Impulse invariance

Impulse invariance 
 Analog‐to‐digital Filter 
Transformations
f
• After discussing different approaches to the design of analog
filters, we are now ready to transform them into digital filters. 
• These transformations are derived by preserving different 
aspects of analog and digital filters.
aspects of analog and digital filters.
– Impulse invariance transformation
• Preserve the shape of the impulse response from A to D filter
– Finite difference approximation technique
Finite difference approximation technique
• Convert a differential eq. representation into a corresponding difference 
eq.
– Step invariance
• Preserve the shape of the step response
– Bilinear transformation
• Preserve the system function representation from A to D domain

P. Mahalakshmi 2
 Impulse Invariance Transformation
Impulse Invariance Transformation
• The digital filter impulse response to look similar to
g p p that of a 
frequency‐selective analog filter.
• Sample ha(t) at some sampling interval T to obtain h(n): 
h( ) ha(nT)
h(n)=h ( T)
jT
w   T or e jw

Since z=ejjw on the unit circle and s=j Ω
• Since z=e
e
on the unit circle and s=j Ω on the imaginary axis, 
on the imaginary axis
we have the following transformation from the s‐plane to the 
z‐plane: z=esT

1   2  Frequency‐domain aliasing formula
H ( z)   H a  s  j k
T k    T 
P. Mahalakshmi 3
Complex‐plane mapping in impulse invariance transformation
p p pp g p

P. Mahalakshmi 4
 Properties:
• Sigma = Re(s):
g ()
– Sigma < 0, maps into |z|<1 (inside of the UC)
– Sigma = 0, maps into |z|=1 (on the UC)
– Sigma >0, maps into |Z|>1 (outside of the UC)
• Many s to one z mapping: many‐to‐one mapping
– Every
Every semi‐infinite left strip (so the whole left plane) maps 
semi infinite left strip (so the whole left plane) maps
to inside of unit circle
• Causality and Stability are the same without 
changing;
• Aliasing occur if filter not exactly band‐limited

P. Mahalakshmi 5
 Given the digital lowpass filter specifications wp,ws,Rp and As, we 
want to determine H(z) by first designing an equivalent analog filter 
() y g g q g
and then mapping it into the desired digital filter. Design Procedure:

1. Choose T and determine the analog frequencies:

Ωp=wp/T, Ωs=ws/T
2. Design
g an analogg filter Ha(s)
( ) using
g the specifications
p with one of the three
prototypes of the previous section.
3. Using partial fraction expansion, expand Ha(s) into

4. Now transform analog poles {pk} into R poles {epkT} to obtain the digital filter
N digital
H a ( s )  k 1 k
s  pk

Rk
H ( z )  k 1
N Ex 8.9
pk T 1
1  e
P. Mahalakshmi
z 6