Lecture 10:

The inverse z-Transform

Instructor: Dr. Ghazi Al Sukkar

Dept. of Electrical Engineering
The University of Jordan
Email: [email protected]

Spring 2014 1
Outline

 Definition
 Inverse z-transform methods:
 Inspection
 Partial fraction expansion
 Power series expansion
 Properties of z-transform

Spring 2014 2
Definition
  Formal inverse z-transform is based on a Cauchy integral

Which is a complex contour integral with C being a closed contour

in the ROC of X(z), encompassing its poles.

 Less formal ways sufficient most of the time

 Inspection method
 Partial fraction expansion
 Power series expansion

3
Spring 2014
Inspection Method
 Becoming familiar with certain transform pair.
 Make use of known z-transform pairs such as
1
anun 
Z
 z  a
1  az1

n
 Example: The inverse
1 z-transform
1 of 1
X  z  z   x n     u n
1 1 2 2
1 z
2
n
1 1 1
X  z  z   x  n     u  n  1
1 2 2
1  z 1
2

4
Spring 2014
Partial Fraction method
 The most practical approach is to use the partial fraction
expansion method.
 It makes use of the z-transform table.
 The z-transform, however, must be a rational function. This
requirement is generally satisfied in digital signal processing.
 General Idea:
 When X(z) is a rational function of z-1, it can be expressed as a
sum of simple (first-order) factors using the partial fraction
expansion.
 The individual sequences corresponding to these factors can then
be written down using the z-transform table.
 Assume that a given z-transform can be expressed as
M

 k
b z k

X z   k 0
N

 k
a
k 0
z k

5
Spring 2014
Inverse Z-Transform by Partial Fraction Expansion
  An equivalent expression is: M
z N
b z k
M k

X  z  k 0
N
z M  a k z N k
k 0
 There will be zeros and poles at nonzero locations in the z-
plane.
 In addition, there will be either poles at if or zeros at if .
 could be expressed as:

Where s are the non-zero zeros of and s the non-zero poles of .

6
Spring 2014
Inverse Z-Transform by Partial Fraction Expansion
  Apply partial fractional expansion
M N N
Rk s Ri ,m
X  z    cr z  r

r 0
    k 1, k  i 1  p k z
     
1

  

m 1 1  p z
  i 
1 m
 
polynomial part proper rational part
 First term exist only if M>N
 is obtained by long division
 Second term represents all first order poles, is the kth pole, is
called the residue
 Third term represents an order s pole at , and is the residue
 There will be a similar term for every high-order pole
 Each term can be inverse transformed by inspection

7
Spring 2014
Partial Fractional Expression
   M N N
Rk s Ri ,m
X  z  c z r
  
r 0
r
k 1, k  i 1  p k z
1

m 1 1  p z
i
1
 m

 Coefficients are given as

 
Rk  1  p k z 1 X  z  z  pk

Ri ,m 
1
 s  m !  pi  sm
 d sm
 s 1 
 s  m 1  p i w  X  w   
 dw  w pi1
 The term correspond to shifted and scaled impulse sequence
i.e.,
 The term corresponds to or depending on the ROC.
 If ROC is then a given pole will correspond to:
 a right-sided exponential if
 a lift-sided exponential if

8
Spring 2014
Example: 2nd Order Z-Transform
1 1
X z   ROC : z 
 1 1  1 1  2
1  z 1  z 
 4  2 
 Order of nominator is smaller than denominator (in terms of z -1)
 No higher order pole

R1 R2
X  z  
 1 1   1 1 
1  z   1  z 
 4   2 

 1  1
R1  1  z 1  X  z    1
 4  1  11  1
z
4 1    
 24 
 
 1 1  1
R2  1  z  X  z   2
 2  1  11 1
z
2 1    
 42 
 
9
Spring 2014
Example Continued
1 2 1
X z    z 
 1 1   1 1  2
1  z   1  z 
 4   2 

 ROC extends to infinity

 Indicates right sided sequence

n n
1 1
xn  2  un -   un
2  4

10
Spring 2014
Example #2

X z  
1  2z  z 1

2
1z 1 2
 z 1
3 1 1

1  z 1  z  2 1  z 1  1  z 1
2 2 2

 
 Long division to obtain co
 1  5z 1
X z   2 
1  2 3 1
z  z 1 z
2
2
 2 z 1  1

 1 
1 1 
2
z  1  z 1  
 
2 2
z  2  3z 1  2
R1 R2
X  z  2  
5 z 1  1 1 1 1  z 1
1 z
2

 1 1 
R1  1  z  X  z   9 R2  1  z 1  X  z  8
 2  z
1 z 1
2

11
Spring 2014
Example #2 Continued
9 8
X z   2   z 1
1 1 1  z 1
1 z
2
 ROC extends to infinity
 Indicates right-sides sequence

n
1
x n  2  n  9  u n  8u n 
2

12
Spring 2014
Example
  Consider a sequence with z-transform

which is a proper rational function because the degree of the

numerator is less than the degree of the denominator.

Therefore:

which (by equating the coefficients of the same degree polynomial

terms) leads to

By solving these equations:

13
Spring 2014
Cont..
  
There are three choices of the ROC for :
Choice 1: ROC :
(anticausal)
Choice 2: ROC :
(causal)
Choice 3: ROC :
(anticausal)

14
Spring 2014
Matlab Implementation
 A Matlab function residuez is available to compute the residue
part and the direct (or polynomial) terms of a rational function
in z-1. Let 1 M
b0  b1 z    bM z B( z )
X ( z)  1 N

1  a1 z    a N z A( z )
M N
N
Rk
 1
  c k z k

k 1 1  p k z
k0  
M N
be a rational function in which the numerator and the denominator
polynomials are in ascending powers of z-1.
 Then [R,p,c] = residuez(b,a),
the returned column vector R contains the residues; Column vector
p contains the poles locations; Row vector c contains the direct
term.

15
Spring 2014
Cont..
 If p(k)=…=p(k+s-1) is a pole of multiplicity s, then the
expansion includes the term of the form

Rk Rk 1 Rk  s 1
1
 1 2

1  pk z (1  p k z ) (1  p k z 1 ) s
Similarly, [b,a]=residuez(R,p,c) can be used to convert the partial
fraction expansion back to polynomials with coefficients in row
vectors b and a.

16
Spring 2014
Inverse Z-Transform by Power Series Expansion
 The z-transform is power series

X  z   x  n  z n

n  
 In expanded form

X  z     x  2 z 2  x  1 z 1  x 0  x1 z 1  x 2 z 2  
 Z-transforms of this form can generally be inversed easily
 Especially useful for finite-length series
 Example (Finite length)

 1 1 
 
X  z   z  1  z  1  z 1 1  z  1
2

2
  1 n  2
 1
 
 2 n  1
1 1 
 z2  z  1  z 1 xn    1 n  0
2 2  1
 2 n1
1 1
xn  n  2  n  1  n  n  1  0
 n2
2 2
17
Spring 2014
Cont..
  Example (power series expressions):
Find the inverse z-transform of
Sol. Using Taylor series expansion for we obtain:

Therefore:

 Example (long division):

The sequence is right-sided.

18
Spring 2014
Cont..
   1  az 1  a 2 z 2  ......
1
1  az 1
1  az 1
az 1
az 1  a 2 z 2
a 2 z 2 ...

 Hence

19
Spring 2014
Properties of z-Transform
  Notation:
Given

And

20
Spring 2014
Z-Transform Properties: Linearity
  
 Linearity
ax1  n  bx 2  n 
Z
aX 1  z   bX 2  z  ROC  at least R x1  R x2
 Proof:

 Note that the ROC of the combined sequence may be larger than
either ROC
 This would happen if some pole/zero cancellation occurs
 Example:

xn  anun - anun - N

 Both sequences are right-sided
 Both sequences have a pole at
 Both have a ROC defined as
 In the combined sequence the pole at cancels with a zero at
 The combined ROC is the entire z plane except .

21
Spring 2014
Z-Transform Properties: Time Shifting

xn  no  
Z
 z no X z  ROC  R x
  Here no is an integer
 If positive the sequence is shifted right
 If negative the sequence is shifted left
 Proof:
, then
Let

 The ROC can change, the new term may

 Add or remove poles at or
 Example  
 1  1
X  z  z 
1
 z 
 1  1 z 1  4
 
 4 
n-1
1
xn    un - 1
 4
22
Spring 2014
Z-Transform Properties: Multiplication by Exponential

zno xn 
Z
 X z / z o  ROC  zo R x
  Proof:

 ROC is scaled by , i.e., if is the set of values of such that ,

then is the set values of such that

 All pole/zero locations are scaled

 If zo is a positive real number: z-plane shrinks or expands
 If zo is a complex number with unit magnitude i.e., it rotates
the locations of the poles and zeros by an angle of .

23
Spring 2014
Cont..
  Example: We know the z-transform pair
1
u n 
Z
ROC : z  1
1 - z -1
 Let’s find the z-transform of

x n  r n cos  o n  u n


1
2
  1
 
re jo u n  re  jo u n
n

2
n

1
 
re jo u n 
n Z 12
, z r
2 1  re jo z 1


1  j o n
re 
u n 
Z 12
 jo 1
, z r
2 1  re z

1/ 2 1/ 2
X  z   z r
1  re jo z 1 1  re  jo z 1

24
Spring 2014
Z-Transform Properties: Differentiation
dX z 
nxn  Z
 z ROC  R x
dz
  Proof:

 Example: We want the inverse z-transform of

Let’s differentiate to obtain rational expression


X z   log 1  az1  z  a

Making use ofdX  z  azproperties

z-transform

2
dX  z ROC
and
 z  az1
1
dz 1  az1 dz 1  az1

nxn  a  a un  1
n 1

25
Spring 2014
Cont..
Therefore
  
an
xn    1 un  1
n 1

 Example:
Let
Then

Hence,

26
Spring 2014
Z-Transform Properties: Conjugation

  
x* n 
Z
 
 X * z* ROC  R x

 Proof:
Let

27
Spring 2014
Z-Transform Properties: Time Reversal

  
 ROC is inverted, i.e., if is the set of such that , then the ROC
of is the set of such that , thus if in the ROC of , then is in
the ROC of
 If is real then

1
x  n 
 Example:
Z
 X1 / z  ROC 
Rx
It is a time reversed
xn  
version  
 anuof n
then
a un
n

1 - a-1z 1
X z    z  a1
1  az 1 - a-1z 1

28
Spring 2014
Z-Transform Properties: Convolution
x1 n  x2 n 
Z
 X1  z  X2  z  ROC : R x1  R x2
  Convolution in time domain is multiplication in z-domain
 Proof:
Let

29
Spring 2014
Cont..

 Example: Let’s calculate the convolution of

x1 n  anun and x2 n  un

1
X1  z   1
ROC : z  a X2  z  
1
ROC : z  1
1  az 1  z 1
 Multiplications of z-transforms is
1
Y  z   X1  z  X2  z  
1  az 1  z 
1 1

 ROC: if |a|<1 ROC is |z|>1 if |a|>1 ROC is |z|>|a|

 Partial fractional expansion of Y(z)
1  1 1 
Y  z   1
 1 
asume ROC : z  1
1  a 1  z 1  az 

yn 
1
1a

un  an1un 
30
Spring 2014
Z-Transform Properties: Initial value theorem
  If is causal, then

Proof:

31
Spring 2014