The z-Transform

© PFH Private Hochschule Göttingen 1

The z-transform
Definition

➢ The z-transform is a tool in describing and analyzing digital systems

➢ It offers the techniques for digital filter design and frequency
analysis of digital signals
➢ The z-transform of a causal sequence x(n) is given by:

X ( z ) or Z  x ( n ) : is defined as the z − transform of x ( n )


X ( z ) = Z  x ( n ) =  x ( n )  z − n = x ( 0 )  z −0 + x (1)  z −1 + x ( 3)  z −3 +
n =0

z is a complex variable

➢ Where a digital signal x(n) is called a causal sequence, if

x (n) = 0 ; n  0
© PFH Private Hochschule Göttingen 2
The z-transform
Definition

▪ First we establish the relationship between the s-plane in Laplace domain

and the z-plane in z-transform domain
▪ The sampled signal, which is not quantized, with a sampling period of Ts is
written as

xs ( t ) =  x ( nTs )   ( t − nTs )
n =0
xs ( t ) = x ( 0 )   ( t ) + x (Ts )   ( t − Ts ) + x ( 2Ts )   ( t − 2Ts ) +

© PFH Private Hochschule Göttingen 3

The z-transform
Definition

▪ Taking the Laplace transform and using the Laplace shift property as

L  ( t − nTs ) → e − nTs s

▪ This leads to:


X s ( s ) =  x ( nTs )  e − nTs s = x ( 0 )  e −0Ts s + x (Ts )  e −1Ts s + x ( 2Ts )  e −2Ts s +
n =0

X ( z ) = Z ( x ( n ) ) =  x ( n )  z − n = x ( 0 )  z −0 + x (Ts )  z −1 + x ( 2Ts )  z −2 +
n =0

▪ The relationship of the sampled system in Laplace domain and its digital
system in z-transform domain by the following mapping

z = e sTs
© PFH Private Hochschule Göttingen 4
The z-transform
Relationship between Laplace transform and z-transform

© PFH Private Hochschule Göttingen 5

z-Transform
Some common z-transform pairs a   (n)

n   (n)

n2   ( n )

an   ( n)

e − na   ( n )

n  an   ( n)

sin ( a  n )   ( n )

cos ( a  n )   ( n )

a n sin ( b  n )   ( n )

a n cos ( b  n )   ( n )

e − an sin ( b  n )   ( n )

e − an cos ( b  n )   ( n )

 (n)

© PFH Private Hochschule Göttingen 6

 z-Transform
Some common z-transform pairs

 (n)

− ( − n − 1)
a   (n)
−b   ( − n − 1)
 a n  sin ( 0 n )    ( n )

 a n  cos ( 0 n )    ( n )

n  an   ( n)
−n  b n   ( − n − 1)

© PFH Private Hochschule Göttingen 7

The z-transform
Sampled-data systems can be described in the frequency domain by using the bilateral z-
transform

X ( z ) Z  x ( n ) =  x (n)  z −n

n =−
The z-transform is a derivation from the Laplace transform
With the definition

z := e sTs
s is the Laplace variable
Ts is the sample time
It is possible to describe the sampled-data system in the frequency domain


X ( j) =  x (k )  e
n =−
− jTs n

© PFH Private Hochschule Göttingen 8

The z-transform
Sampled-data systems can be described in the frequency domain by using the bilateral z-
transform

X ( z ) Z  x ( n ) =  x (n)  z −n

n =−

The Fourier-transform is a special case of the z-transform. With the definition

z = e jTs
it is possible to describe the sampled-data system in the frequency domain

X ( j) =  x (k )  e
n =−
− jTs n

Important is the result, if only one time-delay exists



 x ( k − 1)  z
n =−
−n
= z −1  X ( z )
© PFH Private Hochschule Göttingen 9
The z-transform
Mapping between s-plane and z-plane

▪ Substituting
s = −  j
▪ into
z = e sTs


▪ it follows that

z = e(
−Ts  jTs )

▪ In the polar form, we have

z = e −Ts   Ts

© PFH Private Hochschule Göttingen 10

The z-transform
BIBO stability

The following facts apply to a stable system (bounded-input / bounded-output)

1. If the input to the system is bounded, then the output of the system will also
be bounded, or the impulse response of the system will go to zero in a finite
number of steps
2. An unstable system is one in which the output of the system will grow
without bound due to any bounded input, initial condition, or noise, or its
impulse response will grow without bound
3. The impulse response of a marginally stable system stays at a constant
level or oscillates between two finite values

© PFH Private Hochschule Göttingen 11

z-Transform
Important properties of the z-transform

▪ Linearity: Z k1  x1 ( n ) + k2  x2 ( n ) = k1  X 1 ( z ) + k2  X 2 ( z )

▪ Sample shifting: Z  x ( n − m ) = z − m  X ( z )

z
▪ Frequency shifting:  
Z an  x ( n) = X  
a

1
▪ Folding: Z  x ( −n ) = X  
z

Important is the result from the sample shifting, if only one time-delay exists

Z  x ( n − 1) = z −1  X ( z )
© PFH Private Hochschule Göttingen 12
z-Transform
Inversion of the z-transform

▪ The inverse z-transform computation requires an evaluation of a complex

contour integral that, in general, is a complicated procedure

x ( n ) Z −1  X ( z ) =
1
 X (z) z
n −1
 dz
2   j C

▪ The inverse z-transform may be obtained by at least three methods:

1. Partial fraction expansion and look-up table

2. Power series expansion
3. Residue method

© PFH Private Hochschule Göttingen 13

z-Transform
Inversion of the z-transform

▪ The most practical approach is to use the partial fraction expansion method
▪ The inverse z-transform can be summarized as follows
▪ Given:

b0 + b1  z −1 + b2  z −2 + + bm  z − M
X (z) = Rx−  z  Rx+
1 + a1  z −1 + a2  z −2 + + an  z − N

b0 + b1  z −1 + b2  z −2 + + bm  z −( N −1) M −N
X (z) =
1 + a1  z −1 + a2  z −2 + + an  z − N
+ C
k =0
k  z −k

proper rational part polynomial part if M  N

© PFH Private Hochschule Göttingen 14

z-Transform
Partial fraction expansion on the proper part

b0 + b1  z −1 + b2  z −2 + + bm  z −( N −1) M −N
X (z) =
1 + a1  z −1 + a2  z −2 + + an  z − N
+ C
k =0
k  z −k

proper rational part polynomial part if M  N

N M −N
Rk
X (z) =  −1
+  Ck  z − k
k =1 1 − pk  z k =0
M N

▪ pk is the k-th pole of X(z)

▪ Rk is the residue at pk

© PFH Private Hochschule Göttingen 15

z-Transform
Partial fraction expansion on the proper part

▪ Assumed we have simple real poles

b0 + b1  z −1 + b2  z −2 + + bm  z −( N −1)
Rk = −1 −2
1 + a1  z + a2  z + + an  z −N (
 1 − pk  z −1 ) z = pk

▪ Now we write x(n) as:

N
 1  M −N
x ( n ) =  Rk  Z 
−1
−1 
+  Ck   ( n − k )
k =1 1 − pk  z  k =0
M N

▪ Finally, use the relation from the z-transform table

 z   pkn   ( n )
−1
zk  Rx−
Z  = n
 z − pk  − pk   ( − n − 1) zk  Rx+
© PFH Private Hochschule Göttingen 16
z-transform
Inverse z-transform

➢ The general procedure for using the partial fraction expansion is as follows:

(1) Eliminate the negative powers of z for the z-transform function X(z)

(2) Determine the rational function X(z)/z (assuming it is proper), and apply
the partial fraction expansion to the determined rational function X(z)/z
using the formula in the given table

(3) Multiply the expanded function X(z)/z by z on both sides of the equation
to obtain X(z)

(4) Apply the inverse z-transform using reference tables

© PFH Private Hochschule Göttingen 17
z-transform
Partial fraction and formulas for constant

Partial fraction with the first-order real pole:

X (z)
R = ( z − p)
R
z− p z z= p

Partial fraction with the first-order complex poles:

A z A*  z X (z)
+ A = ( z − P)
z − P z − P* z z=P
*
P : complex conjugate of P
A* : complex conjugate of A
Partial fraction with mth order real pole:
Rm Rm −1 R1 1 d k −1  m X (z)
+ + + Rk =  k −1 ( z − p )  
( z − p ) ( z − p )2 ( z − p)
m
( k − 1) ! dz  z 
z= p

© PFH Private Hochschule Göttingen 18

z-Transform
System representation in the z-Domain

▪ Definition of the System Function (Transfer Function)

The system function is given by:

H ( z ) Z h ( n ) =  h (n)  z −n
; Rh−  z  Rh+
n =−

▪ Using the convolution property of the z-transform, we can calculate the output transform:

Y ( z ) = H ( z )  X ( z ) ; ROC y = ROCh  ROC x

▪ A linear, time-invariant system can be represented in the z-domain by

X (z) H (z) Y (z) = H (z) X (z)

© PFH Private Hochschule Göttingen 19

z-Transform
Difference equation representation

▪ The most important description form of linear, time discrete systems

is the difference equation:
N M
y ( n ) +  ak  y (n − k ) =  bl  x(n − l )
k =1 l =0

▪ The output value y(k) can be calculated by a linear combination of

the last in- and output values and the actual input value weighted by b0

▪ The time-invariant coefficients ak and bl characterize the system

▪ This is a recursive description form

N M
y ( n ) = − ak  y (n − k ) +  bl  x(n − l )
k =1 l =0

© PFH Private Hochschule Göttingen 20

z-Transform
Description form of linear, time invariant systems

▪ Difference equation
N M
y ( n ) +  ak  y (n − k ) =  bl  x(n − l )
k =1 l =0

▪ With the z- transform for both sides you get:

N M
Y ( z ) +  ak  z  Y ( z ) =  bl  z −l  X ( z )
−k

k =1 l =0

▪ From which the transfer function will be calculated

M

b  z b0  z − M  z M + + M 
−l b
Y ( z) l
B(z)  b0 
H ( z) = l =0
= =
X ( z) A( z ) z − N ( z N + + aN )
N
1 +  ak  z − k
k =1
© PFH Private Hochschule Göttingen 21
z-Transform
Description form of linear, time invariant systems

▪ After factorization, we get the result:

 (z − z )
M

H ( z ) = b0  z N − M l =1 l

 (z − z )
N
k =1 k

▪ The zl’s are the system zeros and pk’s are the system poles

▪ So H(z) can also be represented in the z-domain using a pole-zero plot

© PFH Private Hochschule Göttingen 22

z-Transform
Transfer function representation

▪ If the ROC of H(z) includes a unit circle (z=ejω), then H(z) can be evaluated on the unit
circle

 (e  − z )
M j

H ( e j ) = b0  e j ( N − M ) l =1 l

 (e  − p )
N j
k =1 k

▪ The factor (ejω −z) can be interpreted as a vector in

the complex z-plane from a zero zl to the unit circle
at z = ejω

▪ the factor (ejω − pk) can be interpreted as a vector

from a pole pk to the unit circle at z = ejω

© PFH Private Hochschule Göttingen 23

z-Transform
The Bode diagram of the transferfunction

▪ The magnitude response can be calculated by

e j − z1 e j − zM
H ( e j ) = b0 
e j − p1 e j − z N

▪ Or in dB
 e j − z1 e j − zM 
(
AdB ( ) := 20  log H ( e j
)) = 20  log  b0  j
 e − p1 e − z N 
j


▪ The phase response function can be calculated by

(
 H (e ) ) = 0 or 180 + ( N − M )    +   ( e − zk ) −   ( e j − pk )
M M
j j

1 1
constant linear
nonlinear
© PFH Private Hochschule Göttingen 24
z-Transform
Relations between the several representations

© PFH Private Hochschule Göttingen 25

z-Transform
Stability and causality

➢ The z-Domain stability theorem

▪ For discrete LTI-Systems, the BIBO stability is given by



 h(k )  
−

▪ This implies that H(ejω) exists, and therefore the time discrete-Fourier transform

▪ That means, the unit circle must be region of convergence of H(z)

z = 1 is ROC of H ( z )

▪ The dashed paths exist only if the system is stable

© PFH Private Hochschule Göttingen 26

z-Transform
Stability and causality

➢ z-Domain LTI stability theorem

A discrete LTI system is stable if and only if the unit circle is the ROC of H(z)

▪ For LTI causality it‘s required that

h (n) = 0 for n  0 (right − sided sequence)

▪ This implies

ROC of H ( z )  Rh−
➢ z-Domain causal LTI stability theorem

A causal discrete LIT system is stable if and only if the system function H(z) has all
poles and zeroes inside the unit circle
© PFH Private Hochschule Göttingen 27
z-Transform
Some hints

▪ The complex variable z is called the complex frequency given by z = |z|ejω, where |z| is
the magnitude and ω is the real frequency.
▪ Since the ROC is defined in terms of the magnitude |z|, the shape of the ROC is an
open ring.
▪ Note that Rx− may be equal to zero and/or Rx+ could possibly be ∞
If Rx+ < Rx−, then the ROC is a null space and the z-transform does not exist.
▪ The function |z| = 1 (or z = ejω) is a circle of unit radius in the z-plane and is called the
unit circle. If the ROC contains the unit circle, then we can evaluate X(z) on the unit
circle.

( ) =  x (n)  e
X ( z ) z = e j = X e j − j
= F  x ( n ) 
n =−

▪ Therefore the discrete-time Fourier transform X(ejω) may be viewed as a special case
of the z-transform X(z)
© PFH Private Hochschule Göttingen 28
z-Transform
Difference equation representation

▪ The most important description form of linear, time discrete systems

is the difference equation:
N M N M
y ( n ) +  ai  y (n − k ) =  bi  u (n − l ) y ( n ) = − ai  y (n − k ) +  bi  u (n − l )
k =1 l =0 k =1 l =0

▪ The output value y(k) can be calculated by a linear combination of

the last in- and output values and the actual input value weighted by b0

▪ The time-invariant coefficients ai and bi characterize the system

▪ This is a recursive description form

▪ If all ai = 0 (i = 1…N) , we have a non-recursive description form, then

▪ The system order is calculated with: max{M,N}

© PFH Private Hochschule Göttingen 29
z-Transform
▪ The z-transform H(z) of the impulse response h(k) is the transfer function

h(k ) − • H ( z)
▪ Based on the convolution theorem:
 Y (z)
H (z) =  h ( n )  z −n H (z) =
n =− U (z)

▪ This is the fundamental relationship between the input- and output signal of a linear,
time discrete system

▪ The mathematical description of a linear, time discrete system can be given by the
impulse response h(k) or the transfer function H(z) . Both are equivalent!

▪ Note: For many application we can define y(n) = u(n) = 0 for n < 0

© PFH Private Hochschule Göttingen 30

The z-transform
Sampled-data systems can be described in the frequency domain by using the bilateral z-
transform

X ( z ) Z  x ( n ) =  x (n)  z −n

n =−

The Fourier-transform is a special case of the z-transform. With the definition

z = e jTs
it is possible to describe the sampled-data system in the frequency domain

X ( j) =  x (k )  e
n =−
− jTs n

Important is the result, if only one time-delay exists



 x ( k − 1)  z
n =−
−n
= z −1  X ( z )
© PFH Private Hochschule Göttingen 31
z-Transform
Region of convergence (ROC)

▪ All the values of z that make the summation to exist form a region of convergence in the
z-transform domain
▪ The set of z values for which X(z) exists is called the region of convergence (ROC) and
is given by

Rx−  z  Rx+

The inverse z-transform of a complex function X(z) can be calculated by

x ( n ) Z −1  X ( z ) =
1
 X (z) z
n −1
 dz
2   j C

where C is a counterclockwise contour encircling the origin and lying in the ROC

© PFH Private Hochschule Göttingen 32

z-Transform
Region of convergence (ROC)

The set of z values for which X(z) exists is

called the region of convergence (ROC)

Rx−  z  Rx+

© PFH Private Hochschule Göttingen 33

z-Transform
The region of convergence (ROC) example

© PFH Private Hochschule Göttingen 34

z-Transform
Example: find the inverse z-transform

z
Given X (z) =
3 z − 4z +1
2

z z −1
X (z) = =
 2 4 1  4 1 
3   z − z +  3  1 − z −1 + z −2 
 3 3  3 3 

1 −1  
z 1 1
X (z) = 3 = 2 − 2 = 1 1 − 1 1 
−1  1 −1  1 − z −1 1 −1 2  1 − z −1  2  1 −1 

( )
1 − z 1 − z 
 3 
1− z
3
1− z
 3 

© PFH Private Hochschule Göttingen 35

z-Tramform
Example: The region of convergence (ROC)

ROC1 : 1  z   1 1
ROC2 : 0  z  ROC3 :  z  1
z1  Rx− = 1 3 3
z2  1
1 z1  Rx+ = 1
z1  Rx+ =
3 1
1 z2 
1 11
n z2  3
x1 ( n ) =  ( n ) −    ( n ) 3 1  1 
n

x3 ( n ) = −  ( −n − 1) −    ( n ) 
1
2 23 1   1 
n

x2 ( n ) = − ( −n − 1) − −    ( − n − 1) 
1 2  3 
2 
2 2   3  

© PFH Private Hochschule Göttingen 36

z-Transform
Poperties of the region of convergence (ROC)

▪ The ROC is always bounded by a circle since the convergence condition is on the
magnitude |z|
▪ The sequence x1(n) = anσ(n) is a special case of a right-sided sequence, defined as a
sequence x(n) that is zero for some n < n0
▪ The ROC for right-sided sequences is always outside of a circle of radius Rx−
▪ If n0 ≥ 0, then the right-sided sequence is also called a causal sequence
▪ The sequence x2(n) = −bn σ(−n−1) is a special case of a left-sided sequence, defined
as a sequence x(n) that is zero for some n > n0
▪ If n0 ≤ 0, the resulting sequence is called an anti-causal sequence
▪ The ROC for left-sided sequences is always inside of a circle of radius Rx+

© PFH Private Hochschule Göttingen 37

z-transform
ROC

• The sequence x3(n) is a two-sided sequence

• The ROC for two-sided sequences is always an open ring Rx− < |z| < Rx+, if it exists
• The sequences that are zero for n < n1 and n > n2 are called finite-duration sequences
The ROC for such sequences is the entire z-plane
If n1 < 0, then z = ∞ is not in the ROC
If n2 > 0, then z = 0 is not in the ROC
• The ROC cannot include a pole since X(z) converges uniformly in there
• There is at least one pole on the boundary of a ROC of a rational X(z)
▪ The ROC is one contiguous region; that is, the ROC does not come in pieces
▪ In digital signal processing, signals are assumed to be causal since almost every digital
data is acquired in real time.
▪ Therefore the only ROC of interest to us is given in the sequence x1(n) = an σ(n)

© PFH Private Hochschule Göttingen 38

z-Transform
Important properties of the z-transform

▪ Linearity: Z k1  x1 ( n ) + k2  x2 ( n ) = k1  X 1 ( z ) + k2  X 2 ( z ) ROC : ROCx1  ROC x2

▪ Sample shifting: Z  x ( n − m ) = z − m  X ( z ) ROC : ROC x

z
▪ Frequency shifting:  
Z an  x ( n) = X  
a
ROC : ROC x scaled by a

1
▪ Folding: Z  x ( −n ) = X   ROC : inverted ROCx
z

▪ Complex conjugation:  
Z x* ( n ) = X * z* ( ) ROC : ROC x

© PFH Private Hochschule Göttingen 39

z-Transform
Important properties of the z-transform

dX ( z )
▪ Differentiation: Z n  x ( n ) = − z  ROC : ROC x
dz

▪ Multiplication: Z  x1 ( n )  x2 ( n ) =
1
2j  C v( )
 X 1 ( v ) X 2 z  v −1  dv ROC : ROC x1  invertetd ROC x2

▪ Convolution: Z  x1 ( n )  x2 ( n ) = X 1 ( z )  X 2 ( z ) ROC : ROCx1  ROC x2

© PFH Private Hochschule Göttingen 40