
6.64
s
2000 0.3
log
log
p
1000
Ex2: Determine the order of LPF if it has pass band attenuation of - 1dB at 4 KHz and stop
100.1 s 1
log
100.1 p 1
13.0239
N
s
log
p
% IIR2.m
clear all; close all; clc;
% N = ((log10((10.^(0.1*abs(rs))-1)./(10.^(0.1*abs(rp))-1)))/(2*log10(Os/Op)))
N = ceil((log10((10.^(0.1*abs(rs))-1)./(10.^(0.1*abs(rp))-1)))/(2*log10(Os/Op)));
N
z e sT e( j)T e T e jT
And also we know that z re
. Therefore re
jT
.
e T e
Design Procedure:
In impulse invariance method, the IIR filter is designed such that the Unit Impulse Response
h[n] of
digital filter is the sampled version of the Unit Impulse Response of analog filter ha (t ) .
The Analog Filter transfer function is described by
H a ( s)
Ai
i 1 s pi
A1
s p1
A2
s p2
..
Ai e
pit
i 1
Ai e
i 1
pi nT
H [ z]
n0
n0 i 1
h[n] z n Ai e
pi nT n
z
piT 1
n
i
H [ z ] Ai
e
z Ai
e
z
i 1 n0
i 1 n0
N
Ai
pT
i 1 1 e i z 1
Limitations:
(a) Analog filters are band limited, so there will be aliasing due to sampling process. Because of
this aliasing, the frequency response of resulting digital filter will not be identical to the original
frequency response of analog filter.
(b) The change in the value of sampling time has no effect on the amount of aliasing.
(c) The analog frequency is in the range to , which maps into digital frequency
range of
in the
H a ( s)
Ai
i 1 s pi
A1
s p1
A2
s p2
..
Step4: To convert analog low pass filter to digital low pass filter using impulse-invariant
transformation, substitute,
1
1
pk T 1 .
s pk
1 e z
Ai
pT
i 1 1 e i z 1
2
s2
Design digital filter using Impulse Invariance Method. Assume suitable data.
Ans: The sampling period is assumed to be T = 1 sec.
For the given analog filter transfer function the pole is p = -2.
To convert analog low pass filter to digital low pass filter using impulse-invariant transformation,
substitute,
1
1
pk T 1 .
s pk
1 e z
Therefore
H [ z]
1
p T
1 e k z 1
2
1 0.1353z 1
% IIR3.m.
clear all; close all; clc;
% Analog poles and zeros
b = [2]; a = [1 2];
T = 1; fs = 1/T;
Magnitude (dB)
1
0
-1
-2
-3
0.05
0.1
0.15
0.2
0.25
0.3
Frequency (Hz)
0.35
0.4
0.45
0.5
0.05
0.1
0.15
0.2
0.25
0.3
Frequency (Hz)
0.35
0.4
0.45
0.5
Phase (degrees)
0
-2
-4
-6
-8
H a ( s)
b
sa
where s
b
2 1 z 1
T 1 z 1 a
2 1 z 1
1 sT / 2
.
or z
1
1 sT / 2
T 1 z
T
.
tan or 2 tan 1
T
2
2
Warping Effect:
(a) At low frequencies
2
or = T , i.e., the digital
.
T 2 T
Pre-warping: This warping effect can be eliminated by pre-warping the analog frequencies as below:
p
2
2
and s tan s .
tan . Therefore p tan
T
2
T
2
T
2
Design Steps:
Step1: From the given specifications, find the prewarping analog frequencies using the formulas:
p
2
2
and s tan s .
tan
T
2
T
2
Step2: Using the analog frequencies determine the analog filter transfer function H a ( s) .
Step3: Select the sampling rate of the digital filter T sec/sample.
Step4: Determine the digital filter transfer function by substituting s
2 1 z 1
transfer function.
11
Ex4: Design a digital low pass filter using bi-linear transformation for the following analog filter transfer
function.
H a ( s)
0.4225
s 2 0.9192s 0.4225
H [ z ] H a ( s) |
s 2 1 z 1
T 1 z
2 1 z 1
T 1 z 1
0.4225
2
2 1 z 1
2 1 z 1
0.9192
0.4225
1
1
T
T
1
z
1
On simplification, we get
H [ z ] 0.0675
1 2 z 1 z 2
1 1.1428z 1 0.4127z 2
% IIR3.m.
clear all; close all; clc;
b = [0.4225]; a = [1 0.9192 0.4225]; fs =1;
[bz,az] = bilinear(b,a,fs);
%bz = 0.0675
%az = 1.0000
0.1350
-1.1428
0.0675
0.4127
freqz(bz,az,512,fs);
Magnitude (dB)
50
0
-50
-100
-150
0.05
0.1
0.15
0.2
0.25
0.3
Frequency (Hz)
0.35
0.4
0.45
0.5
0.05
0.1
0.15
0.2
0.25
0.3
Frequency (Hz)
0.35
0.4
0.45
0.5
Phase (degrees)
0
-50
-100
-150
-200
12
---------------------------------------------------------------------------------------------------------------------------------------
Convert these digital filter parameters into corresponding analog filter parameters according to the filter
type.
(a) Impulse Invariance Method:
Substitute for analog frequency
. That is p
and s s , where T is sampling
T
T
T
tan .
T
2
p
2
2
and s tan s .
tan
T
2
T
2
100.1 s 1
log
100.1 p 1
N
log s
p
22
log
1
2 1
1
N
s
log
log
N
log s
p
Step3: Formulate the normalized analog low pass Butterworth filter transfer function as
1
H ( s)
Polynomial
where polynomial for various values of N is given below
List of Butterworth polynomial
Denominator of H ( s)
(s 1)
(s 2 2s 1)
(s 1)(s 2 s 1)
Transfer Function
H ( s)
H ( s)
H ( s)
H ( s)
(s 1)
1
(s 2s 1)
1
(s 1)(s 2 s 1)
1
Note: If
10
p
0.1 p
1
2N
p 3dB , c p
s s
c
Step6: Apply transformation technique to convert analog filter into respective digital filter.
(a) Impulse Invariance Technique: Substitute
1
s Pk
1
PT
1 e k z 1
2 1 z 1
T 1 z 1
Step7: The final digital filter transfer function must be presented in the following format.
b b z 1 b2 z 2 ....
H [ z ] 0 1 1
1 a1z a2 z 2 ....
15
Ex5: Design a digital low pass filter Butterworth filter using bilinear transformation with pass
band and stop band cut-off frequencies 800 rad/sec and 1800 rad/sec respectively. The pass
band attenuation is -3 dB and stop band attenuation is -10dB.
Ans: The given filter specification are:
2
tan .
T
2
16
800
3.23 rad / sec
2
Therefore, p 2 tan
1800
s 2 tan
30.12 rad / sec
2
% IIR4.m.
clear all; close all; clc;
% Bilinear transformation
wp = 800; ws = 1800; rp = 3; rs = 10; T = 1;
% Prewarping analog frequencies
Op = (2/T)*tan(wp/2);
Os = (2/T)*tan(ws/2);
Op
Os
100.1 s 1
3
log
100.1 p 1 log
0.997
0.4913 1
s
30.12
log
log
3.23
N = ceil((log10((10.^(0.1*abs(rs))-1)./(10.^(0.1*abs(rp))-1)))/(2*log10(Os/Op)));
N
The transfer function for 1 order normalized low pass filter is given by, H ( s)
The cut-off frequency, c
10
p
0.1 p
1
2N
1
s 1
3.23
3.24 rad / sec
(0.997)
Oc = Op / ((10^(0.1*abs(rp))-1)^(1/(2*N)));
Oc
H a ( s)
b = [1]; a = [1
1
s
1
3.24
s
,
3.24
3.24
s 3.24
[B A]=lp2lp(b,a, Oc); % Desired Analog filter transfer function with cut off
% frequency Oc
17
2 1 Z 1
To convert low pass filter to digital low pass filter using bilinear transformation, s
T 1 Z 1
in the magnitude response H ( s) .
Then the required digital filter transfer function H ( Z )
3.24
0.62(1 Z 1)
1 0.24Z 1
1 Z 1
2
3.24
1
1 Z
Magnitude (dB)
0
-5
-10
-15
-20
0.05
0.1
0.15
0.2
0.25
0.3
Frequency (Hz)
0.35
0.4
0.45
0.5
0.05
0.1
0.15
0.2
0.25
0.3
Frequency (Hz)
0.35
0.4
0.45
0.5
Phase (degrees)
-50
-100
Ex6: Design a digital low pass Butterworth filter using impulse-invariant transformation with pass band
and stop band cut off frequencies 200 Hz and 500 Hz respectively. The pass band and stop band cut
off frequencies 200 Hz and 500 Hz respectively. The pass band and stop band attenuation are -5 dB
and -12 dB respectively. The sampling frequency is 5000 Hz.
Ans: The given specifications of digital filter are:
2 Fstop 2 500
2 Fpass 2 200
% IIR6.m:
clear all; close all; clc;
% wp = 2*pi*(200/5000); ws = 2*pi*(500/5000); rp = 5; rs = 12; T = 1;
wp = 0.08*pi; ws = 0.2*pi; rp = 5; rs = 12; T = 1; fs=1/T;
so
.
T
100.1 s 1
log
100.1 p 1
1.05 2
The order of low pass filter is, N
log s
p
N = ceil((log10((10.^(0.1*abs(rs))-1)./(10.^(0.1*abs(rp))-1)))/(2*log10(Os/Op)));
N
The order of low pass filter for the given magnitude response is 2. The transfer function for II order
normalized low pass filter is given by,
H ( s)
The cut off frequency is given by c
1
s 2s 1
2
p
10
0.1 p
1
2N
Oc = Op / ((10^(.1*abs(rp))-1)^(1/(2*N)));
Oc
The desired analog filter transfer function of low pass filter can be obtained by replacing, s
That is s
s
.
c
s
s
.
c 0.2073
H a ( s)
1
2
s
s
2
1
0.2073
0.2073
0.0430
s 2 0.2932s 0.0430
Impulse-Invariant Transformation:
By converting single pole filters as
0.1466j
0.1466j
H a ( s)
1
1
substitute,
H [ z]
0.0370z 1
1.0000 -1.7088z 1 0.7459z 2
[bz,az] = impinvar(B,A,fs);
bz
az
freqz(bz,az,512,fs);
Magnitude (dB)
0
-10
-20
-30
-40
0.05
0.1
0.15
0.2
0.25
0.3
Frequency (Hz)
0.35
0.4
0.45
0.5
0.05
0.1
0.15
0.2
0.25
0.3
Frequency (Hz)
0.35
0.4
0.45
0.5
Phase (degrees)
0
-50
-100
-150
-200
20
H (e j ) 0.2
0 0.2
0.8 H (e j ) 1
0.6
To design high pass filter, it is desired first to design LPF and transform into HPF. This is
illustrated as below.
The given specifications for high pass filter are:
1
1
0.2; 1
1
1
(i) Impulseinvariant
transformation:
Apply impulse-invariant transformation, T
Since sampling period T is not given, assume T=1 second.
Therefore, p pT 0.2 rad / sec, and s sT 0.6 rad / sec
100.1 s 1
log
log
100.1 p 1
1.706
The order of the low pass filter is, N
log s
log s
p
p
The transfer function for second order normalized low pass filter is given by, H ( s)
The cut off frequency is c
10
p
0.1 p
1
2N
1
s 2 2s 1
21
s sc
The denormalized high pass filter transfer function can be obtained by replacing s by
H a ( s)
1
2
0.7504
0.7504
2
1
s
s
s2
s 2 1.602s 0.5631
0.5129
0.5129
H a ( s) 1 2
2
1
1
pk T 1
s pk
1 e Z
Therefore, H ( z ) 1
0.5129
1 e
0.5129
(0.5131 j 0.5131)T 1
z
1 e
( 0.5131 j 0.5131)T 1
On simplification we get
H ( z)
2
tan
T
2
0.2
0.6
p 2 tan
0.6498 rad / sec , and s 2 tan
2.7528 rad / sec
2
2
The order of low pass filter is
4.89
log
log
0.75 1.2986 2
N
2.7529
log s log
00.6498
p
The transfer function for second order normalized low pass filter is given by, H ( s)
10
p
0.1 p
1
2N
s 2 2s 1
0.6498
0.7503 rad / sec
(0.75)1/2
22
c .
H a ( s)
1
2
0.7503
0.7503
2
1
s
s
s2
s 2 1.061s 0.562
To convert analog high pass filter to digital high pass filter using bilinear transformation
2 1 z 1
, we get the digital HPF transfer function as
T 1 z 1
H [ z]
1 z 1
4
1 z 1
1 z 1
1 z 1
4
1.061z 1 24
0.562
1 z 1
1 z 1
-------------------------------------------------------------------------------------------------------------------------------------
12 l u
1(u l )
and B
22 l u
2 (u l )
log
.
Calculate the order of the filter. N
log r
Find the normalized LPF transfer function
The denormalized band pass filter transfer function H a ( s) can be obtained by replacing
s 2 l u
s(u l )
------------------------------------------------------------------------------------------------------------------------------------23
1 0.1 rad / sec; 2 0.7 rad / sec; l 0.25 rad / sec , u 0.35 rad / sec; p 3 dB and
2 0.7 rad / sec; l 0.25 rad / sec u 0.35 rad / sec; T 1 sample / sec
Apply bilinear transformation,
2
tan
T
2
0.1
0.7
1 2 tan
0.31676 rad / sec; 2 2 tan
3.925 rad / sec
2
2
0.25
0.35
l 2 tan
0.8284 rad / sec; u 2 tan
1.225 rad / sec
2
2
12 l u
9.244
2 (u l )
3.925(1.225 0.8284)
r min A , B 7.28 rad / sec
A
5.5338
log log
0.9976 0.863 1
N
log r
log 7.28
The transfer function for first order normalized low pass filter is given by, H ( s ) 1
s 1
s 2 l u
s(u l )
24
s 2 l u s 2 0.8284 1.225
s( u l ) s(1.225 0.8284)
2
s s 1.048
0.3966s
Therefore, H ( s)
1
0.3966s
2
s 0.3966s 1.0148
s 1.0148
1
0.3966s
2
To convert analog low pass filter to digital band pass filter using bilinear transformation, substitute
2 1 Z 1
,
T 1 Z 1
1 Z 1
0.3966 2
1
1 Z
H (s)
2
1 Z 1
1 Z 1
4
0.3966
2
1.0148
1
1
1 Z
1 Z
On simplification, H ( z )
0.316(1 Z 2 )
1 1.027 Z 1 0.726Z 2
(ii)Impulse-invariant transformation:
Apply impulse-invariant transformation, T . Since sampling period T=1 sec
10.8852
1 ( u l )
0.3141(1.009 0.7854)
8.079
2 ( u l )
2.199(1.099 0.7854)
5.5338
log log
0.9976 0.82 1
N
log r
log 8.079
The transfer function for I order normalized low pass filter is given by H ( s)
1
.
s 1
25
s 2 lu
s ( u l )
s 2 l u s 2 0.7854 1.099
s( u l ) s(1.099 0.7854)
2
s s 0.863
0.3137 s
H ( s)
1
0.3137 s
2
s 0.3137 s 0.863
s 0.863
1
0.3137 s
2
H a ( s)
0.1568 j 0.02686
0.1568 j 0.02686
To convert analog band pass filter to digital band pass filter in impulse-invariant transformation.
1
1
pk T 1
s pk
1 e Z
The poles of equation (1) are p1 0.1568 j 0.9156; p2 0.1568 j 0.9156
0.1568 j 0.02686
0.1568 j 0.02686
0.1568
j
0.9156)
1
1 e
z
1 e( 0.1568 j 0.9156) z 1
0.1568 j 0.02686
0.1568 j 0.02686
j
0.9156
1
1 0.8549e
z
1 0.8549e j 0.9156 z 1
0.3136 0.1996 z 1
1 1.0416 z 1 0.7292 z 2
H ( z)
-------------------------------------------------------------------------------------------------------------------------------------
1 (u l )
12 u l
and B
2 (u l )
22 u l
log
. Find the normalized LPF transfer function
Calculate the order of the filter. N
log r
The denormalized band stop filter transfer function H a ( s) can be obtained by replacing
s(u l )
s 2 l u
------------------------------------------------------------------------------------------------------------------------------------26
p 2dB, s 10dB, and T 1sec .The pass band frequencies are 0.07 and 0.8 and stop band
frequencies are 0.2 and 0.3 .
Use (i) Impulsive-invariant transformation, and (ii) Bilinear transformation
Ans: Given specification for band elimination filter are,
T
1 0.2 rad / sec; 2 0.3 rad / sec; l 0.07 rad / sec; u 0.8 rad / sec
A
1 (u l )
0.2 (0.8 0.07 )
9.125
12 u l (0.2 )2 0.8 0.07
2 (u l )
0.3 (0.8 0.07 )
6.441
2 2 u l
(0.3 )2 0.8 0.07
3
log log
0.7647 0.7338 1
N
log r
log 6.441
The transfer function for I order normalized low pass filter is given by , H ( s ) 1
s 1
s 2 l u
s(u l )
s(0.8 0.07 )
s 2 0.8 0.07
Therefore, H a ( s)
s 2 l u
s(u l )
2.29s
s 0.552
2
1
s 2 0.552
s 2 2.29s 0.552
2.29s
2
1
s 0.552
2.65
0.3598
s 2.0162 s 0.2738
27
H ( z) 1
1
1
p
s pk
1 e k T z 1
2.65
1 e2.0162T z 1
0.36
1 e0.2735T z 1
2.65
0.36
1
1 0.1332 z
1 0.7605z 1
H ( z)
2 1.6011z 1 0.1z 2
1 0.8938 z 1 0.1z 2
0.3
0.6495 rad / sec; 2 2 tan
1.018 rad / sec
2
0.07
0.8
l 2 tan
0.2206 rad / sec; u 2 tan
6.142 rad / sec
2
2
(u l )
0.6495(6.142 0.2206)
A 1 2
4.1216
1 u l (0.6495)2 6.142 0.2206
2
0.2
tan 1 2 tan
T
2
2
2 (u l )
22
u l
1.018(6.142 0.2206)
18.926
(1.018)2 6.142 0.2206
r min A , B 4.1216
The order of the low pass filter is given by,
3
log log
0.7647 0.965 1
N
log r log 4.1214
The transfer function for I order normalized low pass filter is given by , H ( s ) 1
s 1
s ( l )
Denormalized band pass filter can be obtained by replacing s 2 u
s lu
s(u l )
s(6.142 0.2206)
5.9214s
s
2
2
s 6.142 0.2206
s 1.3549
s l u
1
s 2 1.3549
Therefore, H ( s)
s 2 5.921s 1.3549
5.9214s
2
1
s 1.3549
2
To convert analog low pass filter to digital band pass filter using bilinear transformation, substitute
2 1 z 1
,
T 1 z 1
28
H [ z]
On simplification,
1 Z 1
2 1 Z 1 1.3549
1 Z 1
1 Z 1
2
5.9214
1.3549
1 Z 1
1 Z 1
H ( z)
6. Chebyshev Filters:
Type1: All pole filters. Equi-ripple in the passband
monotonic in the stopband.
Type2: Both poles and zeros. Monotonic in the
passband and equiripple in the stopband.
The magnitude response of N-th order filter is expressed as
| H ( j ) | 2
1
2
1 2C N
p
where is ripple parameter and C N ( x) is the N-th order Chebyshev polynomial defined as
C N ( x) C N ( x) , N -odd;
C N ( x) C N ( x) , N -even;
C N (0) 0 , N -odd;
N
0.1
p 1
10
100.1 s 1
cosh 1 s
p
cosh 1
log
22
12
1
1
log s
p
cosh 1
cosh 1
s
p
N1 N1
a p
2
where
N1 N1
b p
2
1 2 1; 100.1 p 1 .
Step3: Calculate the poles of Chebyshev filter which lies on ellipse by using the formula:
(2k 1)
,
2N
k 1, 2,. . . , N
(a) For N-odd, substitute s = 0 in the denominator polynomial and find the value. This value is
equal to the numerator of the transfer function.
(b) For N-even, substitute s = 0 in the denominator polynomial and divide the value by
1 2 .
30
H a ( s)
0.262
s 2 0.512s 0.3277
(b) The Bi-linear transformation analog prototype low pass filter transfer function
H a ( s)
0.2809
s 2 0.53s 0.351
Ans: (a) The single pole filter transfer functions are derived from the given Impulse invariance analog
prototype filter transfer function as below:
H a ( s)
0.256 j
0.256 j
The impulse Invariance transformation for converting an analog low pass filter to a digital low pass filter
is derived by substituting
H ( z)
1
1
p
s pk
1 e k T z 1
0.256 j
1 e(0.256 j 0.512)T z 1
On simplification, we get H ( z )
0.256 j
1 e(0.256 j 0.512)T z 1
0.194 z 1
1 1.3495 z 1 0.6 z 2
(b) The Bi-linear transformation for analog prototype low pass filter transfer function is computed by
substituting s
H ( z)
2 1 z 1
0.2809
in H a ( s)
, we get
1
2
T 1 z
s 0.53s 0.3561
0.2809
2
2 1 z 1
2 1 z 1
0.53
0.3561
1
T 1 z 1
T 1 z
1 2 z 1 z 2
0.052
1 1.349 z 1 0.609 z 2
Ex9: Design a digital Chebyshev high pass filter with the following specifications:
The analog prototype low pass filter transfer function
H a ( s)
0.06175
and the cut-off frequency is c 0.01 .
s 0.06175
Ans: Given that analog prototype low pass filter transfer function H a ( s)
The high pass filter transfer function is computed by substituting s
c
.
s
0.06175
s 0.06175
31
0.06175
s
.
0.01
s
0.5085
0.06175
s
1
1
For Impulse Invariance transformation,
. Here pole pk 1 .
p
s pk
1 e k T z 1
H HPF ( s)
2 1 z 1
T 1 z 1
0.7973(1 z 1 )
H [ z]
1 0.596 z 1
2 1 z 1
0.5085
T 1 z 1
32
Exercise-14
1. Define the ideal digital filters both in mathematical expressions and graphical representations.
Compare the practical filters and with ideal filters.
2. Explain the magnitude response of a digital filter. Illustrate all the filter parameters.
3. Illustrate the steps involved in designing IIR digital high pass filters using Impulse Invariance
transformation.
4. Illustrate the steps involved in designing IIR digital low pass filters using Bi-linear transformation.
5. Design a digital low pass filter Butterworth filter with pass band and stop band cut-off frequencies
200 Hz and 500 Hz respectively. The pass band attenuation is -5 dB and stop band attenuation is 12 dB. The sampling frequency is 5 KHz.
Use (i) Impulse Invariance Method, and (ii) Bilinear transformation.
Hint: First convert analog frequencies into digital frequencies using formula 2
Fp
Fsam
F
Fsam
200
0.08 rad/sec and
5000
Fs
500
2
0.2 rad/sec.
Fsam
5000
0.7 | H [e j ] | 1,
0 0.2
| H [e j ] | 0.3,
0.2 0.6
| H [e j ] | 0.2,
0.8 | H [e j ] | 1,
0 0.2
0.6
33
1 0.2 rad / sec; 2 0.7 rad / sec; l 0.35 rad / sec , u 0.35 rad / sec; p 3 dB
and s 10 dB . Design a digital Butterworth band pass filter Using
0.7 , and the stop band frequencies range is 0.1 0.25 . Use (i) Impulse Invariance
Method, and (ii) Bilinear transformation.
13. Design a digital Chebyshev low pass filter with the following specifications:
0.457
0.31416
and
s 0.31416
34