II The Z Transform

Topics to be covered
1. Introduction

2. The Z transform

3. Z transforms of elementary functions

4. Properties and Theory of z transform

5. The inverse z transform

6. Z transform for solving difference equations

II.1 Introduction
The role of the z transform in discrete-time control system is similar to that of the Laplace
transform in continuous-time system.
Discrete time signals: The sampled signal is x0 , xT , x2T  where T is the sampling period.
The signal can be written as x0, x1, x2 xk  it can be considered as a sampled signal of
xt  where T is 1 sec. xkT  and xk  are interchangeable if it doesn’t make confusion.

II.2 The Z transform

In considering the z transform of a time function xt  , we consider the sampled
value of xt  , that is x0, xT , x2T  The z transform of a time function , where t is
nonnegative, is defined as follows

1
 X ( z)   xt  xkT   xkT z


k 0
k
2.1


xkT   X ( Z )
Definition: Z-transform of a general discrete-time signal xk  is defined as

X ( Z )   xk z k 2.2
k 0


xk   X ( Z )

Note: eq. 2.1 and 2.2 is referred to as the one-sided z transform or unilateral z transform. Z is a
complex variable.

If    t   , or k  0,1,2 , then the z transform will be defined as

X (z)   xt  xkT   xkT z



k 
k
2.3

xkT   X ( Z )
Definition: Z-transform of a general discrete-time signal xk  is defined as

X (Z )   xk z
k  
k
2.4

xk   X ( Z )
Note: eq. 2.3 and 2.4 is referred to as the two-sided z transform or bilateral z transform. Z is a
complex variable. We are only focused on one-sided z transform in this course.

2
 II.3 Z Transforms of elementary
functions
Unit step function

1 0  t
xt   
0 t  0
X ( z)   xt  xkT   xkT z


k 0
k

  z k 
k 0
1
1 z 1

z
z 1
(ROC: z  1 )

Unit ramp function

t 0  t
x t   
0 t  0

X ( z)   xt  xkT   kTz



k 0
k

 T  kz k  T
k 0
z 1
1  z 
1 2

Tz
z  12
k
Polynomial function a
a k k  0,1,2 
x k   
0 k0

X ( Z )   x k z k
k 0

  a k z k
k 0
ROC: z  a

  ( az ) 1 k

k 0

1 z
 1

1  az za

3
Exponential function
e  at 0t
x t   
 0 t0

X ( Z )   x kT z k
k 0

  e akT z k
k 0

  ( e aT z 1 ) k
k 0

1 z
  aT 1

1 e z z  e aT

Example 2.1 Obtain the z transform of

e  at sin wt 0  t
x t   
 0 t0

e at sin wt  e at e  e  jwt  

1 jwt
2j
e
1 a  jw t
2j
 e a  jw t 

z   z  z
  a  jw t   a  jw t
e at   aT
 e    a  jw T , e    a  jw T
ze ze ze
since  1
 z z  e aT z 1 sin wT
 e at sin wt     
2 j  z  e a  jw T z  e a  jw T  1  2e aT z 1 cos wT  e 2 aT z 2

Example 2.2 Obtain the z transform of

x s  
1
s  a 2

x s    x t   te at
1
s  a 2

4
 
X ( Z )   x kT z  k
k 0

  kTe  akT z  k
k 0

 T  k ( e  aT z 1 ) k
k 0

Te  aT z 1

1  e  aT
z 1 
2

Table of z transforms 1

5
Table of z transforms 2

6
II.4 Properties and Theory of z transform
i) Linearity

x1 k   X 1 ( z ), ROC  R1 
  ax1 k   bx 2 k   aX 1 ( z )  bX 2 ( z ), ROC  R1  R2
x 2 k   X 2 ( z ), ROC  R2 

ii) Scaling in the z-Domain

xk   X ( z ), ROC  R  a k xk  X (a 1 z ), ROC  a R

Z

Example 2.3 Determine the z transform and the associated region of convergence for following
function of time: u k  and a k u k 

Since u k   , we have a k u k  
1 1 1

1  a 1 z 
1 
1 z 1
1  az 1

iii) Time Shifting

x t  nT   z  n X ( z ),
 n 1

x t  nT   z n  X ( z )   x kT z k 
 k 0 
for number sequence k. we have

x k  n   z  n X ( z ),
 n 1

x k  n   z n  X ( z )   x k z k 
 k 0 

Example 2.4. Determine the z transform for u k   u k  1

z 1
u k   u k  1 
1
 1
1  z 1 1  z 1

Verify:  k   uk   uk  1  1

iv) Complex translation theorem

x t   X ( z )  e  at x t   X ( ze aT )

7
 v) Differentiation in the z-Domain

x k   X ( z ), ROC  R  kx k    z
Z dX ( z )
, ROC  R
dz

vi) The initial value theorem

If x k   0, k  0 , then x 0  lim X z 

z 

k
1
Example2.5 Determine the z transform for following function of time: x k   k   u k 
2

k
1
x1 k     u k   X 1 z  
1 1
,z 
2 1
1  z 1 2
2

1 1
z
x k   kx1 k  Thus X  z    z X 1  z   
d 2 1
2
,z 
dz  1 1  2
1  z 
 2 
verify the initial value theorem
1 1
z
x 0  lim X z   lim  2
2
0
z  z 
 1 1 
1  z 
 2 

vii) Final Value theorem

If x k   0, k  0 , then lim x k   lim1  z 1 X z 

k  z 1

z
Example2.6 Determine the final value of by using the final value theorem.
z  e aT

lim x k   lim1  z 1 X z   lim1  z 1 

z
0
k  z 1 z 1 z  e aT

8
 e  at 0t
x t    verify that lim x k   0
 0 t0 k 

Property table:

9
 II.5 The inverse z transform
Note: the inverse z transform yields the corresponding time sequence x k  , but doesn’t yield a
unique x t 

If the z transform is given as a ratio of two polynomials in z, then the inverse z transform may be
obtained by several different methods, such as direct division method, the computational method,
the partial-fraction-expansion method, and the inversion integral method.

1) direct division method


X ( Z )   x kT z k  x 0   x T z 1  x 2T z 2   x kT z k   or
k 0

X ( Z )   x k z k  x 0  x 1z 1  x 2 z 2   x k z k  
k 0

z 1
Example2.7 Determine the inverse transform of
1  2 z 1 2

By observation, we can see xk   k 2 k 1

10
 2) The computational method

a) MATLAB approach
z 1
Example 2.8 find the inverse transform of
1  2 z  1 2

z 1 z 1
Let G ( z )  , X ( z)  1 , Y ( z)  input X ( z )  1 is the z transform of
1  2 z 1 2
1  2 z 1 2

the Kronecker delta input. In MATLAB, the Kronecker delta input is given by
x  1 zeros 1, N  , where N corresponds to the end of the discrete time duration of the
process considered.

% example 2.8
% enter the numerator and denominator
num=[0 1 0];
den=[1 -4 4];
% enter the Kronexker delta input and filter
x=[1 zeros(1,40)];
y=filter(num,den,x);
% plot the results
plot(k,y,’o’);
grid
title(‘Response to Kronecker Delta input’);
xlabel(‘k’);
ylabel(‘y(k)’);

b) Difference equation approach

Y z  z 1
G( z)    
 1  4 z 1  4 z  2 Y  z   z 1 X  z 

X z  1  2 z 1 2

 y k   4 y k  1  4 y k  2   xk  1
For above difference equation, x(0)  1, x(k )  0, k  0 y (k )  0, k  0
Thus y (0)  0
y k   4 y k  1  4 y k  2   xk  1
y 1  4 y 0   4 y  1  x0   1
y 2   4 y 1  4 y 0   x1  4
y 3  4 y 2   4 y 1  4 * 4  4  12
y 4   4 y 3  4 y 2   4 * 12  4 * 4  32


This iterative process can be easily programmed.

11
 3) The partial-fraction-expansion method

b0 z m  b1 z m 1    bm 1 z  bm b0 z m  b1 z m 1    bm 1 z  bm
X ( z)  
z n  a1 z n 1   a n 1 z  a n z  p1 z  p 2  z  p n 
Case 1 all pole are distinguished:

a1 a2 an  X z 
, where a i   z  p i 
X ( z)
  
z z  p1  z  p 2  z  p n   z  z  pi

X ( z) c1 c2
Case 2 double pole  
z z  p1  z  p1 
2

 2 X z  d  2 X z  
Then c1   z  p1   , and c 2    z  p1  
 z  z  p1  dz  z   z  p1

1
Example 2.9 find the inverse transform of
z  1z  2

z 1 z 1
X z  
1 1 1
   
z  1z  2   
z  2 z  1 1  2 z 1 1  z 1 
xk   2 k 1  u k  1, k  1,2,3, 

4) The inversion integral method.

xkT   xk  
1
2j C
X ( z ) z k 1 dz

Where C is a circle with its center at the origin of the z plane such that all poles of X ( z ) z k 1 are
inside it.
Using complex variable theory, we have
xkT   xk   K 1  K 2   K m
m
  residue of X  z z k 1 at pole z  z i of X  z z k 1
i 1

Case 1: X ( z ) z k 1 contains simple pole at z  z i , K  lim z  z i X  z z k 1

z  zi
k 1
Case 2 X ( z ) z contains a multiple pole of z  z i of order q .
d q 1
lim q 1 z  z i  X  z z k 1
1
K
q

q  1! i dz
z  z

12
 1
Example 2.10 Find the inverse transform of
z  1z  2

z k 1
Note that X  z z k 1 
z  1z  2
For k=0, X ( z ) z k 1 has three poles at 0,1,2. And for k=1,2,3,…, X ( z ) z k 1 has two poles 1,2.

a) k=0. xk   K 1  K 2  K 3

K 1  residue at simple pole z  0

 lim z  0 X z z 0 1  lim
1 1 1

z 0 z 0 z 1z 2 2

K 2  residue at simple pole z  1

 lim z  1X z z 0 1  lim

1 1
 1
z 1 z 1 z z 2

K 3  residue at simple pole z  2

 lim z  2X z z 0 1  lim

1 1 1

z 2 z z 1
z 2 2
Thus for k=0, xk   K 1  K 2  K 3   1   0
1 1
2 2
b) k=1,2,3…. xk   K 1  K 2

K 1  residue at simple pole z  1

 lim z  1X z z k 1  lim

1
1k 1  1
z 1 z 1 z 2

K 2  residue at simple pole z  2

 lim z  2X z z k 1  lim

1
z k  1  2k  1
z 2 z 2 z 1

 k  0
x k   
0
2
k 1
 u k  1 k  1,2,3,

13
 Example 2.11 Find the inverse transform of

z 1 1  az 1 
1  az 
1 3

X  z z k 1


z 1 1  az 1 z k 1

z k z  a 
1  az 
1 3
z  a 3

For k=0,1,2,… X ( z ) z k 1 has multiple pole a of order 3.

d q 1
lim q 1 z  z i  X  z z k 1
1
K
q

q  1! dz
z  z i

d 31 3 z z  a 
k
lim 31 z  a 
1

3  1! z a dz z  a 3
d2
lim 2 z k  z  a 
1

2 z a dz
 k 2 a k 1

II.6 Z transform for solving difference

equations
Note: difference equations can be solved using digital computer. However, closed form
expressions cannot be obtained from the computer solution. Maple can do some simple ones.

Table: z transform of xk  m  and xk  m 

14
Example 2.12. For following difference equation and associated input and initial
conditions, determine the zero-input and zero-state responses by using the z transform.

y k   3 y k  1  xk ,
k
1
xk     u k 
2
y  1  1
Taking the unilateral z transform of both sides of the given difference equation, we get

Y ( z )  3z 1Y ( z )  3 y[1]  X ( z )
Setting X ( z )  0 , we get
3
Y ( z) 
1  3 z 1

The inverse unilateral z transform gives the zero input response

y zi k   3 3 u k 
k

k
1
Since xk     u k  , we have
2

1 1
X ( z)  , z 
1 2
1  z 1
2

Setting y[1]  0 , we get

   
   1/ 7 
1  1  1 
 6 / 7 
Y ( z )  3 z 1Y ( z )   Y ( z)   1   
1 1  1  3 z  1 1   1  3z 1   1 1 
1 z 1 z  1 z 
2  2   2 

k
1 1
Zero state response: y zs k    3 u k     u k 
6 k

7 7 2

Ex. B-2-2,B-2-5,B-2-8, B-2-13, B-2-17

15