Nguyễn Công Phương

SIGNAL PROCESSING

Design of IIR Filters

 Contents
I. Introduction
II. Discrete – Time Signals and Systems
III. The z – Transform
IV. Fourier Representation of Signals
V. Transform Analysis of LTI Systems
VI. Sampling of Continuous – Time Signals
VII.The Discrete Fourier Transform
VIII.Structures for Discrete – Time Systems
IX. Design of FIR Filters
X. Design of IIR Filters
XI. Random Signal Processing
s i tes.google.com/site/ncpdhbkhn 2
 Design of IIR Filter
1. Introduction to IIR Filter Design
2. Design of Continuous-Time Lowpass Filters
3. Transformation of Continuous-Time Filters to
Discrete-Time IIR Filters
4. Design Examples for Lowpass IIR Filters
5. Frequency Transformations of Lowpass
Filters
6. Design Examples of IIR Filters using Matlab

s i tes.google.com/site/ncpdhbkhn 3
 Introduction to IIR Filter Design
(1)

∑ −k
∏
M M
∞ b z (z − zk )
H ( z ) = ∑ h[n ] z −n = k =0 k =1
k
= b0
1+∑ a z− k ∏
N N
n=0 k =0 k k =1
( z − pk )

∑ ∏
M M
∞ β s k
(s − ζ k )
H c ( s ) = ∫ hc (t )e − st dt = k =0 k =1
k
β
= 0
1 + ∑ k =0 α k s k ∏ k =1 ( s − sk )
0 N N

Given: IIR filter CT filter CT filter CT to DT IIR filter

specifications specifications design transformation

s i tes.google.com/site/ncpdhbkhn 4
 Introduction to IIR Filter Design
(2)
Given: IIR filter CT filter CT filter CT to DT IIR filter
specifications specifications design transformation

Apply CT CT IIR filter

Design CT CT to DT
frequency band
lowpass filter transformation
transformation

Apply DT DT IIR filter

Design CT CT to DT
frequency band
lowpass filter transformation
transformation

s i tes.google.com/site/ncpdhbkhn 5
 Introduction to IIR Filter Design
(3)
FIR vs. IIR:
• Advantages:
– Exactly linear phase, always stable,
– Have design methods that are generally linear in filter
parameters,
– Can have great flexibility in choosing their frequency response,
– Can be realized efficiently in hardware,
– Have finite-duration transients.
• Disadvantages:
– Often require a much higher filter order than IIR filters to
achieve a given level of performance,
– The delay in the output response is often much greater than for
an equal performance IIR filter,
– The design methods often are iterative in nature requiring
computer-aided techniques.

s i tes.google.com/site/ncpdhbkhn 6
 Design of IIR Filter
1. Introduction to IIR Filter Design
2. Design of Continuous-Time Lowpass
Filters
3. Transformation of Continuous-Time Filters to
Discrete-Time IIR Filters
4. Design Examples for Lowpass IIR Filters
5. Frequency Transformations of Lowpass
Filters
6. Design Examples of IIR Filters using Matlab
s i tes.google.com/site/ncpdhbkhn 7
 Design of Continuous-Time
Lowpass Filters (1)
2
H c ( jΩ) = H c ( s) H c (− s) s = jΩ

( s − sk )(− s − sk ) = sk2 − s 2

s = jΩ

→ ( s − sk )( − s − sk ) = ( sk2 + Ω 2 )

2 ( Ω2 + ζ 12 )(Ω 2 + ζ 22 )...(Ω2 + ζ M2 )
→ H c ( jΩ) = G 2
( Ω2 + s12 )( Ω2 + s22 )...( Ω2 + s2N )

s i tes.google.com/site/ncpdhbkhn 8
 Design of Continuous-Time
Lowpass Filters (2)

1, 0 ≤ Ω ≤ Ωc
2
2 H c ( jΩ) = 
0, Ω > Ωc
H ( jΩ)
dB
0 1
Ap
1 2 1
H c ( jΩ) =
1+ε 2 1 + V 2 (Ω)
As
V 2 (Ω) ≪ 1, 0 ≤ Ω ≤ Ωc
 2
V (Ω) ≫ 1, Ω > Ωc
1
A2 Ω
−∞
0 Ω p Ωc Ω s π

s i tes.google.com/site/ncpdhbkhn 9
 Design of IIR Filter
1. Introduction to IIR Filter Design
2. Design of Continuous-Time Lowpass Filters
a) The Butterworth Approximation
b) The Chebyshev Approximation
c) The Inverse Chebyshev or Chebyshev II
Approximation
d) The Elliptic or Cauer Approximation
3. Transformation of Continuous-Time Filters to
Discrete-Time IIR Filters
4. Design Examples for Lowpass IIR Filters
5. Frequency Transformations of Lowpass Filters
6. Design Examples of IIR Filters using Matlab
s i tes.google.com/site/ncpdhbkhn 10
 The Butterworth Approximation
(1)
1
N =1
N =2
0.8
N =3
N =5

2 1, 0 ≤ Ω ≤ Ω c 0.6 N = 10

Hc ( jΩ) =  N = 100

0, Ω > Ω c
0.4

0.2

0
0 0.5 1 1.5 2 2.5 3
2 1
H B ( jΩ) = , N = 1, 2,... / c

1 + (Ω / Ω c ) 2N

1
2
H B ( j 0) = 1 N =1
N =2
0.8
N =3
2 1 N =5
H B ( jΩ c ) = 0.6 N = 10
N = 100
2 0.4
2
H B ( j∞) = 0 0.2

0
0 0.5 1 1.5 2 2.5 3
/ c

s i tes.google.com/site/ncpdhbkhn 11
 The Butterworth Approximation
(2)
1
N =1
N =2
0.8
N =3
N =5
0.6
2 1 N = 10
H B ( jΩ) = N = 100

1 + (Ω / Ω c )2 N 0.4

0.2

0
0 0.5 1 1.5 2 2.5 3
/ c

2 1
Ec2 ( jΩ ) = 1 − H B ( jΩ ) = (Ω / Ω c )2 N − (Ω / Ωc ) 4N + ..., 0 ≤ Ω ≤ Ω c NN == 12
0.8
N =3
N =5
0.6 N = 10

2 1 N = 100

H B ( j 0) ≃ , Ω ≫Ω
0.4
c
(Ω / Ωc ) 2N
0.2

0
0 0.5 1 1.5 2 2.5 3
/ c

s i tes.google.com/site/ncpdhbkhn 12
 The Butterworth Approximation
π
(3) 2N π
N
2 1
H B ( jΩ) = s1
1 + (Ω / Ω c )2 N s2
1 + ( s / jΩ c )2N = 0 → (s / jΩ c )2 N = −1 = e j (2k −1)π s3
k = 1, 2, ..., 2 N s4
s5
→ sk = σ k + jΩ k s6

σ k = Ωc cosθk
Ωk = Ωc sin θk
s1
π
2k − 1 s2
θk = + π , k = 1, 2,..., 2 N
2 2N s3
s4
ΩN s5
H B ( s) = c
(s − s1)( s − s2 )...(s − sN )

s i tes.google.com/site/ncpdhbkhn 13
 The Butterworth Approximation
(4)
 1 1
≥
1 + (Ω / Ω ) 2N 1 + ε 2 2 1
 p c H B ( jΩ) =
 1 + ( Ω / Ωc ) 2 N
 1 1
≤
1 + (Ω / Ω ) 2 N A2
 s c 2
dB H ( jΩ )
(Ω p / Ωc )2 N ≤ ε 2 0 1
→ Ap
(Ω s / Ω c )2 N ≥ A2 − 1 1
1+ε 2
A2 − 1
→ ΩsN ≥ ΩcN 2
A − 1 ≥ Ω Np As
ε
ln β
→ N≥
ln α 1
Ω A2 − 1 10 As /10 − 1 A2 Ω
α= s, β= = −∞
π
Ωp ε 10
Ap /10
−1
0 Ω p Ωc Ωs

− 1) −1/(2 N ) ≤ Ωc ≤ Ω s (10 As /10 − 1) −1/(2 N )

Ap /10
Ω p (10
s i tes.google.com/site/ncpdhbkhn 14
 The Butterworth Approximation
Ex. 1 (5)
−6 dB ≤ 20 log10 H ( jΩ) ≤ 0, 0 ≤ Ω ≤ 2 rad/s
20 log10 H ( jΩ) ≤ −20dB, 3 rad/s ≤ Ω < ∞

− 1 = 100.1×6 −1 = 1.7266
0.1 Ap
ε = 10
A = 100.05 As = 100.05×20 = 10
Ω 3
α = s = = 1.5
Ωp 2

A2 − 1 102 − 1
β= = = 5.7628
ε 1.7266
 ln β   ln(5.7628) 
N=  =  =  4.32  = 5
α
 ln   ln(1.5) 
− 1) −1/(2 N ) ≤ Ωc ≤ Ω s (10 As /10 − 1) −1/(2 N )
Ap /10
Ω p (10

2(106/10 −1)−1/( 2×5) = 1.7931 ≤ Ωc ≤ 3(1020/10 − 1)−1/(2×5) = 1.8948

Ω c = 1.8948 rad/s
s i tes.google.com/site/ncpdhbkhn 15
 The Butterworth Approximation
Ex. 1 (6)
−6 dB ≤ 20 log10 H ( jΩ) ≤ 0, 0 ≤ Ω ≤ 2 rad/s
20 log10 H ( jΩ) ≤ −20dB, 3 rad/s ≤ Ω < ∞

Ω c = 1.8948 rad/s; N = 5
π 2k − 1
θk = + π , k = 1, 2,...,2 N
2 2N
π 2k − 1
= + π = 0, 4π + 0, 2kπ , k = 1,2,3, 4,5
2 2.5
sk = Ωc (cos θ k + j sin θ k )
= 1.8948[cos(0.4π + 0.2 kπ ) + j sin(0.4π + 0.2kπ )], k = 1,2,3, 4,5

ΩcN
H B (s) =
( s − s1 )( s − s2 )...( s − s N )

24.42
=
s 5 + 6.13s 4 + 18.80s 3 + 35.61s 2 + 41.71s + 24.42
s i tes.google.com/site/ncpdhbkhn 16
 The Butterworth Approximation
Ex. 1 (7)
1

0.8

0.6

0.4

0.2

0
1 2 3 4 5 6 7 8 9 10
Frequency (rad/s)

200

100

-100

-200
1 2 3 4 5 6 7 8 9 10
Frequency (rad/s)

s i tes.google.com/site/ncpdhbkhn 17
 The Butterworth Approximation
Ex. 2 (8)
Fp = 40 Hz; Ap = 1 dB; Fs = 50 Hz; As = 30 dB.

s i tes.google.com/site/ncpdhbkhn 18
 Design of IIR Filter
1. Introduction to IIR Filter Design
2. Design of Continuous-Time Lowpass Filters
a) The Butterworth Approximation
b) The Chebyshev Approximation
c) The Inverse Chebyshev or Chebyshev II
Approximation
d) The Elliptic or Cauer Approximation
3. Transformation of Continuous-Time Filters to
Discrete-Time IIR Filters
4. Design Examples for Lowpass IIR Filters
5. Frequency Transformations of Lowpass Filters
6. Design Examples of IIR Filters using Matlab
s i tes.google.com/site/ncpdhbkhn 19
 The Chebyshev Approximation
(1)
1
, TN ( x ) = cos( N cos −1 x), x ≤ 1
2
H C ( j Ω) =
1 + ε 2TN2 ( Ω / Ωc )

1
ε ≪ 1 → EC2 (Ω / Ωc ) = 1 − ≈ ε 2TN2 (Ω / Ω c ), Ω ≤ Ωc
1 + ε 2TN2 (Ω / Ωc )
s i tes.google.com/site/ncpdhbkhn 20
 The Chebyshev Approximation
(2)
1
, TN ( x ) = cos( N cos −1 x), x ≤ 1
2
H C ( j Ω) =
1 + ε 2TN2 ( Ω / Ωc )

EC2 ( Ω / Ωc ) ≈ ε 2TN2 (Ω / Ωc ), Ω ≤ Ωc

TN ( x) = cosh( N cosh −1 x), x > 1

2 1 TN ( x) = cos( N cos−1 x ), Ω ≤ Ω c

H C ( j Ω) = , 
1 + ε TN (Ω / Ωc )
2 2 −1
TN ( x) = cosh( N cosh x), Ω > Ωc
s i tes.google.com/site/ncpdhbkhn 21
 The Chebyshev Approximation
(3)
2 1 TN ( x) = cos( N cos −1 x ), Ω ≤ Ωc
HC ( jΩ) = , 
1 + ε TN ( Ω / Ωc ) TN ( x) = cosh( N cosh −1 x), Ω > Ωc
2 2

Ω c = 0.5; ε = 0.75

2 1, N even
HC ( j 0) = 
1 / (1 + ε ), N odd
2

2 1
H C ( j Ωc ) =

|Hc (j )|
1+ε 2
−1
H C ( jΩ) ≃ ε 2 22(N −1) (Ω / Ω c )2 N 
2
 
( Ω ≫ Ωc )

s i tes.google.com/site/ncpdhbkhn 22
The Chebyshev Approximation
(4)
2 1 TN ( x) = cos( N cos −1 x ), Ω ≤ Ωc
HC ( jΩ) = , 
1 + ε TN ( Ω / Ωc ) TN ( x) = cosh( N cosh −1 x), Ω > Ωc
2 2

sk = σ k + jΩk

σ k = Ωc a cos θk , Ω k = Ω cb sin θ k
π 2k − 1
θk = + π , k = 1,2,..., 2 N
2 2N
−1 −1 1/ N
γ −γ γ +γ 1 1 
a= , b= γ
, =  + 1 + 2 
2 2 ε ε 

G N 1/ 1 + ε 2 , N ch½n
Hc (s) = , G = ∏ (− sk ) × 
∏k =1 (s − sk )
N
k =1 1, N lÎ

s i tes.google.com/site/ncpdhbkhn 23
 The Chebyshev Approximation
(5)
2 TN ( x) = cos( N cos −1 x ), Ω ≤ Ωc
1
HC ( jΩ) = , 
1 + ε TN ( Ω / Ωc ) TN ( x) = cosh( N cosh −1 x), Ω > Ωc
2 2

Ωc = Ω p
2
1 1 H ( jΩ )
≤ dB
1 + ε 2TN2 ( Ω s / Ω p) A 2 0 1
Ap
1
A −1
2
→ TN (Ωs / Ω p ) ≥ 1+ε 2
ε
As
A −1
2
→ cosh  N cosh −1 (Ωs / Ω p ) ≥
  ε
cosh −1 (β ) ln(β + β 2 −1) 1
→N≥ = ,
cosh −1 (α ) ln(α + α 2 − 1) −∞
A2 Ω
0
Ω p Ωc Ωs
π
Ωs A −1
2
10 As /10
−1
α= , β= =
Ωp ε 10
A p /10
−1
s i tes.google.com/site/ncpdhbkhn 24
 The Chebyshev Approximation
Ex. 1 (6)
−6 dB ≤ 20 log10 H ( jΩ) ≤ 0, 0 ≤ Ω ≤ 2 rad/s
20 log10 H ( jΩ) ≤ −20dB, 3 rad/s ≤ Ω < ∞

− 1 = 100.1×6 −1 = 1.7266
0.1 Ap
ε = 10

A = 100.05 As = 100.05×20 = 10

Ωs 3
α= = = 1.5
Ωp 2

A2 − 1 102 − 1
β= = = 5.7628
ε 1.7266

 ln( β + β 2 − 1)   ln(5.7628 + 5.76282 − 1) 

N= =  = 2.53  = 3
 ln(α + α − 1)  
2
ln(1.5 + 1.5 − 1)
2 
  

s i tes.google.com/site/ncpdhbkhn 25
 The Chebyshev Approximation
Ex. 1 (7)
−6 dB ≤ 20 log10 H ( jΩ) ≤ 0, 0 ≤ Ω ≤ 2 rad/s
20 log10 H ( jΩ) ≤ −20dB, 3 rad/s ≤ Ω < ∞
ε = 1.7266; N = 3
sk = Ωc a cos θk + j Ωcb sin θ k = Ω p a cos θ k + j Ω pb sin θ k = 2(a cosθ k + jb sin θ k )

π 2k − 1 π 2k − 1
θk = + π= + π , k = 1, 2,..., 2.3
2 2N 2 2.3
1/ N 1/3
1 1   1 1 
γ =  + 1 + 2  =  + 1+ 2 
= 1.2016
ε ε   1.7266 1.7266 

γ − γ −1 1.2016 − 1/ 1.2016
a= = = 0.1847
2 2

γ + γ −1 1.2016 + 1/ 1.2016
b= = = 1.0169
2 2

s i tes.google.com/site/ncpdhbkhn 26
 The Chebyshev Approximation
Ex. 1 (8)
−6 dB ≤ 20 log10 H ( jΩ) ≤ 0, 0 ≤ Ω ≤ 2 rad/s
20 log10 H ( jΩ) ≤ −20dB, 3 rad/s ≤ Ω < ∞
π 2k −1
ε = 1.7266; N = 3; a = 0.1847; b = 1.0169; θk = + π
2 6

sk = 2(a cos θk + jb sin θk ) = 2(0,1847cos θk + j1, 0169sin θk )

π π  π π 
s1 = 2 × 0.1847cos  +  + j 2 × 1.0169sin  +  = −0.1847 − j1.7613
2 6 2 6

 π 3π   π 3π 
s2 = 2 × 0.1847 cos  +  + j 2 × 1.0169sin  +  = −0.3693
2 6  2 6 

 π 5π   π 5π 
s3 = 2 × 0.1847 cos  +  + j 2 × 1.0169 sin  2 + 6  = −0.1847 + j1.7613
 2 6   

s i tes.google.com/site/ncpdhbkhn 27
 The Chebyshev Approximation
Ex. 1 (9)
−6 dB ≤ 20 log10 H ( jΩ) ≤ 0, 0 ≤ Ω ≤ 2 rad/s
20 log10 H ( jΩ) ≤ −20dB, 3 rad/s ≤ Ω < ∞
ε = 1.7266; N = 3

s1 = −0.1847 − j1.7613; s2 = −0.3693; s3 = −0.1847 + j1.7613

G N 1/ 1 + ε 2 , N even
Hc (s) = , G = ∏ (− sk ) × 
∏k =1 (s − sk )
N
k =1 1, N odd

G = ( − s1)( −s2 )(− s3 ) = (0.1847 + j1.7613)(0.3693)(0.1847 − j1.7613) = 1.1584

1.1584
Hc (s) =
( s + 0.1847 + j1.7613)( s + 0.3693)(s + 0.1847 − j1.7613)
1.1584
=
s 3 + 0.7387 s 2 + 3.2728s + 1.1584
s i tes.google.com/site/ncpdhbkhn 28
 The Chebyshev Approximation
Ex. 1 (10)
−6 dB ≤ 20 log10 H ( jΩ) ≤ 0, 0 ≤ Ω ≤ 2 rad/s 1.1584
Hc ( s) =
20 log10 H ( jΩ) ≤ −20dB, 3 rad/s ≤ Ω < ∞ s3 + 0.7387s 2 + 3.2728s + 1.1584

s i tes.google.com/site/ncpdhbkhn 29
 The Chebyshev Approximation
Ex. 2 (11)
Fp = 40 Hz; Ap = 1 dB; Fs = 50 Hz; As = 30 dB.

s i tes.google.com/site/ncpdhbkhn 30
 Design of IIR Filter
1. Introduction to IIR Filter Design
2. Design of Continuous-Time Lowpass Filters
a) The Butterworth Approximation
b) The Chebyshev Approximation
c) The Inverse Chebyshev or Chebyshev II
Approximation
d) The Elliptic or Cauer Approximation
3. Transformation of Continuous-Time Filters to
Discrete-Time IIR Filters
4. Design Examples for Lowpass IIR Filters
5. Frequency Transformations of Lowpass Filters
6. Design Examples of IIR Filters using Matlab
s i tes.google.com/site/ncpdhbkhn 31
 The Inverse Chebyshev/Chebyshev II
Approximation (1)

2
A −1
A2

1
2
A

2 1 1 1
HC ( jΩ ) = 1−
1 + ε 2TN2 ( Ω / Ωc ) 1 + ε 2TN2 (Ω C / Ω) 1 + ε 2TN2 ( ΩC / Ω)

s i tes.google.com/site/ncpdhbkhn 32
 The Inverse Chebyshev/Chebyshev II
Approximation (2)
2 1 ε 2TN2 (ΩC / Ω)
Hc (Ω ) = 1 − =
1 + ε TN (Ω C / Ω ) 1 + ε 2TN2 ( ΩC / Ω)
2 2

Ωc
ε TN ( jΩC / s ) = 0 → ζ k = j
cos uk

2k − 1 π
uk = , k = 1, 2,..., 2 N
N 2 1
2
A
Ω2c
1 + ε 2TN2 (ΩC / Ω) = 0 → sk =
σ k + j Ωk

σ k = Ωc a cos θk , Ωk = Ωc b sin θk
π 2k −1
θk = + π , k = 1, 2,..., 2 N
2 2N
γ − γ −1 γ + γ −1
( )
1/ N 1
a= , b= , γ = A + A −1 2 1−
2 2 1 + ε 2TN2 ( ΩC / Ω)

s i tes.google.com/site/ncpdhbkhn 33
 The Inverse Chebyshev/Chebyshev II
Approximation (3)
ε 2TN2 ( Ω / Ωc )
2
HC ( jΩ) =
1 + ε 2TN2 ( Ω / Ωc )

Ωc = Ωs 2
dB H ( jΩ)
1 A −12 0 1
≤ Ap
1 + ε 2TN2 ( Ωs / Ω p ) A2 A2 − 1
A2
A2 − 1
→ TN (Ω s / Ω p ) ≥ As
ε
A2 − 1
→ cosh  N cosh −1 ( Ω s / Ω p )  ≥
ε
1
cosh −1 ( β ) ln( β + β 2 −1) −∞
A2 Ω
→N≥ = , π
cosh −1 (α ) ln(α + α 2 − 1)
0 Ω p Ωc Ωs

Ω A2 −1 10 As /10 −1
α= s , β= =
Ωp ε 10
A p /10
−1

s i tes.google.com/site/ncpdhbkhn 34
 The Inverse Chebyshev/Chebyshev II
Approximation (4)
ε 2TN2 ( Ω / Ωc )
2
HC ( jΩ) =
1 + ε 2TN2 ( Ω / Ωc )

Ωc = Ωs 2
dB H ( jΩ)
−1
0 1
cosh ( β ) ln( β + β −1)
2
Ap
→N≥ −1
= , A2 − 1
cosh (α ) ln(α + α − 1)
2
A2
Ω A2 −1 10 As /10 −1 As
α= s , β= =
Ωp ε 10
A p /10
−1
 ln( β + β 2 − 1) 
N=  1
 ln(α + α 2 − 1)  A2 Ω
  −∞
π
0 Ω p Ωc Ωs

 cosh −1 ( β ) 
→ Ω′c = Ω p cosh  
 N 

s i tes.google.com/site/ncpdhbkhn 35
 The Inverse Chebyshev/Chebyshev II
Ex. 1 Approximation (5)
−6 dB ≤ 20 log10 H ( jΩ) ≤ 0, 0 ≤ Ω ≤ 2 rad/s
20 log10 H ( jΩ) ≤ −20dB, 3 rad/s ≤ Ω < ∞

ε = 10 − 1 = 100.1×6 − 1 = 1.7266
0.1Ap

A = 100.05 As = 100.05× 20 = 10
Ωs 3
α= = = 1.5
Ωp 2

A2 − 1 10 2 − 1
β= = = 5.7628
ε 1.7266
 ln( β + β 2 − 1)   ln(5.7628 + 5.76282 − 1) 
N= =  = 2.53  = 3
 ln(α + α − 1)  
2
ln(1.5 + 1.5 − 1)
2 
  
 cosh −1 ( β )   cosh −1 (5.7628) 
Ω′c = Ω p cosh   = 2cosh   = 2.6970 = Ωc
 N   3 
s i tes.google.com/site/ncpdhbkhn 36
 The Inverse Chebyshev/Chebyshev II
Ex. 1 Approximation (6)
−6 dB ≤ 20 log10 H ( jΩ) ≤ 0, 0 ≤ Ω ≤ 2 rad/s
20 log10 H ( jΩ) ≤ −20dB, 3 rad/s ≤ Ω < ∞

ε = 1.7266; A = 10; α = 1.5; β = 5.7628; Ω′c = 2.6970

( ) = (10 + )
1/ N 1/3
γ = A + A −1 2
10 − 1
2
= 2.7121

γ − γ −1 2.7121 − 2.7121−1
a= = = 1.1717
2 2

γ + γ −1 2.7121 + 2.7121−1
b= = = 1.5404
2 2

2k − 1 π 2k − 1 π π
uk = = = (2k −1)
N 2 3 2 6
s i tes.google.com/site/ncpdhbkhn 37
 The Inverse Chebyshev/Chebyshev II
Ex. 1 Approximation (7)
−6 dB ≤ 20 log10 H ( jΩ) ≤ 0, 0 ≤ Ω ≤ 2 rad/s
20 log10 H ( jΩ) ≤ −20dB, 3 rad/s ≤ Ω < ∞
Ω′c = 2.6970; γ = 2.7121; a = 1.1717; b = 1.5404; uk = (2 k − 1)π / 6

Ωc Ω′c 2.6970
ζk = j = j =j
cos uk cos uk (2k −1)π / 6

2.6970
ζ1 = j = j3.1141
(2 × 1 − 1)π / 6

2.6970
ζ2 = j = j∞
(2 × 2 −1)π / 6

2.6970
ζ3 = j = − j3.1141
(2 × 3 − 1)π / 6
s i tes.google.com/site/ncpdhbkhn 38
 The Inverse Chebyshev/Chebyshev II
Ex. 1 Approximation (8)
−6 dB ≤ 20 log10 H ( jΩ) ≤ 0, 0 ≤ Ω ≤ 2 rad/s
20 log10 H ( jΩ) ≤ −20dB, 3 rad/s ≤ Ω < ∞
Ω′c = 2.6970; γ = 2.7121; a = 1.1717; b = 1.5404; uk = (2 k − 1)π / 6

Ω2c π 2k −1
sk = , σ k = Ωc a cos θk , Ω k = Ωc b sin θk , θk = + π
σ k + jΩk 2 2N

2.67902
s1 = = −0.7443 − j1.6948
2.6790[1.17 cos(π / 2 + π / 6) + j1.5404 sin(π / 2 + π / 6)]

2.6790 2
s2 = = −2.3017
2.6790[1.17 cos(π / 2 + 3π / 6) + j1,5404sin(π / 2 + 3π / 6)]

2.67902
s3 = = −0.7443 + j1.6948
2.6790[1.17 cos(π / 2 + 5π / 6) + j1.5404 sin(π / 2 + 5π / 6)]
s i tes.google.com/site/ncpdhbkhn 39
 The Inverse Chebyshev/Chebyshev II
Ex. 1 Approximation (9)
−6 dB ≤ 20 log10 H ( jΩ) ≤ 0, 0 ≤ Ω ≤ 2 rad/s
20 log10 H ( jΩ) ≤ −20dB, 3 rad/s ≤ Ω < ∞
ζ1,3 = ± j 3.1141; ζ 2 = j∞; s1,3 = −0.7443 ± j1.6948; s2 = −2.3017

( s − ζ 1)( s − ζ 3 )
H IC ( s) =
( s − s1)( s − s2 )( s − s3 )
( s − j 3.1141)(s + j 3.1141)
=
(s + 0.7443 + j1.6948)( s + 2.3017)(s + 0.7443 − j1.6948)

s 2 + 9.6981
=
s3 + 3.7903s 2 + 6.8524 s + 7.8861

(s 2 + 9.6981) / 7.8861 0.8132 s 2 + 7.8861

H IC (0) = 1 → H IC ( s) = =
s + 3.7903s + 6.8524s + 7.8861
3 2
s3 + 3.7903s 2 + 6.8524 s + 7.8861

s i tes.google.com/site/ncpdhbkhn 40
 The Inverse Chebyshev/Chebyshev II
Ex. 1 Approximation (10)
−6 dB ≤ 20 log10 H ( jΩ) ≤ 0, 0 ≤ Ω ≤ 2 rad/s 0.8132 s2 + 7.8861
H c ( s) = 3
20 log10 H ( jΩ) ≤ −20dB, 3 rad/s ≤ Ω < ∞ s + 3.7903s 2 + 6.8524 s + 7.8861

s i tes.google.com/site/ncpdhbkhn 41
 The Inverse Chebyshev/Chebyshev II
Ex. 2 Approximation (11)
Fp = 40 Hz; Ap = 1 dB; Fs = 50 Hz; As = 30 dB.

s i tes.google.com/site/ncpdhbkhn 42
Butterworth & Chebyshev

s i tes.google.com/site/ncpdhbkhn 43
 Design of IIR Filter
1. Introduction to IIR Filter Design
2. Design of Continuous-Time Lowpass Filters
a) The Butterworth Approximation
b) The Chebyshev Approximation
c) The Inverse Chebyshev or Chebyshev II
Approximation
d) The Elliptic or Cauer Approximation
3. Transformation of Continuous-Time Filters to
Discrete-Time IIR Filters
4. Design Examples for Lowpass IIR Filters
5. Frequency Transformations of Lowpass Filters
6. Design Examples of IIR Filters using Matlab
s i tes.google.com/site/ncpdhbkhn 44
 The Elliptic/Cauer
Approximation (1)

2 1
H E ( j Ω) =
1 + ε 2 RN2 (Ω / Ωc )

 2 Ω12 − Ω 2 Ω32 − Ω2 Ω 22 N −1 − Ω2
v ... , N ch½n
 1 − Ω 1Ω
2 2
1 − Ω32Ω 2 1 − Ω 22 N −1Ω 2
RN (Ω) = 
 2 Ω2 − Ω Ω 24 − Ω2 Ω 22 N − Ω2
2 2
v ... , N lÎ
 1 − Ω 2Ω
2 2
1 − Ω 4Ω 1 − Ω 2 N Ω
2 2 2 2

s i tes.google.com/site/ncpdhbkhn 45
 The Elliptic/Cauer
Ex. 1 Approximation (2)
−6 dB ≤ 20 log10 H ( jΩ) ≤ 0, 0 ≤ Ω ≤ 2 rad/s
20 log10 H ( jΩ) ≤ −20dB, 3 rad/s ≤ Ω < ∞
[N,Wc]=ellipord(2,3,6,20,’s’);
[C,D]=ellip(N,6,20,Wc,’s’);

s i tes.google.com/site/ncpdhbkhn 46
 The Elliptic/Cauer
Ex. 2 Approximation (3)
Fp = 40 Hz; Ap = 1 dB; Fs = 50 Hz; As = 30 dB.

s i tes.google.com/site/ncpdhbkhn 47
 The Elliptic/Cauer
Ex. 3 Approximation (4)
Fp = 40 Hz; Ap = 0.1 dB; Fs = 50 Hz; As = 50 dB.

s i tes.google.com/site/ncpdhbkhn 48
 Design of IIR Filter
1. Introduction to IIR Filter Design
2. Design of Continuous-Time Lowpass Filters
3. Transformation of Continuous-Time
Filters to Discrete-Time IIR Filters
4. Design Examples for Lowpass IIR Filters
5. Frequency Transformations of Lowpass
Filters
6. Design Examples of IIR Filters using Matlab

s i tes.google.com/site/ncpdhbkhn 49
Transformation of Continuous-Time Filters
to Discrete-Time IIR Filters (1)

s = f ( z)
s-plane z-plane
jΩ
H c ( s) → H ( z ) Im{z}

z = e jω

σ ω 1
0 0 Re{ z}

{s = jΩ | −∞ < Ω < ∞} → { z = e jω | −π < ω < π }

{s | Re{s} < 0} → {z | s < 1}

s i tes.google.com/site/ncpdhbkhn 50
 Design of IIR Filter
1. Introduction to IIR Filter Design
2. Design of Continuous-Time Lowpass Filters
3. Transformation of Continuous-Time
Filters to Discrete-Time IIR Filters
a) Impulse-Invariance Transformation
b) Bilinear Transformation
4. Design Examples for Lowpass IIR Filters
5. Frequency Transformations of Lowpass
Filters
6. Design Examples of IIR Filters using Matlab
s i tes.google.com/site/ncpdhbkhn 51
Impulse-Invariance
Transformation (1)

h[n] = Td hc (nTd )

∞  ω
jω 2π 
H (e )= ∑ c T T
H j + j k
k =−∞  d d 

H c ( jΩ ) = 0, Ω ≥ π / Td

jω  ω 
→ H (e ) = H c  j , ω ≤ π
 Td 

s i tes.google.com/site/ncpdhbkhn 52
Impulse-Invariance
Transformation (2)
h[n] = Td hc (nTd )
N
Ak
H c (s ) = ∑
k =1 s − sk
N
→ hc (t ) = ∑ Ak e k u (t )
st

k =1

( )
N
skTd n
→ h[n] = ∑ Td Ak e u[ n]
k =1
∞ ∞ N
→ H ( z) = ∑ h[n]z −n
= ∑ ∑ Td Ak ( e )
skTd n

n =0 n= 0 k =1
N
Td Ak
=∑ skTd −1
k =1 1 − e z

s i tes.google.com/site/ncpdhbkhn 53
 Impulse-Invariance
Transformation (3) s-plane
jΩ
N
Ak
H c (s ) = ∑
k =1 s − sk
N σ
Td Ak
H (z ) = ∑
0
skTd −1
k =1 1 − e z
1 Td Td skTd
→ = −1
, p = e
1 − e k d z −1 1 − pk z
k
s − sk sT
s = f (z)
ln z
→ z = e sTd → s = = f (z) z-plane
Td Im{ z}

jω z = e jω
s = σ + jΩ, z = re
ω 1
 r = eσ Td 0 Re{ z}
→
ω = ΩTd

s i tes.google.com/site/ncpdhbkhn 54
 Impulse-Invariance
Ex. 3 Transformation (4)
ωp = 0.25π rad; Ap = 1 dB; ωs = 0.4π rad; As = 30 dB, Td = 0.1s.
ωp 0.25π ω 0.4π
Ωp = = = 7.8540; Ω s = s = = 12.5664
Td 0.1 Td 0.1
[N,Wc]=buttord(7.854,12.5664,1,30,’s’);

[C,D]=butter(N,Wc,’s’);

[B,A]=impinvar(C,D,1/Td);

s i tes.google.com/site/ncpdhbkhn 55
 Impulse-Invariance
Ex. 3 Transformation (5)
ωp = 0.25π rad; Ap = 1 dB; ωs = 0.4π rad; As = 30 dB, Td = 0.1s.
1 0
Butterworth, N = 9

0.8
-20

0.6
-40
0.4

-60
0.2

0 -80
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9
/ /

0.3
150 Continuous
Discrete
100 0.2

50

0 0.1

-50
0
-100

-150
-0.1
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 0 10 20 30 40 50
/ Time (s)

s i tes.google.com/site/ncpdhbkhn 56
 Impulse-Invariance
Ex. 3 Transformation (6)
ωp = 0.25π rad; Ap = 1 dB; ωs = 0.4π rad; As = 30 dB, Td = 0.1s.
ωp 0.25π ω 0.4π
Ωp = = = 7.8540; Ω s = s = = 12.5664
Td 0.1 Td 0.1
[N,Wc]=cheb1ord(7.854,12.5664,1,30,’s’);

[C,D]=cheby1(N,1,Wc,’s’);

[B,A]=impinvar(C,D,1/Td);

s i tes.google.com/site/ncpdhbkhn 57
 Impulse-Invariance
Ex. 3 Transformation (7)
ωp = 0.25π rad; Ap = 1 dB; ωs = 0.4π rad; As = 30 dB, Td = 0.1s.
1 0
Butterworth, N = 9 Butterworth
Chebyshev I, N = 5 Chebyshev I
0.8
-20

0.6
-40
0.4

-60
0.2

0 -80
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9
/ /

0.3
150 Butterworth Chebyshev I, continuous
Chebyshev I Chebyshev I, discrete
100 0.2 Butterworth, continuous

50

0 0.1

-50
0
-100

-150
-0.1
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 0 10 20 30 40 50
/ Time (s)

s i tes.google.com/site/ncpdhbkhn 58
 Impulse-Invariance
Ex. 3 Transformation (8)
ωp = 0.25π rad; Ap = 1 dB; ωs = 0.4π rad; As = 30 dB, Td = 0.1s.
1.2 0
Butterworth, N = 9
1 Chebyshev I, N = 5
Chebyshev II, N = 5 -20
0.8
Butterworth
0.6 -40 Chebyshev I
Chebyshev II
0.4
-60
0.2

0 -80
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9
/ /

0.3
150 Butterworth Chebyshev II, continuous
Chebyshev I Chebyshev II, discrete
100 Chebyshev II 0.2 Chebyshev I, continuous
Butterworth, continuous
50

0 0.1

-50
0
-100

-150
-0.1
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 0 10 20 30 40 50
/ Time (s)

s i tes.google.com/site/ncpdhbkhn 59
 Design of IIR Filter
1. Introduction to IIR Filter Design
2. Design of Continuous-Time Lowpass Filters
3. Transformation of Continuous-Time
Filters to Discrete-Time IIR Filters
a) Impulse-Invariance Transformation
b) Bilinear Transformation
4. Design Examples for Lowpass IIR Filters
5. Frequency Transformations of Lowpass
Filters
6. Design Examples of IIR Filters using Matlab
s i tes.google.com/site/ncpdhbkhn 60
Bilinear Transformation (1)
s = f ( z)
s-plane z-plane
jΩ
Im{z}

z = e jω

σ ω 1
0 0 Re{ z}

ln z
s= = f (z)
Td
2 1 − z −1
s=
Td 1 + z −1
s i tes.google.com/site/ncpdhbkhn 61
 Bilinear Transformation (2)
2 1 − z −1
s=
Td 1 + z −1
H ( z ) = H c ( s) s = 2 1− z −1
Td 1+ z −1

∏
M
(s − ζ k )
k =1
H c ( s ) = β0
∏k =1 (s − sk )
N

(1 + z −1) N −M ∏ k =1(1 − zk z −1)

M
→ H ( z) = G
∏ pk z −1)
N
k =1
(1 −
N −M M
T   Td 
1 + Td ζ k / 2 1 + Td sk / 2
β0  d 
 2 
∏  k 2 
1 − ζ
k =1
zk = , pk = , G=
1 − Td ζ k / 2 1 − Td sk / 2 N
 Td 
∏  k 2 
1 − s
k =1
s i tes.google.com/site/ncpdhbkhn 62
 Bilinear Transformation (3)
Ex. 4
5 s + 10 5(s + 2)
Hc (s) = =
s2 + 7s + 12 ( s + 3)( s + 4)
1 + Td ζ1 / 2 1 + 2(−2) / 2 1
z1 = = =−
1 − Tdζ 1 / 2 1 − 2(−2) / 2 3
1 + Td ζ 2 / 2 1 + Td ζ 2 / 2
z2 = = lim = −1
1 − Td ζ 2 / 2 ζ 2 →∞ 1 − Td ζ 2 / 2
1 + Td s1 / 2 1 + 2( −3) / 2 1
p1 = = =−
1 − Td s1 / 2 1 − 2( −3) / 2 2
1 + Td s2 / 2 1 + 2( −4) / 2 3
p2 = = =−
1 − Td s2 / 2 1 − 2(−4) / 2 5
N −M M 3−2
T   Td 
β0  d 
 2
∏ 1 − ζ k 2  5  2  1 − ( −2) 22  3
 2 
G= k =1
=     =
N
 Td   2  2 4
∏  k 2 
1 − s 1 − ( − 3)
2  
1 − ( − 4)
2 
k =1

s i tes.google.com/site/ncpdhbkhn 63
 Bilinear Transformation (4)
Ex. 4
5 s + 10 5(s + 2)
Hc (s) = = z1 = −1 / 3; z2 = −1; p1 = −1/ 2; p2 = −3 / 5; G = 3 / 4
s2 + 7s + 12 ( s + 3)( s + 4)

−1 −1
(1 + z − 1) N − M ∏ k =1(1 − zk z −1 ) 3 1 − (−1 / 3) z  1 − (−1) z 
M

H (z) = G =
4 1 − ( −1 / 2) z −1  1 − ( −3 / 5) z −1 
∏ pk z −1 )
N
(1 −
k =1   
0,75 + z −1 + 0, 25 z −2
=
1 + 1,1z − 1 + 0,3z −2
[B,A] = bilinear([5,10],[1,7,12],1/Td)

s i tes.google.com/site/ncpdhbkhn 64
 Bilinear Transformation (5)
s = f ( z)
s-plane z-plane
jΩ
Im{z}

z = e jω

σ ω 1
0 0 Re{z}

2 1 − z −1
s=
Td 1 + z −1  (2 / Td + σ )2 + Ω 2
r = z =
2 / Td + s 2 / Td + σ + j Ω  (2 / Td − σ ) 2 + Ω 2
→z= = = re jω →
2 / Td − s 2 / Td − σ − j Ω  −1  Ω  −1  Ω 
ω = tan   + tan  
  2 / Td + σ   2 / Td − σ 
s i tes.google.com/site/ncpdhbkhn 65
Bilinear Transformation (6)
s-plane
jΩ
 (2 / Td + σ ) 2 + Ω2
r = z =
 (2 / Td − σ ) 2 + Ω2

 −1  Ω  −1  Ω  σ
ω = tan   + tan   0
  2 / Td + σ   2 / Td − σ 

σ < 0 → r < 1

σ = 0 → r = 1
σ > 0 → r > 1 s = f (z)

z-plane
 ΩTd  2 ω 
σ = 0 → ω = 2 tan −1 
Im{ z}
 → Ω = tan  
 2  Td 2 z = e jω

Ω = 0 ⇒ ω = 0 ω 1
 0 Re{ z}
Ω → ∞ ⇒ ω → π

Ω → −∞ ⇒ ω → −π
s i tes.google.com/site/ncpdhbkhn 66
 Bilinear Transformation (7)
 ΩTd  2 ω 
σ = 0 → ω = 2 tan −1   → Ω = tan  
 2  Td 2

π
3

-1

-2

-3
−π
-10 -8 -6 -4 -2 0 2 4 6 8 10

s i tes.google.com/site/ncpdhbkhn 67
 Bilinear Transformation (8)
10 10
Ω Ω
9 9

8 8

7 7

6 6

5 5

4 4

3 3

2 2

1 1

0 0 ω/π
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
H ( jΩ)

H (e jω )
ω /π
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1

s i tes.google.com/site/ncpdhbkhn 68
 Bilinear Transformation (9)
Ex. 3
ωp = 0.25π rad; Ap = 1 dB; ωs = 0.4π rad; As = 30 dB.
2  ωp  2  0.25π 
Ωp = tan   = tan  2  = 0.4142
Td  2  2  

2 ω  2  0.4π 
Ωs = tan  s  = tan   = 0.7265
Td  2  2  2 

[N,Ωc] = buttord(Ωp,Ωs,Ap,As,’s’)

[C,D] = butter(N,Ωc,’s’)

[B,A] = bilinear(C,D,1/Td)

s i tes.google.com/site/ncpdhbkhn 69
 Bilinear Transformation (10)
Ex. 3
ωp = 0.25π rad; Ap = 1 dB; ωs = 0.4π rad; As = 30 dB.

s i tes.google.com/site/ncpdhbkhn 70
 Bilinear Transformation (11)
Ex. 3
ωp = 0.25π rad; Ap = 1 dB; ωs = 0.4π rad; As = 30 dB.

s i tes.google.com/site/ncpdhbkhn 71
 Design of IIR Filter
1. Introduction to IIR Filter Design
2. Design of Continuous-Time Lowpass Filters
3. Transformation of Continuous-Time Filters to
Discrete-Time IIR Filters
4. Design Examples for Lowpass IIR Filters
5. Frequency Transformations of Lowpass
Filters
6. Design Examples of IIR Filters using Matlab

s i tes.google.com/site/ncpdhbkhn 72
 Design Examples for Lowpass
Ex. 5 IIR Filters (1)
ωp = 0.2π rad; Ap = 0.1 dB; ωs = 0.3π rad; As = 50 dB.
[N,Ωc] = buttord(0.2,0.3,0.1,50)
[B,A] = butter(N,Ωc)

s i tes.google.com/site/ncpdhbkhn 73
 Design Examples for Lowpass
Ex. 6 IIR Filters (2)
ωp = 0.3π rad; Ap = 0.5 dB; ωs = 0.4π rad; As = 60 dB.
[N,Ωc] = cheb1ord(0.3,0.4,0.5,60)
[B,A] = cheby1(N,0.5,Ωc)

s i tes.google.com/site/ncpdhbkhn 74
 Design Examples for Lowpass
Ex. 7 IIR Filters (3)
ωp = 0.35π rad; Ap = 0.1 dB; ωs = 0.45π rad; As = 60 dB.
[N,Ωc] = cheb2ord(0.35,0.45,0.15,60)
[B,A] = cheby2(N,60,Ωc)

s i tes.google.com/site/ncpdhbkhn 75
 Design Examples for Lowpass
Ex. 8 IIR Filters (4)
ωp = 0.4π rad; Ap = 1 dB; ωs = 0.55π rad; As = 80 dB.
[N,Ωc] = ellipord(0.4,0.55,1,80)
[B,A] = ellip(N,1,80,Ωc)

s i tes.google.com/site/ncpdhbkhn 76
 Design of IIR Filter
1. Introduction to IIR Filter Design
2. Design of Continuous-Time Lowpass Filters
3. Transformation of Continuous-Time Filters to
Discrete-Time IIR Filters
4. Design Examples for Lowpass IIR Filters
5. Frequency Transformations of Lowpass
Filters
6. Design Examples of IIR Filters using Matlab

s i tes.google.com/site/ncpdhbkhn 77
 Frequency Transformation of
Lowpass Filters
Given: IIR filter CT filter CT filter CT to DT IIR filter
specifications specifications design transformation

Apply CT CT IIR filter

Design CT CT to DT
frequency band
lowpass filter transformation
transformation

Apply DT DT IIR filter

Design CT CT to DT
frequency band
lowpass filter transformation
transformation

s i tes.google.com/site/ncpdhbkhn 78
 Design of IIR Filter
1. Introduction to IIR Filter Design
2. Design of Continuous-Time Lowpass Filters
3. Transformation of Continuous-Time Filters to
Discrete-Time IIR Filters
4. Design Examples for Lowpass IIR Filters
5. Frequency Transformations of Lowpass
Filters
a) Continuous-Time Frequency Transformations
b) Discrete-Time Frequency Transformations
6. Design Examples of IIR Filters using Matlab
s i tes.google.com/site/ncpdhbkhn 79
 Continous-Time Frequency
Transformation (1)
Apply CT CT IIR filter
Design CT CT to DT
frequency band
lowpass filter transformation
transformation

Filter type Transformation, Ωc = 1 rad/s Design parameters

s
Lowpass s→ Ωc: cutoff frequency
Ωc

Ωc
Highpass s→ Ωc: cutoff frequency
s
s 2 + Ω1Ω2 Ω1: lower cutoff frequency
Bandpass s→
s(Ω 2 − Ω1) Ω2: upper cutoff frequency

s(Ω 2 − Ω1) Ω1: lower cutoff frequency

Banstop s→
s 2 + Ω1Ω2 Ω2: upper cutoff frequency

s i tes.google.com/site/ncpdhbkhn 80
 Continous-Time Frequency
Ex. 9 Transformation (2)
A bandpass filter with Fp1 = 20 Hz; Fp2 = 60 Hz; Ap = 1 dB; As = 40 dB.
Omegap=1; Omegas=1.5;
[N,Omegac]=ellipord(Omegap,Omegas,Ap,As,’s’);
[C,D]=ellip(N,Ap,As,Omegac,’s’);
Omega0=sqrt(Omegap1*Omegap2);
BW=Omegap2-Omegap1;
[B,A]=lp2bp(C,D,Omega0,BW);

0 0

-10 -10

-20 -20

-30 -30

-40 -40

-50 -50

-60 -60
0 1 2 3 4 5 0 20 40 60 80 100
(rad/s) s i tes.google.com/site/ncpdhbkhn F (Hz) 81
 Design of IIR Filter
1. Introduction to IIR Filter Design
2. Design of Continuous-Time Lowpass Filters
3. Transformation of Continuous-Time Filters to
Discrete-Time IIR Filters
4. Design Examples for Lowpass IIR Filters
5. Frequency Transformations of Lowpass
Filters
a) Continuous-Time Frequency Transformations
b) Discrete-Time Frequency Transformations
6. Design Examples of IIR Filters using Matlab
s i tes.google.com/site/ncpdhbkhn 82
 Frequency Transformation of
Lowpass Filters
Given: IIR filter CT filter CT filter CT to DT IIR filter
specifications specifications design transformation

Apply CT CT IIR filter

Design CT CT to DT
frequency band
lowpass filter transformation
transformation

Apply DT DT IIR filter

Design CT CT to DT
frequency band
lowpass filter transformation
transformation

s i tes.google.com/site/ncpdhbkhn 83
 Discrete-Time Frequency
Transformation (1)
Apply DT DT IIR filter
Design CT CT to DT
frequency band
lowpass filter transformation
transformation

Filter type Transformation Design parameters

ωc: cutoff frequency

z −1 − α
Lowpass z −1 → sin[(θ c − ωc ) / 2]
1 − α z −1 α=
sin[(θ c + ωc ) / 2]

ωc: cutoff frequency

z −1 + α
Highpass z −1 → cos[(θc + ω c ) / 2]
1 + α z −1 α =−
cos[(θ c − ωc ) / 2]

s i tes.google.com/site/ncpdhbkhn 84
 Discrete-Time Frequency
Transformation (2)
Filter type Transformation Design parameters

ω1: lower cutoff frequency

ω2: upper cutoff frequency
cos[(ω2 + ω1 ) / 2]
−1 z − 2 − α1 z −1 + α 2 α=
Bandpass z → cos[(ω2 − ω1 ) / 2]
α 2 z −2 − θ1 z −1 + 1 K = cot[(ω2 – ω1)/2]tan(θc/2)
α1 = 2αK/(K+1)
α2 = (K – 1)/(K + 1)
ω1: lower cutoff frequency
ω2: upper cutoff frequency
cos[(ω2 + ω1) / 2]
−1 z − 2 − α1 z −1 + α 2 α=
Bandstop z → cos[(ω 2 − ω1) / 2]
α 2 z −2 − θ1 z −1 + 1 K = tan[(ω2 – ω1)/2]tan(θc/2)
α1 = 2αK/(K+1)
α2 = (1 – K)/(K + 1)

s i tes.google.com/site/ncpdhbkhn 85
Discrete-Time Frequency
Transformation (3)

−1 z −1 + α
z →
1 + α z −1

− jθ e− jω + α
e →
1 + α e − jω

 −1  (1 − α 2 ) sin θ 
ω = tan  − 
 2α + (1 + α ) cos θ 
2

→
 −1  (1 − α 2 ) sin ω 
 θ = tan − 
 2α + (1 + α ) cos ω 
2


s i tes.google.com/site/ncpdhbkhn 86
 Discrete-Time Frequency
Ex. 9 Transformation (4)
Highpass filter with ωs = 0.45π rad; Ap = 15 dB; ωp = 0.6π rad; As = 1 dB.

θp =1
cos[(θ c + ωc ) / 2] cos[(1 + 0, 6π ) / 2]
α =− =− = −0.1416
cos[(θc − ωc ) / 2] cos[(1 − 0, 6π ) / 2]

−1
 (1 − α 2 ) sin ω s 
θ s = tan  − 
 2α + (1 + α 2
) cos ω s

−1  (1 − ( −0.1416)2 )sin(0.45π ) 
= tan  −  = 1.4437 rad
 2(−0.1416) + (1 + ( −0.1416) ) cos(0.45π ) 
2

θp=1; θs=1.4437;
[N,ωc]=cheb1ord(θp/π,θs/π,Ap,As);
[Bp,Ap]=cheby1(N,Ap,ωc)

B p ( z − 1)
0.0403 + 0.1208 z − 1 + 0.1208 z −2 + 0.0403z − 3
H LP ( z ) = −1
=
Ap ( z ) 1 − 1.4726 z −1 + 1.1715z −2 − 0.3767 z −3
s i tes.google.com/site/ncpdhbkhn 87
 Discrete-Time Frequency
Ex. 9 Transformation (4)
Highpass filter with ωs = 0.45π rad; Ap = 15 dB; ωp = 0.6π rad; As = 1 dB.

0.0403 + 0.1208 z − 1 + 0.1208 z −2 + 0.0403z − 3

H LP ( z ) = −1 −2 −3
, α = −0.1416
1 − 1.4726 z + 1.1715 z − 0.3767 z
−1 z −1 + α z −1 − 0.1416
z → −1
=
1 +α z 1 − 0.1416 z −1

0.0736 − 0.2208 z −1 + 0.2208 z −2 − 0.0736 z −3

→ H HP ( z ) =
1 + 0.9761z −1 + 0.8568z −2 + 0.2919 z −3

s i tes.google.com/site/ncpdhbkhn 88
 Discrete-Time Frequency
Ex. 9 Transformation (5)
Highpass filter with ωs = 0.45π rad; Ap = 15 dB; ωp = 0.6π rad; As = 1 dB.

s i tes.google.com/site/ncpdhbkhn 89
 Design of IIR Filter
1. Introduction to IIR Filter Design
2. Design of Continuous-Time Lowpass Filters
3. Transformation of Continuous-Time Filters to
Discrete-Time IIR Filters
4. Design Examples for Lowpass IIR Filters
5. Frequency Transformations of Lowpass
Filters
6. Design Examples of IIR Filters using
Matlab
s i tes.google.com/site/ncpdhbkhn 90
 Design Examples of IIR Filters
Ex. 1 using Matlab (1)
Highpass, ωs = 0.5π rad; Ap = 1 dB; ωp = 0.7π rad; As = 40 dB.
[N,Ωc] = buttord(0.7,0.5,1,40)
[B,A] = butter(N,Ωc,’high’)

s i tes.google.com/site/ncpdhbkhn 91
 Design Examples of IIR Filters
Ex. 2 using Matlab (2)
Lower stopband: [0; 0.25π]; Asl = 40 dB
Bandpass, Passband: [0.3π; 0.5π]; Ap = 0.5 dB
Upper stopband: [0.6π; π]; Asu = 50 dB
[N,Ωc] = cheb1ord([0.3,0.5],[0.25,0.6],0.5,50)
[B,A] = cheby1(N,0.5,Ωc)

s i tes.google.com/site/ncpdhbkhn 92
 Design Examples of IIR Filters
Ex. 3 using Matlab (3)
Lower passband: [0; 0.2π]; Apl = 0.5 dB
Bandstop, Stopband: [0.3π; 0.6π]; As = 40 dB
Upper passband: [0.75π; π]; Apu = 0.5 dB
[N,Ωc] = cheb2ord([0.2,0.75],[0.3,0.6],0.5,40)
[B,A] = cheby2(N,40,Ωc,’stop’)

s i tes.google.com/site/ncpdhbkhn 93
 Design Examples of IIR Filters
Ex. 4 using Matlab (4)
Lower stopband: [0; 0.2π]; Asl = 60 dB
Bandpass, Passband: [0.3π; 0.6π]; Ap = 0.1 dB
Upper stopband: [0.7π; π]; Asu = 60 dB
[N,Ωc] = ellipord([0.3,0.6],[0.2,0.7],0.1,60)
[B,A] = ellip(N,0.1,60,Ωc)

s i tes.google.com/site/ncpdhbkhn 94