Topic 3: The 𝒛 −Transform

Session 2024/04

1
Topic Outline
• 𝑧 −transform
• Region of convergence (ROC)
• 𝑧 −transform Properties
• Inverse 𝑧 −transform
• Application of 𝑧 −transform

2
Introduction to 𝑍 −transform
• Recall: A continuous signal such as 𝑥(𝑡) or y(t) when sampled every
𝑇 seconds will yield 𝑥(𝑛𝑇) and 𝑥(𝑛𝑇). A discrete signal is usually
represented as just 𝑥(𝑛) and 𝑦(𝑛).

Digitized
continuous signal
… … … -4𝑇 -3𝑇 -2𝑇 -𝑇 𝑇 2𝑇 3𝑇 4𝑇 ………

• Z-transform converts difference equations into algebraic equations,

which simplifying the analysis.
• It’s a generalized form of discrete time Fourier Transform which
can be used to analyze a broader range of signals.
3
𝑧 −transform
• Let define the DTFT pair 𝑥 𝑛 ← 𝐷𝑇𝐹𝑇 → 𝑋(Ω) as follow:
• DTFT synthesis equation
1 𝜋
𝑥𝑛 = න 𝑋 Ω 𝑒 𝑗Ω𝑛 𝑑(Ω)
2𝜋 −𝜋
• DTFT analysis equation
∞

𝑋 Ω = ෍ 𝑥[𝑛]𝑒 −𝑗Ω𝑛
𝑛=−∞
• Based on DTFT analysis equation,
∞
bilateral z-transform
∞

𝒵 𝑥 𝑛 𝑒 −𝜎𝑛 = ෍ 𝑥 𝑛 𝑒 −𝜎𝑛 𝑒 −𝑗Ω𝑛 = ෍ 𝑥[𝑛]𝑒 −(𝜎+𝑗Ω)𝑛 .

𝑛=−∞ 𝑛=−∞
• Substituting 𝑒
𝜎+𝑗Ω
= 𝑧, then
• z-analysis equation
∞

𝑋 𝑧 = 𝒵 𝑥 𝑛 𝑒 −𝜎𝑛 = ෍ 𝑥 𝑛 𝑧 −𝑛 .
𝑛=−∞ 4
Unilateral 𝑧 −transform
• Most of physical systems in signals processing are casual.
• By assuming 𝑥 𝑛 = 0 for 𝑛 < 0, the analysis equation can be
simplified as: ∞

𝑋 𝑧 = ෍ 𝑥 𝑛 𝑧 −𝑛
𝑛=0
• Unilateral used to simplify the analysis of causal system.
• For a causal signal and system, unilateral and bilateral z-transforms are
the same.

5
𝑧 −transform in polar form
• 𝑧 is complex valued, and the polar form is written as
𝑧 = 𝑟𝑒 𝑗Ω
with 𝑟 is magnitude and Ω is the angle of 𝑧.
• It’s convenient to interpret the z-transform using complex z-
𝐼𝑚
plane.
𝑧 = 𝑒 𝑗Ω
• In the z-plane, the contour 𝑈𝑛𝑖𝑡 𝐶𝑖𝑟𝑐𝑙𝑒

corresponding to |𝑧| = 1 is a circle of Ω

𝑅𝑒
unit radius known as the unit circle. 1

6
Zeros and Poles of 𝑋(𝑧)
• The most important and useful z-transform are those for which
𝑋(𝑧) is a rational function
𝑃 𝑧
𝑋 𝑧 =
𝑄 𝑧
• where 𝑃(𝑧) and 𝑄(𝑧) are polynomials in 𝑧 −1
• The values of z for which 𝑋(𝑧) = 0 are called the zeros of 𝑋(𝑧),
which are the roots of 𝑃(𝑧) = 0.
• The values of z for which 𝑋(𝑧) = ∞ are called the poles of 𝑋(𝑧),
which are the roots of 𝑄(𝑧) = 0.

7
Zeros and Poles of 𝑋(𝑧)
Example:
Find the poles and zeros of 𝑋(𝑧) and plot the pole zero plot for 𝑋(𝑧).
𝑧 −1 − 0.5𝑧 −2
𝑋 𝑧 =
1 + 1.2𝑧 −1 + 0.45𝑧 −2
Solution:
Multiplying the numerator and the denominator of 𝑋(𝑧) with 𝑧 2 .
𝑧 − 0.5
𝑋 𝑧 = 2
𝑧 + 1.2𝑧 + 0.45 −𝑏 ± 𝑏2 − 4𝑎𝑐
𝑧 − 0.5 𝑧 =
2𝑎
𝑋 (𝑧) =
𝑧 + 0.6 − 𝑗0.3 𝑧 + 0.6 + 𝑗0.3
8
Zeros and Poles of 𝑋(𝑧) Study

𝑧 − 0.5
𝑋 𝑧 =
𝑧 + 0.6 − 𝑗0.3 𝑧 + 0.6 + 𝑗0.3 poles
• Find the roots of 𝑃(𝑧) = 0 and 𝑄(𝑧) = 0.
𝑃 𝑧 = 𝑧 − 0.5 = 0
𝑧 = 0.5 is the zero of 𝑋 𝑧
𝑄(𝑧) = (𝑧 + 0.6 − 𝑗0.3)(𝑧 + 0.6 + 𝑗0.3) = 0
𝑧1 = −0.6 + 𝑗0.3 and 𝑧2 = −0.6 − 𝑗0.3 are the poles of 𝑋 𝑧
• Zero is marked with the symbol ‘O’ on the z-plane.
• Poles are marked with the symbol ‘X’ on the z-plane. zero

9
Region of Convergence (ROC)

10
Region of Convergence (ROC)
• The range value of the complex variable 𝑧 for which the z-transform
converges is called the region of convergence (ROC).
• The condition for existence of
∞
the z-transform is:
−𝑛
෍ 𝑥𝑛 𝑧 <∞
𝑛=−∞
• The complete specification of z-transform is includes the ROC associated
with 𝑋[𝑧].
• An alternative condition for the existence of the DTFT is the inclusion of the
unit circle within the ROC of the z-transform.

ROC is centered on origin and consists of a set of rings.

11
Region of Convergence (ROC)
Example 1:
• Let consider the sequence
𝑥 𝑛 = 𝑎𝑛 𝑢 𝑛 𝑎 is real
• Then, the z-transform analysis of 𝑥 𝑛 is ∞ ∞

𝑋 𝑧 = ෍ 𝑎𝑛 𝑢 𝑛 𝑧 −𝑛 = ෍ 𝑎𝑧 −1 𝑛
Infinite geometric series:
∞
𝑛=−∞ 𝑛=0 𝑎
𝑛
• For the convergence of 𝑋(𝑧) we require that∞ 𝑆∞ = ෍ 𝑎𝑟 =
1−𝑟
𝑛=0
෍ 𝑎𝑧 −1 𝑛 <∞
Convergence condition,
𝑛=0
• if |𝛼𝑧 −1 | ≥ 1 the bilateral z-transform has an infinite value 𝑟 <1

• Thus the ROC is the of values of 𝑧 for which 𝑎𝑧 −1 < 1 or equivalently, 𝑧 > |𝑎|.
• Then ∞
1
𝑋 𝑧 = ෍ 𝑎𝑧 −1 𝑛 = 𝑧 > |𝑎|
1 − 𝑎𝑧 −1
𝑛=0
• Alternatively,
𝑧
𝑋 𝑧 = 𝑧 > |𝑎|
𝑧−𝑎 12
Region of Convergence (ROC)
• From previous example 1:
𝑧
𝑋 𝑧 = 𝑧 > |𝑎|
𝑧−𝑎
• The 𝑋(𝑧) has 1 zero at 𝑧 = 0 and 1 pole at 𝑧 = 𝑎.

0<𝑎<1 𝑎>1 −1 < 𝑎 < 0 𝑎 < −1

13
Region of Convergence (ROC)
Example 2:
• Let consider the sequence
𝑥 𝑛 = −𝑎𝑛 𝑢 −𝑛 − 1
• The z-transform
∞ −1

𝑋 𝑧 = ෍ −𝑎𝑛 𝑢 −𝑛 − 1 𝑧 −𝑛 = − ෍ 𝑎𝑧 −1 𝑛

𝑛=−∞ ∞ 𝑛=−∞

= 1 − ෍ 𝑎−1 𝑧 𝑛

𝑛=0
• If |𝑎 𝑧| < 1 or, equivalently, 𝑧 < 𝑎 ,
−1

• Then ,
1 1
𝑋 𝑧 =1− = 𝑧 < |𝑎|
1 − 𝑎−1 𝑧 1 − 𝑎𝑧 −1
• Or alternatively:
𝑧
𝑋 𝑧 = 𝑧 < |𝑎|
𝑧−𝑎

14
Region of Convergence (ROC)
• From previous example 2:
𝑧
𝑋 𝑧 = 𝑧 < |𝑎|
𝑧−𝑎
• The 𝑋(𝑧) has 1 zero at 𝑧 = 0 and 1 pole at 𝑧 = 𝑎

0<𝑎<1 𝑎>1 −1 < 𝑎 < 0 𝑎 < −1 15

Region of Convergence (ROC)
• The ROC as a ring in the z-plane.
• For specific cases, the inner boundary can extend
inward to the origin, and the ROC becomes a disc.
• For other cases, the outer boundary can extend
outward to infinity.

Disc Ring/annulus Circle exterior

|𝑧| < 𝑟 𝑟0 < |𝑧| < 𝑟1 𝑧 > 𝑟 16

Properties of ROC
• The properties of the ROC depend on the nature of the signal.
Property 1
The ROC does not contain any poles
Rational sequence
Property 2
ROC is entire z-plane except 𝑧 = 0 and ∞
Finite duration signal
Property 3 ROC extends outward from outermost finite pole.
Right sided sequence 𝑧 > 𝑟𝑚𝑎𝑥 or ∞ > 𝑧 > 𝑟𝑚𝑎𝑥
Property 4 ROC extends inward from innermost non zero pole.
Left sided sequence 𝑧 < 𝑟𝑚𝑖𝑛 or 0 < 𝑧 < 𝑟𝑚𝑖𝑛
Property 5 ROC appears as ring centered at the origin bounded by
Both-sided sequence an interior and exterior pole.
𝑟1 < 𝑧 < 𝑟2

17
Example 3
• Calculate the unilateral z-transform for the following sequences:
i. Unit impulse sequence, 𝑥1 𝑛 = 𝛿 𝑛
ii. Unit step sequence, 𝑥2 𝑛 = 𝑢 𝑛
iii. Exponential sequence, 𝑥3 𝑛 = 𝛼 𝑛 𝑢 𝑛
iv. First order, time-rising, exponential sequence, 𝑥4 𝑛 = 𝑛𝛼 𝑛 𝑢 𝑛 ;
1 𝑛 = 0, 1
v. Time-limited sequence, 𝑥5 𝑛 = ቐ 2 𝑛 = 2, 5
0 𝑒𝑙𝑠𝑒𝑤ℎ𝑒𝑟𝑒

18
i. Unit impulse sequence, 𝑥1 𝑛 = 𝛿 𝑛
Solution:
By definition, ∞

𝑋1 𝑧 = ෍ 𝛿 𝑛 𝑧 −𝑛
𝑛=0
= 𝛿 0 𝑧0
𝑋1 𝑧 = 1.
Thus, the ROC: entire z-plane.

The z-transform pair for an impulse sequence given by

𝛿 𝑛 ←𝑧→1
𝑅𝑂𝐶: 𝑒𝑛𝑡𝑖𝑟𝑒 𝑧 − 𝑝𝑙𝑎𝑛𝑒.
19
ii. Unit step sequence, 𝑥2 𝑛 = 𝑢 𝑛
Solution:
By definition, ∞ ∞

𝑋2 𝑧 = ෍ 𝑢[𝑛] 𝑧 −𝑛 = ෍ 𝑧 −𝑛
𝑛=0 𝑛=0 Infinite geometric series:
1 −1 < 1
∞
𝑎
= ቐ1 − 𝑧 −1 ; for 𝑧 𝑆∞ = ෍ 𝑎𝑟 𝑛 =
1−𝑟
𝑛=0
undefined elsewhere.
The z-transform pair for a unit step sequence is given by,
1
𝑢 𝑛 ←𝑧→ −1
;
1−𝑧
𝑅𝑂𝐶: 𝑧 > 1.
Note: ROC 𝑧 −1 < 1 is equivalent to 𝑧 > 1
20
iii. Exponential sequence, 𝑥3 𝑛 = 𝛼 𝑛 𝑢 𝑛
Solution:
By definition, ∞ ∞

𝑋3 𝑧 = ෍ 𝛼 𝑛 𝑢[𝑛] 𝑧 −𝑛 = ෍ 𝛼 𝑛 𝑧 −𝑛
𝑛=0 𝑛=0
1 −1 < 1
= ቐ1 − 𝛼𝑧 −1 ; for 𝛼𝑧
undefined elsewhere.
The z-transform pair for a unit step sequence is given by,
𝑛
1
𝛼 𝑢 𝑛 ←𝑧→ −1
;
1 − 𝛼𝑧
𝑅𝑂𝐶: 𝑧 > |𝛼|.
Note: ROC 𝛼𝑧 −1 < 1 is equivalent to 𝑧 > 𝛼
21
iv. First order, time-rising, exponential sequence, 𝑥4 𝑛 = 𝑛𝛼 𝑛 𝑢 𝑛 ;
Solution:
By definition, ∞ ∞ ∞

𝑋4 𝑧 = ෍ 𝑛𝛼 𝑛 𝑢[𝑛] 𝑧 −𝑛 = ෍ 𝑛𝛼 𝑛 𝑧 −𝑛 = ෍ 𝑛(𝛼𝑧 −1 )𝑛
𝑛=0 𝑛=0 𝑛=0
Using the following result:
∞
𝑟
෍ 𝑛𝑟 𝑛 = 2
; 𝑝𝑟𝑜𝑣𝑖𝑑𝑒𝑑 𝑟 < 1
1−𝑟
𝑛=0
Then, ∞
𝛼𝑧 −1
෍ 𝑛𝑟 𝑛 = ; 𝑝𝑟𝑜𝑣𝑖𝑑𝑒𝑑 𝛼𝑧 −1 < 1
1 − 𝛼𝑧 −1 2
𝑛=0
The z-transform pair for a time-rising, complex exponential is given by
−1
𝑛
𝛼𝑧 𝛼𝑧
𝑛𝛼 𝑢 𝑛 ← 𝑧 → −1 2
; 𝑜𝑟
1 − 𝛼𝑧 𝑧−𝛼 2
𝑅𝑂𝐶: 𝑧 > |𝛼|.
22
 1 𝑛 = 0, 1
v. Time-limited sequence, 𝑥5 𝑛 = ቐ 2 𝑛 = 2, 5
0 𝑒𝑙𝑠𝑒𝑤ℎ𝑒𝑟𝑒
Solution:
Since the input sequence 𝑥5 [𝑛] is zero outside the range 0 ≤ 𝑛 ≤ 5,
∞

𝑋5 𝑧 = ෍ 𝑥[𝑛] 𝑧 −𝑛
𝑛=0
= 𝑥 0 + 𝑥 1 𝑧 −1 + 𝑥 2 𝑧 −2 + 𝑥 3 𝑧 −3 + 𝑥 4 𝑧 −4 + 𝑥 5 𝑧 −5 .
Substituting the values of 𝑥5 𝑛 for the range 0 ≤ 𝑛 ≤ 5 we obtain
𝑋5 𝑧 = 1 + 𝑧 −1 + 2𝑧 −2 + 2𝑧 −5
𝑅𝑂𝐶: entire z − plane, except z = 0.

23
Table 1: Some Common
z-Transform Pairs

24
𝑧 −transform Properties

25
Properties of z-transform
1. Linearity
If 𝑥1 𝑛 and 𝑥2 [𝑛] are two DT sequences with the following z-transform pairs:
𝑥1 𝑛 ← 𝓏 → 𝑋1 𝑧 , ROC: R1
and
𝑥2 𝑛 ← 𝓏 → 𝑋2 𝑧 , ROC: R 2
Then, 𝑎1 𝑥1 𝑛 + 𝑎2 𝑥2 𝑛 ← 𝓏 → 𝑎1 𝑋1 𝑧 + 𝑎2 𝑋2 𝑧 , ROC: at least R1 ∩ 𝑅2
Proof: ∞

𝒵 𝑎1 𝑥1 𝑛 + 𝑎2 𝑥2 𝑛 = ෍ 𝑎1 𝑥1 𝑛 + 𝑎2 𝑥2 𝑛 𝑧 −𝑛
∞ 𝑛=0 ∞

= 𝑎1 ෍ 𝑥1 𝑛 𝑧 −𝑛 + 𝑎2 ෍ 𝑥2 𝑛 𝑧 −𝑛
𝑛=0 𝑛=0
26
Properties of z-transform
2. Time Scaling
Decimation: the z-transform of 𝑥[𝑛] and its decimated sequence 𝑦 𝑛 =
𝑥[𝑘𝑛] (scaled by 𝑘) are not related to each other.
Interpolation:
𝑛
𝑘 𝑥 if 𝑛 is a multiple of interger 𝑘
𝑥 𝑛 =ቐ 𝑘
0 otherwise
If 𝑥 𝑛 ← 𝒵 → 𝑋(𝑧) with ROC 𝑅𝑥 , then the z-transform 𝑋 𝑘 (𝑧) of 𝑥 𝑘 [𝑛] is
given by
1
𝑥𝑘 𝑛 ← 𝓏 → 𝑋𝑘 𝑧 = 𝑋 𝑧𝑘 , 𝑅𝑂𝐶: 𝑅𝑥𝑘
For 2 < 𝑘 < ∞.
27
Properties of z-transform
3. Time-shifting
If 𝑥 𝑛 ← 𝑏𝑖𝑙𝑎𝑡𝑒𝑟𝑎𝑙 𝓏 → 𝑋(𝑧) with ROC:𝑅𝑥 , then
𝑥 𝑛 − 𝑘 ← 𝑏𝑖𝑙𝑎𝑡𝑒𝑟𝑎𝑙 𝓏 → 𝑧 −𝑘 𝑋 𝑧
ROC: 𝑅𝑥
For causal sequence 𝑥 𝑛 𝑢[𝑛],
𝑥 𝑛 𝑢 𝑛 ← 𝓏 → 𝑋 𝑧 𝑅𝑂𝐶: 𝑅𝑥
For positive integer 𝑘
a) 𝑥 𝑛 − 𝑘 𝑢 𝑛 − 𝑘 ← 𝓏 → 𝑧 −𝑘 𝑋 𝑧 ;
b) 𝑥 𝑛 + 𝑘 𝑢 𝑛 ← 𝓏 → 𝑧 𝑘 𝑋 𝑧 − 𝑧 𝑘 σ𝑘−1
𝑛=0 𝑥 𝑛 𝑧 −𝑛 ;

c) 𝑥 𝑛 − 𝑘 𝑢 𝑛 ← 𝓏 → 𝑧 −𝑘 𝑋 𝑧 + 𝑧 −𝑘 σ𝑘𝑛=1 𝑥 −𝑛 𝑧 𝑛 ;
28
Properties of z-transform
4. Time differencing
If 𝑥 𝑛 ← 𝓏 → 𝑋(𝑧) with ROC:𝑅𝑥 ,
𝑥 𝑛 − 𝑥[𝑛 − 1] ← 𝓏 → (1 − 𝑧 −1 )𝑋(𝑧)
ROC: 𝑅𝑥
Example 4:
Calculate the z-transform of impulse function 𝑥 𝑛 = 𝛿[𝑛] using the time differencing
property.
Solution 4:
Using the time-differencing property, the z-transform of 𝑢 𝑛 − 𝑢 𝑛 − 1 is given by
𝑢 𝑛 − 𝑢 𝑛 − 1 ← 𝓏 → 1 − z −1 . 𝑍 𝑢 𝑛 , 𝑅𝑂𝐶: 𝑧 > 1
1
𝑍 𝑢𝑛 =
1 − 𝑧 −1
Then, 𝑢 𝑛 − 𝑢 𝑛 − 1 = 1
Equivalent to, 𝛿 𝑛 ← 𝓏 → 1
29
Properties of z-transform
5. Z-domain differentiation
If 𝑥 𝑛 ← 𝓏 → 𝑋(𝑧) with ROC: 𝑅𝑥 then
𝑑𝑋 𝑧
𝑛𝑥 𝑛 ← 𝓏 → −𝑧 , 𝑅𝑂𝐶: 𝑅𝑥
𝑑𝑧
Proof:
By differentiation, ∞

𝑋 𝑧 = ෍ 𝑥 𝑛 𝑧 −𝑛 .
𝑛=0
Differentiating both sides with respect
∞
to 𝑧, ∞
𝑑𝑋 𝑧 𝑑𝑧 −𝑛
= ෍𝑥 𝑛 = ෍ 𝑥 𝑛 −𝑛 𝑧 −𝑛−1 .
𝑑𝑧 𝑑𝑧
𝑛=0 𝑛=0
Multiplying both size with −𝑧, ∞
𝑑𝑋 𝑧
−𝑧 = ෍ 𝑛 𝑥 𝑛 𝑧 −𝑛 .
𝑑𝑧 30
𝑛=0
Properties of z-transform
6. Time convolution
If 𝑥1 [𝑛] and 𝑥2 [𝑛] are two arbitrary functions with the following z-
transform pairs:
𝑥1 𝑛 ← 𝓏 → 𝑋1 𝑧 , 𝑅𝑂𝐶: 𝑅1
and
𝑥2 𝑛 ← 𝓏 → 𝑋2 𝑧 , 𝑅𝑂𝐶: 𝑅2
Then the convolution property
𝑥1 𝑛 ∗ 𝑥2 𝑛 ← 𝓏 → 𝑋1 𝑧 𝑋2 𝑧 𝑅𝑂𝐶: 𝑅1 ∩ 𝑅2

Note: convolution of discrete sequence is convolution sum

31
Properties of z-transform
7. Time accumulation
If 𝑥 𝑛 ← 𝓏 → 𝑋(𝑧) with𝑛 ROC: 𝑅𝑥 then
𝑧
෍𝑥 𝑘 ← 𝓏 → 𝑋 𝑧 , 𝑅𝑂𝐶: 𝑅𝑥 ∩ 𝑧 > 1 .
𝑧−1
𝑘=0
Proof:
Let the convolution result 𝑛

෍𝑥 𝑘 = 𝑥 𝑛 ∗𝑢 𝑛 .
𝑘=0
z-transform on left side, and apply conv.
𝑛
property,
෍ 𝑥 𝑘 ← 𝓏 → 𝑋 𝑧 𝑍{𝑢[𝑛]}
𝑘=0
1
The z-transform of unit step function is, 𝑍 𝑢 𝑛 =
1−𝑧 −1
Thus 𝑛
1 𝑧
෍𝑥 𝑘 ←𝓏 →𝑋 𝑧 = 𝑋(𝑧)
1 − 𝑧 −1 𝑧−1 32
𝑘=0
Properties of z-transform
8. Initial- and final-value theorems
If 𝑥 𝑛 ← 𝓏 → 𝑋(𝑧) with ROC: 𝑅𝑥 then
Initial-value theorem,
𝑥 0 = lim 𝑋 𝑧 , provided 𝑥 𝑛 = 0 for 𝑛 < 0;
𝑧→∞
Final-value theorem,
𝑥 ∞ = lim 𝑥 𝑛 = lim 𝑧 − 1 𝑋 𝑧 , provided 𝑥 ∞ exists.
𝑛→∞ 𝑧→1

Note:
Initial value theorem valid only for unilateral z-transform.
Final value theorem valid for both unilateral and bilateral.

33
Initial- and final-value theorems
Example 5:
Determine the initial and final values for z-transform of right sided
sequence,
𝑧
𝑋1 𝑧 = 2 ;
𝑧 − 3𝑧 + 2
solution:
Using the initial-value theorem,
𝑧 1
𝑥1 0 = lim 𝑋1 𝑧 = lim 2 = lim −1
=0
𝑧→∞ 𝑧→∞ 𝑧 − 3𝑧 + 2 𝑧→∞ 𝑧 − 3 + 2𝑧

Using the final-value theorem,

𝑧 𝑧−1 𝑧
𝑥1 ∞ = lim 𝑧 − 1 𝑋1 𝑧 = lim 2 = lim = −1
𝑧→1 𝑧→1 𝑧 − 3𝑧 + 2 𝑧→1 𝑧 − 2

34
Initial- and final-value theorems
Example 6:
Determine the initial and final values for z-transform of right sided
sequence,
1
𝑋2 𝑧 = ;
(𝑧 − 0.1)(𝑧 − 0.5)(𝑧 + 0.2)
solution:
Using the initial-value theorem,
1
𝑥2 0 = lim 𝑋2 𝑧 = lim = 0.
𝑧→∞ 𝑧→∞ (𝑧 − 0.1)(𝑧 − 0.5)(𝑧 + 0.2)

Using the final-value theorem,

𝑧−1
𝑥2 ∞ = lim 𝑧 − 1 𝑋2 𝑧 = lim =0
𝑧→1 𝑧→1 (𝑧 − 0.1)(𝑧 − 0.5)(𝑧 + 0.2)

35
Inverse 𝑧 −transform

36
Inverse z-transform
• Inversion of z-transform is to find 𝑥[𝑛] from its 𝑋(𝑧) is called
inverse z-transform.
• There are four common methods used to evaluate the inverse z-
transform:
1. Inversion formula
2. Table look-up
3. Partial-fraction expansion
4. Power series method.

37
Inverse z-transform

1. Inversion formula Practice from here

• The expression for inverse z-transform in term of an integration in

the z-plane given as,
1
𝑥𝑛 = ර𝑋 𝑧 𝑧 𝑛−1 𝑑𝑧
2𝜋𝑗 𝑐
• Where 𝑐 is the counterclockwise contour integration enclosing the
origin.
• Rarely used because of its complexity.

38
Inverse z-transform

2. Table look-up
• In this method, we find the z-transform function 𝑋(𝑧) that is matched
with the entries in Table 1.
• For example,
1
𝑋 𝑧 = −1
, 𝑅𝑂𝐶: 𝑧 > 0.3
1 − 0.3𝑧
• By looking at Table 1, we find that
𝑛
1
𝑎 𝑢 𝑛 ←𝓏→ −1
, 𝑅𝑂𝐶: 𝑧 > 𝑎.
1 − 𝑎𝑧
• Substituting 𝑎 = 0.3, the inverse z-transform of 𝑋(𝑧) is given by
𝑥 𝑛 = 0.3𝑛 𝑢 𝑛 .
• However, this method still limited since not all functions are available.
39
Inverse z-transform

3. Partial fraction expansion

• The z-transform of 𝑥(𝑛) generally takes the following rational form:
𝑁 𝑧 𝑏𝑚 𝑧 𝑚 + 𝑏𝑚−1 𝑧 𝑚−1 + ⋯ + 𝑏1 𝑧 + 𝑏0
𝑋 𝑧 = =
𝐷 𝑧 𝑧 𝑛 + 𝑎𝑛−1 𝑧 𝑛−1 + ⋯ + 𝑎1 𝑧 + 𝑎0
• Determine the roots of polynomial 𝑁(𝑧) and 𝐷(𝑧).
• Assuming 𝑛 ≥ 𝑚, divide both side with 𝑧 𝑛
𝑋 𝑧 𝑐0 𝑐1 𝑐2 𝑐𝑛 𝑐0 𝑐𝑘
= + + + ⋯+ = +෍
𝑧 𝑧 𝑧 − 𝑝1 𝑧 − 𝑝2 𝑧 − 𝑝𝑛 𝑧 𝑧 − 𝑝𝑘
𝑘=1
𝑋 𝑧
• Where, 𝑐0 = 𝑋 𝑧 |𝑧=0 𝑐𝑘 = 𝑧 − 𝑝𝑘 |𝑧=𝑝𝑘
𝑧
• Multiply both side with 𝑧, then we obtain, 𝑛
𝑧 𝑧 𝑧
𝑋 𝑧 = 𝑐0 + 𝑐1 + ⋯ + 𝑐𝑛 = 𝑐0 + ෍ 𝑐𝑘
𝑧 − 𝑝1 𝑧 − 𝑝𝑛 𝑧 − 𝑝𝑘
𝑘=1
• If 𝑋(𝑧) has multiple-order poles, 𝑝1 is the multiple pole with multiplicity 𝑟,
𝜆1 𝜆2 𝜆𝑟
+ + ⋯ +
𝑧 − 𝑝𝑖 𝑧 − 𝑝𝑖 2 𝑧 − 𝑝𝑖 𝑟
• Where
1 𝑑𝑘 𝑟
𝑋 𝑧
𝜆𝑟−𝑘 = 𝑧 − 𝑝 ቚ
𝑘! 𝑑𝑧 𝑘 𝑧 𝑧=𝑝𝑖 40
Inverse z-transform

3. Partial fraction expansion

Example 7:
Calculate the inverse z-transform.
𝑧
𝑋1 𝑧 = 2
𝑧 − 3𝑧 + 2
Solution 7:
Denominator equal to zero, 𝑧 2 − 3𝑧 + 2 = 0, 𝑧 − 1 𝑧 − 2 = 0
Therefore the z-transform 𝑋1 (𝑧) can be expressed as
𝑋1 𝑧 1 𝑐1 𝑐2
= 2 = + .
𝑧 𝑧 − 3𝑧 + 2 𝑧 − 1 𝑧 − 2
Using Heaviside partial fraction
1 1
𝑐1 = 𝑧 − 1 = = −1
𝑧 − 1 𝑧 − 2 𝑧=1 𝑧 − 2 𝑧=1
1 1
𝑐2 = 𝑧 − 2 = = 1.
𝑧 − 1 𝑧 − 2 𝑧=2 𝑧 − 1 𝑧=2
Thus,
𝑧 𝑧 1 1 Table look-up −𝑢 𝑛 + 2𝑛 𝑢(𝑛)
𝑋1 𝑧 = − + =− +
𝑧−1 𝑧−2 1 − 𝑧 −1 1 − 2𝑧 −1 41
ROC: 𝑧 > 1 ROC: 𝑧 > 2
Inverse z-transform

4. Power Series Expansion

• It can be obtained by performing long division.
• The power series given as ∞

𝑋(𝑧) = ෍ 𝑥(𝑛)𝑧 −𝑛
𝑛=−∞
= ⋯ + 𝑥 −2 𝑧2
+ 𝑥 −1 𝑧 + 𝑥 0 + 𝑥 1 𝑧 −1 + 𝑥 2 𝑧 −2 + ⋯
• where the sequence values 𝑥(𝑛) are the coefficients of 𝑧 −𝑛 .
• This method is very useful for a finite-length sequence where 𝑋(𝑧)
may have no simpler form than a polynomial in 𝑧 −1

42
Inverse z-transform

4. Power Series Expansion

Example 8:
Using the power series expansion technique, find the inverse z-transform of the
following 𝑋(𝑧)
1
𝑋 𝑧 = −1
, 𝑅𝑂𝐶: 𝑧 > 𝑎
1 − 𝑎𝑧
Solution 8:
Since the ROC is 𝑧 > 𝑎 , that is exterior of a circle, 𝑥[𝑛] is a right-sided sequence.
1 + 𝑎𝑧 −1 + 𝑎2 𝑧 −2 + ⋯ 1
𝑋 𝑧 =
1 − 𝑎𝑧 −1 | 1 1 − 𝑎𝑧 −1
1 − 𝑎𝑧 −1 = 1 + 𝑎𝑧 −1 + 𝑎2 𝑧 −2 + ⋯ + 𝑎𝑘 𝑧 −𝑘 + ⋯
𝑎𝑧 −1 𝑥 𝑛 =0 𝑛<0
𝑎𝑧 −1 − 𝑎2 𝑧 −2 𝑥 0 = 1 𝑥 1 = 𝑎 𝑥 2 = 𝑎2 … 𝑥 𝑘 = 𝑎𝑘
𝑎2 𝑧 −2 Thus ,
43
𝑥 𝑛 = 𝑎𝑛 𝑢[𝑛]
Application of 𝑧 −transform

44
1. Solution of Difference Equations
• Z-transform can be used to solve linear, constant coefficients
differential equations (LCCDE).

Example 9:
A causal system is represented by the following difference equation:
𝑦 𝑛 + 2 − 5𝑦 𝑛 + 1 + 6𝑦 𝑛 = 3𝑥 𝑛 + 1 + 5𝑥 𝑛 .
Calculate the output 𝑦[𝑛] for the input 𝑥 𝑛 = 2−𝑛 𝑢[𝑛] and the
initial conditions 𝑦 −1 = 11/6, 𝑦 −2 = 37/36.

45
1. Solution of Difference Equations
Solution 9:
• Substituting 𝑛 − 2 for 𝑛 in the question, we obtain
𝑦 𝑛 − 5𝑦 𝑛 − 1 + 6𝑦 𝑛 − 2 = 3𝑥 𝑛 − 1 + 5𝑥 𝑛 − 2 .
Note that the input sequence 𝑥 𝑛 = 2−𝑛 𝑢[𝑛] is causal, hence 𝑥 −2 =
𝑥 −1 = 0.
• Using the time-shifting property, on right-hand equation,
3𝑥 𝑛 − 1 + 5𝑥 𝑛 − 2 ← 𝓏 → 3𝑧 −1 𝑋 𝑧 + 5𝑧 −2 𝑋 𝑧 .
• and
−𝑛 𝑛
1
𝑥 𝑛 = 2 𝑢 𝑛 = 0.5 𝑢 𝑛 ← 𝓏 → 𝑋 𝑧 =
1 − 0.5𝑧 −1
• Then,
−1
1 −2
1
3𝑥 𝑛 − 1 + 5𝑥 𝑛 − 2 ← 𝓏 → 3𝑧 −1
+ 5𝑧 −1
1 − 0.5𝑧 1 − 0.5𝑧
3𝑧 −1 + 5𝑧 −2
=
1 − 0.5𝑧 −1 46
1. Solution of Difference Equations
Solution 9: (continued)
• Since initial condition, 𝑦[−1] and 𝑦[−2] are non-zero, the output is non-causal.
• Using time-shifting property, z-transform left-hand side is
−1
11 −2
11 −1 37
𝑦 𝑛 − 5𝑦 𝑛 − 1 + 6𝑦 𝑛 − 2 ← 𝓏 → 𝑌 𝑧 − 5 𝑧 𝑌 𝑧 + +6 𝑧 𝑌 𝑧 + 𝑧 +
6 6 36
−1 −2 −1
= 𝑌 𝑧 1 − 5𝑧 − 6𝑧 + 11𝑧 − 3.
• Thus we get
3𝑧 −1 + 5𝑧 −2
𝑌 𝑧 1 − 5𝑧 −1 − 6𝑧 −2 + 11𝑧 −1 − 3 =
−1 −1 0.5𝑧 −1
1 −−2
3 − 11𝑧 3𝑧 + 5𝑧
𝑌 𝑧 = +
1 − 5𝑧 −1 + 6𝑧 −2 (1 − 0.5𝑧 −1 )(1 − 5𝑧 −1 + 6𝑧 −2 )
3 − 11𝑧 −1 1 − 0.5𝑧 −1 + 3𝑧 −1 + 5𝑧 −2
=
1 − 0.5𝑧 −1 1 − 5𝑧 −1 + 6𝑧 −2
3 − 9.5𝑧 −1 + 10.5𝑧 −2
=
1 − 0.5𝑧 −1 1 − 2𝑧 −1 1 − 3𝑧 −1
• Using partial fraction expansion, 𝑌(𝑧)
26 1 7 1 18 1
𝑌 𝑧 = × − × + × .
15 1 − 0.5𝑧 −1 3 1 − 2𝑧 −1 5 1 − 3𝑧 −1
• The inverse transform,
26 7 18
𝑦 𝑛 = 0.5𝑛 − 2𝑛 + 3𝑛 𝑛>0
15 3 5
47
2. 𝑍 −transfer function of LTID systems
• An alternative representation for an LTID system is obtained by taking the z-
transform of impulse response:
ℎ 𝑛 ←𝓏→𝐻 𝑧 .
• The 𝐻(𝑧) can be used to determine the output 𝑦 𝑛 of an LTID system.
𝑦 𝑛 =𝑥 𝑛 ∗ℎ 𝑛
Thus,
𝑌 𝑧 =𝑋 𝑧 𝐻 𝑧 .
Then,
𝑌 𝑧
𝐻 𝑧 = .
𝑋 𝑧
Let consider the difference equation of LTID system given as
𝑦 𝑛 + 𝑘 + 𝑎𝑘−1 𝑦 𝑛 + 𝑘 − 1 + ⋯ + 𝑎0 𝑦 𝑛
= 𝑏𝑚 𝑥 𝑛 + 𝑚 + 𝑏𝑚−1 𝑥 𝑛 + 𝑚 − 1 + ⋯ + 𝑏0 𝑥 𝑛 .
• 𝑘-number of output coefficients
• 𝑚-number of input coefficients. 48
2. 𝑍 −transfer function of LTID systems
• The z-transform of both sides, we obtain,
𝑧 𝑘 + 𝑎𝑘−1 𝑧 𝑘−1 + ⋯ + 𝑎0 𝑌 𝑧
= 𝑏𝑚 𝑧 𝑚 + 𝑏𝑚−1 𝑧 𝑚−1 + ⋯ + 𝑏0 𝑋(𝑧)
• Which reduces to following transfer function
𝑌 𝑧 𝑏𝑚 𝑧 𝑚 + 𝑏𝑚−1 𝑧 𝑚−1 + ⋯ + 𝑏0
𝐻 𝑧 = =
𝑋 𝑧 𝑧 𝑘 + 𝑎𝑘−1 𝑧 𝑘−1 + ⋯ + 𝑎0
Or
−1 −𝑚
𝑚−𝑘 𝑚
𝑏 + 𝑏𝑚−1 𝑧 + ⋯ + 𝑏0 𝑧
𝐻 𝑧 =𝑧
1 + 𝑎𝑘−1 𝑧 −1 + ⋯ + 𝑎0 𝑧 −𝑘

49
2. Z-transfer function of LTID systems

Characteristic equation, poles and zeros

• Characteristic equation
• The characteristic of equation of transfer function is
𝐷 𝑧 = 𝑎𝑘 𝑧 𝑘 + 𝑎𝑘−1 𝑧 𝑘−1 + ⋯ + 𝑎0 = 0
• Zeros
• 𝐻 𝑧 = 0, finite location in the complex z-plane.
• The zeros location can be obtained by solving,
𝑁 𝑧 = 𝑏𝑚 𝑧 𝑚 + 𝑏𝑚−1 𝑧 𝑚−1 + ⋯ + 𝑏0 = 0
• Poles
• |𝐻(𝑧)| has an infinite value.
• Solving, 𝐷 𝑧 = 𝑎𝑘 𝑧 𝑘 + 𝑎𝑘−1 𝑧 𝑘−1 + ⋯ + 𝑎0 = 0
• To calculate the zeros and poles, a transfer function is factorized as
𝑁 𝑧 𝑧 − 𝑧1 𝑧 − 𝑧2 … (𝑧 − 𝑧𝑚 )
𝐻 𝑧 = =
𝐷 𝑧 𝑧 − 𝑝1 𝑧 − 𝑝2 … (𝑧 − 𝑝𝑘 )
• Or
1 − 𝑧 𝑧 −1 1 − 𝑧 𝑧 −1 … (1 − 𝑧 𝑧 −1 )
1 2 𝑚
𝐻 𝑧 = 𝑧 𝑚−𝑘 .
1 − 𝑝1 𝑧 −1 1 − 𝑝2 𝑧 −1 … (1 − 𝑝𝑘 𝑧 −1 ) 50
2. Z-transfer function of LTID systems

Characteristic equation, poles and zeros

Example 10:
Determine the poles and zeros of the following LTID systems:
𝑧
i) 𝐻1 𝑧 = 2
𝑧 −3𝑧+2
;
1
ii) 𝐻2 𝑧 = (𝑧−0.1)(𝑧−0.5)(𝑧+0.2)
;
𝑧 2 2𝑧−1.5
iii) 𝐻3 𝑧 = 𝑧+0.4 𝑧−0.5 2
;
𝑧 2 +0.7𝑧+1.6
iv) 𝐻4 𝑧 = (𝑧 2 −1.2𝑧+1)(𝑧+0.3)
.

51
2. Z-transfer function of LTID systems

Characteristic equation, poles and zeros

Solution 10:
𝑧
i) 𝐻1 𝑧 = 𝑧 2 −3𝑧+2
𝑧
𝐻1 𝑧 = 2
𝑧 𝑧− 3𝑧 + 2
Im 𝑧
=
𝑧−1 𝑧−2
∴ 𝑧 = 0, one zero and
𝑧 = 1, 𝑧 = 2, two poles.
Re 𝑧

52
2. Z-transfer function of LTID systems

Characteristic equation, poles and zeros

1
ii) 𝐻2 𝑧 =
(𝑧−0.1)(𝑧−0.5)(𝑧+0.2)
1
𝐻2 𝑧 =
(𝑧 − 0.1)(𝑧 − 0.5)(𝑧 + 0.2)
∴no zero, but three poles, 𝑧 = 0.1, 0.5, −0.2. Im 𝑧

Re 𝑧

53
2. Z-transfer function of LTID systems

Characteristic equation, poles and zeros

Solution 10:
𝑧 2 2𝑧−1.5
iii. 𝐻3 𝑧 = 𝑧+0.4 𝑧−0.5 2
;

𝑧 2 2𝑧 − 1.5 Im 𝑧
𝐻3 𝑧 = 2
𝑧 + 0.4 𝑧 − 0.5

∴ 𝑧 = 0,0,0.75, three zeros and

𝑧 = −0.4, 𝑧 = 0.5,0.5, three poles. Re 𝑧

54
2. Z-transfer function of LTID systems

Characteristic equation, poles and zeros

Solution 10:
𝑧 2 +0.7𝑧+1.6
iv. 𝐻4 𝑧 = (𝑧 2 −1.2𝑧+1)(𝑧+0.3)
𝑧 2 + 0.7𝑧 + 1.6
𝐻4 𝑧 = 2
(𝑧 − 1.2𝑧 + 1)(𝑧 + 0.3) Im 𝑧
𝑧 − 0.5 𝑧 + 1.2
=
(𝑧 − 0.6 + 𝑗0.8)(𝑧 − 0.6 − 𝑗0.8)(𝑧 + 0.3)
∴ 𝑧 = 0.5, −1.2 two zeros and
𝑧 = 0.6 − 𝑗0.8, 0.6 + 𝑗0.8, −0.3, three poles.
Re 𝑧

55
2. Z-transfer function of LTID systems

Determine of impulse response

Example 11:
The input-output relationship of an LTID system is given by the
following difference equation:
3 1
𝑦 𝑛 + 2 − 𝑦 𝑛 + 1 + 𝑦 𝑛 = 2𝑥 𝑛 + 2 .
4 8
Determine the transfer function and the impulse response of the
system.

56
2. Z-transfer function of LTID systems

Determine of impulse response

Solution 11:
Substituting 𝑚 = 𝑛 + 2, 𝑠𝑜 𝑛 = 𝑚 − 2
3 1
𝑦 𝑚 − 𝑦 𝑚 − 1 + 𝑦 𝑚 − 2 = 2𝑥 𝑚
4 8
Z-transform for both side
3 1
𝑌 𝑧 − 𝑧 −1 𝑌 𝑧 + 𝑧 −2 𝑌 𝑧 = 2𝑋 𝑧 ,
4 8
Thus the transfer function
𝑌 𝑧 2
𝐻 𝑧 = = .
𝑋 𝑧 3 1
1 − 𝑧 −1 + 𝑧 −2
4 8
Consider the partial fraction expansion of 𝐻(𝑧) as
2 4 2
𝐻 𝑧 = = − .
3 −1 1 −2 1 −1 1 −1
1− 𝑧 + 𝑧 1− 𝑧 1− 𝑧
4 8𝑛 2 𝑛 4
1 1
ℎ 𝑛 =4 𝑢 𝑛 −2 𝑢𝑛
2 4
57
Summary
✓𝑧 −transform
• generalized form of discrete time Fourier Transform
• 𝑧 = 𝑟𝑒 𝑗Ω
✓Region of convergence (ROC)
• Range value of the complex variable 𝑧, for which z-transform converge.
✓𝑧 −transform Properties
✓Inverse 𝑧 −transform
✓Application of 𝑧 −transform

58