DSP6

DSP-G 6.
1
z TRANSFORMS
z Transform Basics
Transfer Functions
Back to the Time Domain
Transfer Function and Stability
DSP-G 6.2
z Transform Basics
The definition of the z z transform transform for a digital
signal is:
z is a complex variable
The z transform of x[n], called X(z) is said to lie in the z
domain
The z transform may not be defined for all value of z
Region of convergence (ROC) Region of convergence (ROC) the values of z for
which the z transform defines
=
=
0
] [ ) (
n
-n
z n x z X
DSP-G 6.3
z Transform Basics (cont.)
The z-transform is considered an operator that
transforms a digital signal into its z domain:
Z{} indicates a z transform is taken
The inverse z transform inverse z transform
See Examples 6.1 ~ 6.5
) ( ] [ } ] [ {
0
z X z n x n x
n
-n
= =
=
Z
} ) ( { ] [ z X n x
-1
Z =
DSP-G 6.4
For any signal with a finite number of samples,
the ROC is z 0, i.e., all z except z = 0
See Table 6.1 for the z transform of basic signals
DSP-G 6.5
DSP-G 6.6
Time shifting property Time shifting property if Z{x[n]} = X(z), then
Z{x[n Z{x[n- -1]} = 1]} = z z
- -1 1
X(z) X(z)
A factor z
-1
in the z domain
corresponds to a one
sample delay in the time
domain
General form: General form:
Z{x[n Z{x[n- -k]} = k]} = z z
- -k k
X X(z) (z)
) (
} ] [ { ] [
] [ ] [
] 1 [ } ] 1 x[ {
1
1
0
1
1
1
1
) 1 (
0
z z
n x z z m x z
z z m x z m x
z n x n
-
-
m
-m -
m
- -m
m
m -
n
-n
X
Z
Z
=
= =
= =
=
=
+
=
DSP-G 6.7
See Examples 6.6 & 6.7
Delay x[n] x[n-1]
Z
-1
x[n] x[n-1]
DSP-G 6.8
Transfer Functions
Let X(z) be the z transform of the input x[n], Y(z)
be the z transform of the output y[n], the transfer transfer
function H(z) function H(z) is defined as the ratio of Y(z) to X(z):
impulse Response
h[n]
x[n] y[n] = x[n]h[n]
Transfer Function
H(z)
X(z) Y(z) = H(z) X(z)
) (
) (
) (
z X
z Y
z H =
DSP-G 6.9
Transfer Functions (cont.)
Given a general difference equation:
or
Take z transform on both sides:
or

= =
=
M
k
k
N
k
k
k n x b k n y a
0 0
] [ ] [
] [ ] 1 [ ] [
] [ ] 1 [ ] [
1 0
1 0
M n x b n x b n x b
N n y a n y a n y a
M
N
+ + + =
+ + +
L
L
) ( ) ( ) (
) ( ) ( ) (
1
1 0
1
1 0
z X z b z X z b z X b
z Y z a z Y z a z Y a
M
M
N
N

+ + + =
+ + +
L
L
( ) ( ) ) ( ) (
1
1 0
1
1 0
z X z b z b b z Y z a z a a
M
M
N
N

+ + + = + + + L L
DSP-G 6.10
The transfer function is:

=
+ + +
+ + +
= =
N
k
k
k
M
k
k
k
N
N
M
M
z a
z b
z a z a a
z b z b b
z X
z Y
z H
0
0
1
1 0
1
1 0
) (
) (
) (
L
L
DSP-G 6.11
From the digital convolution
Take z transform on both sides:
The transfer function is the z transform of the The transfer function is the z transform of the
impulse response impulse response
=
= =
k
k n x k h n x n h n y ] [ ] [ ] [ ] [ ] [

=
=
= =
k k
z k h z X z X z k h z Y
k - k -
] [ ) ( ) ( ] [ ) (
{ } ] [ ] [
) (
) (
) (
k -
n h Z z k h
z X
z Y
z H
k
= = =

=
DSP-G 6.12
The impulse response is the inverse z transform The impulse response is the inverse z transform
of the transfer function of the transfer function:
See Example 6.13
{ } ) ( ] [
1
z H Z n h

=
DSP-G 6.13
Finding filter outputs
Time domain
Computed from the
difference equation
Computed by using
digital convolution
z domain
Computed by using the transfer function
) ( ) ( ) (
) (
) (
) ( z X z H z Y
z X
z Y
z H = =
impulse Response
h[n]
x[n] y[n] = x[n]h[n]
Difference Equation x[n] y[n]
Time Domain
Transfer Function
H(z)
X(z) Y(z) = H(z) X(z)
z Domain
DSP-G 6.14
Summary
The z transform of the output, Y(z), is the product of
the transfer function, H(z), in the z domain and the z
transform of the input, X(z)
The output of the filter can be taken by the inverse z
transform of Y(z):
Convolution theorem Convolution theorem convolution in the time convolution in the time
domain corresponds to multiplication in the z domain corresponds to multiplication in the z
domain domain
{ } ) ( ] [ y
1
z Y Z n

=
DSP-G 6.15
Cascade combinations of filters
Parallel combinations of filters
DSP-G 6.16
Back to the Time Domain
Standard form Standard form of the transfer function
All exponents of z in the z transform be positive
The coefficient of the highest power term in both
numerator and the denominator be one
+ + +
+ + +
=
+ + +
+ + +
=
+ + +
+ + +
=

0
1
0
1
0
1
0
1
0
0
0
1
0
1
0
0
1
0
1
0
1
1 0
1
1 0
1
1

) (
a
a
z
a
a
z
z
b
b
z
b
b
z
a
b
z
a
a
z
a
a
a
z
b
b
z
b
b
b
z a z a a
z b z b b
z H
N N N
M N M N N
N N
M M
N
N
M
M
L
L
L
L
L
L
DSP-G 6.17
Back to the Time Domain (cont.)
A transfer function expressed in standard form is a
rational function consisting of a numerator polynomial
divided by a denominator polynomial
Degree Degree of a polynomial the highest power in a
polynomial
Proper rational function Proper rational function the degree of the numerator
polynomial is less than or equal to the degree of the
denominator polynomial
Strictly proper rational function Strictly proper rational function the degree of the
numerator is less than the degree of the denominator
Improper rational function Improper rational function the degree of the numerator
is larger than the degree of the denominator
DSP-G 6.18
Inverse z transform
Inspection method
Power series expansion long division
Partial fraction expansion
Inspection method using basic transforms
listed in Table 6.1
DSP-G 6.19
Power series expansion Power series expansion using long division
If the z-transform is given as a power series of the
form:
The sequence value x The sequence value x[ [ n n] ] are the are the coefficients coefficients of z of z
- -n n
L L + + + + + + =
=

=
2 1 1 2
] 2 [ ] 1 [ ] 0 [ 1] [ [-2]
] [ ) (
- -
n
-n
z x z x x z x z x
z n x z X
DSP-G 6.20
Partial fraction expansion Partial fraction expansion useful for a strictly
proper rational function in standard form
An example: x[n] = u[n-1], h[n] = (-0.25)
n
u[n], y[n]?
Step 1:
Step 2:
Step 3:
1
} ] [ {
1
1
1
} ] 1 [ { } ] [ { ) (
1
= = =

z
z
n u
z z
z
z n u n x z X Z Z Z Q
=
+
= =
z
z
n u
z
z
n h z H
n
} ] [ {
25 . 0
} ] [ { ) ( Z Z Q
( )( ) 1 25 . 0
) ( ) ( ) (
+
= =
z z
z
z X z H z Y
DSP-G 6.21
Step 4: represent Y(z) by partial fraction expansion
Cover Cover- -up method up method for finding the coefficients A and B
A multiply both side by (z+0.25) and then set z = 0.25
B multiply both side by (z-1) and then set z = 1
( )( ) 1 25 . 0 1 25 . 0
) (
+
+
=
+
=
z
B
z
A
z z
z
z Y
( )
( )
2 . 0
1 25 . 0
25 . 0
: 25 . 0 Set
1
25 . 0
1
) ( 25 . 0
=

= =

+
+ =
= +
A z
z
z B
A
z
z
z Y z
( )
8 . 0
25 . 0 1
1
: 1 set ,
25 . 0
1
25 . 0
=
+
= = +
+

=
+
B z B
z
z A
z
z
DSP-G 6.22
Step 4: (cont.)
Step 5: take the inverse z transform of Y(z)
+
+
=
+
+
=

1
8 . 0
25 . 0
2 . 0
1
8 . 0
25 . 0
2 . 0
) (
1
z
z
z
z
z
z z
z Y
] 1 [ 8 . 0 ] 1 [ ) 25 . 0 ( 2 . 0 ] [
1
+ =

n u n u n y
n
DSP-G 6.23
Transfer Functions and Stability
Poles Poles the values of z that make the denominator
of a transfer function zero
Zeros Zeros the values of z that make the numerator
of a transfer function zero
Given a general form of a transfer function
M zeros: z
1
, z
2
, , z
M
N poles: p
1
, p
2
, , p
N
N
N
M
M
z a z a a
z b z b b
z X
z Y
z H

+ + +
+ + +
= =
L
L
1
1 0
1
1 0
) (
) (
) (
DSP-G 6.24
Transfer Functions and Stability (cont.)
The transfer function can be written as
z
i
: zeros of the filter
p
i
: poles of the filter
K: gain of the filter
z z- -plane plane the complex plane on which the poles
and zeros of the transfer function are plotted
0
0
2 1
2 1
,
) ( ) )( (
) ( ) )( (
) (
a
b
K
p z p z p z
z z z z z z K
z H
N
M
=

=
L
L
DSP-G 6.25
Stable system every bounded input every bounded input ( (finite in finite in
size size) ) produces bounded output produces bounded output
| x[n] | < B
1
| y[n] | < B
2
, n
If a filter is unstable, output grows without bound
The output from an unstable filter can change
dramatically even when the input changes by only
the smallest amount
All useful filters are stable and an important
aspect of filter design is to guarantee stability
DSP-G 6.26
If the input x[n] is bounded, i.e., | x[n] | < B, n
then
If the impulse response is absolutely summable,
i.e., if
The system is stable

=
=
=
k k k
k h B k n x k h k n x k h n y ] [ ] [ ] [ ] [ ] [ ] [
<
= k
k h ] [
DSP-G 6.27
Fourier transform:
z transform:
If z = e
j
with real
(i.e., |z|=1), the z transform
of h[n] corresponds to the
discrete-time Fourier
transform of h[n]
Unit circle Unit circle a circle with radius one centered at
the origin of the z plane
=
=
n
n -j j
e n h e H

] [ ) (
=
=
n
-n
z n h z H ] [ ) (
Re
Im
1
z-plane
z=e
j
unit circle
DSP-G 6.28
If the Fourier transform of h[n] converges
The ROC of H(z) must include the unit circle
For a causal system with rational transfer
function, the ROC is outside the outermost pole
If the ROC includes the unit circle, |z|=1, all of the
poles must be inside the unit circle
A causal digital filter with rational transfer function

H(z) is stable if and only if all of the poles of H(z) lie all of the poles of H(z) lie
inside the unit circle inside the unit circle i.e., they must all have
magnitude smaller than 1
DSP-G 6.29
Summary
Stable Stable all the poles of the filter are inside the unit
circle
Marginally stable Marginally stable with some poles on the unit circle
Unstable Unstable with some poles outside the unit circle
The ROC for a stable transfer function must The ROC for a stable transfer function must include the include the
unit circle unit circle
DSP-G 6.30
DSP-G 6.31
First order system
A simple first order system is:
Has just one pole at z =
Requirement for stability: | | < 1
Impulse response:
When | | > 1, h[n] grows without bound as n increases
When | | < 1, h[n] settles down to zero
Difference equation:
The step response settles to a constant value y
ss
in steady
state:
+
=
+
=

z
z
z
z H
1
1
1
) (
] [ ) ( ] [ n u n h
n
=
] [ ] 1 [ ] [ n x n y n y = +
+
= = +
1
1
1
ss ss ss
y y y
DSP-G 6.32
DSP-G 6.33
> 0 < 0
DSP-G 6.34
> 0 < 0
DSP-G 6.35
Second order system
The transfer function of a simple second order system
is:
p
1
and p
2
are the two poles of the transfer function
Has two zeros at z = 0
Requirement for stability: | p
1
| < 1 and | p
2
| < 1
) )( ( 1
1
) (
2 1
2
2
2
2 1
p z p z
z
z z
z
z z
z H

=
+ +
=
+ +
=

DSP-G 6.36
Second order system
Difference equation:
The steady state value y
ss
can be predicted by:
] [ ] 2 [ ] 1 [ ] [ n x n y n y n y = + +

+ +
= = + +
1
1
1
ss ss ss ss
y y y y
DSP-G 6.37
DSP-G 6.38
DSP-G 6.39
DSP-G 6.40
DSP-G 6.41
DSP-G 6.42
DSP-G 6.43
DSP-G 6.44
DSP-G 6.45
The magnitudes of the poles have a large impact
on the time it takes the system to settle to its final
value
The closer a pole is to the edge of the unit circle, the
longer it takes for the output to settle
The closer a pole is to the center of the unit circle, the
faster the output settles
The magnitudes of the zeros can modify the
behavior of the output dramatically
The closer the zeros are to the poles, the greater their
effect on system behavior
DSP-G 6.46
DSP-G 6.47
DSP-G 6.48
DSP-G 6.49

DSP6

Uploaded by

Document Information

Original Description:

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

DSP6

Uploaded by

Copyright:

Available Formats

DSP-G 6.

A causal digital filter with rational transfer function

You might also like