Additional Examples of Chapter 9:

IIR Digital Filter Design

Example E9.1: The peak passband ripple and the minimum stopband attenuation in dB of a
digital filter are p = 0.15 dB and s = 41 dB. Determine the corresponding peak passband and
stopband ripple values p and s .
/10

/10
0.15/ 20
Answer: p = 1 10 p
and s = 10 s
. Hence p = 1 10
= 0.017121127 and
s = 1041/ 20 = 0.0089125 .
______________________________________________________________________________

Example E9.2: Determine the peak passband ripple p and the minimum stopband attenuation
s in dB of a digital filter with peak passband ripple p = 0.035 and peak stopband ripple
s = 0.023.
Answer: p = 20 log10 (1 p ) and s = 20log10 (s ) . Hence,
p = 20 log10 (1 0.035) = 0.3094537 dB and s = 20log10 (0.023) = 32.76544 dB.
______________________________________________________________________________
Example E9.3: Determine the digital transfer function obtained by transforming the causal
analog transfer function
16(s + 2)
Ha (s) =
(s + 3)(s2 + 2s + 5)
using the impulse invariance method. Assume T = 0.2 sec.
Answer: Applying partial-fraction expansion we can express

1 / 8 0.0625 j0.1875 0.0625 + j 0.1875

Ha (s) = 16
+
+

s + 1+ j 2
s + 1 j2
s+3

1
1
s + 7
8
2(s + 1)
2
62
2
2s + 14
8
8
= 16
+ 2
+
=
+
.
=
2
2
2
2
s + 3 (s + 1) + 2 (s + 1)2 + 22
s + 3 s + 2s + 5 s + 3 (s + 1) + 2

Using the results of Problems 9.7, 9.8 and 9.9 we thus arrive at
2 z2 z e 2 T cos(2T)
2
6z e 2 T sin(2T)
G(z) =
+
+
. For
1 e 3Tz 1 z 2 2ze 2T cos(2T) + e 4 T z 2 2ze 2T cos(2T) + e 4 T
T = 0.2, we then get
2 z 2 z e 0.4 cos(0.4)
2
6z e 0.4T sin(0.4)
G(z) =
+
+
1 e 0.6 z 1 z 2 2ze 0.4 cos(0.4) + e 0.8 z 2 2z e 0.4 cos(0.4) + e 0.8
2 z 2 0.6174z
2
1.5662z
=
+ 2
1 + 2
1 0.5488z
z 1.2348z + 0.4493 z 1.2348z + 0.4493

-1-

Additional Examples of Chapter 9:

IIR Digital Filter Design
2
2 1.2348z 1
1.5662 z1
+
+
1
1
2
1
2
1 1.2348z + 0.4493z
1 1.2348z + 0.4493z
1 0.5488z
2
2 + 0.3314 z 1
=
+
1
1
2 .
1 1.2348z + 0.4493z
1 0.5488z
______________________________________________________________________________
=

Example E9.4: The causal digital transfer function

2z
3z
G(z) =
0.9 +
1.2
ze
ze
was designed using the impulse invariance method with T = 2. Determine the parent analog
transfer function.
Answer: Comparing G(z) with Eq. (9.59) we can write
2
3
2
3
G(z) =
. Hence, = 3 and = 4.
0.9 1 +
1.2 1 =
T 1 +

1e
1 e
1e
z
z
z
1 e T z 1
2
3
Therefore, Ha (s) =
+
.
s+3 s+4
______________________________________________________________________________
Example E9.5: The causal IIR digital transfer function
5z 2 + 4z 1
G(z) =
2
8z + 4z
was designed using the bilinear transformation method with T = 2. Determine the parent analog
transfer function.
1 + s
1 + s 2

+ 4
5

1
2 + 3s
Answer: Ha (s) = G(z) z= 1+s = 1 s 2 1 s
= 2
.
1 + s
1 + s
s + 4s + 3

1s
8
+4
1 s
1 s
______________________________________________________________________________
Example E9.6: A lowpass IIR digital transfer function is to be designed by transforming a
lowpass analog filter with a passband edge Fp at 0.5 kHz using the impulse invariance method
with T = 0.5 ms. What is the normalized passband edge angular frequency p of the digital
filter if the effect of aliasing is negligible? What is the normalized passband edge angular
frequency p of the digital filter if it is designed using the bilinear transformation method with
T = 0.5 ms?
Answer: For the impulse invariance design p = pT = 2 0.5 103 0.5 103 = 0.5 .
p T
For the bilinear transformation method design p = 2tan 1

= 2 tan

(FpT )= 2 tan1(0.25) = 0.4238447331.

______________________________________________________________________________

-2-

Additional Examples of Chapter 9:

IIR Digital Filter Design
Example E9.7: A lowpass IIR digital filter has a normalized passband edge at p = 0.3. What
is the passband edge frequency in Hz of the prototype analog lowpass filter if the digital filter
has been designed using the impulse invariance method with T = 0.1 ms? What is the passband
edge frequency in Hz of the prototype analog lowpass filter if the digital filter has been designed
using the bilinear transformation method with T = 0.1 ms?
Answer: For the impulse invariance design 2Fp =

0.3

or Fp = 1.5 kHz. For the

T 10 4
bilinear transformation method design Fp = 10 4 tan(0.15) / = 1.62186 kHz.
______________________________________________________________________________

Example E9.8: The transfer function of a second-order lowpass IIR digital filter with a 3-dB
cutoff frequency at c = 0.42 is
0.223(1+ z 1)2
G LP (z) =
1
2 .
1 0.2952 z + 0.187z
? c = 0.57 by
Design a second-order lowpass filter H LP(z) with a 3-dB cutoff frequency at
transforming G LP (z) using a lowpass-to-lowpass spectral transformation. Using MATLAB plot
the gain responses of the two lowpass filters on the same figure.

? c = 0.57 we have
Answer: For c = 0.42 and
?c

sin c

2 sin(0.075)
=
=
= 0.233474.
?c
sin(0.495)
c +
sin

Thus, H LP( ?z) = GLP (z) z 1 = z?1 =

1 ?z 1

z?1
0.223 1+

1 z?1

?z1
z?1
+
0.187
1 0.2952

1 z?1
1 z?1

0.360454(1 + z?1)2
=
1
2 .
1 + 258136 z? + 0.1833568 z?

-3-

Additional Examples of Chapter 9:

IIR Digital Filter Design

Gain, dB

-10
GLP(z)

-20

H (z)
LP

-30
-40
-50

0.2

0.4

0.6

0.8

______________________________________________________________________________
Example E9.9: Design a second-order highpass filter H HP (z) with a 3-dB cutoff frequency at
? c = 0.61 by transforming G LP (z) of Example E9.8 using the lowpass-to-highpass spectral

transformation. Using MATLAB plot the gain responses of the both filters on the same figure.

? c = 0.61 we have
Answer: For c = 0.42 and
+ ? c
cos c

cos(0.515)
2
=
=
= 0.0492852.
cos(0.95)
c ? c
cos

H HP (?z) = GLP (z) z 1 = z? 1 + =

1+ z?1

z?1 +
z? 1 +
+
0.187
1 + 0.2952

1 + z?1
1+ z? 1

0.19858(1 z?1 )2
1
2 .
1 + 0.4068165 ?z + 0.200963 z?

0
-10
Gain, dB

z? 1 +
0.223 1

1+ z? 1

GLP(z)

HHP(z)

-20
-30
-40
-50

0.2

0.4

-4-

0.6

0.8

Additional Examples of Chapter 9:

IIR Digital Filter Design
______________________________________________________________________________
Example E9.10: The transfer function of a second-order lowpass Type 1 Chebyshev IIR digital
filter with a 0.5-dB cutoff frequency at c = 0.27 is
0.1494(1 + z 1 )2
G LP (z) =
1
2 .
1 0.7076 z + 0.3407z
? o = 0.45 by
Design a fourth-order bandpass filter H BP (z) with a center frequency at
transforming G LP(z) using the lowpass-to-bandpass spectral transformation. Using MATLAB
plot the gain responses of the both filters on the same figure.
Answer: Since the passband edge frequencies are not specified, we use the mapping of Eq.
(9.44) to map = 0 point of the lowpass filter G LP(z) to the specified center frequency
? o = 0.45 of the desired bandpass filter H BP (z) . From Eq. (9.46) we get

? o ) = 0.1564347. Substituting this value of in Eq. (9.44) we get the desired lowpass-to = cos(
1
0.1564347z 1 z 2
1
1 z
bandpass transformation as z z
1 =
1 .
1 z
1 0.1564347 z
Then, H BP (z) = G LP (z)
0.1564347 z 1 z 2
z 1

1 0.423562z

10.1564347 z 1
0.1494(1 z2 )2

+ 0.757725z

0.217287z

4 .

+ 0.3407 z

Gain, dB

-10

GLP(z)

H (z)

0.6

0.8

BP

-20
-30
-40
-50

0.2

0.4

______________________________________________________________________________
Example E9.11: A third-order Type 1 Chebyshev highpass filter with a passband edge at
p = 0.6 has a transfer function

0.0916(1 3z 1 + 3z 2 z 3 )
GHP (z) =
1
2
3.
1+ 0.7601 z + 0.7021 z + 0.2088z
? p = 0.5 by transforming using the lowpassDesign a highpass filter with a passband edge at
to-lowpass spectral transformation. Using MATLAB plot the gain responses of the both filters
on the same figure.

-5-

Additional Examples of Chapter 9:

IIR Digital Filter Design

sin p

? p = 0.5, Thus, =
Answer: p = 0.6, and
p
sin

Therefore, H HP (z) = GHP (z)

=

z 1

?p

2
?p
+

sin(0.05 )
= 0.15838444.
sin(0.55)

z 1 0.15838444

10.15838444 z
1 3

0.15883792(1 z )
1
2
3
1 + 0.126733z + 0.523847z + 0.125712 z

Gain, dB

-10
GHP(z)

HHP(z)

-20
-30
-40
-50

0.2

0.4

0.6

0.8

______________________________________________________________________________
Example E9.12: The transfer function of a second-order notch filter with a notch frequency at
60 Hz and operating at a sampling rate of 400 Hz is
0.954965 1.1226287z 1 + 0.954965z 2
GBS (z) =
1
2
1 1.1226287z + 0.90993z
Design a second-order notch filter H BS(z) with a notch frequency at 100 Hz by transforming
GBS (z) using the lowpass-to-lowpass spectral transformation. Using MATLAB plot the gain
responses of the both filters on the same figure.
60
= 0.3 . The desired
Answer: The above notch filter has notch frequency at o = 2
400
? o = 2 100 = 0.5 . The lowpass-to-lowpass
notch frequency of the transformed filter is
400
?o

sin o

z
2
1
transformation to be used is thus given by z
=
1 where =
?o
o +
1 z
sin

2
0.32492.

The desired transfer function is thus given by H BS(z) = GBS (z) z 1 = z 1

1 z 1

-6-

Additional Examples of Chapter 9:

IIR Digital Filter Design

z 1
z 1
0.954965 1.1226287
+
0.954965
1 z 1
1 z 1
z1
z 1
1 1.1226287
+
0.90993

1 z 1
1 z 1

0.9449 0.1979 10 7 z 1 + 0.9449 z 2

=
.
7 1
2
1 0.1979 10 z + 0.8898z
5

Gain, dB

0
-5
GBS(z)

-10

H (z)
BS

-15
-20
-25

0.2

0.4

0.6

0.8

______________________________________________________________________________

-7-