Signals and Systems Chapter 4 – The z-Transform and Discrete-Time LTI Systems

CHAPTER FOUR
THE z-TRANSFORM AND DISCRETE-TIME LTI SYSTEMS

4.1 Introduction

In Chapter 3 we introduced the Laplace transform. In this chapter we present the

z-transform, which is the discrete-time counterpart of the Laplace transform. The
z-transform is introduced to represent discrete-time signals (or sequences) in the z-domain
(z is a complex variable), and the concept of the system function for a discrete-time LTI
system will be described. The Laplace transform converts integrodifferential equations into
algebraic equations. In a similar manner, the z-transform converts difference equations into
algebraic equations, thereby simplifying the analysis of discrete-time systems.

The properties of the z-transform closely parallel those of the Laplace transform. However,
we will see some important distinctions between the z-transform and the Laplace transform.

4.2 The z-Transform

In Sec. 2.8 we saw that for a discrete-time LTI system with impulse response h[n], the
output y[n] of the system to the complex exponential input of the form zn is

 
yn   z n  H z z n 4.1


where H z    hnz
n  
n
4.2

4.2.1 Definition
The function H(z) in Eq. 4.2 is referred to as the z-transform of h[n]. For a general discrete-
time signal x[n], the z-transform X(z) is defined as


X z    xnz
n  
n
4.3

The variable z is generally complex-valued and is expressed in polar form as

z  re j 4.4

where r is the magnitude of z and Ω is the angle of z. The z-transform defined in Eq. 4.3 is
often called the bilateral (or two-sided) z-transform in contrast to the unilateral (or one-
sided) z-transform, which is defined as

University of Mines and Technology, Tarkwa 49 Solomon Nunoo MPhil, BSc, MIET, MASEE, MIAENG
Signals and Systems Chapter 4 – The z-Transform and Discrete-Time LTI Systems


X1 z    xnz  n 4.5
n 0

Clearly the bilateral and unilateral z-transforms are equivalent only if x[n] = 0 for n < 0.
The unilateral z-transform is discussed in Sec. 4.8. We will omit the word "bilateral" except
where it is needed to avoid ambiguity.

As in the case of the Laplace transform, Eq. 4.3 is sometimes considered an operator that
transforms a sequence x[n] into a function X(z), symbolically represented by

X z   xn 4.6

The x[n] and X(z) are said to form a z-transform pair denoted as

xn  X z  4.7

4.2.2 The Region of Convergence:

As in the case of the Laplace transform, the range of values of the complex variable z for
which the z-transform converges is called the region of convergence. To illustrate the
z-transform and the associated ROC let us consider some examples.

Example 4.1 Find the z-transform of the sequence in Eq. 4.8.

xn  anun a real 4.8

Solution 4.1 By Eq. 4.3 the z-transform of x[n] is

 
 
X z    anunz n   az 1
n

n   n 0

For the convergence of X(z) we require that


1 n
 az
n 0


Thus, the ROC is the range of values of z for which |az–1| < 1 or, equivalently, |z| > |a|.
Then

 

1
X z    az 1
n
 |z| > |a| 4.9
n 0 1  az 1

University of Mines and Technology, Tarkwa 50 Solomon Nunoo MPhil, BSc, MIET, MASEE, MIAENG
Signals and Systems Chapter 4 – The z-Transform and Discrete-Time LTI Systems

Alternatively, by multiplying the numerator and denominator of Eq. 4.9 by z, we may write
X(z) as

z
X z   |z| > |a| 4.10
za

Both forms of X(z) in Eqs. 4.9 and 4.10 are useful depending upon the application. From
Eq. 4.10 we see that X(z) is a rational function of z. Consequently, just as with rational
Laplace transforms, it can be characterized by its zeros (the roots of the numerator
polynomial) and its poles (the roots of the denominator polynomial). From Eq. 4.10 we see
that there is one zero at z = 0 and one pole at z = a. The ROC and the pole-zero plot for
this example are shown in Fig. 4.1. In z-transform applications, the complex plane is
commonly referred to as the z-plane.

Figure 4.1 ROC of the form |z| > |a|

University of Mines and Technology, Tarkwa 51 Solomon Nunoo MPhil, BSc, MIET, MASEE, MIAENG
Signals and Systems Chapter 4 – The z-Transform and Discrete-Time LTI Systems

4.2.3 Properties of the ROC

As we saw in Examples 4.1, the ROC of X(z) depends on the nature of x[n]. The properties
of the ROC are summarized below. We assume that X(z) is a rational function of z.

Property 1: The ROC does not contain any poles.

Property 2: If x[n] is a finite sequence (that is, x[n] = 0 except in a finite interval
N1 ≤ n ≤ N2, where N1 and N2 are finite) and X(z) converges for some value of z, then the
ROC is the entire z-plane except possibly z = 0 or z = ∞.

Property 3: If x[n] is a right-sided sequence (that is, x[n] = 0 for n < N1 < ∞) and X(z)
converges for some value of z, then the ROC is of the form

z  rmax or   z  rmax

where rmax equals the largest magnitude of any of the poles of X(z). Thus, the ROC is the
exterior of the circle |z|= rmin in the z-plane with the possible exception of z = 0.

Property 4: If x[n] is a left-sided sequence (that is, x[n] = 0 for n > N2 > –∞) and X(z)
converges for some value of z, then the ROC is of the form

z  rmin or 0  z  rmin

where rmin is the smallest magnitude of any of the poles of X(z). Thus, the ROC is the
interior of the circle |z|= rmin in the z-plane with the possible exception of z = 0.

Property 5: If x[n] is a two-sided sequence (that is, x[n] is an infinite-duration sequence that
is neither right-sided nor left-sided) and X(z) converges for some value of z, then the ROC is
of the form

r1  z  r2

where r1 and r2 are the magnitudes of the two poles of X(z). Thus, the ROC is an annular
ring in the z-plane between the circles |z|= r1 and |z| = r2 not containing any poles.

Note that Property 1 follows immediately from the definition of poles; that is, X(z) is infinite
at a pole.

University of Mines and Technology, Tarkwa 52 Solomon Nunoo MPhil, BSc, MIET, MASEE, MIAENG
Signals and Systems Chapter 4 – The z-Transform and Discrete-Time LTI Systems

4.3 z-Transforms of Some Common Sequences

4.3.1 Unit Impulse Sequence δ[n]

From definition (1.45) and (4.3)


X z     nz
n  
n
 z 0  1 all z 4.14

Thus,

 n  1 all z 4.15

4.3.2 Unit Step Sequence u[n]

Setting a = 1 in Eqs. (4.8) to (4.10), we obtain

1 z
un  1
 |z| > 1 4.16
1 z z 1

4.3.3 z-Transform Pairs

The z-transforms of some common sequences are tabulated in Appendix 3.

4.4 Properties of the z-Transform

Basic properties of the z-transform are presented in the following discussion.

4.4.1 Linearity
If

x1 n  X1 z  ROC = R1

x2 n  X2 z  ROC = R2

then

a1 x1 n  a2 x2 n  a1 X1 z   a2 X2 z  R  R1  R2 4.17

where a1 and a2 are arbitrary constants.

4.4.2 Time Shifting

If
xn  X z  ROC = R

University of Mines and Technology, Tarkwa 53 Solomon Nunoo MPhil, BSc, MIET, MASEE, MIAENG
Signals and Systems Chapter 4 – The z-Transform and Discrete-Time LTI Systems

then

xn  n0   z n0 X z  R  R  0  z   4.18

Special Cases

xn  1  z 1 X z  R  R  0  z  4.19

xn  1  zX z  R  R  z   4.20

Because of these relationships (Eqs. 4.19 and 4.20), z–1 is often called the unit-delay
operator and z is called the unit-advance operator. Note that in the Laplace transform the
operators s–1 = 1/s and s correspond to time-domain integration and differentiation,
respectively (Eqs. 3.22 and 3.20).

4.4.3 Multiplication by z0n

If

xn  X z  ROC = R

then

 z 
z0 xn  X  
n
R  z0 R 4.21
 z0 

In particular, a pole (or zero) at z = zk in X(z) moves to z = z0zk after multiplication by z0n
and the ROC expands or contracts by the factor |z0|.

Special Case


e j0 n xn  X e  j0 z  R  R 4.22

In this special case, all poles and zeros are simply rotated by the angle Ω0 and the ROC is
unchanged.

4.4.4 Time Reversal

If

xn  X z  ROC = R

then

University of Mines and Technology, Tarkwa 54 Solomon Nunoo MPhil, BSc, MIET, MASEE, MIAENG
Signals and Systems Chapter 4 – The z-Transform and Discrete-Time LTI Systems

1 1
x n  X   R  4.23
z R

Therefore, a pole (or zero) in X(z) at z = zk moves to 1/zk after time reversal. The
relationship R' = 1/R indicates the inversion of R, reflecting the fact that a right-sided
sequence becomes left-sided if time-reversed, and vice versa.

4.4.5 Multiplication by n (or Differentiation in z)

If

xn  X z  ROC = R

then

dXz 
nxn   z R  R 4.24
dz

4.4.6 Accumulation
If

xn  X z  ROC = R

then

n
1 z
 xk   1  z 1
X z   X z  R  R  z  1 4.25
k   z 1

n
Note that  xk  is the discrete-time counterpart to integration in the time domain and is
k  

called the accumulation. The comparable Laplace transform operator for integration is 1/s.

4.4.7 Convolution
If

x1 n  X1 z  ROC = R1

x2 n  X2 z  ROC = R2

then

x1 n x2 n  X1 z   X2 z  R  R1  R2 4.26

University of Mines and Technology, Tarkwa 55 Solomon Nunoo MPhil, BSc, MIET, MASEE, MIAENG
Signals and Systems Chapter 4 – The z-Transform and Discrete-Time LTI Systems

This relationship plays a central role in the analysis and design of discrete-time LTI systems,
in analogy with the continuous-time case.

4.4.8 Summary of Some z-Transform Properties

For convenient reference, the properties of the z-transform presented above are
summarized in Appendix 4.

4.5 The Inverse z-Transform

Inversion of the z-transform to find the sequence x[n] from its z-transform X(z) is called the
inverse z-transform, symbolically denoted as

xn  1X z  4.27

4.5.1 Inversion Formula

As in the case of the Laplace transform, there is a formal expression for the inverse
z-transform in terms of an integration in the z-plane; that is,

1
xn   X z z n1dz 4.28
2j C

where C is a counterclockwise contour of integration enclosing the origin. Formal

evaluation of Eq. 4.28 requires an understanding of complex variable theory.

4.5.2 Use of Tables of z-Transform Pairs

In the second method for the inversion of X(z), we attempt to express X(z) as a sum

X z   X1 z   ...  Xn z  4.29

where X1(z), ..., Xn(z) are functions with known inverse transforms x1[n], ..., xn[n]. From the
linearity property 4.17 it follows that

xn  x1 n  ...  xn n 4.30

4.5.3 Power Series Expansion

The defining expression for the z-transform [Eq. (4.3)] is a power series where the sequence
values x[n] are the coefficients of z−n. Thus, if X(z) is given as a power series in the form

X z    xnz n

n  

 ...  x 2z 2  x 1z  x0  x1z 1  x2z 2  ... 4.31

University of Mines and Technology, Tarkwa 56 Solomon Nunoo MPhil, BSc, MIET, MASEE, MIAENG
Signals and Systems Chapter 4 – The z-Transform and Discrete-Time LTI Systems

we can determine any particular value of the sequence by finding the coefficient of the
appropriate power of z−1. This approach may not provide a closed-form solution but is very
useful for a finite-length sequence where X(z) may have no simpler form than a polynomial
in z−1. For rational z-transforms, a power series expansion can be obtained by long division.

4.5.4 Partial-Fraction Expansion

As in the case of the inverse Laplace transform, the partial-fraction expansion method
provides the most generally useful inverse z-transform, especially when X(z) is a rational
function of z. Let

N z  z  z1 ...z  zm 
X z   k 4.32
Dz  z  p1 ...z  pm 

Assuming n ≥ m and all poles pk are simple, then

X z  c0 c1 c2 cn c n
c
    ...   0  k 4.33
z z z  p1 z  p2 z  pn z k 1 z  pk

where

X z 
c0  X z  z  0 ck  z  pk  4.34
z z  pk

Hence, we obtain

z z n
z
X z   c0  c1  ...  cn  c 0   ck 4.35
z  p1 z  pn k 1 z  pk

Inferring the ROC for each term in Eq. (4.35) from the overall ROC of X(z) and using Table
4.1, we can then invert each term, producing thereby the overall inverse z-transform.

If m > n in Eq. 4.32, then a polynomial of z must be added to the right-hand side of Eq.
4.35, the order of which is (m − n). Thus for m > n, the complete partial-fraction expansion
would have the form

mn n
z
X z    bq z q   ck 4.36
q0 k 1 z  pk

If X(z) has multiple-order poles, say pi is the multiple pole with multiplicity r, then the
expansion of X(z)/z will consist of terms of the form

University of Mines and Technology, Tarkwa 57 Solomon Nunoo MPhil, BSc, MIET, MASEE, MIAENG
Signals and Systems Chapter 4 – The z-Transform and Discrete-Time LTI Systems

1 2 r
  ...  4.37
z  pi z  pi 
2
z  pi r

where

r  k 
1 dk 
k 
z  pi r X z  4.38
k! dz  z  z  pi

4.6 The System Function of Discrete-Time LTI Systems

4.6.1 The System Function

In Sec. 2.6 we showed that the output y[n] of a discrete-time LTI system equals the
convolution of the input x[n] with the impulse response h[n]; that is Eq. 2.351,

yn  xn  hn 4.39

Applying the convolution property (4.26) of the z-transform, we obtain

Y z   X z H z  4.40

where Y(z), X(z), and H(z) are the z-transforms of y[n], x[n], and h[n], respectively.
Equation 4.40 can be expressed as

Y z 
H z   4.41
X z 

The z-transform H(z) of h[n] is referred to as the system function (or the transfer function)
of the system. By Eq. 4.41 the system function H(z) can also be defined as the ratio of the
z-transforms of the output y[n] and the input x[n]. The system function H(z) completely
characterizes the system. Figure 4.3 illustrates the relationship of Eqs. 4.39 and 4.40.

Figure 4.3 Impulse response and system function

4.6.2 Characterization of Discrete-Time LTI Systems

Many properties of discrete-time LTI systems can be closely associated with the
characteristics of H(z) in the z-plane and in particular with the pole locations and the ROC.

University of Mines and Technology, Tarkwa 58 Solomon Nunoo MPhil, BSc, MIET, MASEE, MIAENG
Signals and Systems Chapter 4 – The z-Transform and Discrete-Time LTI Systems

Causality
For a causal discrete-time LTI system, we have [Eq. 2.44]

hn  0 n<0

since h[n] is a right-sided signal, the corresponding requirement on H(z) is that the ROC of
H(z) must be of the form

z  rmax

That is, the ROC is the exterior of a circle containing all of the poles of H(z) in the z-plane.
Similarly, if the system is anticausal, that is,

hn  0 n≥0

then h[n] is left-sided and the ROC of H(z) must be of the form

z  rmin

That is, the ROC is the interior of a circle containing no poles of H(z) in the z-plane.

Stability
In Sec. 2.7 we stated that a discrete-time LTI system is BIB0 stable if and only if [Eq. 2.49]

 hn  
n  

The corresponding requirement on H(z) is that the ROC of H(z) contains the unit circle
(that is, |z| = 1).

Causal and Stable Systems

If the system is both causal and stable, then all of the poles of H(z) must lie inside the unit
circle of the z-plane because the ROC is of the form |z| > rmax and since the unit circle is
included in the ROC, we must have rmax < 1.

4.6.3 System Function for LTI Systems Described by Linear Constant-Coefficient

Difference Equations
In Sec. 2.9 we considered a discrete-time LTI system for which input x[n] and output y[n]
satisfy the general linear constant-coefficient difference equation of the form

N M

 a yn  k    b xn  k 
k 0
k
k 0
k 4.42

University of Mines and Technology, Tarkwa 59 Solomon Nunoo MPhil, BSc, MIET, MASEE, MIAENG
Signals and Systems Chapter 4 – The z-Transform and Discrete-Time LTI Systems

Applying the z-transform and using the time-shift property (Eq. 4.18) and the linearity
property (Eq. 4.17) of the z-transform, we obtain

N M

 ak z k Y z    bk z k X z 
k 0 k 0

or

N M
Y z  ak z  k  X z  bk z  k 4.43
k 0 k 0

Thus,

Y z  
k
bz k
H z    k 0
4.44
X z  N

a z
k 0
k
k

Hence, H(z) is always rational. Note that the ROC of H(z) is not specified by Eq. 4.44 but
must be inferred with additional requirements on the system such as the causality or the
stability.

4.6.4 Systems Interconnection

For two LTI systems (with h1[n] and h2[n], respectively) in cascade, the overall impulse
response h[n] is given by

hn  h1 n  h2 n 4.45

Thus, the corresponding system functions are related by the product

H z   H1 z H2 z  R  R1  R2 4.46

Similarly, the impulse response of a parallel combination of two LTI systems is given by

hn  h1 n  h2 n 4.47

and

H z   H1 z   H2 z  R  R1  R2 4.48

University of Mines and Technology, Tarkwa 60 Solomon Nunoo MPhil, BSc, MIET, MASEE, MIAENG
Signals and Systems Chapter 4 – The z-Transform and Discrete-Time LTI Systems

4.7 The Unilateral z-Transform

4.7.1 Definition
The unilateral (or one-sided) z-transform X1(z) of a sequence x[n] is defined as [Eq. 4.5]


X1 z    xnz  n 4.49
n0

and differs from the bilateral transform in that the summation is carried over only n ≥ 0.
Thus, the unilateral z-transform of x[n] can be thought of as the bilateral transform of
x[n]u[n]. Since x[n]u[n] is a right-sided sequence, the ROC of X1(z) is always outside a
circle in the z-plane.

4.7.2 Basic Properties

Most of the properties of the unilateral z-transform are the same as for the bilateral
z-transform. The unilateral z-transform is useful for calculating the response of a causal
system to a causal input when the system is described by a linear constant-coefficient
difference equation with nonzero initial conditions. The basic property of the unilateral
z-transform that is useful in this application is the following time-shifting property which is
different from that of the bilateral transform.

4.7.3 Time-Shifting Property

If x[n] ↔ X1(z), then for m ≥ 0,

xn  m  z m X1 z   z m1 x 1  z m 2 x 2  ...  x m 4.50

xn  m  z m X1 z   z m x0  zm1 x1  ...  zxm  1 4.51

4.7.4 System Function

Similar to the case of the continuous-time LTI system, with the unilateral z-transform, the
system function H(z) = Y(z)/X(z) is defined under the condition that the system is relaxed,
that is, all initial conditions are zero.

University of Mines and Technology, Tarkwa 61 Solomon Nunoo MPhil, BSc, MIET, MASEE, MIAENG