11 ZTransform Part 01

Outline Introduction z-Transform Properties
Analytical Methods for Multivariable and

Discrete-Time Systems
The z-Transform: Part 1
Dr. Adam Panagos
Dr. Adam Panagos

Analytical Methods for Multivariable and Discrete-Time Systems
1 Introduction
2 z-Transform
Preliminaries
Representation
Other Details
Region of Convergence
3 Properties
Dr. Adam Panagos

Introduction
In This Chapter We Introduce The z-Transform

The z-Transform Is The Discrete-Time Counterpart to the
Laplace Transform
The Fourier Transform Had Limitations We Overcame with
the Laplace Transform
The DTFT Has Limitations We’ll Overcome with the
z-Transform
Dr. Adam Panagos

Introduction
We’ll Be Interested In Many of the Same Types Of

Computations/Operations, But Now With the z-Transform:
Transfer Function H(z)
Input/Output Relationship Y (z) = H(z)X (z)
System Characteristics from H(z)
Solving Difference Equations
Dr. Adam Panagos

Preliminaries
Complex Exponential Signal
z = re jΩ Is A Complex Number of Magnitude r and Angle Ω

The Signal x[k] = z k Is Complex Exponential Signal
x[k] = z k = r k e jΩk = r k cos(Ωk) + jr k sin(Ωk)
Real/Imaginary Part of x[k] Is A Damped Cosine/Sine

The Magnitude r Determines the Damping Factor
The Angle Ω Determines The Frequency Of The Cosine/Sine
Dr. Adam Panagos

Preliminaries
Complex Exponential Signal

z = rejΩ
Re(z) 2
−2
−5 0 5 10 15
k
2
Im(z)
−2
−5 0 5 10 15
k
Figure : Real and Imaginary Parts of z k

Dr. Adam Panagos
Preliminaries
Eigenfunction Property
Apply x[k] = z k To An LTI System With Impulse Response

h[k]
y [k] = h[k] ∗ x[k]

∞
X
= h[m]x[k − m]
m=−∞
∞
X
= h[m]z k−m
m=−∞
∞
!
X
= zk h[m]z −m
m=−∞
Dr. Adam Panagos

Preliminaries
Definition
The Transfer Function of the discrete-time LTI system with
impulse response h[k] is
X∞
H(z) = h[k]z −k
k=−∞
Dr. Adam Panagos

Preliminaries
Continuing Math From The Previous Page Yields

∞
X
y [k] = z k h[m]z −m
m=−∞
k
= H(z)z
So, An Input of z k Gives The Same Output Scaled By The

Transfer Function, i.e. H(z)z k
z k Is An Eigenfunction of the System with Associated
Eigenvalue H(z)
Dr. Adam Panagos

Preliminaries
Transfer Function
The Transfer Function H(z) Is A Complex Valued Quantity

Written in Polar Form as H(z) = |H(z)|e jφ(z)
An Input Of x[k] = z k Is Modified In Amplitude And Phase
Amplitude Is Changed By |H(z)| = |H(re jΩ )|
Phase Is Changed By ∠H(z) = ∠H(re jΩ )
System Does Not Change Ω
Dr. Adam Panagos

Representation
z-Transform Representation
Text Development
The z-Transform
Your Textbook Derives The Inverse z-Transform Equation

Stated on the Following Chart
Dr. Adam Panagos

Representation
Definition
The Bilateral z-Transform of the discrete-time signal x[k] is
X∞
X (z) = x[k]z −k
k=−∞
Definition
The Inverse Bilateral z-Transform
I of X (z) is
1
x[k] = X (z)z k−1 dz
2πj
Dr. Adam Panagos

Representation
z
We Use The Notation x[k] ⇐⇒ X (z)
We Use Bilateral z-Transform For “General“ Signals
Can Use Simpler Unilateral z-Transform For Causal Signals
Direct Evaluation Of The Inverse z-Transform Is Difficult
(need complex-variable theory for contour integral)
We’ll Discuss Other Simpler Ways To Do This For A
Relatively Large Class of Signals
Dr. Adam Panagos

Other Details
Convergence
z-Transform Exists When Infinite Sum Converges

Necessary Condition for This to Occur Is Absolute Summable
x[k]z −k
We Have |x[k]z −k | = |x[k]r −k e −jΩk | = |x[k]r −k |, So Require
∞
X
|x[k]r −k | < ∞
k=−∞
Values Of r For Which This Occurs Is Called the Region of

Convergence
Dr. Adam Panagos

Other Details
Convergence
x[k]=αku[k], α > 1
10
5
0
−5 0 5 10
r−k
20
10
0
−5 0 5 10
x[k]r−k, r > α
1.5
1
0.5
0
−0.5
−5 0 5 10
Figure : DTFT of x[k] Does Not Exist, But z-Transform Does For r > α
Dr. Adam Panagos

Other Details
The z-Plane
Notes Development
The z-Plane
If x[k] Is Absolutely Summable We Can Get X (Ω) From X (z)

By Letting r = 1
For r = 1, z = e jΩ Is A Unit Radius Circle
The DTFT Is Given By z-Transform Evaluated on the Unit
Circle
Dr. Adam Panagos

Other Details
Examples
Notes Example
z-Transform Example #1
Notes Example
Notes Example
In General, Must Specify X (z) and ROC For Unique

Correspondance To A Signal x[k]
Dr. Adam Panagos
The ROC
Let’s Take A Closer Look At the Region Of Convergence

Try and Show How The ROC Is Based On Signal
Characteristics
Recall, The ROC Is The z-Plane Region Where The
z-Transform X (z) Exists/Converges
Dr. Adam Panagos

Poles
If d Is A Pole of X (z), Then |X (d)| = ∞ and Thus X (z)

Does Not Converge At z = d
The Region Of Convergence Cannot Contain Any Poles
Dr. Adam Panagos

Finite Duration Signals
Assume x[k] Is Non-Zero On Interval k1 ≤ k ≤ k2

z-Transform Is Written As
k2
X
X (z) = x[k]z −k
k=k1
This Sum Converges For All z (except z = 0, |z| = ∞, or

both) As Long as Each Term is Finite
Dr. Adam Panagos

Finite Duration Signals
If Signal Is Causal (k2 > 0), X (z) Contains z −1 , z −2 , . . .

Terms → ROC Cannot Include z = 0
If Signal Is Non-Causal (k1 < 0), X (z) Contains z, z 2 , ...
Terms → ROC Cannot Include z = ∞
If Signal Has k2 < 0 ROC Can Include z = 0
If Signal Has k1 ≥ 0 ROC Can Include |z| = ∞
Dr. Adam Panagos

Infinite Duration Signals

Mathematically, Convergence Requires |X (z)| < ∞
We Can Bound |X (z)| As
∞
X
|X (z)| = x[k]z −k
k=−∞
X∞ ∞
X
≤ |x[k]z −k | = |x[k]||z −k |
k=−∞ k=−∞
−1
X ∞
X
= |x[k]||z −k | + |x[k]||z −k |
k=−∞ k=0
= S− (z) + S+ (z)
Dr. Adam Panagos
We’ve Shown That |X (z)| ≤ S− (z) + S+ (z)

If S− (z) and S+ (z) Are Both Finite, Then |X (z)| Is Finite
Dr. Adam Panagos

Suppose We Can Find Constants A+ , A− , r+ , and r− That

Satisfy
|x[k]| ≤ A− (r− )k , k < 0
|x[k]| ≤ A+ (r+ )k , k ≥ 0
Then The Signal x[k] Grows No Faster Than (r+ )k For
Positive k and (r− )k For Negative k
Dr. Adam Panagos

S− (z) Converges For |z| < r−

Proof:
−1
X
S− (z) = |x[k]||z −k |
k=−∞
−1
X
≤ A− (r− )k |z|−k
k=−∞
−1 ∞
r− k |z| n
X X
= A− = A−
|z| r−
k=−∞ n=1
Dr. Adam Panagos

S+ (z) Converges For |z| > r+ .

Proof:
∞
X
S+ (z) = |x[k]||z −k |
k=0
∞
X
≤ A+ (r+ )k |z|−k
k=0
∞ k
X r+
= A+
|z|
k=0
Dr. Adam Panagos

Conclusion
Definition
A Left-Sided Signal has x[k] = 0 for k ≥ 0, a Right-Sided Signal
has x[k] = 0 for k < 0, and a Two-Sided Signal extends infinitely
in both directions.
The ROC of a left-sided signal has the form |z| < r−

The ROC of a right-sided signal has the form |z| > r+
The ROC of a two-sided signal has the form r+ < |z| < r−
Dr. Adam Panagos

Conclusion
Figure : Time Extent and ROC Relationship
Dr. Adam Panagos

Example
Notes Example
ROC of Two-Sided Signal
Dr. Adam Panagos

Introduction
We Did Detailed Derivations When Discussing Properties Of

Fourier Representations
The Approach For Deriving Properties Of The z-Transform Is
Very Similar
We’ll Simply State The Properties, Many of These Should
Look Familiar/Similar
Dr. Adam Panagos

Introduction
The Following Properties Assume We Have

z
x[k] ⇐⇒ X (z)
with ROC Rx and

z
y [k] ⇐⇒ Y (z)
with ROC Ry .
Dr. Adam Panagos

Properties
Definition
The Linearity Property of the z-Transform States that
z
ax[k] + by [k] ⇐⇒ aX (z) + bY (z).
with ROC of at least Rx ∩ Ry
The ROC Can Be Larger Than Intersection of Original

Regions if Pole-Zero Cancellation Occurs
Dr. Adam Panagos

Properties
Definition
The Time Reversal Property of the z-Transform States that

z 1
x[−k] ⇐⇒ X
z
1
with ROC Rx
If Original ROC Was a < |z| < b, then ROC of Reversed

Signal is a < 1/|z| < b ⇒ 1/b < |z| < 1/a
Dr. Adam Panagos

Properties
Definition
The Time-Shift Property of the z-Transform States that
z
x[k − k0 ] ⇐⇒ z −k0 X (z)
with ROC Rx except possibly z = 0 or |z| = ∞
If k0 > 0, Introduces Poles at z = 0

If k0 < 0, Introduces Poles at Infinity
Unless Canceled By Zeros, Cannot Include These Points
Dr. Adam Panagos

Properties
Definition
The Exponential Sequence Property of the z-Transform States that
z
z
αk x[k] ⇐⇒ X
α
with ROC |α|Rx
If Rx is a < |z| < b, then new ROC is |α|a < |z| < |α|b
Poles And Zeros Of X (z) Have Their Radii Changed by |α|
and Angles Changed by ∠(α)
Dr. Adam Panagos

Properties
Figure : Impact of αn Multiplication on Poles/Zeros
Dr. Adam Panagos

Properties
Definition
The Convolution Property of the z-Transform States that
z
x[k] ∗ y [k] ⇐⇒ X (z)Y (z).
with ROC of at least Rx ∩ Ry
The ROC Can Be Larger Than Intersection if Pole-Zero

Cancellation Occurs
Dr. Adam Panagos

Properties
Definition
The z-Domain Differentation Property of the z-Transform States
that z d
kx[k] ⇐⇒ −z X (z)
dz
with ROC Rx
Multiplication by k In Time Domain Does Not Change ROC
Dr. Adam Panagos

Properties
Definition
The Initial Value Theorem of the z-Transform States that for
causal x[k]
x[0] = lim X (z)
z→∞
Proof:
∞
X
lim X (z) = lim x[k]z −k
z→∞ z→∞
k=0

x[1] x[2]
= lim x[0] + + 2 + ···
z→∞ z z
= x[0]
Dr. Adam Panagos
Properties
Definition
The Final Value Theorem of the z-Transform States that if
(z − 1)X (z) has no poles outside the unit circle, then
lim x[N] = lim (z − 1)X (z)
N→∞ z→1
Dr. Adam Panagos

Examples
Notes Development
Pole-Zero Cancellation
Notes Example
Using z-Transform Properties #1
Notes Example
Using z-Transform Properties #2
Dr. Adam Panagos


11 ZTransform Part 01

Uploaded by

Copyright:

Available Formats

11 ZTransform Part 01

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

11 ZTransform Part 01

Uploaded by

Copyright:

Available Formats

Outline Introduction z-Transform Properties

Analytical Methods for Multivariable and

Dr. Adam Panagos

Dr. Adam Panagos

Dr. Adam Panagos

In This Chapter We Introduce The z-Transform

Dr. Adam Panagos

We’ll Be Interested In Many of the Same Types Of

Dr. Adam Panagos

Complex Exponential Signal

z = re jΩ Is A Complex Number of Magnitude r and Angle Ω

x[k] = z k = r k e jΩk = r k cos(Ωk) + jr k sin(Ωk)

Real/Imaginary Part of x[k] Is A Damped Cosine/Sine

Dr. Adam Panagos

Complex Exponential Signal

Figure : Real and Imaginary Parts of z k

Apply x[k] = z k To An LTI System With Impulse Response

y [k] = h[k] ∗ x[k]

Dr. Adam Panagos

Dr. Adam Panagos

Continuing Math From The Previous Page Yields

So, An Input of z k Gives The Same Output Scaled By The

Dr. Adam Panagos

The Transfer Function H(z) Is A Complex Valued Quantity

Dr. Adam Panagos

Your Textbook Derives The Inverse z-Transform Equation

Dr. Adam Panagos

Dr. Adam Panagos

Dr. Adam Panagos

z-Transform Exists When Infinite Sum Converges

Values Of r For Which This Occurs Is Called the Region of

Dr. Adam Panagos

Dr. Adam Panagos

If x[k] Is Absolutely Summable We Can Get X (Ω) From X (z)

Dr. Adam Panagos

In General, Must Specify X (z) and ROC For Unique

Let’s Take A Closer Look At the Region Of Convergence

Dr. Adam Panagos

If d Is A Pole of X (z), Then |X (d)| = ∞ and Thus X (z)

Dr. Adam Panagos

Finite Duration Signals

Assume x[k] Is Non-Zero On Interval k1 ≤ k ≤ k2

This Sum Converges For All z (except z = 0, |z| = ∞, or

Dr. Adam Panagos

Finite Duration Signals

If Signal Is Causal (k2 > 0), X (z) Contains z −1 , z −2 , . . .

Dr. Adam Panagos

Infinite Duration Signals

Infinite Duration Signals

We’ve Shown That |X (z)| ≤ S− (z) + S+ (z)

Dr. Adam Panagos

Infinite Duration Signals

Suppose We Can Find Constants A+ , A− , r+ , and r− That

Dr. Adam Panagos

Infinite Duration Signals

S− (z) Converges For |z| < r−

Dr. Adam Panagos

Infinite Duration Signals