Вы находитесь на странице: 1из 31

61

Solutions - Chapter 3 The z- Transform

63 3.1. (a)

Z
(b)

2 2 iz-1 [(I)" urn]] = ~ (I)" z-" = ~ (1)" = 1


00 00

2z

1_

Izl > 2

=-

L
-1
0>=-00

(I)" 2

z-ft

= - ~)2z)"
n=l

00

= ---=
(c)

2z 1- 2z

1
11 ,Z-1

Izl < 2

z [(~) ul-n]] = 100


n

(2zt

1~ 2z

Izi < 2

(d)

Z[o[n]]

= =1
zO

all z

(e)
Z[o[n - 1]] = Z-1
(f)

Izl > 0
0::; Izi <

Z[o[n
(g)

+ 1]] = z+1

00

Z 3.2.

[(I)" (u[n1- urn - 10])] =" (1)" = 1;0 _


9

2z

(2 )-10 z 1- (2Z)-1

Izl > 0

x[n]

={ n N,
<=>

O<n<N-l

N-"S n-

= n u[n1-

(n - N)u[n - N]

xln]

d d 1 -z dz X(z) => n urn] <=> -z d.z 1 _ %-1


::-1

Izi > 1

n urn) <=>

(1- Z-I)2.
X(z).

1::1 ~

1
-N-l

xln - no]
therefore

<=>

z-ftO => (n - N)u[n - N1 ~ (1_::-1)2 z

Izi > 1

64

3.3. (a) 0< [e] < 1

Xo(z)

=
::=

L
go

-1

a-ftz-"
00

La"z-"
ex> n=O

":-00

n=1

La"z" + La"z-"
n=O

--+ 1- QZ

az

1
1- az-1

= (l -z(l- ( az){z

2)

- a)

la1 < Izl < la1

Xdz)

(b)

%b

N <n { 0, n <0
0,

I,

05 n :S N - 1
. :::?

X.(z) =

N-1
n=O

L z-"

1_

z-N

1 - z-1

-=...."...,--.
zN-1{Z -

zN _

1)

(e)

Xc{z)

=Z

-1 (

zN-l(z

zN _1)2 = 1)
_

1
z2N-l

(zN _ 1)2 -;=-}

Z:F 0,1

Xdz)
__ pole zero cancel

-~~-~~--T~-

65 3.4. The pole-zero plot of X(z) appears below. X(z}

(a) For the Fourier transform of z[n1 to exist, the z-transform of z[n] must have an ROC which includes the unit circle, therefore, Iii < Izi < 121· Since this ROC lies outside this pole contributes a right-sided sequence. Since the ROC lies inside 2 and 3, these poles contribute left.-sided sequences. The overall z[n} is therefore two-sided.

l,

(b) T~sided

sequences have ROC's which look like washers. There are two possibilities. The ROC's to these are: Iii < Izi < 121 and 121< Izi < 131· (c) The ROC must be a connected region. For stability, the ROC must contain the unit circle. For causality the ROC must be outside the outermost pole. These conditions cannot be met. by any of the possible ROC's of this pole-zero plot.

corresponding

3.5.
X(z)

= (l + 2z)(1 + 3..-1)(1- z-l) = 2z + 5 - 4z-1 - 3z-2


=

L z[nJz-n
DOl

.... -00

Therefore, 3.6. (a.)

~~=u~+~+~~-~~-~-u~-~
X(z)

==

1+ iZ==

Izi > -2

Partial fractions: one pole -+ inspection, z[n]

(-!)nu[n1

Long division:
1 1
1 - iZ -1

.%

1 -2

+ ...

_ 1z-1
2 2

iz-1

_ 1%-1

1 + JZ -3

z[n] =

(-i)

urn]

66 (b)
X (z)

==

-1-+"'""'!'-Z--"'""1
2

Izi < 2

1 -1]

Partial Fractions: one pole ~ inspection, Long division:

%[n]= _(_j.)nu[-n

iz-1 + 1

2z
I

- 4z2

+ 8z3

+ ...

+ 2z
- 2z - 2z - 4z2

+ 4z2
+ 4z2

+ 8r
=> z[n] =

_ (_~)

u[-n -1]

(c) X(z) - 1+ Partial Fractions: X(z)


1-1 1 - -z 2 !Z-l + 1z-2 4 a

Izl> 2"

Long division: 1 + :!z-l 4

+ 1z-2 a

!1

(-136

+ l)z-2

+ ...

1 (_:_~)Z-l (_~ _ ~)z-l

(d) X(z) Partial Fractions: X(z) l_1z-1


2

- l-1z-2

Izl> 2"

1%1> 2"

Long division: see part (i) above. (e)

Partial Fractions:

Long division;

-a+z-1

3.7. (a) xln] = u[-n - 1] +


~ X(z)

(D
1

u[n1

= 1_

-1
z-1

1_

1%-1 2

2 < 1%1 < 1

Now to find H(z) we simply use H(z) = Y(z)/X(:);


H
% _

Le.,
. (1- :-1)(1-

()-

Y(z) _ z-1 X(z) - (l-tz-1)(I+z-1)

-i

-tz-1

iz-

)_

1- z-1 -1+:-1

H(z) causal ~ ROC Izi > 1. (b) Since one of the poles of X(z), which limited the ROC of X(z) to be less thaD 1, is cancelled by the zero of H(z), the ROC of Y(z) is the region in the z·plane that satisfies the remaining two constraints [z] > and [zl > 1. Hence Y(z) converges on Iz[ > 1.

(c) Y(z) Therefore,

-3 1 - 1Z-1
2

1 + z-1

Izl> 1

yIn] = -3.8. The causal system has system function

1(I)" uln] + -(-I)",,[n) 1 2 3


1- Z-1

B(z)- ..:....~- 1+ lz-1

..

and the input is %[n J X(z)

= (!)" urn 1 + u[ -n - 1J. Therefore the z·traDsform of the input is = 1_


1

lz-l -

1 1 - Z-1

= (1 ~ lz-I)(1 -

-lz-

z-1)

'3 < 1:1 < 1

68 (a) h{nJ causal ~

h{n] =
{b} Y(z}

( -43)" urn] - ( -. 3)"-1urn -1]


:( < Izi
3

Therefore the output is

y[n) :: -13

8 (I)" uln1 + 13 ( - i urn) 8 3)" 3


i < Izl which
2 includes the unit circle. Therefore, 1 --~~ tz-1) - (1- ~z-l) (1+ 1z-1)

(c) For h[nJ to be causal the ROC of H(z} must be . h[n] absolutely summable.

3.9.
H(z)-

1 + z-1 - (1- !z-1}(1

(a) h[n] causal ~ ROC outside (b) ROC includes (c)

!zl == 1

"* stable.

lzj::

!~Izl> i.
-3 -4
_1 1+

y[nJ ==
Y(z)
::

1(1)"
..
3 1Z-1

u[n] - 3(2)"u[-n - 1]

..
1- 2z-1
i

=
X(z)
::

l+z-1
1)

(1+ lz-l)(l- 2z-1)


Y(z) __ = (1- 1z11

"4 < Izi < 2

H(z)

(1 -

2z-1) 2

Izi < 2

x[nJ
(d)

::

-(2}"u[-n

- 1] + !{2)"-lul-n}

h(nJ
3.10. (a)

= 2 (~)

u[nJ- (_~)

u[n]

x[n]

==

G)
(~)"

uln -10] +

(D "urn

-10J

::

u(n1 + (~) n ufn]

- [( (~) ..+

(~y)urnJ - u[n - 11))] (


= O.

The last term is finite length and couverges everywhere except at z Therefore, ROC outside largest pole < 1%1·

69
(b)

%[n] == { 1, -10:5.n:5 10
0, otherwise Finite length but has positive and negative powers at
z in

its X (z). Therefore the ROC is 0

<

Izi < 00.


(e)

%[n1 %[-n)
(~) n

= 211u[-n) = ( 2"
H

1)-11

u[-nJ
1

X(l/z)

2" (1)-n u[-n]


(d)

u[n] =* ROC is =* ROC


is

Izl> 2" Izi <


2

:r[n I
(e)

non-sere

is right-sided,

at

n = -1,

so its ROC extends outward from the outermost pole e1" /3. But since it is the ROC does DOt include 00. So the ROC is 1 < Izi < 00.

:r[n1

== uln + 10J + 5J = {I, -10:5 n :5 -6


0, otherwise So the ROC is

urn

:r[n] is finite-length
(f)

and has only positive powers of z in its X{z).

Izl < 00.

:r[n] =

2" (1)"-1 urn) +

(2 + 3j)"-2u[-R

- I]

:tin] is two-sided, with two poles. Its ROC is the ring between the two poles:

1 < Izl <~.


3.11. :r[n] causal =* X (z)

!< Izl < 1213; I, or

= L :r[n]z-n
n=O

co

. which means this summation will include no positive powers of z, This means that the closed form of X(z) must converge at z == 00, Le.• z = oc must be in the ROC of X(z), or lim~~co X(z) 1: 00.

(a)
z--+OI:>(1 - !z-1)

lim

(1-Z-1)2

=1

could be causal

70
(b)

r-+oo

lim

(z _1)2

(z -

!)
4

=00

could DOt be causal

(c) lim (z _ !}5

HOO(Z

-i)6

=0

could be causal

(d)

" .... 00

lim

(z - '2)5

(Z - 1)6 ~

=00

couJd.not be causal

3.12.

(a) X1(z) The pole is at -2, and the zero is at 1/2. (b)
1- 1z-1

= 1 + ~z-l

1-

!Z-l

X2(Z) = (I + 1:-1)(1_1Z-1)
.23

The poles are at -1/2 and 2/3, and the zero is at 1/3. Since from the outermost pole: Izl > 2/3.

%2[n1 is eaasal, the ROC is extends

Xl

(z)

(c) 1+ z-l - 2z-2 X3(z) = 1 _ llZ-l + z-2


6

The poles ate at 3/2 and 2/3, and the zeros are at 1 and -2. Since %3[n] is absolutely the ROC must include the unit circle: 2/3 < Izl < 3/2.

summable,

X3 (z)

3/2

71 3.13. G(z)

= sin(z-1)(1 = (Z
-1

+ 3z-2 + 2z-4) +Z-$

--

Z-3

3!

5!

-)(1

Z-7

7!

+ 3z + 2z

-2-4

Lg[n1z-n
ft

g[l1] is simply the coefficient in front of

Z-l1

in this power series expansion

of G(z):

gIll] = -ill + 9i - 11!'


3.14.

H(z)

=
=

1
1- 1Z-2 4 (1 1-

1.

z-1)(1

+ !z-l)
1 + 1.12

=
Taking the inverse z-Transform:

0.5
!Z-1 2

0.5 +-.....,....-1

h[nJ = -{ - )nu[nJ
So, A2 3.15. Using long division, we get

11 22

+ -( - - )nu[n]
1 2

1 2

1 2

=-j

1 __

H(z)

1014 z

1_ -10

Taking the inverse z-transfOlm,

h[n] = {

(i)",
0,

otherwise

n = 0,1,2, ... ,9

Since hln] is 0 for n

< 0, the

system is

ca.usaJ.
1

3.16.

(a) To determine H(z), we first find X(z) aDd Y(z): X(z)

1
1- 2z-1

3' < Izi < 2

72 Y(z)

= =

5 _1z-1
3

1- tz-l

1- !Z-l 3

(1-lZ-1)(1 - iZ-1)'

Izl> '3

Now
H(z)

=
=

Y(z) X(z) 1 - 2z-1 1- ~Z-l


3

Izl > '3

The pole-zero plot of H(z)

is plotted below.

(b) Taking the inverse z-transform

of H{z),

we get

h[n]

(c) Since

H(z) = Y(z}
X(z) we can write

1- 2z- , 1- ~z-l

Y(z)(lwhose inverse z-transform leads


to

~z-l) = X{z)(l2 -yIn -1] 3

2z-1),

y[n]-

= %(n]- 2%[n - 1]

(d) The system is stable because the ROC includes the unit circle. It is also causal since the impulse response h[n] = 0 for n < O. 3.17. We solve this problem by finding the system function H(z) of the system, and then" looking at the differeIlt impulse responses which can result &om our choice of the ROC. Taking the z.-transform of the dilierence equatioD, we get Y(z)(l and thus

iz-1 +

z-') = X(z)(I-

z-1),

H(z) =

Y(z) X(z)

1-

1- Z-1 z-1 + z-2

73 = =
If the ROC is
1Z-1

(1- 2z-1)(1-iz-1) 2/3 1/3 1- 2z-1 + 1 - 1Z-1 2

(a) Izl <

j:

(b)

j < Izi < 2:


h[n1

= -~2"u[-n-l] 332
==>
2

+ !(!Y'u[nJ
h[O] =

3'

(c) Izl> 2:

h[nJ = 32"u[n]

+ 3 ('2)"u[n}
h[O] = 1.

11

==>
(d) Izi

> 2 or Izl <

t:

3.1S.

(a)

H(z)

Taking the inverse z-transform: h[n]

= -26[nl
yIn]

+ 3 (-'2)"u[n] + 3u[n].

118

(b) We use the eigenfunction property of the input:

= H(e1"Il)z[n],
1

where

-2+

1+ fe-j .. j2 1 - IJ

+
I

3" 1- e-i'rr /2

= -2+~1 1·+J_1 .
+J

=~.
Putting it together,

-2j

2+i

74
3.19. The ROC(Y(z» (a) Y{z) _ 1 - (1 + iz-1)(1 - tz-l) includes the intersection of ROC(H(z)) and ROC(X(z)).

lzl > t· (b) The ROC of Y(z) is exactly the .intersection of ROes of H(z) and X(z): ! [e] < 2. <
The intersection of ROCs of H(z) and X(z) is

1%1 >

i. So the

ROC of Y(z) is

(c) 1 Y(z) = (1- !z-l}(l The ROC is

+ iz-l)
includes the intersection of

1%1 > j.

3.20·. In both cases, the ROC of H(z) has to be chosen such that ROC(Y(z» ROC(R(z)) and ROC(X(z)). (a)

H(z) The ROC is (b)

1- ~z-l
4

- 1+ lz-1
3

Izi >

i.
H(z) :::;1_ !z-l
6

3.21.

(a) The ROC is

Izi > t·
=0
"
n<0

y[n] y[n1 y[n]


(b)

:::; Lz[kJh[n - kJ :: La"-I: = a" 1 _ a-I


1:=0 N-l

"

1- a-{"+l}

1 - a"+l 1_ a
I-N

0~ n< N - 1

1=0
N-l

:::; L:z[k]h[n - k]:: La"-I: :::; " c


1=0 1=0

1-=-:-1 :::;0"+1 : ~ l'


-N

n?. N

H(z)

= =

~anz-"
n=O

00

'-'

1 = -_-,.. _
1_;,-N

1- az-1

Izi > lal Izi > 0

X(z)
Therefore, Y(z) Now,

~z-"
n=O

'-'

- 1- ;,-1

= (1_ Oz-l )(1- Z-1) = (1- (1%-1 )(1- ;,-1) - ~(l:-_-(1%---:-l7"':)(--l-_-z--':":"l)

1- z-N

z-N

75
So

yIn]

= =

(_1_) [u[n]I-a
1-

on+lu[n] - urn - N] - o,,-N+lu[n


1_ o,,-N+l

N])

1_ a 0

0"+1

u[n] -

1_ a

urn - N]

n<O 05n5N-l

yIn]

1_....+1 1- ..

0..+1 (l_O-N) «e» ..-1


yrn]

3.22.

(a)

'=-CIQ

CIQ

h[k}:r[n - k]

=
= =
(b) Y(z)

,f;co
t=o

(3

n L3 (If3 --

(-~r
(-1),,+1), 3
3

U[k]) urn - k]

{ t (10,

n'?,O
otherwise

=
=

H{z)X(z)

=
yIn}

= = =

I + 1! + .. 1+ !z-l 13( 3 9 - -- 1)" urn} + -u[n] 4 4 ~(1+ 3(-3) n) urn] 9 :4 ( 1- (1) n+l) urn] -3
1z-1
%-1

z-1

3.23.

(a) H(z)

= (1-iz- }(I-iz5 + 1z= -4+ 1- 1 Z+ lz-2 t •


1 1) 1 Z-1

l- 1z-2 2

=
h[n]
:::

-4-

1-

1%-1

1- 1
t

Z-1

-4<S[n]- 2 (~)"

urn] + 7 (~) n urn]

76 (b)

3 1 1I[n1- -y[n - 1] + -lI[n - 2] 482 3.24. The plots of the sequences are shown below. {a.} Let

= z[n]

1 - -:tIn - 21

GIn] ==
Then
A(z} (b)

"=-00
=

co

6[n - 4k],

L
1:=-00

co

Z-4"

bIn]

= ~ [ei

nl

+ cos

(in)

+ sin

(i + 21m)] urn]

= ~ [(-1)" + cos (in)


=
B(z) 2"'

= =

{ 0, L £z-4" + L ~z-(4n+2)
00 00

f,

+ 1]uIn]

== 4k, k ~ 0 n = 4k + 2, k 2: 0 otherwise
n
n=0

R=O

3/2 + 1/2z-2
1Z-4

[s]

>1

.-

OJ

?u •••
11.4

o~~~~~~~~~=
... 0 1 ..... .... .... ....t

.'0

1.5

C li u

...
.10


3.25.

... ...
Z2

..

-a

IJ
II

II

to

X(z} - ~--:---:"~~ - (z - G)(Z - b) - z2 - (0 + b)z Obtain a. proper fraction:

z2

+ Gb

77

Z2 -

{c + b)z

+ ab r
%2 _

(c + b)z (4+ b)z

+ cb
- ab

z X(Z) = 1 + (Za -+a)(z -- b)

b)

cb

= 1+

(o+&o).. -co6co-6

z- a
(

(coH)6--a6 6- ..

Z- 6
_

= =

1+

z- a

0-6

.. 2

z- b

6" .. -6

= 1+ _1_

c- 6

aZ 1- az-1

2 -1

z 1 - 6z-1

f!2

-1

zln] = J[n] + ~an-1u[n a-b


6[n]

(_1_)
a-b

- 11- ~b"-lu[n a-b (cn+1


_

- 1]

bn+1)u[n

-1]

X(z) Long division:

1- 1z-1

- 1+

tz-l

+1.9Z-2
2 -1

+ ...

- 3z 2 -1 - 3z Therefore, (b)

9Z

2 -2

+
zIn] = 2(-t)nu[n]

iZ

2 -2

J[n]
_

x ( )_
zPoles at

z _ i _ Iz-1
% [n

3z-1
(1 - iZ-1)(1

+ iz-1) -

4 1-

iz-1 -

4 1-

iZ-1

!, and - i·

1 stable,

:;.

IZ I >

i :;. causal.
n urn]

Therefore,

zIn] = 4 (~)
(c) X(z)

_ 4 (_~)

uIn}

10(1- 4=)

Izi <:(

Therefore,

78
(d) X(z)

1 1 3 1- iZ-

Izl>

(3)-1 => causal

By long division:
1
1- 1z-3
3

+ iz-3

1%-&

+ ...

1 1

xln] = { (l)t,
0,
3.27. (a) X(z)

= 0,3,6, ...

otherwise

(1+ !z-I )2(1- 2z-1)(1 - 3z-1)


.!.

2' < Izi < 2


!!!!
1225

=
Therefore,

(1 + iz-2)2

35

..!L

(1 + iZ-1)

1225

(1 - 2z-1)

(1 - 3rl)

1225

1 ( -1 ) n+l 58 (-1)" 1568 x[n] = ~5(n+l)""2 u[n+lJ+ (35)2 2" u(n]+ (35)2 (2}"u[-n-l](b)

(35)2 (3}"u[-n-l]

2700

1 Therefore, :t[n1 = n! urn). (e) X(z)=~=z Therefore,


Z3 -

2z

+2z+1_2z-1

Izl <2

x[n]
3.28.

= OIn + 2) + Urn
nz[n)
¢>

+ 11- 2(2)nu[-n - 1]
d -z dz X(z)

(a)

xln - no] ¢> z-no X(z)


X{z)

3zd = (l-lz-1)2 = 12z -2 [-z dz

1 )] 1- tz-I

xln]

is left-sided.

Therefore, X(z) c:omsponds to:

xln) = -12(n - 2)

4 (1)"-2 u[-n + 1)

79
(b) X(z} = sin(z) Therefore, (-1)" = L (2k + 1)1 ~H1 t=O
00

ROC includes

Izi =

Which is stable.
(c) X{z)

= 1_ Z-7 = Z7 X(%)

Z7

-2

1
l-z-7
00

1%1

>1

= Z7 - L z-7n
tI=O

Therefore,

:tin] = 5[n + 7] -

L 51n tI=O

00

7k]

3.29.

3.30. X(z) = log2( - z) 2 (a)

1 1%1 < 2"

X{z) = log(! _ 2%) = _,,_"


~

00

(2z)'
j

s=1

=_ " ~

-1

_(2z}-t

1=-00

1 -l

= '" ~l
l=-oo

(l)t _ 2
%

Therefore,

z[n]
(b)

=;;1 c)n "2


(

1&[-n - 1]

1U'[n] nz[n1

d -z-log(I-2%) dz

= -z

-1

1)

2%

(-2) = z-1 ( 1 -

-1 ) ,
!%-1 2

Izi <"2

zIn]

(D nu[-n-"I] ;; C)n u[-n-l] "2


1

80 3.31. (a)

%[n)
X(z)

= 4"u[n] + b"u[n] + e"u[-n - 1] =


1 + 1- bz-1 - 1- cz 1- 2cz-1 + (be + ClC - Ab)z-2 (1- 4% 1)(1 _ bz-1 )(1- a-I)
Z2

141< Ibl < lei


Ibl < Izi <

1 1- az-1

lei

X(z)
Poles: a, b, c,

=
and

Ibl < Izi < lei

Zeros:

ZlJ Z2, 00

where

Zl

are roots of numerator

quadratic.

pole-zero pi t (0)

(b)

xln] x1ln] x2[n] x[n]

=
=

n2a"u[n] anu[n] <:}X1(z) = 1- 4Z-1


d 1

Izi > a
4Z-1

n%1

In] = na"u[n].<:} X2{z) = -z dz Xl (z) = (1 _ az-1)2


d d az-1 = n2a"u[n] <:} -z dz X2(z) = -z dz ( (1 _ 4Z-1)2
)

[z] > a

= nX2[n]
X(z)

Izl >4

-4Z-1(1

+ az-l)

(1 -

4Z-1)3

Izj > a

(c)

xln] = e"· (cos ;;n)


Therefore, X(z) = 1 for 3.32. From the pole-zero diagram

urn] -

e"· (cos ~ n) uln -

1]

= e"· (cos ~ n) (u[n] - uln - 1]) = 6[n]

alllzl.
Izl > i
3

yIn]

= x[-n + 3] = xl-(n - 3)]

81
Poles at 0,

-i, 1 ± i. zeros at

00

z[n] causal => z[ -n + 3J is left-sided => ROC is 0 < Izl < 4/3.

3.33. From pole-zero diagram

<a)
y[n]::=; zeros poles

(I)"
2

z[nJ => Y(z)

= X(2z) = 4z2 -+21 1 2z

Y(z)

(b)

w[n1
W(z)

= cos (~)

z[n]

= i(e1"n/2

+ e-j" ../2)z[nj

= ~X(e-j'l</2z) + ~X(ei·/2z) = ~X(-jZ)


=2

W(z} poles at ±t; zeros at ±I

1 (_%2, + 1') + -1 (-z'l + 1) = ,


-1% - 2
1

+ ~X(jz)
2
(2
Z

JZ -

z2 -

+i

1)

82 Y(z)

3.34. H(z)
1 = 3 - 7z- + 5z- = 5+ ~-::--~1 1- iZ-1 + z-2 1- 2z1 2

3
1 -lz-1

h[n] stable:::} (a)

h[n]

= M[n]-

2"u[-n - 1] - 3

G)" urn]

yfn1

h[n] * :r[nl

..

h[k] n<O

= n~O

(b) Y(z)

1-

Z-1

H{z) = -2

1- z-l

+2

yIn] = -2u[n]- 2(2)"u[-n -1] + 3


3.35. H(z)

G)
z

1- 2z-1
n

+3

1-

!z-l
2

2 < Izl < 2

urn]

l-z

1- z4 1 1- z-1
Z-l

- z-1

CZ-1)

1-

z3 -)
Z-4

Izl> 1

uIn] ~

= z -1
_

Izl> 1

U(z)H(z) =

z-4

(1- z-4)(Iz-1

=
u[n1- h[n]

1- z-l

z-4 l-z-4
GO

Izi > 1

uln - 1]- L6[n - 4. - 4k1 ~

fI
83 3.36.

z[n]
y[n1 =
(a)

= urn] ~
1] = 4

X(z)

= 1_ Z-l
Y(z)

Izl > 1

2' (1)"-1 urn +

(1)"+1urn + 1] ~ 2'

1_

4z

!Z-1
1

Izl> 2'

H(z) = Y(z) = 4%(1 - z-l) X(Z) . 1- 1z-1 2

Izl> 2'

(b) H(z)

= = =

C 1- 1z-1 2

hIn]

2' 2' (1)"+1u[n+l]-4 (I)" uIn]

4 1_ 1Z-1
2

[z]

> 2'

4O[n + 1] - 2
stabJe.

(~r
urn]
1

(c) The ROC of H(z)

includes

1z1 = 1 =>

(d) From part (b) we see that h[n] starts at n 3.37. X(z}

= -1
1 2

=> not causal

1_

L-1 + ;Z-1
1-

has poles at z and z 2. Since the unit circle is in the region of convergence X(z) and :tIn] have both a causal and an anticausal part. The ca.usal part is "outside" the pole at The antica.usal part is "inside" the pole at 2, therefore,

=!

i.

:rIO] is the sum of the two parts


%(O}
3.38. Y(z) - ---:-----:-- (1- ~z-1)(1 + 1z-1)
Z-l

lim
& .....

00

1 , 1 - "2Z-1

+ z-O _i.:_2 lim Z2

1z

= -3 + 0 = -3
lzl>

+ Z-2

1- z-1

84 Therefore using a contour C that lies outside of Izl

= 1 we get

Y[l] = =
=
3.39. (a)

11 2(z + l)zndz 27rj!e (z - !)(z + l)(z - 1) 2(i + 1)(1) 2(-i + I)(-t) <!+!)(!-1) + (-!- H(-!-I) 18 2 -5 - '5 +6=2 2(1+ 1)(1) + (1-!)(I+i)

X (z)

= -:---~-:--"""",-:-:-=-:--",-::-:--......-:--:--"""";7 + i) (z - !)(z - j)lO(z + 1)2(% + j)(z l < 1%1< j.


z+2)(z+2
::; 7) inside unit circle] inside C), where C is contour in ROC (say the unit circle). l){ 3 )10 z

Z10

Stable => ROC includes Izi = 1. Therefore, the ROC is (b) %[-8}

= E[residues

of X(z)z-S

%[8] = E [residues of (
First order pole a.t z

z-2

z-2

(z+2

3 )2(

= !is only
x[-8]

one inside the unit circle. Therefore !

= (1 - i)10C!+ !)2(!
V(z) W{z)

+ jH! +

i) = 96

3.40.

(a) After writing the following equalities:

= X(z) - W(z) = V(z)H(z) + E(z)


1

we solve for W(z):

W(z)

H(z) 1+ H(z) X(z)

+ 1+ H(z)

E(z)

(b)

H1(z)

H(z) H{ ) 1+ z

I-rT

6-1

1+

x-I
1-,&-1

=z

-1

H..(z) ..
(e) H(:) is not stable due to its pole at 3.41.

=1
%

+ 1-rl

,&-1

= 1- %-1

= 1, but Hl(Z) and H2{Z) are. <1

(a) Yes, hin] is BIBO stable if its ROC includes the unit circle. Beuce, the system is stable if rmin and r_u: > 1. (b) Let's consider the system step by step. (i) First, vln] Q-nz[n]. By taking the z-transform of both sides, V(z) = X(az). (ti) Second, u[n] is filtered to get wIn]. So W(z) = H(z}V(z) = H(z)X(az}. [iii] Finally, yIn] Qnw[n]. In the z..transform domain, Y(z) = W(zjo) = H(zjo)X(z).

In conclusion, the system is tTl, with system function G(z) = H(z/o)

and g[n]

= onh[nJ.

85

(c) The ROC of G(z) is Qrmin < Izl < Qr,_". We want rmin < 1/0 and r .....z > 1/0 for the
to be stable. 3.42. (a) h[n] is the response of the system when %[nJ
10

system

= 6[n]. Hence,
-'=1

hln]

+ L ath[n pI

= 6[n] + p6[n -

1],

FUrther, since the system is causal,

hln] = 0 for n < O. Therefore,

h[O]
(b) At n = 1,

+ L o..h[-kl.=
1:=1

10

h[O] == 6[0] == 1.

h[l]

+ olh[O] = 6[1} + 136[0]

=>

01

fJ - hIll
h[O]

= f3 - hll]

(c) How can we extend hln] for n > 10 and still have it compatible with the difference equation for S? Note that the difference equation can describe systems up to order 10. If we choose

h[n1 = (O.9)ft cos( 4n)u[n],


we only need a second order difference equation:

,..

The z-transform

of h{n] can be found from the z-uansfotm table:

H(z)

(1 -

- :;'2 O.ge 1l";'z-1}(1 - O.ge-J"i'


j

0.'

z-1)

3.43.

(a)
X(z)

= 1- !Z-l
2

1- 2%-1' 1- 1z-1' ..

2 < Izl < 2


lzl> 4
3

Y(z) = 1-

6
1Z-1

B(z)

86

1.5r----.-------"1

~1~_'-,---,~:-:--O~-:'05O:---1-',...,.s,---..,-'2

(b) hIn] = (c)

-2

4 (3)

ft

u[n1-

(3) 4

ft-l

urn -

1]

3 y[nJ- -yIn - 1] = x[n] - 2x[n - 1]


4

(d) The system is stable because the ROC includes the unit circle. It is also causal since h[nl n < O. 3.44. (a)

= 0 for

The ROC is

i < Izl < 2.

(b) The following figure shows the pole-zero plot of Y(z). Since X(z) has poles at 0.5 and 2, the poles at 1 and ~0.5 are due to H(z). Since H(z) is causal, its ROC is [e] > 1. The ROC of Y(z) must contain the intersection of the ROC of X(z) and the ROC of H(z). Hence the ROC of Y(z) is 1 < Izi < 2.
__ ..... dy!Z)

1.sr----...------,

-1.5c....,_,--0,5""""''-+O --=05-=-"",'---"",.s:--."..J2 (c)

-JIM

H(z)

Y(z)

X(z)
(l-",-l)(1+ ",-IHI-2z-1,
2z-1) 1 (1-12z-1)(1

87

Taking t.he inverse z-t.ransform, we find

(d) Since H{z) has a pole on the unit circle, the system is not. stable.

.45.

(a)

ny[n1

== %[n]
== X(z) ==
- / z-1 X(z)dz

=rz:
Y(z)

dY(z)

(b) To apply the results of part (a), we let %[nJ == W(z)

urn /

IJ, and wIn]

== yIn].

= =
=

z-1

1_ 1

Z-1 Z-1

dz

-/ -

=
3.46. • (b) Since the ROC is a ring all the z-plane,
00. ROC; Izl > (d) Since the ROC of %[nl includes n < O. Therefore :t[n] is causal.

-1 1 -+--dz z z- 1 lD(z) -In(z - 1)

z{z - 1)

dz

(a) Since yIn] is stable, its ROC contains the unit-circle. (c) x[n1 is stable, so its ROC contains the unit-circle.

Hence, Y(z) converges for!

< Izl < 2.

yIn] is a tWl>sided sequence.


Also, it has a zero at
DO 00

so the ROC includes

00,

X(z) contains

positive powers of z, and so xln]

= 0 for

(e) :t[01

= X(Z)I"'::DD ==

(1 0

+ 1Z-1)(1
00.

A(l -iZ-1)

- iz-1) I&=DD

(f) H(z) has zeros at -.75 and 0, and poles at 2 and

Its ROC is

Izi < 2.

88

-1.5·~_I~-CU""""""'O!--U=-....,,--:":t.5~Z~

(g) Since the ROC of h[n] includes 0, H(z) contains h[nl 0 for n > O. Therefore h[n] is anti-causal.

.....

DO

negative powers of z, which implies that

3.47.

(a) X(z)

=
=

OC>

~>[nlz-n
n=O
gO

X(oo)
Therefore, X(oo) at z 00.

.-"00 ~ R:O

lim

"z[nlz-n

= z[o]

= :r[01 #- 0 and

finite by assumption.

Thus, X(z) has neither a pole nor a zero Then the most general .

(b) Suppose X (z) has finite numbers of poles and zeros in the finite z-plaae. form for X(z) is

X(z)

L z[n]z-n =
n=O

ee

IT (z KzL";:"N~=.:..l __

1I

c.)
-

II(z - d,.)
=

where K is a constant and M and N integer representing the net number X(oo) = :rIO] #- 0 and < 00 we must z-plane must equal the total number 3.48.

are finite positive integen and L is a finite positive or negative of poles (L < 0) or zeros (L > 0) at % O. Clearly, since have L + M N; i.e., the total Dumber of zeros in the finite of poles in the finite s-plaae.

X( )
Z

=:

P(z) Q(z)

where P(z) and Q(z) are polynomials and roots of Q{z) inside Izi 1-

in z. Sequence is absolutely summable => ROC c:ootains 1%1

=1

These conditions do not necessarily imply that :tIn] is ca.usal. A shift of a caasal sequence would only add more zeros at z = 0 to P{z). For example, consider

X(z)

=
=

--1 Z-

.r
J

Izi >-

__'-=Z·

.I-I

=>

:r[n] =

2 (1)"+1 u[n+

1 1 1-2.1-1

1] ~ right-sided but non-causal.

89
3.49.

z{n]
X(z}

= 6[n) + a6(n - N) lal < 1 = 1 + az-N


= logX(z) == loge!

.f(z)
Therefore,

+ az-N) = 4Z-N

aZz-2N
2

a3z-3N
3

- .. ,

3.50.

(a)

:z:{n]= :z:[-n] =* X(z)


Therefore,

=X

CD

i.e., (b)

1/Zo

is also a zero of X {z}.

Therefore X(Zo} 3.51. (a)

= 0 = X(zo)

i.e., zQ is also a zero and by part (a) so is 1/ zO.

(b)

Z{z(-nlJ
(c)

=L
00

%f-njz-n

L
n..e-oo

co

:z:fnJ(z-l)-"

= X(z-l)

n=-oo

Z[R.e{zfn]}J
(d)

= Z [:z:{n]~ z'"[n]] = ~ [X(z)

+ X·(z·)]

3.52.

%l(n}:: (-l}"z(n) ~ Xl(z) =

L
CICI
fP'-QO

(-I)"z(n)z-· = X(-z)

The poles and zeros are rotated ISO degrees about the origin.

90
N-l

3.53.

(a)

. 8( )
:t

- L%[n]sin(nw)
=:}

= tan

-1

(Im{X(e''''}») Xe{X(e "')}


J

() tanS: W =

...."N~-~l;__---

n=O

L%[n]C06(nw)
N-l

tan9:(w)
tan 8: (w}%[O]+

L %[n1cos(wn)
N-l »=1

N-l

=-

L %[n]siD(nw)
+ SiD(nw)) = 0

n=(J

n=1

L %[n](ta.n 9:&-(",)cos(nw)

tan 8z(Wl)

+ %[0] ~

N-l

%[nJ(tan8:{wl} COS(nwl) + sin(nw.)) = 0

for N - 1 values of (b) %[n] = 6[n]

Wk in the range 0 < "".I: < 11". + Urn -I) + U[n - 2] => X(z} =- 1 + 2z-1 + 3z-:l

8,.(",)
Consider the values 8z

=- taD-l
9z

1 + 2 cos(w) + 3 cos(~)

-2si.D(w) - 3sin(2w)

(j)

= fI: and
(~)

(2;) = 5;, which

give the equations

tan8z

+ %;od%[l] ( tan 8.. (i) cos ~ + siD~)


+
%[2) (tan 9",

tan

8",C;)

(i) cos

1r

+ sin 1r)] =- 0
cos 2;

+ %: [%[1] ( tanS:&+

(2;) %[21 (tan9,.(2;)


1

+ sin 2;)

COS ~

+ sin ~)]

=0
=0 =0

1 + %[0] (%[1] . 1 + %(2]· -1)

-~ +

%to]

(%[1] + 2~

+ ~)

+ %[2] (2~

- ~)

%[0] + %[1]- %[2] -%[0]


Therefore where %[0] is undetermined. 3.54.:r[n]

O}
===>

{%[1] %[21

= 2z[0]

+ 2%[1] -

%[2] = 0

=-

3%[01

:tin] =- z[O}(6[n] + urn - 1] + U[n - 2})

= 0 for n < 0 implies:


00 .J'~OD

lim X(z) == lim "z[n]z-n


.J~OO

n~

= :rIO] +

lim
,iI-flOO

L :r{nJz-n
00

=- %[0]

~l

91 For the case %[n]

= 0 for n > 0,
.-0

lim X(z)

= .-0 __" ~ %[nJz-" = %[0]+ • """I lim lim "%[-nJz" ~ £....


....0

= %[0]

S.55.

Ca>
Css[n] =

L z[~}z(n + k] = L z[-k]z[n
~ oc

- k] = z[-n).

zfn]

Csz(z)

= X(z-l)X(z)

= X(z)X(z-l)
has ROC:

X(z) has ROC: rR < Izi < rL and thereCore X(Z-l) bas ROC: maz[ri-!,rR] < Izi < miD(rill,TL1 (b) z[n] o",,[nJ is stable if 101< 1. In this case

ri1 < 1%1 ril. <

Therefore Cu(z}

X(z) = 1_ oz-l
Therefore

101<

Izi

and

X(z-l):::::--

1 l-oz

C.s(z)

_4%-1

= 1-

0%-1

1- az ::::: 1- oz-l)(1 - a-1z-l) (


-

=
This implies that

~ 1- 4%-1

1-

'-::-;:
0-1%-1

es.[n1 =

1':02 [anu[n] + a-nul -n - 11]

Thus, in summary, the poles an at a and a-I; the zeros are at 0 and 00; and the ROC of Cu(z) is lal < JzI < la-II·

(e) Clearly, %I[n] = z[-n]

will have the same autocorrelatioD function. For example,


1 1-0

Xl(z)=--

(d) Also, any delayed versioD. of z[n] will have the same autocorTela.tioD function; e.1" %2[n] :: %[n-mJ implies

92
3.56. In order to be a z-tra.osform, X(z) must be analytic in some annular region of the z-plane. To determine if X(z) = z· is aJ?alytic we examine the existence of X'(:) by the Cauchy Riemann conditions. If X(z)

= X(~

+ j,)

= u(~,,,) + itl(z,,)

then for the derivative to aist at z, we must have

In our cue,

X(:t and thus,

+ i,) = z - i,

8u

8~ -1 #
unless ~ and II are zero. Thus, X'(z) exists oalyat ~[nJ 3.57. U X(z) has a pole at z

= ~j

a, =-1 z = o. X{z)

8v

is DOt ualytic

Ulywhere. Therefore,

X(z)z·-Jdz

does Dot exist.


&S

= zo then A{z) cao be expressed


A(z)= A(zo)
00

a Taylor's series about z

= ZOo

+ L -,-(z n.
_1

.4"(Zo)

- %0)"

where A(zo) = O. Thus Res [X(z) at z ...

%0] = X(z)(z - %o)I=.te == B(Z)j A(z) ==


-=1

B(z)(z ~

zo)

.4"(zo) (z _ .to)" "=aa


n!

.. ".te-