Sei sulla pagina 1di 29

1

Filter:
A Filter is a network or system which extracts information from
the signal by removing noise.
Broad classification of Filters: 1.Analog Filters 2.Digital Filters

UNIT III IIR FILTER DESIGN


3.1 Structures of IIR
3.2 Analog filter design
3.3 Discrete time IIR filter from analog
filter
3.4 IIR filter design
(i)by Impulse Invariance Digital Filters:
Digital filters play very important role in Digital Signal Processing.
(ii)Bilinear transformation
They are used in data transmission, data compression, echo
(iii)Approximation of derivatives cancellation, image processing, speech processing, digital audio,
3.5Filter design using frequency Biomedical signal processing etc.,
Advantages of Digital filters:
translation 1.The performance of digital filters do not vary with environmental
changes.
2.Filtered & unfiltered data can be saved for future use.
3.They are cheap & consume low power.
4.Precision is high compared to analog filters.
5.The frequency response can be automatically adjusted using
programmable processor.
2

Disadvantages:
1.The time taken for the design &hardware development is longer .

2.They are subject to ADC noise due to quantization & round off noise.

Classification of Digital Filters:


1.IIR Filter 2.FIR Filter
3

Design of IIR Filters:

Impulse Invariant Technique: (many to one mapping)

Two different poles map on same location of Z plane called many


to one mapping.(Limitation of this technique)
4
Imaginary
real

(i)

(ii) If Then the analog pole s lies on left half of the S-plane.In this case, the
corresponding digital pole Z will lie inside the unit circle.

By taking inverse laplace transform,


5

problem

Practice:

Ans:
6

Practice:

Ans

LIMITATIONS OF IIT METHOD:


In IIT method, many poles on S-plane are marked on same location
in Z-plane called many to one mapping.

Applying partial fraction method


(One to one mapping)
7
8
Comparing eqn (a) and (5)

W.K.T

-------------------(7)

-------------------(7)
9

problem

Problem:
10

PRACTICE:

ANS:

PRACTICE:

PROBLEM
ANS:

PROBLEM:
11

S^2N = 1= e^j2π k (or) Sk = e^j2πk/2N for k= 1,2…2N

For N= even, ex for N=2, s^4= -1, Sk = e^j(2k-1)π/2N

DESIGN OF ANALOG BUTTERWORTH FILTER:


12
=

= 1= (or) Sk = (or) for k= 1,2…2N

Thus, = = cos π/3 + j sin π/3= 0.5+j0.866

S2= = cos 2π/3 + j sin2 π/3=-0.5+j0.866


S3= -1, S4= -0.5-j0.866, S5= 0.5-j0.866,S6= 1

Similarly,
13

The pass band cutoff frequency is denoted by with maximum attenuation factor dB

The stop band cutoff frequency is with maximum attenuation


14

the equn (14) can be simplified as

here

STEPS FOR DESIGN OF BUTTERWORTH FILTER


15

Find N

Find transfer function of normalized LPF H(s)


(Refer table)

Find denormalised transfer function

In H(s)
Find Ω
16

The transfer function for II order filter

To find roots , denominator term equal to =0

S2+ 1465.78 S + 1072.25Х103 = 0


17

To get H(s) = Form Apply partial fraction method

There fore

We know
18

problem
19
20

problem
21

For N=2
22

IIR SYSTEM REALIZATION


23
24
25
26
27
28
29

Potrebbero piacerti anche