1

8. z-Domain Analysis of Discrete-Time

Signals and Systems
8.1. Definition of z-Transform (10.0-10.3)
8.2. Properties of z-Transform (10.5)
8.3. System Function (10.7)
8.4. Classification of a Linear Time-Invariant Discrete-Time
System by its System Function (10.7)
8.5. Linear Constant-Coefficient Difference Equations (10.7)
8.6. Implementation of Discrete-Time Systems (10.8)
8.1. Definition of z-Transform
The z-transform of a sequence x(n) is defined as
where z is a complex variable. x(n) can be recovered from X(z) by
, z ) n ( x ) z ( X
n
n

(8.1)
. dz z ) z ( X
j 2
1
) n ( x
) r | z (|
1 n

(8.2)
(8.2) is called the inverse z-transform.
8.1.1. Derivation of Inverse z-Transform
Letting z=re
j
in (8.1), we obtain
( ) . e r ) n ( x re X
n
n j n j

(8.3)
It can be seen from (8.3) that X(re
j
) is actually the Fourier transform
of x(n)r
-n
. Thus,
( ) . d e re X
2
1
r ) n ( x
2
n j j n

(8.4)
Multiplying both the sides of (8.4) by r
n
, we obtain
( )( )

2
n
j j
d re re X
2
1
) n ( x (8.5)
and thus (8.2).
8.1.2. Relation of z-Transformto Fourier Transform
The z-transform of x(n) is the Fourier transform of x(n)r
-n
.
The Fourier transform of x(n) is the z-transform of x(n) when z=
e
j
. If X() and X(z) are the Fourier transform and the z-transform
of x(n), respectively, then
X()=X(z)|
z=exp(j)
=X(e
j
). (8.6)
8.1.3. Zeros and Poles of z-Transform
z
0
is called a zero of X(z) if X(z
0
)=0. z
0
is called a pole of X(z) if
X(z
0
)=.
Zeros and poles are indicated with o and in the complex plane,
respectively.
If it is a rational function, the algebraic expression of X(z) can be
specified by the zeros and the poles except for a scale factor.
8.1.4. Region of Convergence of z-Transform
The set of values of z for which X(z) converges is referred to as
the region of convergence (ROC) of X(z). Note that only the
algebraic expression of X(z) cannot specify x(n) uniquely. The ROC
of X(z) is also required.
Different types of signals correspond to different types of ROCs.
2
(1) If x(n) is of finite duration, then the ROC of X(z) is the entire
plane, possibly excluding z=0, z= or both (If x(n) has nonzero
values for n>0, then z=0 is excluded. If x(n) has nonzero values for
n<0, then z= is excluded).
(2) If x(n) is right-sided, then the ROC of X(z) is the exterior of a
circle centered about the origin (|z|>r
0
), possibly excluding z= (If
x(n) has nonzero values for n<0, then z= is excluded), or empty.
(3) If x(n) is left-sided, then the ROC of X(z) is the interior of a
circle centered about the origin (|z|<r
0
), possibly excluding z=0 (If
x(n) has nonzero values for n>0, then z=0 is excluded), or empty.
(4) If x(n) is two-sided, then the ROC of X(z) is a ring centered
about the origin (r
1
<|z|<r
2
) or empty.
Furthermore, if X(z) is rational, then the ROC of X(z) is bounded
by poles or extends to infinity.
Example. Find the z-transforms of the following signals, including
the algebraic expressions and the ROCs.
(1) x(n)=a
n
u(n).
(2) x(n)=-a
n
u(-n-1).
(3) x(n)=7(1/3)
n
u(n)-6(1/2)
n
u(n).
(4) x(n)=(1/3)
n
sin(n/4)u(n).
(5) x(n)=(n).
(6) x(n)=(n-1).
(7) x(n)=(n+1).
.
otherwise 0,
1 - N n 0 , a
x(n) (8)
n

'

(9) x(n)=a
|n|
, a>0.
8.2. Properties of z-Transform
8.2.1. Linearity
If x
1
(n)X
1
(z) and x
2
(n)X
2
(z), then
a
1
x
1
(n)+a
2
x
2
(n)a
1
X
1
(z)+a
2
X
2
(z), (8.7)
where a
1
and a
2
are two arbitrary constants.
8.2.2. Differentiation
If x(n)X(z), then
.
dz
) z ( dX
z ) n ( nx (8.8)
8.2.3. Shifting
If x(n)X(z), then
x(n-m)X(z)z
-m
, (8.9)
where m is an arbitrary integer.
Example. Find the inverse z-transforms of the following signals:
(1) X(z)=1/(1-az
-1
)
m
, |z|>|a|.
(2) X(z)=1/(1-az
-1
)
m
, |z|<|a|.
8.2.4. Scaling
If x(n)X(z), then
( ), z X ) z ( Y
m of multiple a n , 0
m of multiple a n ), m / n ( x
) n ( y
m

'

(8.10)
where m is a nonzero integer.
Letting m=-1 in (8.10), we can obtain the time-reversal property of
the z-transform, i.e.,
x(-n)X(z
-1
). (8.11)
3
.
a
z
X ) n ( x a
n

,
_

(8.12)
From (8.11), the following conclusions can be drawn:
(1) x(n) even X(z)=X(z
-1
).
(2) x(n) odd X(z)=-X(z
-1
).
If x(n)X(z), then
Letting a=exp(j
0
) in (8.12), we can obtain the rotation property of
the z-transform, i.e.,
x(n)exp(j
0
n)X[zexp(-j
0
)]. (8.13)
8.2.5. Conjugation
If x(n)X(z), then
x
*
(n)X
*
(z
*
). (8.14)
From (8.14), the following conclusions can be drawn:
(1) x(n) real X(z)=X
*
(z
*
).
(2) x(n) purely imaginary X(z)=-X
*
(z
*
).
8.2.6. Convolution
If x
1
(n)X
1
(z) and x
2
(n)X
2
(z), then
x
1
(n)*x
2
(n)X
1
(z)X
2
(z). (8.15)
Example. Let x
1
(n)=a
n
u(n), x
2
(n)=b
n
u(n) and ab. Find x
1
(n)*x
2
(n).
8.2.7. Initial-Value Theorem and Final-Value Theorem
Assume x(n)=0 for n<0. If x(n)X(z), then
). z ( X lim ) 0 ( x
z
(8.16)
Proof. X(z) can be expressed as
. z ) n ( x ) 0 ( x z ) n ( x z ) n ( x X(z)
1 n
n
0 n
n
n
n

+ (8.17)
Letting z, we obtain
[ ]. ) z ( X ) 1 z ( lim ) ( x
1 z

(8.19)
Proof. (z-1)X(z) can be expressed as
. z )] n ( x ) 1 n ( x [ z ) 0 ( x z )] n ( x ) 1 n ( x [
z ) n ( x z ) 1 n ( x ) z ( X ) z ( zX ) z ( X ) 1 z (
0 n
n
n
n
n
n
n
n

+ + +
+
(8.20)
). 0 ( x ) z ( X lim
z

(8.18)
Assume x(n)=0 for n<0. If x(n)X(z), then
Letting z1, we obtain
). ( x ) 1 N ( x lim )] n ( x ) 1 n ( x [ ) 0 ( x lim
)] n ( x ) 1 n ( x [ ) 0 ( x )] z ( X ) 1 z [( lim
N
N
0 n
N
0 n
1 z
+

'

+ +
+ +

(8.21)
Example. Find the inverse z-transform of X(z)=ln(1-az
-1
), |z|>|a|.
8.2.8. Inverse of Rational z-Transforms
Let us consider how to find x(n) if X(z) is a rational function. This
can be carried out by the partial-fraction expansion.
(1) By the long division or inspection, expand X(z) into
X(z)= P
0
(z
-1
)+ P
1
(z
-1
)/P
2
(z
-1
), (8.22)
where P
0
(z
-1
), P
1
(z
-1
) and P
2
(z
-1
) are polynomials of z
-1
, and the order
of P
1
(z
-1
) is lower than that of P
2
(z
-1
).
(2) Expand P
1
(z
-1
)/P
2
(z
-1
) into partial fractions of z
-1
. Let z=a be an
4
Mth-order root of P
2
(z
-1
)=0. Then, it corresponds to M terms in the
expansion, i.e.,
( )
( )
( )
,
az 1
A
z P
z P
M
1 m
m
1
m
1
2
1
1
+

(8.23)
where A
m
is determined by coefficient comparison or the elimination
method (multiplying (8.23) by (1-az
-1
)
M
, differentiating the result M-
m times with respect to z
-1
and substituting a for z to obtain A
m
).
(3) Determine x(n) according to the basic z-transform pairs and the
properties of the z-transform. The basic z-transform pairs are
(n)1, (8.24)
( )
|, a | | z | ,
az 1
1
) n ( u a C
m
1
n 1 m
1 m n
>

+
(8.25)
( )
. | a | | z | ,
az 1
1
) 1 n ( u a C
m
1
n 1 m
1 m n
<

+
(8.26)
When m=1, (8.25) and (8.26) become, respectively,
|, a | | z | ,
az 1
1
) n ( u a
1
n
>

(8.27)
. | a | | z | ,
az 1
1
) 1 n ( u a
1
n
<

(8.28)
Example. Find the inverse z-transform of
.
z
3
1
1 z
4
1
1
z
6
5
3
) z ( X
1 1
1

,
_

,
_

8.3. System Function

A linear time-invariant discrete-time system can be characterized
by the system function. It is defined as the z-transform of the impulse
response.
(8.29)
The input-output relation of a linear time-invariant discrete-time
system can be expressed by the system function, i.e.,
Y(z)=X(z)H(z), (8.30)
where X(z) and Y(z) are the z-transforms of the input and the output,
respectively, and H(z) is the system function.
8.4. Classification of a Linear Time-Invariant Discrete-Time
System by its System Function
8.4.1. Memoryless Systems versus Systems with Memory
Assume that H(z) is the system function of a linear time-invariant
discrete-time system. The system is memoryless if and only if H(z) is
a constant.
8.4.2. Causal Systems versus Noncausal Systems
Assume that H(z) is the system function of a linear time-invariant
discrete-time system. The system is causal if and only if the ROC of
H(z) is the exterior of a circle centered about the origin, including z=
.
Example. Determine the causality of the following systems:
. 2 | z | ,
z 2 1
1
z
2
1
1
1
) H(z) 2 (
.
8
1
z
4
1
z
z z 2 z
) H(z) 1 (
1
1
2
2 3
>

+ +
+

8.4.3. Stable Systems versus Unstable Systems

Let H(z) be the system function of a linear time-invariant discrete-
time system. Then the system is stable if and only if the ROC of H(z)
includes the unit circle centered about the origin.
5
Example. Determine the stability of the following systems:
. causal is system the and ,
z r z ) cos r 2 ( 1
1
) H(z) 3 (
. causal is system the and ,
az 1
1
) H(z) 2 (
.
z 2 1
1
z
2
1
1
1
) H(z) 1 (
2 2 1
1
1
1

8.4.4. Invertible Systems versus Noninvertible Systems

Assume that two linear time-invariant discrete-time systems A and
B have the system functions G(z) and H(z), respectively. Then A and
B are mutually inverse if and only if
G(z)H(z)=1. (8.31)
(8.31) can be used to construct the inverse of a given system.
8.5. Linear Constant-Coefficient Difference Equations
The zero-state response of a linear constant-coefficient difference
equation can be found using the z-transform.
Example. A causal discrete-time system is given by
y(n)-0.5y(n-1)=x(n), (8.32)
where x(n)=0.25
n
u(n) and y(-1)=2. Find the zero-input response, the
zero-state response and the total response.
First, using the method in section 5.5, we can obtain the zero-input
response
y
zi
(n)=0.5
n
. (8.33)
Next we consider the zero-state response. Like the total response, the
zero-state response satisfies (8.32), i.e.,
y
zs
(n)-0.5y
zs
(n-1)=x(n). (8.34)
Taking the z-transform of (8.34), we obtain
.
z 25 . 0 1
1
) z ( Y z 5 . 0 ) z ( Y
1 zs
1
zs

(8.35)
From (8.35), we obtain
.
z 25 . 0 1
1
z 5 . 0 1
2
) z ( Y
1 1 zs

(8.36)
The ROC of 2/(1-0.5z
-1
) is determined by the system. The system is
causal, and thus the ROC of 2/(1-0.5z
-1
) is |z|>0.5. The ROC of -1/(1-
0.25z
-1
) is determined by the input. The input is 0.25
n
u(n), and thus
the ROC of -1/(1-0.25z
-1
) is |z|>0.25. Hence, the inverse z-transform
of Y
zs
(z) is
y
zs
(n)=(20.5
n
-0.25
n
)u(n). (8.37)
The total response is the sum of the zero-input response and the zero-
state response, i.e.,
y(n)=0.5
n
+(20.5
n
-0.25
n
)u(n). (8.38)
Example. A causal discrete-time system is given by
y(n)-0.5y(n-1)=x(n). (8.39)
Find the impulse response.
The zero-state response satisfies (8.39), i.e.,
y
zs
(n)-0.5y
zs
(n-1)=x(n). (8.40)
Taking the z-transform of (8.40), we obtain
.
z 5 . 0 1
1
) z ( X
) z ( Y
1
zs

(8.41)
Thus,
.
z 5 . 0 1
1
) z ( H
1

(8.42)
6
Since the system is causal, the inverse z-transform of H(z) is
h(n)=0.5
n
u(n). (8.43)
Example. Consider a linear time-invariant discrete-time system. If
the input is
x
1
(n)=(1/6)
n
u(n), (8.44)
then the output is
y
1
(n)=[a(1/2)
n
+10(1/3)
n
]u(n). (8.45)
If the input is
x
2
(n)=(-1)
n
, (8.46)
then the output is
y
2
(n)=(7/4)(-1)
n
. (8.47)
Use a difference equation to characterize the system.
Figure 8.1. Elements of a Block Diagram.
+
X
1
(z)
X
2
(z)
X
1
(z)+X
2
(z)
An Adder
X(z) aX(z)
A Coefficient Multiplier
X(z)
A Delayer
z
-1
z
-1
X(z)
a
8.6. Implementation of Discrete-Time Systems
Let us discuss the implementation of a causal linear time-invariant
discrete-time system characterized by a rational system function. We
only consider the implementation using adders, coefficient
multipliers and delayers. In the implementation, various structures
can be chosen. These structures are different in accuracy, speed, cost,
and so on.
A structure can be described by a block diagram. Figure 8.1 gives
the descriptions of an adder, a coefficient multiplier and a delayer in
a block diagram.
Example. Suppose that a causal linear time-invariant discrete-time
system is characterized by
.
z
8
1
z
4
1
1
z
2
1
z
4
7
1
H(z)
2 1
2 1


+

(8.48)
z
-1
z
-1
-7/4
X(z)
-1/2
+
+
z
-1
z
-1
+
+
1/8
Y(z)
-1/4
Figure 8.2. Direct Form I.
From (8.48), we obtain
Y(z)=-(1/4)z
-1
Y(z)+(1/8)z
-2
Y(z)+
X(z)-(7/4)z
-1
X(z)-(1/2)z
-2
X(z). (8.49)
The direct-form I structure of the system is based on (8.49). Draw the
block diagram.
7
The block diagram is shown in figure 8.2.
Example. From (8.48), we obtain
(8.50)
(8.50) is also written as
Y(z)=[1-(7/4)z
-1
-(1/2)z
-2
]W(z), (8.51)
where
). z ( X
z
8
1
z
4
1
1
z
2
1
z
4
7
1
Y(z)
2 1
2 1


+

(8.52) ). z ( X
z
8
1
z
4
1
1
1
W(z)
2 1
+

Thus,
z
-1
z
-1
X(z) Y(z)
Figure 8.3. Direct Form II.
+ -1/4
1/8
+
W(z)
-7/4
-1/2
+
+
W(z)=-(1/4)z
-1
W(z)+(1/8)z
-2
W(z)+X(z), (8.53)
Y(z)=W(z)-(7/4)z
-1
W(z)-(1/2)z
-2
W(z). (8.54)
The direct-form II structure of the system is obtained from (8.53) and
(8.54). Draw the block diagram.
The block diagram is shown in figure 8.3.
Example. (8.48) is also written as
.
z
4
1
1
z 2 1
z
2
1
1
z
4
1
1
H(z)
1
1
1
1

,
_

,
_

+
+

(8.55)
A cascade-form structure of the system is obtained from (8.55). The
system is expressed as the cascade of 2 subsystems. Each subsystem
is implemented in direct form I or II. Draw the block diagram.
Example. (8.48) is also written as
.
z
4
1
1
3 / 14
z
2
1
1
3 / 5
4 H(z)
1 1

+
+
(8.56)
A parallel-form structure of the system is obtained from (8.56). The
system is expressed as the parallel of 3 subsystems. Each subsystem
is implemented in direct form I or II. Draw the block diagram.