Chapter V

The Z-Transform and Application to

Discrete-Time System Analysis

University of Engineering and Technology - VNU Hanoi

Signals and Systems 2010 1 / 29

 Z-Transform of Discrete-Time Signals The bilateral Z-transform

The bilateral Z-transform of a discrete-time

signal x(n) is defined as:
+∞
X
X (z) = Z[x(n)] = x(n)z −n
n=−∞

in which, z is a complex variable → the

Z-transform transforms a discrete-time signal
into a complex space (the Z-plane).
The Z-transform of x(n) exists if the above
transform series converges

Signals and Systems 2010 2 / 29

 Z-Transform of Discrete-Time Signals The unilateral Z-transform

The unilateral Z-transform of a discrete-time

signal x(n) is defined as:
+∞
X
1
X (z) = Z [x(n)] = 1
x(n)z −n
n=0

The unilateral and bilateral transforms of causal

signals are identical.

Signals and Systems 2010 3 / 29

 Z-Transform of Discrete-Time Signals Region of convergence of Z-transform

The region of convergence (ROC) of the

Z-transform is the setPof all values of Z for which
the transform series x(n)z −n converges.
The criteria of convergence of the Z-transform
are derived from the Cauchy criterion:
+∞
X
1/n
lim |x(n)| < 1 ⇐⇒ x(n) < ∞
n→∞
n=0

Signals and Systems 2010 4 / 29

 Z-Transform of Discrete-Time Signals Region of convergence of Z-transform

The following criteria of convergence are

obtained for the Z-transform by using the
Cauchy criterion:
Rx− < |z| < Rx+
in which:
Rx− = lim |x(n)|1/n
n→∞

Rx+ = 1/ lim |x(−n)|1/n

n→∞

ROC of the Z-transform is the region bounded

by two circles having the radii of Rx− and Rx+
respectively in the Z-plane.
Signals and Systems 2010 5 / 29
 Z-Transform of Discrete-Time Signals Region of convergence of Z-transform

ROC of the Z-transform of some special types of

signals:
Finite-length signals: ROC is the entire Z-plane
except the origin (Rx− = 0, Rx+ = ∞).
Infinite-length causal signals: ROC is the entire
space outside the circle with the radius of Rx−
(Rx+ = ∞).
Infinite-length anti-causal signals: ROC is the entire
region inside the circle with the radius of Rx+ except
the origin (Rx− = 0).
ROC of the unilateral Z-transform is same as
the ROC of the bilateral Z-transform of causal
signals.

Signals and Systems 2010 6 / 29

 Z-Transform of Discrete-Time Signals Properties of Z-transform

Linearity:

Z[αx1 (n) + βx2 (n)] = αZ[x1 (n)] + βZ[x2 (n)]

Time shifting:

Z[x(n − n0 )] = z −n0 X (z)

Scaling in Z -plane:

Z[an x(n)] = X (a−1 z)

with the ROC being |a|Rx− < |z| < |a|Rx+ .

Signals and Systems 2010 7 / 29

 Z-Transform of Discrete-Time Signals Properties of Z-transform

Reflection:
Z[x(−n)] = X (z −1 )
with the ROC being 1/Rx+ < |z| < 1/Rx− .
Z-plane differentiation:
dX (z)
Z[nx(n)] = −z
dz
Convolution:
Z[x1 (n) ∗ x2 (n)] = X1 (z)X2 (z)
Correlation:
Z[rx1 x2 (n)] = X1 (z)X2 (z −1 )
Signals and Systems 2010 8 / 29
 Z-Transform of Discrete-Time Signals Properties of unilateral Z-transform

Time delaying:
k
X
1 −k 1
Z [x(n−k)] = z X (z)+ x(−m)z m−k (k > 0)
m=1

Time advancing:
k−1
X
1
Z [x(n + k)] = z X (z) − k 1
x(m)z −m (k > 0)
m=0

Signals and Systems 2010 9 / 29

 Z-Transform of Discrete-Time Signals Properties of unilateral Z-transform

The final value theorem:

lim x(n) = lim (z − 1)X 1 (z)

n→∞ z→1

if the ROC of (z − 1)X 1 (z) contains the unit

circle in the Z-plane.

Signals and Systems 2010 10 / 29

 The Inverse Z-transform Derivation of inverse Z-transform

The Cauchy integral theorem:


1
I 1 (n = 0)
n−1
z dz =
j2π C 
0 (n 6= 0)

in which, C is a positively oriented simple closed

contour around the origin in the Z-plane.
The following formula was obtained for the
inverse Z-transform by applying the Cauchy
integral theorem:
I
1
x(n) = X (z)z n−1 dz
j2π C
Signals and Systems 2010 11 / 29
 The Inverse Z-transform Computing the inverse Z-transform

Using the Cauchy residue theorem (1)

Denote {zpk } the poles of X (z)z n−1 lying inside
the closed contour C, then:
X
x(n) = Res[X (z)z n−1 |z=zpk ]
k

If the pole zpk is a distinct pole, then the residue:

Res[X (z)z n−1 |z=zpk ] = (z − zpk )X (z)z n−1 |z=zpk

Signals and Systems 2010 12 / 29

 The Inverse Z-transform Computing the inverse Z-transform

Using the Cauchy residue theorem (2)

If the pole zpk is a repeated pole with the degree
of sk , then the residue:

Res[X (z)z n−1 |z=zpk ] =

1 d sk −1
[(z − zpk )sk X (z)z n−1 ]|z=zpk
(sk − 1)! dz sk −1

Signals and Systems 2010 13 / 29

 The Inverse Z-transform Computing the inverse Z-transform

Power series expansion method

If X (z) can be expanded into a power series of
z −1 such as:
+∞
X
X (z) = αn z −n
n=−∞

then we have x(n) = αn .

Method: using the division of polynomials.
Note: ROC of X (z) decides the form of the
power series.

Signals and Systems 2010 14 / 29

 The Inverse Z-transform Computing the inverse Z-transform

Partial-fraction expansion method (1)

Without loss of generalization, assume that
X (z) is represented in the form of a proper
rational function N(z)/D(z) (N(z) and D(z) are
polynomials and the degree of N(z) is less than
the degree of D(z)).
Denote {zpk } the poles of X (z): {zpk } are roots
of the equation D(z) = 0.

Signals and Systems 2010 15 / 29

 The Inverse Z-transform Computing the inverse Z-transform

Partial-fraction expansion method (2)

If all {zpk } are distinct, the partial-fraction
expansion of X (z) is:
X Ak
X (z) =
z − zpk
k

in which, the coefficients {Ak } are computed by:

Ak = (z − zpk )X (z)|z=zpk

Signals and Systems 2010 16 / 29

 The Inverse Z-transform Computing the inverse Z-transform

Partial-fraction expansion method (3)

In case X (z) has repeated poles, denote sk the
number of repetitions of the pole zpk , then we
have the following expansion of X (z):
sk
XX Aks
X (z) =
(z − zpk )s
k s=1

in which, the coefficients {Aks } are computed by:

1 d sk −s (z − zpk )sk X (z)

Ak s =
(sk − s)! dz sk −s z=zpk

Signals and Systems 2010 17 / 29

 The Inverse Z-transform Computing the inverse Z-transform

Inverse Z-transform of partial fractions (1)


  αn u(n) (|z| > |α|)
−1 z 
Z =
z−α 
−αn u(−n − 1) (|z| < |α|)
 n−1

1
  α u(n − 1) (|z| > |α|)
−1
Z =
z−α
−αn−1 u(−n) (|z| < |α|)


Signals and Systems 2010 18 / 29

 The Inverse Z-transform Computing the inverse Z-transform

Inverse Z-transform of partial fractions (2)

 
z
Z −1 =
(z − α)m+1
n(n−1)...(n−m+1) n−m

 m! α u(n) (|z| > |α|)

− n(n−1)...(n−m+1) αn−m u(−n − 1) (|z| < |α|)


m!

Note: it is usually easier to compute the inverse

Z-transform if we expand X (z)/z instead of X (z).

Signals and Systems 2010 19 / 29

 Relation with Fourier transform

The Fourier transform of a discrete-time signal

x(n) is the Z-transform on the unit circle in the
Z-plane → the Fourier transform of x(n) exists if
the ROC of its Z-transform contains the unit
circle.
Application: computing Fourier and inverse
Fourier transforms of discrete-time signals via
Z-transform and inverse Z-transform.

Signals and Systems 2010 20 / 29

Transfer Functions of Discrete-Time LTI Systems Definition of transfer function

Consider a discrete-time LTI system having the

impulse response h(n), i.e.:

y(n) = h(n) ∗ x(n)

Perform the Z-transform for both sides of the

above equation and apply the convolution
property of the Z-transform to obtain:

Y (z)
Y (z) = H(z)X (z) → H(z) =
X (z)

H(z) is called the transfer function of the system.

Signals and Systems 2010 21 / 29

Transfer Functions of Discrete-Time LTI Systems Definition of transfer function

An discrete-time LTI system is generally

represented by a constant-coefficient linear
difference equation of the following form:
N
X M
X
ak y(n − k) = br x(n − r )
k=0 r =0

Perform the Z-transform for both sides of the

above equation to obtain:
N
X M
X
−k
ak z Y (z) = br z −r X (z)
k=0 r =0

Signals and Systems 2010 22 / 29

Transfer Functions of Discrete-Time LTI Systems Definition of transfer function

The transfer function of the system is then

determined by:
PM −r
PM M−r
Y (z) r =0 br z N−M r =0 br z
H(z) = = PN =z PN
X (z) k=0 a k z −k
k=0 ak z
N−k

Transfer function determines a system, and is

based on the solution of the difference equation
using Z-transform and inverse Z-transform:

y(n) = Z −1 [H(z)X (z)]

Signals and Systems 2010 23 / 29

Transfer Functions of Discrete-Time LTI Systems Transfer functions of combined systems

Cascade connection:

H(z) = H1 (z)H2 (z)

Parallel connection:

H(z) = H1 (z) + H2 (z)

System with negative feedback:

H(z) = H1 (z)/[1 + H1 (z)H2 (z)]

System with positive feedback:

H(z) = H1 (z)/[1 − H1 (z)H2 (z)]

Signals and Systems 2010 24 / 29
 Solving the linear difference equations

Given a discrete-time LTI system represented by

a constant-coefficient linear difference equation
of the following form:
N
X M
X
ai y(n − i) = bj x(n − j)
i=0 j=0

Perform the unilateral Z-transform for both sides

of the above equation, we obtain:
N
X M
X
−i 1
ai z Y (z) + I(z) = bj z −j X 1 (z)
i=0 j=0

in which, I(z) is composed of the initial

conditions at n = −1, −2, ..., −N.
Signals and Systems 2010 25 / 29
 Solving the linear difference equations

Recall that the response y(n) of a discrete-time

LTI system is the general solution of a
constant-coefficient linear difference equation,
composed of the initial response and the
zero-state response:

y(n) = y0 (n) + ys (n)

or:
Y 1 (z) = Y01 (z) + Ys1 (z)

Signals and Systems 2010 26 / 29

 Solving the linear difference equations

The zero-state response to a causal input (i.e.,

X 1 (z) = X (z)):
" PM #
−j
j=0 bj z
ys (n) = Z −1 PN X (z) = Z −1 [H(z)X (z)]
−i
i=0 ai z

The initial response:

" #
I(z)
y0 (n) = −Z −1 PN
−i
i=0 ai z

Signals and Systems 2010 27 / 29

 System Analysis Analyzing the stability

A discrete-time LTI system is stable if and only if

its transfer function converges when |z| = 1 →
the ROC of H(z) must contain the unit circle,
i.e., Rh− < 1 < Rh+ .
For causal systems: Rh+ = ∞ → the condition
for the system to be stable is Rh− < 1 → all the
poles of H(z) must lies inside the unit circle.

Signals and Systems 2010 28 / 29

 System Analysis Analyzing the stability

Another method for analyzing the stability of a

discrete-time causal system represented by
transfer function: using the Jury criteria
There is no need to solve the characteristic equation
to find the poles.
A Jury table is created from the coefficients of the
characteristic polynomial. This table is used to
determine the stability of the system.

Signals and Systems 2010 29 / 29