100% found this document useful (1 vote)
37 views14 pages

16 Karlekar Purva DTSP Assi 3

This document contains the design of an IIR Chebyshev low-pass filter and an FIR high-pass filter. 1) The IIR filter is designed to have a passband ripple of 1 dB from 0 to 0.2T and a stopband ripple of 15 dB above 0.3T using bilinear transformation. The order is found to be N=73. 2) The FIR filter is an 11-tap high-pass filter with a cutoff frequency of T/2 designed using a Hamming window. The impulse response is calculated.

Uploaded by

Purva
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
100% found this document useful (1 vote)
37 views14 pages

16 Karlekar Purva DTSP Assi 3

This document contains the design of an IIR Chebyshev low-pass filter and an FIR high-pass filter. 1) The IIR filter is designed to have a passband ripple of 1 dB from 0 to 0.2T and a stopband ripple of 15 dB above 0.3T using bilinear transformation. The order is found to be N=73. 2) The FIR filter is an 11-tap high-pass filter with a cutoff frequency of T/2 designed using a Hamming window. The impulse response is calculated.

Uploaded by

Purva
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 14

DATE 29 lo 2)

PAGE NO.

Name:Purva Prakash Karlekar


RolLND 16 CLasS:TE Ex TC

ASSIGNMENT NO. 3 D1SPJ

Design IIR chebyshev LPF with passband ripple 1dB in


passband 0 u a0.2T and 15 dB nipple in stophand
oVer 03n w TT uinq bilinear transformation
Che byshev I_LPF
1om giuen specification

5 8

o.2T O.3 Tt

Using BL Prearp the_analog frequen ctes


p 0:2T
p1dR_
Let : 1
1 Prewarping the dnaloq trequencies:
Lan
p 2 tan T=1

p _ 2 tan (o:| T) =0.6S

anws 2 tan
DATE.
PAGE NO.

s 1.02

2 find order0t_chebyshev-I LPF

O.1 SS
N Cos h 10 1
100 pP
1

Cosh s

- D-1C15)
Cosh -
1
o-l C1) 4

COsh :021
o 65 7

N 73 N3
3]fales of cheb- I LPF usIng
P_ -4 0.51
-2

E + 1 =4-1 62.

1/N -1/N
Minoz qxis otE
ellip.se

+73 4-1623
O =_0.654162
2

E -321
1/N -/N
major a S of_
b P 2
-/3 ellipse
b o.65| 4-162 4/3 +4162
2

b -725
Now P T +2 T k=1,2,3
2N

T tST
2

S a cos + jksin ó-
0 321 cos
25 sin(

Sa a cos tbsin_da
o32i cos CTD t o:125 Csin T)
-0:321 +(o

S2=0301

S cos +bcos da
a 0 3 21 co SS
(a 0:25sin

- o 6o5 0.6239
-o16 0S-0-:6279

Deno minator of HCS) is


CS+ 0:16 O5-06219) S t0:321) ls t o-16 o5t
io.629) =(sto 321) (S+0:16os)'+ (a-6239)
= (s+ 0:321) (S+ 0321 St 0:42o )
PAGE NO.

As Nis odd, Numerator value of H(s) is found by putting


SO
SEO in
n denominator polynomial of HCS)
=
(o 321) (o420) =0:13 48

HCS)= 0:13 48 Ha Cs)


(S+o 321) Cs+0 3 21S t 0.420)

HCS) - H (s)

Find t C2) usin a BiL


HCz) = H,Cs)
1
12
CLet T=1 asin_preuarping step

H) = 0 1348
Cs + 0.321) (s*+ 0 32 l S +0.420)

Luith S 24
1+
O 1348

t0:321

0-32 (1-z|to:420
C1tz

O 134 88
2
t 0 - 32 t 0 : 3 2 1 z 2 2z +
( 1tz
1+z

0.G42 0:642 z +0420 +0-420 z7


(1+ z )
PAGE NO.

O.1348
1 (2321-1679z ) ( 2 - 2 z ) 4 0 - 0 6 2 -0.2
z)*
Citz)3 3
0-1348 (1+z)
(2:321-|G79z 4-8 z 4z2 t128-0-425z
O042-
2

HCz)= 0 13 48 C1tz
(2:321-679 z) (5-128 - 8 425 z"4 - 04 z )
- -Ans

2Desiqn FIRHiah.Paas Filterwith cutoff_freuency T/2


using Hamninguindou af lenqth N=11
Giuen Nai11
w = T/2

Gain Czi
N 2
11-1
5

For a HPF, Desized requen.cu


L
responseis

H4L) = -Jw
Ce we w | s T

Jw
H (e )_ -JSw
1 e T/2Slw<T

Desired Împulse_esp 0nse


h (n) a Inu. DIET Halu)
Wn
H CU)_e
2T
PAGE NO.
- T/2 -jSw e du +
TY 2 Jwn
Oe
2T
-T -TY2
T

+
5w
w
T/2

n-5)jw n-5)
e dw 0t| e dw O
2TT T
TT72

Cn-S) Tl2

2
L (n-5)j.
TC
n-s)jJ/a

-jh -5) T 7
2T
Cn-5
LeJto-s)r Co -5) T/2
e

Cn-5 T n -5)TT JCn-5)TYa


e e
1
-Cn-S) T/2
2TC (h-5)j te

JCn-ST -jth-5TT Cn s) TU/2


1
2TT
Cn-s)j -JCh-5)TYa
e

2sin (n-5 1-sin Cn -5) T (n#s)


21
(n-5
Put ne5 in e9
- T2
e dw tot e du
hS) 2T TL2
-

1
2
L-n]+[iq
PACE NO.

haS) =1x
2T
T

h (s) =0-SS

Desired impulse Tesponse


Sin (n-s)TU2
(o)=)SinCn-s)Tt( h s T nE5
-

h
n=5

b (n)is af in finite duration To make i finite we


_multiply ith_bamminq indom functlon
w (n)= 054-0:4 6 cos Tn ld 0SnaN -
oth erw'se

wCn) 0 54 -04 Co $ Tn_/s_ 0Sn S 10


0therwise

Finite impulse respanse,


hch)= h n ) x wn)

tor linear phase P I R £flker is


b(n):h [NH-n) = h (u - -n)
h (n) ha (1o-n)

hLo) =h (o) h C3) h C); hdC8) =h4C2)


h4(t) h C3) bac6) = ba C4)_h G)=h4s)
from(2) C3) &(4) the tflEer coefffcients_are

bdn) Wn) h Cn)


-3
0.0 636 6 0 08 -5-09 x 1o
1 0-1C78
2 o.1061 o.3 978 0-04 22
DATE

PAGE NO.

3 0 6 8 2 1
- 0 3183
0 912 - 0 23003

0 5 1
6 - 0 318B3 O.9121 -0290b3

8 0:3948B O 04 222
0:1628 O
3
0. 063G 0.08 -5 0gx 1o
'Transfer functlon of flter. - b

HC) zhtn) 2 h Cn) -z

-3 -4
=hCo)+ h ) Z +h(2) z tha) z (4)z
S -6
+h (5)z + h (6)z*+hl) z th (8) z
+hC9)z -9 b Cio ) z

0 00.50 9 0 04 22 z _ t 0 t0 - 0:29o3 z_toS z


- !o
6
-0 29 o3 2 + 0 t 0 0 422 z t 0 - 0 o0So2 z

0 00 5og(z 2 to: 0422 (+2


0-2903(z_

the fIR filHer Summttric and N = l iso d a


equen cy resp.onse

H Ce = b )+2 h(h)

CoS d-n)w

e h) +
2h(n)cas (Sen).u
-J5
hs)+ 2 hlo) cos (s)wbu)
tb(s Co SS-3)w
cOs(5-D w+b_(2) cos (S-2)w
+h(4co5 ( 5-4)w

js
e
0:5 t 2_Fo.oo 509 COs 5 w t O-
+0.0 4 22 CO S_3w t O
0 2 9 3 Cos

Phase= - 5 w and
Hmplitude = 0 . 5 t 2 | - 0 0 o509 cos 5 w t
0:0422 Los 3w_-0:29 03 CoS u

0:50:01018 Los 5 w t 0 844 Los3


-
0 5 8 06
CoS w

Amplitude = _0:5-aaLoi8 cos Su t0.0R44_Los3w


0:58 O 6 CosW

3 ConVert the transfer tunction HCs)_of analoe4


butterworth LPF of orcler N=3_jn tto
equivalent digital ffLLter transfer functfon
usinqimpulse ín v ariance method. Consider
T 1 Sec
g.al for N=3.
Hc9)
Cs+1)Cs2+S +i)

HCS) =
CS+1 Cs+ 0 5 +0 86 G) (S A o: 5
J0.8 66)_
Usinqp.artia tuntion exp ansioh,-
PAGE No.

HCS)= A C

CS+1) C5+o:5 + jo.86<) 5+05-jo 86 6)


Sfmplify
A 1 B - o:5 +o.283
HCs)= 1 + 0 5 t0 288 o 5 - J 0 2888
1+S
Csto5 +jo.866) S +0 5 j 0 8666
N
C then HC2) 2 CK
fHCS)= -1
k- S-PR k -e Pr T z

Here P -1, Pa 0:S o:866


o'5 + o8 666
& let 1-1sec
-0 S + j0*8 G6
HCz) -I . eJ 0*866 -
1-e
1 -e
+-0-S-jo86
- e
0 5 o-B66-1

-1 +0:66 z
1
HCz)= 1 0 4 86 z '+ o:3 68 z
- 0 3 68z-

0 5 t a i n parallel realization of the IIRtilter Sustem


-2
given by +2z
1t
+z

H C2) = -2
1 - 0 7 Sz + o - 125 z
2
HC2) 1+2zt
1- 75z+012 5z
H(zJH(2)-
PACENC..

HCz)=

H2)

ytn)

H2) = _H(2) tH,Cz) t H(z)


HC2) = 1 2z
1-0:7S z + o-125

HC) (1+ 2z+ ) z+2z t1


z(1-o+5z+0.{25z2) z*-0 7Szt o.125

H()= z+2 zti


Cz-05) Cz-0-25)

Hcz) z 2 +z
z-o5) Cz-o-25)

Takinq- Partial fra ction expanston


H) A B
z
( z-0: 5) (z-o.25)_
A= (z 0:5) H2)|
z 2=0-5

( z 5 ) Ztz t2
(z 5) (2-o:25) Z o-5
O N
U

N
o N
N O

O
N
N
N
N
N
N
N
25
N

N N
PAGE NO.

05 HCz)
- yCh)

21
Paratie
Structue 0-25
L

5 Explain z ero input limi cycle oscillatfon effect in


diqital tlters
Whena
WihenL stable TIR diaital filter is extited by a
finite input sequen cethat is Constant, the output
of i idea lly decay to_zero
Ho ever thenonlin.earities due_to tinite precisio.n
arith metic opprations ofter cause periodicascillatio0s to
o Ccur jo the Output
Such OS cilloatioDs in recursiv eSystem are called 2ero
limit cu cle o sallations
Consider a first orde IIRflker with dittereh.ce eq

Le E & data reqiste Lenqth_is 3 -bits plusia.


Stgh bit
Tf the input_is x(n)= 0 8 +5 for n=0
Otheruise
and Soun din g applied attet arith mekfc_oper ation, the
table belaca desceibes the limit cucle behavious
Here . C repre sents the raunded-operation
PAGE NO.

yCn-1) ytn-) Cocy cn-)I|ytn)= in)


+0yCn+)|
O
1
1/6 O l00

2 4
l4 1/8
0 0 0
1l4 1/8
/16 O-00 1/8
18 1/1 0.00 1/8
5

rom table we can fin d that for n 23,


the0/p_pemain
Constant å_ qives a s_as steady stuteOlp
Ca using
limit_cucle behaviour. -
When - 1 w e can see that the outpu
0Scillat or between o-125 and - 0:125

nxnn y tn-1 &9-DLdy Cn )yn)= xCn)+0Ky(n1)]


0-85 o 0-000
-/h6 110o -/2
2

3 J4 1/8

0 4/8 1/16 18
5 18 -116 1.001

Noie that beyondn the valu e y (n -D1


is+ i n binary 1. 000100 &0-ao ol 0 0- &
1G
when uuoded ives .00i& O g01
erhibitinq oscillatory behavîour.

You might also like