CHAPTER-3

The 𝑍-Transform

Addisalem H.
December 25, 2020

Topics:
𝑍-transform, Region of Convergence, Properties of 𝑍-transform, Inverse 𝑍-
transform, 𝑍-transforms and LTI systems.
𝑍 -Transform

– The 𝑍- transform is a very important tool for describing and analyzing discrete
time systems.

– You can think of the 𝑍-transform as simply an alternative way of expressing

discrete time signals.

– It plays the same role in the analysis of discrete time signals and linear time
invariant systems as the Laplace transform does in the analysis of continuous
time signals and linear time invariant systems.

1
𝑍 -Transform

Definition
The 𝑍-transform of a discrete time signal 𝑋 [𝑛] is defined as:
∞
∑︁
𝑋 (𝑧) = 𝑥 [𝑛]𝑧 −𝑛
𝑛=−∞

Where 𝑧 is a complex variable.

𝑥 [𝑛] ⇐⇒ 𝑋 (𝑧)
 𝑛
Example: The 𝑧 transform of 𝑥 [𝑛] = 12 𝑢[𝑛] is
∞  2 ∞  𝑛
∑︁ 1 −𝑛
∑︁ 1 −1 1
𝑋 (𝑧) = 𝑢[𝑛]𝑧 = 𝑧 = , |𝑧| > 0.5
𝑛=−∞
2 𝑛=0
2 1 − 0.5𝑧 −1

2
𝑍 -Transform

x[n] Im(z)

X(z)
2 3 4
Direct z-transform We may obtain the Fourier
-1 0 1 5 6 n Re(z)
transform from the 𝑍-transform by
Inverse z-transform
making the substitution 𝑧 = 𝑒 𝑗 𝜔
Time Domain Signal
Complex Signal representation

For |𝑧| = 1, the 𝑍-transform corresponds to the Fourier transform.

Also with 𝑧 = 𝑟𝑒 𝑗 𝜔
∞
∑︁
𝑋 (𝑟𝑒 𝑗 𝑤 ) = 𝑥 [𝑛] (𝑟𝑒 𝑗 𝜔 ) −𝑛
𝑛=−∞
That is, the 𝑧-transform is the Fourier transform of the sequence 𝑥 [𝑛]𝑟 −𝑛 , for 𝑟 = 1 this
becomes the Fourier transform of 𝑥 [𝑛].
The Fourier transform therefore corresponds to the 𝑧-transform evaluated on the unit circle.
3
Region of Convergence(ROC)

– The ROC defines the values of 𝑧 for which the 𝑍-transform sum will converge.
– The 𝑍-transform is therefore, uniquely characterized by:
∗ Expression of 𝑋 (𝑧)
∗ ROC of 𝑋 (𝑧)
– The 𝑗𝜔-axis in the 𝑠-plane maps to the unit circle in the 𝑧-plane using the
transformation 𝑧 = 𝑒 𝑗 𝜔 .
s-plane Im z-plane Im Just fold the j⍵-axis s-plane z-plane
into a circle

Re Re

s-plane z-plane

Since the ROC’s for continuous-time signals look like

half-planes or strips, the ROC’s in the 𝑧-plane
generally look like circles, holes, or donuts. 4
Region of Convergence(ROC)

Rules for the ROC

− The ROC never contains any poles (the values of 𝑧 that makes 𝐻 (𝑧) = ∞)
− If 𝑥 [𝑛] is right sided, the ROC is the area outside a circle, i.e. |𝑧| > 𝑎, where 𝑎
is the |outermost pole|.
− If 𝑥 [𝑛] is left sided, the ROC is the area inside of a circle, i.e. |𝑧| < 𝑎, where 𝑎
is the |innermost pole|.
− If 𝑥 [𝑛] is two-sided or the sum of a left and right sided signal, the ROC is either
a donuts (𝑎 < |𝑧| < 𝑏), or else the individual ROC’s will not overlap, producing
the null set.
− If 𝑥 [𝑛] is of finite duration, then the rock is the entire 𝑧-plane, except possibly
𝑧 = 0 and 𝑧 = ∞.
5
Region of Convergence(ROC)

Example: Determine the 𝑍-transform of the following sequence

𝒙[𝒏] = {1, 2, 3, 4, 5, 0, 7}
Solution: The samples of 𝑥 [𝑛] are;
𝑥 [0] = 1, 𝑥 [1] = 2, 𝑥 [2] = 3, 𝑥 [3] = 4, 𝑥 [4] = 5, 𝑥 [5] = 0, 𝑥 [6] = 7
The 𝑍-transform of this sequence can be obtained as
∞
∑︁
𝑋 (𝑧) = 𝑥 [𝑛]𝑧 −𝑛
𝑛=−∞
𝑥 [𝑛] has the value from 𝑛 = 0 to 𝑛 = 6. Hence the above equation becomes,
6
∑︁
𝑋 (𝑧) = 𝑥 [𝑛]𝑧 −𝑛
𝑛=0
Putting values of 𝑥 [𝑛]
𝑋 (𝑧) = 1 + 2𝑧 −1 + 3𝑧 −2 + 4𝑧 −3 + 5𝑧 −4 + 0𝑧 −5 + 2𝑧 −1 + 7𝑧 −6
6
Region of Convergence(ROC)

=⇒ 𝑋 (𝑧) = 1 + 2𝑧 + 𝑧32 + 𝑧43 + 𝑧54 + 𝑧76

𝑋 (𝑧) has finite values except at 𝑧 = 0 (𝑋 (𝑧) is convergent for all values of 𝑧 ), Hence the
ROC is the entire 𝑧-plane except 𝑧 = 0.
Exercise: Determine the 𝑍-transform, including the ROC, for the following signals.
1. 𝑥 [𝑛] = {2, 4, 5, 7, 6, 5}
↑
2. 𝑥 [𝑛] = 𝑎 𝑛 𝑢[𝑛] + 𝑏 𝑛 𝑢[−𝑛 − 1]
 𝑛
3. 𝑥 [𝑛] = 12 𝑢[𝑛] − 2𝑛 𝑢[−𝑛 − 1]
 𝑛  𝑛
4. 𝑥 [𝑛] = 12 𝑢[𝑛] + − 13 𝑢[𝑛]
 𝑛  𝑛
5. 𝑥 [𝑛] = − 31 𝑢[𝑛] − 21 𝑢[−𝑛 − 1]
 𝑛
6. 𝑥 [𝑛] = 12 (𝑢[𝑛] − 𝑢[𝑛 − 10])
7. 𝑥 [𝑛] = (−1) 𝑛 (𝑐𝑜𝑠 𝜋3 𝑛)𝑢[𝑛] 7
𝑍 -transform Pairs

𝒙[𝒏] 𝑿 (𝒛) ROC

𝛿[𝑛] 1 all 𝑧
1
𝑢[𝑛] 1−𝑧 −1
|𝑧| > 1
1
𝑎 𝑛 𝑢[𝑛] 1−𝑎𝑧 −1
|𝑧| > |𝑎|
1
−𝑎 𝑛 𝑢[−𝑛 − 1] 1−𝑎𝑧 −1
|𝑧| < |𝑎|
𝑎𝑧 −1
𝑛𝑎 𝑛 𝑢[𝑛] (1−𝑎𝑧 −1 ) 2
|𝑧| > |𝑎|
𝑎𝑧 −1
−𝑛𝑎 𝑛 𝑢[−𝑛 − 1] (1−𝑎𝑧 −1 ) 2
|𝑧| < |𝑎|
1−𝑐𝑜𝑠 ( 𝜔0 ) 𝑧 −1
𝑐𝑜𝑠(𝜔0 𝑛)𝑢[𝑛] 1−2𝑐𝑜𝑠 ( 𝜔0 ) 𝑧 −1 +𝑧 −2
|𝑧| > 1
𝑠𝑖𝑛( 𝜔0 ) 𝑧 −1
𝑠𝑖𝑛(𝜔0 𝑛)𝑢[𝑛] 1−2𝑐𝑜𝑠 ( 𝜔0 ) 𝑧 −1 +𝑧 −2
|𝑧| > 1 8
𝑍 -transform Properties

Property Time Domain 𝑍-Domain New ROC

Linearity 𝑎𝑥 [𝑛] + 𝑏𝑦[𝑛] 𝑎𝑋 (𝑧) + 𝑏𝑌 (𝑧) ROC ⊇ ROC(𝑥)∩ROC(𝑦)

Time shift 𝑥 [𝑛 − 𝑘] 𝑧 −𝑘 𝑋 (𝑧) ROC(𝑥)(but watch 𝑧 = 0, ∞)

Exponential Scaling 𝑎 𝑛 𝑥 [𝑛] 𝑋 (𝑧/𝑎) {|𝑎|.𝑧} s.t. 𝑧 ∈ ROC(𝑥)

𝑑
Linear scaling 𝑛𝑥 [𝑛] −𝑧 𝑑𝑧 𝑋 (𝑧) ROC(𝑥)(but watch 𝑧 = 0)

Time Reversal 𝑥 [−𝑛] 𝑋 (𝑧 −1 ) {1/𝑧} s.t. 𝑧 ∈ ROC(𝑥)

Convolution 𝑥 [𝑛] ∗ 𝑦[𝑛] 𝑋 (𝑧)𝑌 (𝑧) At least ROC(𝑥)∩ROC(𝑦)

Conjugation 𝑥 ∗ [𝑛] 𝑋 ∗ (𝑧 ∗ ) ROC(𝑥)

9
The Inverse 𝑍 -transform

The general formula for recovering 𝑥 [𝑛] from 𝑋 (𝑧) is the complex contour integral:
∮
𝑥 [𝑛] = 2 𝜋1 𝑗 𝑋 (𝑧)𝑧 𝑛−1 𝑑𝑧 (DO NOT USE THIS FORMULA!!!)

Rather, to perform the inverse 𝑍-transform, we will merely manipulate the given expression until we
see patterns we recognize from the 𝑍-transform table.
Note: Knowing the ROC is critical to performing an inverse 𝑍-transform.
(
1 𝑎 𝑛 𝑢[𝑛], ROC: |𝑧| > |𝑎|
For Example: 𝑋 (𝑧) = 1−𝑎𝑧 −1 −→ 𝑥 [𝑛] =
−𝑎 𝑢[−𝑛 − 1], ROC: |𝑧| < |𝑎|
𝑛

Is 𝒙[𝒏] 𝒂 𝒏 𝒖[𝒏] or −𝒂 𝒏 𝒖[−𝒏 − 1]?

If it is known that the ROC is |𝑧| > |𝑎|, then we know that 𝑥 [𝑛] = 𝑎 𝑛 𝑢[𝑛] or equivalently we could
know that 𝑥 [𝑛] is right sided(or causal).

10
Methods of evaluation for the Inverse 𝑍 -transform

Inspection Method

This method consists of becoming familiar with, or recognizing “by inspection,” certain
transform pairs.
For example, if we need to find the inverse 𝑍-transform of
1 1
𝑋 (𝑧) = 1 −1
, |𝑧| >
1− 2𝑧
2

we recall the 𝑍-transform

  𝑛 pairs, and we would recognize “by inspection” the associate
sequence as 𝑥 [𝑛] = 12 𝑢[𝑛].
Exercise: What would be the sequence if the ROC associated with 𝑋 (𝑧) is |𝑧| < 12 ?

11
Methods of evaluation for the Inverse 𝑍 -transform

Partial Fraction Expansion

If 𝑋 (𝑧) is not given explicitly in an available table, it is possible to obtain an alternative

expression for 𝑋 (𝑧) as a sum of simpler terms, each of which is tabulated.
𝑀
𝑏 𝑘 𝑧 −𝑘
Í

Assume 𝑋 (𝑧) is expressed as a ratio of polynomials in 𝑧 −1 ; i.e., 𝑋 (𝑧) = 𝑘=0

𝑁
𝑎 𝑘 𝑧 −𝑘
Í
𝑘=0
𝑀
(1−𝑐𝑘 𝑧 −1 )
Î
𝑏0
𝑋 (𝑧) could be expressed in the form, 𝑋 (𝑧) = 𝑘=1
𝑁
(1−𝑑𝑘 𝑧 −1 )
Î
𝑎0
𝑘=1
Where the 𝑐 𝑘 s are the nonzero zeros of 𝑋 (𝑧) and the 𝑑 𝑘 s are the nonzero poles of 𝑋 (𝑧).
𝑁
Í 𝐴𝑘
If 𝑀 < 𝑁 and the poles are all 1𝑠𝑡 -order, then 𝑋 (𝑧) = 1−𝑑 𝑧 −1
.
𝑘
𝑘=1
Multiplying the above equation by 1 − 𝑑 𝑘 𝑧 −1
and evaluating for 𝑧 = 𝑑 𝑘 shows that the
coefficients,𝐴 𝑘 , can be found from 𝐴 𝑘 = (1 − 𝑑 𝑘 𝑧 −1 ) 𝑋 (𝑧)| 𝑧=𝑑𝑘
12
Methods of evaluation for the Inverse 𝑍 -transform

Partial Fraction Expansion

1
Example: 𝑋 (𝑧) = (1− 41 𝑧 −1 ) (1− 12 𝑧 −1 )
, |𝑧| > 12 , Find 𝑥 [𝑛].
𝐴1 𝐴2
Solution: 𝑋 (𝑧) can be expressed as 𝑋 (𝑧) = 1− 14 𝑧 −1
+ 1− 12 𝑧 −1
(1− 14 𝑧 −1 )
𝐴1 = (1 − 41 𝑧 −1 ) 𝑋 (𝑧) 𝑧= 14
= (1− 4 𝑧 ) (1− 12 𝑧 −1 )
1 −1 = −1
𝑧= 14
(1− 12 𝑧 −1 )
𝐴2 = (1 − 12 𝑧 −1 ) 𝑋 (𝑧) 𝑧= 12
= (1− 4 𝑧 ) (1− 12 𝑧 −1 )
1 −1 =2
𝑧= 12
−1 2
=⇒ 𝑋 (𝑧) = 1− 14 𝑧 −1
+ 1− 12 𝑧 −1
From the 𝑍-transform pairs and linearity property of 𝑍-transform, it then follows that

𝒏
𝒏
𝒙[𝒏] = 2 21 𝒖[𝒏] − 14 𝒖[𝒏]
The ROC extends outward from the outermost pole, since 𝑥 [𝑛] is right sided.
13
Methods of evaluation for the Inverse 𝑍 -transform

Power Series Expansion

The defining expression for the 𝑍-transform is a Laurent series where the sequence values
𝑥 [𝑛] are the coefficients of 𝑧−𝑛 .
If the 𝑍-transform is given as a power series in the form
∞
𝑥 [𝑛]𝑧 −𝑛 = ... + 𝑥 [−2]𝑧 2 + 𝑥 [−1]𝑧 + 𝑥 [0] + 𝑥 [1]𝑧 −1 𝑥 [2]𝑧 −2 + ...
Í
𝑋 (𝑧) =
𝑛=−∞
Any particular value of the sequence can be determined by finding the coefficient of the
power of 𝑧 −1 .  
Example: 𝑋 (𝑧) = 𝑧 2 1 − 21 𝑧 −1 (1 + 𝑧 −1 ) (1 − 𝑧−1 ), Find 𝑥 [𝑛].
Solution:𝑋 (𝑧) can be expressed as
1 1
𝑋 (𝑧) = 𝑧 2 − 𝑧 − 1 + 𝑧 −1
2 2
14
Methods of evaluation for the Inverse 𝑍 -transform

Power Series Expansion

Therefore by inspection, 𝑥 [𝑛] is seen to be

1, 𝑛 = −2





1


 , 𝑛 = −1
2



𝑥 [𝑛] = −1, 𝑛=0


 1,

𝑛=1


 2

 0,
 otherwise

Equivalently,
𝒙[𝒏] = 𝜹[𝒏 + 2] − 21 𝜹[𝒏 + 1] − 𝜹[𝒏] + 12 𝜹[𝒏 − 1]

15
Exercise

1. Determine the 𝑍-transform and the ROC of the signal 𝑥 [𝑛] = 𝑢[𝑛] − (0.5) 𝑛 𝑢[𝑛]

2. Determine the 𝑍-transform and the ROC of the signal 𝑥 [𝑛] = [3(2) 𝑛 − 4(3) 𝑛 ]𝑢[𝑛].

3. Determine 𝑍-transform of the signal 𝑥 1 [𝑛] = 𝑥 [𝑛 − 2], where 𝑥 [𝑛] = 𝑎 𝑛 .

4. Determine the 𝑍-transform and the ROC of the signal 𝑥 [𝑛] = 𝑛2 𝑎 𝑛 𝑢[𝑛].

5. Compute the convolution 𝑦[𝑛] of signals

𝑥 1 [𝑛] = {1, −2, 1}

 1, 0 ≤ 𝑁 ≤ 5


𝑥 2 [𝑛] =
 0, elsewhere



16
Exercise

6. Find 𝑥 [𝑛] with the 𝑍-transform 𝑋 (𝑧) = 𝑧−1 + 3 − 2𝑧.

1−2𝑧 −1 +3𝑧 −2
7. 𝑋 (𝑧) = 1−𝑧 −1
, Find 𝑥 [𝑛], given that it is right sided.

1 1
8. 𝑋 (𝑧) = 1− 53 𝑧 −1 − 23 𝑧 −2
, ROC: 3 < |𝑧| < 2, Find 𝑥 [𝑛].

1
9. 𝑋 (𝑧) = 1−1.5𝑧 −1 +0.5𝑧 −2
, Find 𝑥 [𝑛].
1
10. Determine the partial fraction expansion of 𝑋 (𝑧) = (1+𝑧 −1 ) (1−𝑧 −1 ) 2
.

1+2𝑧 −1 +𝑧 −2
11. 𝑋 (𝑧) = 1− 32 𝑧 −1 + 12 𝑧 −2
, Determine 𝑥 [𝑛].

17
𝑍 -transforms and LTI systems

Recall that an LTI system can be represented as the convolution 𝑦[𝑛] = 𝑥 [𝑛] ∗ ℎ[𝑛] of the
input 𝑥 [𝑛] with ℎ[𝑛], where ℎ[𝑛] is the response of the system to the unit impulse
sequence 𝛿[𝑛].
From the convolution property, it follows that the 𝑍-transform of 𝑦[𝑛] is
𝑌 (𝑧) = 𝐻 (𝑧) 𝑋 (𝑧)
Where 𝐻 (𝑧) and 𝑋 (𝑧) are the 𝑍-transforms of ℎ[𝑛] and 𝑥 [𝑛] respectively.
Example: Use the 𝑍-transform to evaluate the convolution 𝑦[𝑛] = 𝑥 [𝑛] ∗ ℎ[𝑛], Where,
𝑥 [𝑛] = 𝐴𝑢[𝑛] and ℎ[𝑛] = 𝑎 𝑛 𝑢[𝑛].
Solution: Begin by finding the corresponding 𝑍-transforms as
∞
1
𝑎 𝑛 𝑧 −𝑛 = 1−𝑎𝑧
Í
𝐻 (𝑧) = −1 , |𝑧| > |𝑎|.
𝑛=0
and
∞
𝐴𝑧−𝑛 =
Í 𝐴
𝑋 (𝑧) = 1−𝑧 −1
|𝑧| > 1
𝑛=0 18
𝑍 -transforms and LTI systems

The 𝑍-transform of the convolution 𝑦[𝑛] = 𝑥 [𝑛] ∗ ℎ[𝑛] is therefore

𝐴 𝐴𝑧 2
𝑌 (𝑧) = = (𝑧−𝑎) (𝑧−1) , |𝑧| > 1
( 1−𝑎𝑧 −1 ) ( 1−𝑧 −1 )
Where we assume that |𝑎| < 1 so that the overlap of the ROC is |𝑧| > 1.
The sequence 𝑦[𝑛] can be obtained by determining the inverse 𝑍-transform using partial
fraction expansion of 𝑌 (𝑧).
 
𝐴 1 𝑎
𝑌 (𝑧) = 1−𝑎 1−𝑧 −1 − 1−𝑎𝑧 −1

∴ Taking the inverse 𝑍-transform of each term yields

𝐴


𝑦[𝑛] = 1−𝑎 1 − 𝑎 𝑛+1 𝑢[𝑛]

19