71
Chapter 1: Introduction to z- Transform 1.1 z-Transform 1.2 Properties 1.2.1 Linearity 1.2.2 Shifting 1.2.3 Time Reversal 1.2.4 Multiplication by Exponential Sequence 1.2.5 Differentiation in the z-domain 1.2.6 Discrete Convolution 1.3 Inverse z-transform 1.3.1 Relationship between the z-transform and the Laplace Transform 1.4 Frequency Response Estimation 1.5 Pole-Zero Description of Discrete-Time Systems 1.6 A Second-Order Resonant System 1.7 Problem Sheet B1 Professor E. Ambikairajah UNSW, Australia

Part B Chapter1 Z-Transform-modeemedia.ee.unsw.edu.au/contents/elec3104/LectureNotes/DSP3.pdf · Clearly, the z-transform is a power series with an infinite number of terms and so

  • Upload
    others

  • View
    0

  • Download
    0

Embed Size (px)

Citation preview

Page 1: Part B Chapter1 Z-Transform-modeemedia.ee.unsw.edu.au/contents/elec3104/LectureNotes/DSP3.pdf · Clearly, the z-transform is a power series with an infinite number of terms and so

Chapter 1: Introduction to z-Transform

1.1 z-Transform1.2 Properties

1.2.1 Linearity1.2.2 Shifting1.2.3 Time Reversal1.2.4 Multiplication by Exponential Sequence1.2.5 Differentiation in the z-domain1.2.6 Discrete Convolution

1.3 Inverse z-transform1.3.1 Relationship between the z-transform and the

Laplace Transform1.4 Frequency Response Estimation1.5 Pole-Zero Description of Discrete-Time Systems1.6 A Second-Order Resonant System1.7 Problem Sheet B1Prof

esso

r E. A

mbikair

ajah

UNSW, A

ustra

lia

Page 2: Part B Chapter1 Z-Transform-modeemedia.ee.unsw.edu.au/contents/elec3104/LectureNotes/DSP3.pdf · Clearly, the z-transform is a power series with an infinite number of terms and so

Chapter 1: Introduction to z-Transform [1,3]

s

aa f

fT πωθ 2==

Analogue Domain(Continuous-time Domain)

x(t)

X(s)

LaplaceTransform

s=jω

X(ω)

ω-analoguefrequency

jωS-PlaneStable

Region

Discrete-time Domain

x[n]

X(z)

z-transform

z=ejθ

X(ejθ)

-π≤θ≤π

t = nT

Im(z) z-Plane

θ = πθ = -π Re(z)

stable region

|z|=1(unit circle)

θσ

θ

-digital

frequency

θ -digital frequency

Profes

sor E

. Ambik

airaja

h

UNSW, A

ustra

lia

Page 3: Part B Chapter1 Z-Transform-modeemedia.ee.unsw.edu.au/contents/elec3104/LectureNotes/DSP3.pdf · Clearly, the z-transform is a power series with an infinite number of terms and so

The primary role of the Laplace transformin engineering is transient and stability analysis of causal LTI system described by differential equations.

The primary roles of the z-transform are the study of system characteristics and derivation of computational structures for implementing discrete-time systems on computers. The-transform is also used solve difference equations.

Profes

sor E

. Ambik

airaja

h

UNSW, A

ustra

lia

Page 4: Part B Chapter1 Z-Transform-modeemedia.ee.unsw.edu.au/contents/elec3104/LectureNotes/DSP3.pdf · Clearly, the z-transform is a power series with an infinite number of terms and so

1.1 z-TransformThe z-transform of a discrete-time signal

In causal systems, x[n] may be zero when n < 0

[ ]∑∞

−∞=

−=n

nznxzX )( z = rejθ, θ = digital frequencywhere z is a complex variable

two-sided transform

[ ]∑∞

=

−=0

)(n

nznxzX one-sided transform

Profes

sor E

. Ambik

airaja

h

UNSW, A

ustra

lia

Page 5: Part B Chapter1 Z-Transform-modeemedia.ee.unsw.edu.au/contents/elec3104/LectureNotes/DSP3.pdf · Clearly, the z-transform is a power series with an infinite number of terms and so

Clearly, the z-transform is a power series with an infinite number of terms and so may not converge for

all values of z.

The region where the z-transform converges is known as the region of convergence (ROC) and in this region the values of X(z) are finite

The step sequence: [ ]⎩⎨⎧

<∞≤≤

=00

01n

nnx

...11)( 21

0+++=⋅= −−

=

−∑ zzzzXn

n

This is a geometric series with a common ratio of z-1.The series converges if |z-1| < 1 or equivalently if |z| > 1Prof

esso

r E. A

mbikair

ajah

UNSW, A

ustra

lia

Page 6: Part B Chapter1 Z-Transform-modeemedia.ee.unsw.edu.au/contents/elec3104/LectureNotes/DSP3.pdf · Clearly, the z-transform is a power series with an infinite number of terms and so

In this case, the z-transform is valid everywhere outside a circle of unit radius whose centre is at the origin (see below)

111...1)( 1

21

−=

−=+++= −

−−

zz

zzzzX

Im(z)

region of convergenceRe(z)

|z|=1 is a circle of unit radius referred to as the ‘unit circle’

|z|=1

Profes

sor E

. Ambik

airaja

h

UNSW, A

ustra

lia

Page 7: Part B Chapter1 Z-Transform-modeemedia.ee.unsw.edu.au/contents/elec3104/LectureNotes/DSP3.pdf · Clearly, the z-transform is a power series with an infinite number of terms and so

221

1...221)(2 121 =

−=+++=⇒= −

−−zXzLet

( ) ( ) ( ) KK +++=+++=⇒= −− 42115.0 1211

21zXzLet

⎭⎬⎫

⎩⎨⎧

−=

1)(

zzzX

|z| > 1 then X(z) converges and |z| < 1 then it diverges

So the Region Of Convergence (ROC) is seen to be bounded

by the circle |z| = 1, the radius of the pole of

Profes

sor E

. Ambik

airaja

h

UNSW, A

ustra

lia

Page 8: Part B Chapter1 Z-Transform-modeemedia.ee.unsw.edu.au/contents/elec3104/LectureNotes/DSP3.pdf · Clearly, the z-transform is a power series with an infinite number of terms and so

The delta sequence δ[n]:

[ ] [ ] 1}{0

== ∑∞

=

n

nznnZ δδ

11

1)(0

<−

=−

==∑∞

=

zafor

azz

zazazX

n

nn

1)(

−=

zzzX

The geometric sequence: x[n] = an

or equivalently, |z| > |a|.

when a = 1, x[n] = 1 for n ≥ 0 ie. x[n] = u[n]

ROC: |z|>1Profes

sor E

. Ambik

airaja

h

UNSW, A

ustra

lia

Page 9: Part B Chapter1 Z-Transform-modeemedia.ee.unsw.edu.au/contents/elec3104/LectureNotes/DSP3.pdf · Clearly, the z-transform is a power series with an infinite number of terms and so

The complex exponential sequence: x[n] = ejnθ

{ }

1cos2)sincos(

1)()(

))(()(

1

1

2

2

0

+−+−

=

++−−

=−−

−=

−−

×−

=−

=−

==

−∞

=

−∑

θθθ

θθ

θ

θθ

θ

θ

θ

θθθθθ

zzjzz

zeezezz

ezezezz

ezez

ezz

ezz

ze

zeez

jj

j

jj

j

j

j

jjjn

nnjnj

1cos2sin

1cos2)cos(}sin{cos

1cos2sin

1cos2)cos(}{

22

22

+−+

+−−

=+

+−+

+−−

=

θθ

θθθθ

θθ

θθθ

zzzj

zzzznjnZ

zzzj

zzzzeZ nj

Profes

sor E

. Ambik

airaja

h

UNSW, A

ustra

lia

Page 10: Part B Chapter1 Z-Transform-modeemedia.ee.unsw.edu.au/contents/elec3104/LectureNotes/DSP3.pdf · Clearly, the z-transform is a power series with an infinite number of terms and so

Exploiting the linearity property

1cos2sin

1cos2)cos(}sin{cos

1cos2sin

1cos2)cos(}{

22

22

+−+

+−−

=+

+−+

+−−

=

θθ

θθθθ

θθ

θθθ

zzzj

zzzznjnZ

zzzj

zzzzeZ nj

1cos2sin}{sin

1cos2)cos(}{cos

2

2

+−=

+−−

=∴

θθθ

θθθ

zzznZ

zzzznZ

Profes

sor E

. Ambik

airaja

h

UNSW, A

ustra

lia

Page 11: Part B Chapter1 Z-Transform-modeemedia.ee.unsw.edu.au/contents/elec3104/LectureNotes/DSP3.pdf · Clearly, the z-transform is a power series with an infinite number of terms and so

1.2 Properties

1.2.1 Linearity [ ] [ ] )()( zbYzaXnbynax Z +⎯→←+

1.2.2 Shifting Property (Delay Theorem):

[ ] )(zXzknx kZ −⎯→←−

A very important property of the z-transform is the delay theorem. [ ]{ } ( )

[ ]{ } ( )zXznxZzXznxZ

2

1

21

=−

=−

z-1

Unit delay

x[n]

X(z)

x[n-1]

z-1X(z) Tx[n] x[n-1]

One sample delay

Profes

sor E

. Ambik

airaja

h

UNSW, A

ustra

lia

Page 12: Part B Chapter1 Z-Transform-modeemedia.ee.unsw.edu.au/contents/elec3104/LectureNotes/DSP3.pdf · Clearly, the z-transform is a power series with an infinite number of terms and so

1.2.3 Time reversal:

[ ] ( )11 −⎟⎠⎞

⎜⎝⎛⎯→←− zXor

zXnx z

Example: Find the z-transform of x[n] = -u[n].

[ ] [ ]zz

zznuz

zznu−

=−

⎯→←−−

⎯→← −

11

1;

1 1

1

ROC: |z| < 1

Profes

sor E

. Ambik

airaja

h

UNSW, A

ustra

lia

Page 13: Part B Chapter1 Z-Transform-modeemedia.ee.unsw.edu.au/contents/elec3104/LectureNotes/DSP3.pdf · Clearly, the z-transform is a power series with an infinite number of terms and so

1.2.4 Multiplication by exponential sequence:

[ ] )( 1zaXnxa zn −⎯→←

[ ] )( zeXznxe jnj θθ −⎯→←

As a special case if x[n] is multiplied by ejnθ

1.2.5 Differentiation in the z-domain:

[ ] )(zXdzdzznnx −⎯→←

Profes

sor E

. Ambik

airaja

h

UNSW, A

ustra

lia

Page 14: Part B Chapter1 Z-Transform-modeemedia.ee.unsw.edu.au/contents/elec3104/LectureNotes/DSP3.pdf · Clearly, the z-transform is a power series with an infinite number of terms and so

1.2.6 Discrete Convolution:If y[n] = x[n] * h[n] then Y(z) = X(z) . H(z)

[ ] [ ] [ ] [ ] [ ] [ ]

[ ] [ ] [ ] [ ]

[ ] [ ] [ ] [ ]

[ ] [ ] [ ] [ ]⎭⎬⎫

⎩⎨⎧

=

−=⎭⎬⎫

⎩⎨⎧

−−=

⎭⎬⎫

⎩⎨⎧

−−⋅=

⎭⎬⎫

⎩⎨⎧

−−⋅=

∑∑

∑∑

∑ ∑

−=

−−∞

−∞=

=

−∞

−∞=

=

−∞

−∞=

−∞=

km

km

k

n

n

k

n

n

k

k

zmhmukxku

knmLetzknhknukxku

zknhknukxku

knhknukxkuZnhnxZ

0

0

}*{

Profes

sor E

. Ambik

airaja

h

UNSW, A

ustra

lia

Page 15: Part B Chapter1 Z-Transform-modeemedia.ee.unsw.edu.au/contents/elec3104/LectureNotes/DSP3.pdf · Clearly, the z-transform is a power series with an infinite number of terms and so

[ ] [ ] [ ] [ ]

[ ] [ ]

)()()()(

0 0

zHzXzHzX

zmhzkx

zmhmuzkxku

k m

mk

km

m

k

k

⋅=

⎭⎬⎫

⎩⎨⎧

=

⎭⎬⎫

⎩⎨⎧

=

∑ ∑

∑∑∞

=

=

−−

−=

−∞

−∞=

y[n] = x[n] * h[n]Y(z) = X(z) ⋅ H(z)

h[n]H(z)

x[n]

X(z)

y[n]

Y(z)

Profes

sor E

. Ambik

airaja

h

UNSW, A

ustra

lia

Page 16: Part B Chapter1 Z-Transform-modeemedia.ee.unsw.edu.au/contents/elec3104/LectureNotes/DSP3.pdf · Clearly, the z-transform is a power series with an infinite number of terms and so

Example :

Concept of the transfer function

]2[]1[]2[]1[][][ 21210 −+−+−+−+= nybnybnxanxanxany

Take the z-transform of both sides:

( ) ( ) ( ) ( ) ( ) ( )( )[ ] ( )[ ]( ) ( )

( ) 22

11

22

110

22

110

22

11

22

11

22

110

1

1

−−

−−

−−−−

−−−−

−−++

==

++=−−

++++=

zbzbzazaa

zXzYzH

zazaazXzbzbzY

zzYbzzYbzzXazzXazXazY

Profes

sor E

. Ambik

airaja

h

UNSW, A

ustra

lia

Page 17: Part B Chapter1 Z-Transform-modeemedia.ee.unsw.edu.au/contents/elec3104/LectureNotes/DSP3.pdf · Clearly, the z-transform is a power series with an infinite number of terms and so

Example :Find the difference-equation of the following transfer function

First rewrite H(z) as a ratio of polynomials in z-1

2325)( 2 ++

+=

zzzzH

( ) ( ) ( ) ( ) ( ) 2121

21

21

2523231

25)()()(

−−−−

−−

−−

+=++

+++

==

zzXzzXzzYzzYzYzz

zzzHzXzY

Take inverse z-transform

[ ] [ ] [ ] [ ] [ ]22132215 −−−−−+−=⇒ nynynxnxny

[ ] [ ] [ ] [ ] [ ]22152213 −+−=−+−+ nxnxnynyny

Profes

sor E

. Ambik

airaja

h

UNSW, A

ustra

lia

Page 18: Part B Chapter1 Z-Transform-modeemedia.ee.unsw.edu.au/contents/elec3104/LectureNotes/DSP3.pdf · Clearly, the z-transform is a power series with an infinite number of terms and so

Example : If x[n] = u[n] – u[n-10], find X(z)?

1

109

0 11)1()( −

=

−−

== ∑ zzzzX

n

n

Example:

[ ]

k

k

k

kn

nn

nn

z

zz

znuzY

∑∑

=

=

−∞=

−∞=

⎟⎠⎞

⎜⎝⎛−=

⎟⎠⎞

⎜⎝⎛−=⎟

⎠⎞

⎜⎝⎛−=

−−−=

0

1

1

1

1)(

α

αα

α

If y[n] = -αnu[-n-1], find Y(z)?

Profes

sor E

. Ambik

airaja

h

UNSW, A

ustra

lia

Page 19: Part B Chapter1 Z-Transform-modeemedia.ee.unsw.edu.au/contents/elec3104/LectureNotes/DSP3.pdf · Clearly, the z-transform is a power series with an infinite number of terms and so

-1

u[n+1]

n -1

u[-(n+1)]

n

1<αz

||||

||||1

11)( 1

αα

αα

<−

=

<−

−= −

zz

z

zz

zY

The sum converges provided ie. |z| < |α|

Profes

sor E

. Ambik

airaja

h

UNSW, A

ustra

lia

Page 20: Part B Chapter1 Z-Transform-modeemedia.ee.unsw.edu.au/contents/elec3104/LectureNotes/DSP3.pdf · Clearly, the z-transform is a power series with an infinite number of terms and so

Depict the ROC and pole and zero locations in the z-plane

Im(z)

Re(z)α

Profes

sor E

. Ambik

airaja

h

UNSW, A

ustra

lia

Page 21: Part B Chapter1 Z-Transform-modeemedia.ee.unsw.edu.au/contents/elec3104/LectureNotes/DSP3.pdf · Clearly, the z-transform is a power series with an infinite number of terms and so

1.2.6 Discrete ConvolutionCompute the convolution y[n] of the digital signals given byx1[n] = [1, -2, 1];x2[n] = 1 for 0 ≤ n ≤ 5, x2[n] = 0 elsewherey[n] = x1[n] * x2[n] ⇒ Y(z) = X1(z) ⋅X2(z)X1(z) = 1 –2z-1 + z-2

X2(z) = 1 + z-1 + z-2 + z-3 + z-4 +z-5

Y(z) = X1(z) ⋅ X2(z)= 1-z-1 – z-6 + z-7

Inverse z-transform

y[n] = [1, -1, 0,0,0,0,-1,1]Profes

sor E

. Ambik

airaja

h

UNSW, A

ustra

lia

Page 22: Part B Chapter1 Z-Transform-modeemedia.ee.unsw.edu.au/contents/elec3104/LectureNotes/DSP3.pdf · Clearly, the z-transform is a power series with an infinite number of terms and so

Example :

Determine the system function H(z) of the system shown below:

+

T

x[n]y[n]

aay[n-1]

y[n] = x[n] + ay[n-1]

+

z-1

X(z) Y(z)

aaz-1Y(z)

Y(z) = X(z) + az-1Y(z)

111

)()()( −−==

azzXzYzH Prof

esso

r E. A

mbikair

ajah

UNSW, A

ustra

lia

Page 23: Part B Chapter1 Z-Transform-modeemedia.ee.unsw.edu.au/contents/elec3104/LectureNotes/DSP3.pdf · Clearly, the z-transform is a power series with an infinite number of terms and so

Profes

sor E

. Ambik

airaja

h

UNSW, A

ustra

lia

Page 24: Part B Chapter1 Z-Transform-modeemedia.ee.unsw.edu.au/contents/elec3104/LectureNotes/DSP3.pdf · Clearly, the z-transform is a power series with an infinite number of terms and so

Profes

sor E

. Ambik

airaja

h

UNSW, A

ustra

lia

Page 25: Part B Chapter1 Z-Transform-modeemedia.ee.unsw.edu.au/contents/elec3104/LectureNotes/DSP3.pdf · Clearly, the z-transform is a power series with an infinite number of terms and so

Basic z-Transforms

|z|>rrnsin(nθ) u[n]

|z| > rrncos(nθ) u[n]

|z| > 1sin(nθ) u[n]

|z|>1cos(nθ) u[n]

|z| > |α|nαn u[n]

|z| > |α|αn u[n]

|z| > 1u[n]

all z1δ[n]

ROCTransformSignal

111

−− z

111

−− zα

21

1

)1( −

− zzαα

21

1

cos21cos1

−−

+−−

zzz

θθ

21

1

cos21sin

−−

+− zzz

θθ

221

1

cos21cos1

−−

+−−

zrrzrzθ

θ

221

1

cos21sin1

−−

+−−

zrrzrzθ

θProfes

sor E

. Ambik

airaja

h

UNSW, A

ustra

lia

Page 26: Part B Chapter1 Z-Transform-modeemedia.ee.unsw.edu.au/contents/elec3104/LectureNotes/DSP3.pdf · Clearly, the z-transform is a power series with an infinite number of terms and so

z-Transform Properties

nx[n]

X(z)⋅Y(z)x[n]*y[n]

x[-n]

anx[n]

z-kX(z)X[n-k]

aX(z) + bY(z)ax[n] + by[n]

X(z)X[n]

Transformsignal

⎟⎠⎞

⎜⎝⎛

azX

⎟⎠⎞

⎜⎝⎛

zX 1

)(zXdzdz−Prof

esso

r E. A

mbikair

ajah

UNSW, A

ustra

lia

Page 27: Part B Chapter1 Z-Transform-modeemedia.ee.unsw.edu.au/contents/elec3104/LectureNotes/DSP3.pdf · Clearly, the z-transform is a power series with an infinite number of terms and so

Note:

If N →∞ , |z-1| < 1

1)1(21

11...1 −

−−−−−

−−

=++++=zzzzzS

NN

N

111

−∞ −=

zS

Profes

sor E

. Ambik

airaja

h

UNSW, A

ustra

lia

Page 28: Part B Chapter1 Z-Transform-modeemedia.ee.unsw.edu.au/contents/elec3104/LectureNotes/DSP3.pdf · Clearly, the z-transform is a power series with an infinite number of terms and so

1.3 Inverse z-Transform

The inverse z-transform-Partial fraction

The inverse z-transform allows us to recover the discrete-time sequence.

x[n] = Z-1[X(z)]

where X(z) is the z-transform of x[n].

Profes

sor E

. Ambik

airaja

h

UNSW, A

ustra

lia

Page 29: Part B Chapter1 Z-Transform-modeemedia.ee.unsw.edu.au/contents/elec3104/LectureNotes/DSP3.pdf · Clearly, the z-transform is a power series with an infinite number of terms and so

Example :Find x[n] for the following:

⎥⎥⎥⎥

⎢⎢⎢⎢

+

⎟⎠⎞

⎜⎝⎛

−−

⎟⎠⎞

⎜⎝⎛

=

⎥⎥⎥⎥

⎢⎢⎢⎢

+

⎟⎠⎞

⎜⎝⎛−

+−

⎟⎠⎞

⎜⎝⎛

=

⎥⎦⎤

⎢⎣⎡

++

−=

+−=

−−=

−−= −−

5.054

75.054

5.054

75.054

5.075.0)5.0)(75.0(375.025.0

375.025.01)(

2

21

1

z

z

z

z

zzz

zB

zAz

zzz

zzz

zzzzX

[ ] 0,)5.0()75.0(54

5.054

75.054][ 11

>−−=

⎥⎦⎤

⎢⎣⎡

+−⎥⎦

⎤⎢⎣⎡

−= −−

n

zzz

zzznx

nnProfes

sor E

. Ambik

airaja

h

UNSW, A

ustra

lia

Page 30: Part B Chapter1 Z-Transform-modeemedia.ee.unsw.edu.au/contents/elec3104/LectureNotes/DSP3.pdf · Clearly, the z-transform is a power series with an infinite number of terms and so

Power series Method

Residue Method - evaluating the contour integral

1.3 Inverse z-Transform

Profes

sor E

. Ambik

airaja

h

UNSW, A

ustra

lia

Page 31: Part B Chapter1 Z-Transform-modeemedia.ee.unsw.edu.au/contents/elec3104/LectureNotes/DSP3.pdf · Clearly, the z-transform is a power series with an infinite number of terms and so

1.3.1 Relationship between the z-transform and Laplace transform.

If we let z = esT, then z = e(σ +jω)T (T is the sampling period)

z = eσ T⋅ ejωT = eσ T⋅ ejθ -π ≤ θ ≤ π

Thus |z| = eσ T and

θπω ===∠sf

fTz 2 (θ = digital frequency)

Profes

sor E

. Ambik

airaja

h

UNSW, A

ustra

lia

Page 32: Part B Chapter1 Z-Transform-modeemedia.ee.unsw.edu.au/contents/elec3104/LectureNotes/DSP3.pdf · Clearly, the z-transform is a power series with an infinite number of terms and so

As ω varies from -∞ to ∞ the s-plane is mapped to the z-plane as shown in Figure 2.10.

The entire jω axis in the s-plane is mapped onto the unit circle. The left-hand s-plane is mapped to the inside of the unit circle and the right-hand s-plane maps to the outside of the unit circle.

Profes

sor E

. Ambik

airaja

h

UNSW, A

ustra

lia

Page 33: Part B Chapter1 Z-Transform-modeemedia.ee.unsw.edu.au/contents/elec3104/LectureNotes/DSP3.pdf · Clearly, the z-transform is a power series with an infinite number of terms and so

s-plane

σ

z-plane

θ = 0

If θ = πUnit circle|z|=1

θ = πθ = -π

(Half thesampling frequency)

Figure 2.10. Mapping of the s-plane to the z-plane.

Profes

sor E

. Ambik

airaja

h

UNSW, A

ustra

lia

Page 34: Part B Chapter1 Z-Transform-modeemedia.ee.unsw.edu.au/contents/elec3104/LectureNotes/DSP3.pdf · Clearly, the z-transform is a power series with an infinite number of terms and so

In terms of frequency response, the jω axis is the most important in the s-plane. In this

case σ = 0 and the frequency points in the s-plane are related to points on the z-plane unit circle by z = eσT.ejωT = 1 ejθ

θjez =

Profes

sor E

. Ambik

airaja

h

UNSW, A

ustra

lia

Page 35: Part B Chapter1 Z-Transform-modeemedia.ee.unsw.edu.au/contents/elec3104/LectureNotes/DSP3.pdf · Clearly, the z-transform is a power series with an infinite number of terms and so

1.4 Frequency Response Estimation

There are many instances when it is necessary to evaluate the frequency response of discrete-time systems. The frequency response of a system can be readily obtained from its z-transforms.

For example, if we set z = ejθ, that is evaluate the z-transform around the unit circle, we obtain the Fourier Transform of the system.

πθπθ θ <<==

- , |)()( jezzHH

Fourier transform of the discrete-time systemProfes

sor E

. Ambik

airaja

h

UNSW, A

ustra

lia

Page 36: Part B Chapter1 Z-Transform-modeemedia.ee.unsw.edu.au/contents/elec3104/LectureNotes/DSP3.pdf · Clearly, the z-transform is a power series with an infinite number of terms and so

Example :

Find H(θ). {H(θ)-the frequency response}

Fourier transform of the discrete-time system.

6.0say 01 , 1

1)( 1 =<<−

= − aaaz

zH

( ) πθπθ θ ≤≤−==

|)( jezzHH

Profes

sor E

. Ambik

airaja

h

UNSW, A

ustra

lia

Page 37: Part B Chapter1 Z-Transform-modeemedia.ee.unsw.edu.au/contents/elec3104/LectureNotes/DSP3.pdf · Clearly, the z-transform is a power series with an infinite number of terms and so

( ) ( ) ( ) θθθ θθ sincos1

11

1|jaaae

zHH jez j+−

=−

== −=

( ) ( ) 222 cos211

sincos1

1)(aaaa

H+−

=+−

=θθθ

θ

θ = ωT; θ = 2π f/fs:; θ = π ⇒ f = fs/2fs= sampling frequency

|H(θ)|

-π πθ

-fs/2 fs/20 f (analogue frequency)

digital frequency

a−11

a+11

Profes

sor E

. Ambik

airaja

h

UNSW, A

ustra

lia

Page 38: Part B Chapter1 Z-Transform-modeemedia.ee.unsw.edu.au/contents/elec3104/LectureNotes/DSP3.pdf · Clearly, the z-transform is a power series with an infinite number of terms and so

1.5 Pole-Zero Description of Discrete-Time Systems

The zeros of a z-transform H(z) are the values of zfor which H(z)=0. The poles of a z-transform are

the values of z for which H(z)=∞ . If H(z) is a rational function , then

))......()(()).......()((

......1...

)()()(

21

210

11

110

L

M

LL

MM

pzpzpzzzzzzza

zbzbzazaa

zXzYzH

−−−−−−

=

++++++

== −−

−−

Profes

sor E

. Ambik

airaja

h

UNSW, A

ustra

lia

Page 39: Part B Chapter1 Z-Transform-modeemedia.ee.unsw.edu.au/contents/elec3104/LectureNotes/DSP3.pdf · Clearly, the z-transform is a power series with an infinite number of terms and so

The complex quantities (or may be real) z1, z2, z3 ….

are called zeros of H(z) and the complex quantities

(or may be real) p1, p2, p3 … are called the poles of

H(z). We thus see that H(z) is completely

determined , except for the constant a0, by the

values of poles and zeros

The information contained in the z-transform can be conveniently displayed as a poles-zero diagram (see figure in the next slide)

Profes

sor E

. Ambik

airaja

h

UNSW, A

ustra

lia

Page 40: Part B Chapter1 Z-Transform-modeemedia.ee.unsw.edu.au/contents/elec3104/LectureNotes/DSP3.pdf · Clearly, the z-transform is a power series with an infinite number of terms and so

0.75

|z|=1 0.5

-0.5

Im(z)

Re(z)-1

In the diagram, ‘X’ marks the position of a pole and ‘O’denotes the position of a zero.

The poles are located at z = 0.5 ± 0.5j and z = 0.75, a single zero is at z = -1.Prof

esso

r E. A

mbikair

ajah

UNSW, A

ustra

lia

Page 41: Part B Chapter1 Z-Transform-modeemedia.ee.unsw.edu.au/contents/elec3104/LectureNotes/DSP3.pdf · Clearly, the z-transform is a power series with an infinite number of terms and so

An important feature of the pole-zero diagram is the unit circle |z|=1. The pole-zero diagram provides an insight into the properties of a given discrete-time system.

From the locations of the poles d zeros we can infer the frequency response of the system as well as its degree of stability.

For a stable system, all the poles must lie inside the unit circle. Zeros may lie inside, on, or outside the unit circle.Prof

esso

r E. A

mbikair

ajah

UNSW, A

ustra

lia

Page 42: Part B Chapter1 Z-Transform-modeemedia.ee.unsw.edu.au/contents/elec3104/LectureNotes/DSP3.pdf · Clearly, the z-transform is a power series with an infinite number of terms and so

Example :Determine the transfer function H(z) of a discrete-time system with the pole-zero diagram shown below:

0.5|z| =1 0.5

-0.5

Im(z)

Re(z)-1

21

2

5.01)1(

)5.05.0()5.05.0()1()1()(

−−

−−+

=

+−−−+−

=

zzzK

jzjzjzjzKzH

Profes

sor E

. Ambik

airaja

h

UNSW, A

ustra

lia

Page 43: Part B Chapter1 Z-Transform-modeemedia.ee.unsw.edu.au/contents/elec3104/LectureNotes/DSP3.pdf · Clearly, the z-transform is a power series with an infinite number of terms and so

Example :

Determine the pole-zero plot:

azzzH−

=)(

|z|=a

Im(z)

Re(z)

Profes

sor E

. Ambik

airaja

h

UNSW, A

ustra

lia

Page 44: Part B Chapter1 Z-Transform-modeemedia.ee.unsw.edu.au/contents/elec3104/LectureNotes/DSP3.pdf · Clearly, the z-transform is a power series with an infinite number of terms and so

Example :

Determine the pole-zero plot :

121

1

)().........)(()(

)( −−

−−−=

−−

= MM

MM

zMzzzzz

azzazzH

The zero z = a cancels the pole at z = a. Thus, H(z)has M-1 zeros and M-1 poles as shown in the diagram below for M = 8.

|z|=a

Im(z)

Re(z)

z= a

8 polesProfes

sor E

. Ambik

airaja

h

UNSW, A

ustra

lia

Page 45: Part B Chapter1 Z-Transform-modeemedia.ee.unsw.edu.au/contents/elec3104/LectureNotes/DSP3.pdf · Clearly, the z-transform is a power series with an infinite number of terms and so

Consider a system, H(z) with two complex conjugate poles in the z-plane :

|z|=1

Im(z)

Re(z)θθ

r p1

p2

θ

θ

j

j

rep

rep−=

=

2

1 Poles

)(01 zeroz =

Profes

sor E

. Ambik

airaja

h

UNSW, A

ustra

lia

Page 46: Part B Chapter1 Z-Transform-modeemedia.ee.unsw.edu.au/contents/elec3104/LectureNotes/DSP3.pdf · Clearly, the z-transform is a power series with an infinite number of terms and so

A typical transfer function might be:

( ) ( )

⎥⎥⎥⎥

⎢⎢⎢⎢

−−

−=

⎥⎦⎤

⎢⎣⎡

−+

−=

−−=

−−

θθ

θθθθ

θθjj

jjjj

rezjr

rezjrz

rezB

rezAz

rezrezzzH

sin21

sin21

)(

⎥⎦⎤

⎢⎣⎡

−−

−= −−− 11 1

11

1)sin2(

1)(zrezrerj

zH jj θθθProfes

sor E

. Ambik

airaja

h

UNSW, A

ustra

lia

Page 47: Part B Chapter1 Z-Transform-modeemedia.ee.unsw.edu.au/contents/elec3104/LectureNotes/DSP3.pdf · Clearly, the z-transform is a power series with an infinite number of terms and so

( ) ( )[ ][ ]njnj

n

njnj

eerjr

rererj

nh

θθ

θθ

θ

θ

−=

−=

sin2

sin21)(

θ = frequency of oscillation

[ ]θθ

nrnh n sinsin

1)( 1−=

This is the impulse response of the 2nd order system with complex poles

Profes

sor E

. Ambik

airaja

h

UNSW, A

ustra

lia

Page 48: Part B Chapter1 Z-Transform-modeemedia.ee.unsw.edu.au/contents/elec3104/LectureNotes/DSP3.pdf · Clearly, the z-transform is a power series with an infinite number of terms and so

We note that the impulse response will decay away to zero provided r is less than one. [ r < 1]

Recall that r is also the distance from the origin in

the z-plane to the poles p1 or p2, so that system will be stable if the poles in the z-plane lie inside the unit circle.

Profes

sor E

. Ambik

airaja

h

UNSW, A

ustra

lia

Page 49: Part B Chapter1 Z-Transform-modeemedia.ee.unsw.edu.au/contents/elec3104/LectureNotes/DSP3.pdf · Clearly, the z-transform is a power series with an infinite number of terms and so

Example :Poles inside unit circle

θ-θ

Exponential decay sinewave (r<1)Stable system

rn

[ ]θθ

nrnh n sinsin

1)( 1−=

Poles on theunit circle

θ

r = 1

-1

1

Profes

sor E

. Ambik

airaja

h

UNSW, A

ustra

lia

Page 50: Part B Chapter1 Z-Transform-modeemedia.ee.unsw.edu.au/contents/elec3104/LectureNotes/DSP3.pdf · Clearly, the z-transform is a power series with an infinite number of terms and so

Poles outside unit circleθ

Exponential increasing (r>1)

rn

[ ]θθ

nrnh n sinsin

1)( 1−=

one real Pole inside the unit circle rn, θ = 0

n

Profes

sor E

. Ambik

airaja

h

UNSW, A

ustra

lia

Page 51: Part B Chapter1 Z-Transform-modeemedia.ee.unsw.edu.au/contents/elec3104/LectureNotes/DSP3.pdf · Clearly, the z-transform is a power series with an infinite number of terms and so

Note :

A system that is both stable and casual must have all its poles inside that unit circle within the z-plane.

We cannot have a pole outside the unit circle, since the inverse transform of a pole located outside the circle will contribute either a right sided increasing exponential term, which is not stable, or a left-sided decaying exponential term that is not causal.

Profes

sor E

. Ambik

airaja

h

UNSW, A

ustra

lia

Page 52: Part B Chapter1 Z-Transform-modeemedia.ee.unsw.edu.au/contents/elec3104/LectureNotes/DSP3.pdf · Clearly, the z-transform is a power series with an infinite number of terms and so

1.6 A second order Resonant System (Complex Poles)

z -1

z -1

+y[n]x[n]

-b1

-b2

θ0r

002

001

sincos

sincos0

0

θθ

θθθ

θ

jrrrep

jrrrepj

j

−==

+==−

(A) 1

1)(21

2

2

22

11 bzbz

zzbzb

zH++

=++

= −−Profes

sor E

. Ambik

airaja

h

UNSW, A

ustra

lia

Page 53: Part B Chapter1 Z-Transform-modeemedia.ee.unsw.edu.au/contents/elec3104/LectureNotes/DSP3.pdf · Clearly, the z-transform is a power series with an infinite number of terms and so

All pole system has poles only (without counting the zeros at the origin)

(B) cos2)(

)(

))(())(()(

20

2

2

22

2

2

21

2

21

12

2

00

00

rzrzz

rzeerzzzH

rezrezz

pzpzz

bzbzzzH

jj

jj

+−=

++−=

−−=

−−=

++=

−−

θθθ

θθ

01 cos2 θrb −= 22 rb =

2

10 2 b

bCos −=∴ θ

sff0

02πθ =

Comparing (A) and (B), we obtain

θ0 = resonant frequencyProfes

sor E

. Ambik

airaja

h

UNSW, A

ustra

lia

Page 54: Part B Chapter1 Z-Transform-modeemedia.ee.unsw.edu.au/contents/elec3104/LectureNotes/DSP3.pdf · Clearly, the z-transform is a power series with an infinite number of terms and so

We can derive H(θ) and the magnitude from

22

111

1)( −− ++=

zbzbzH

0.99-1.41

0.9-1.34

0.7-1.16

0.5-0.94

b2 = r2b1

⎥⎥⎦

⎢⎢⎣

⎡ −=

−==

2

110

01

22

2cos

)cos(2

bb

rbrb

θ

θ

Profes

sor E

. Ambik

airaja

h

UNSW, A

ustra

lia

Page 55: Part B Chapter1 Z-Transform-modeemedia.ee.unsw.edu.au/contents/elec3104/LectureNotes/DSP3.pdf · Clearly, the z-transform is a power series with an infinite number of terms and so

40πθ =4

π− ππ−

IV

I

θ

dB

Profes

sor E

. Ambik

airaja

h

UNSW, A

ustra

lia

Page 56: Part B Chapter1 Z-Transform-modeemedia.ee.unsw.edu.au/contents/elec3104/LectureNotes/DSP3.pdf · Clearly, the z-transform is a power series with an infinite number of terms and so

-3 -2 -1 0 1 2 3-20

-10

0

10

20

30

40

50dB

theta

Magnitude response

I II IIIIV

Profes

sor E

. Ambik

airaja

h

UNSW, A

ustra

lia

Page 57: Part B Chapter1 Z-Transform-modeemedia.ee.unsw.edu.au/contents/elec3104/LectureNotes/DSP3.pdf · Clearly, the z-transform is a power series with an infinite number of terms and so

Example :sketch the magnitude response for the system having the transfer function

The system has a zero at z = -1 & poles at

)9.01()9.01(

1)(1414

1

−−−

−−

+=

zeze

zzHjj ππ

49.0πj

ez−

=

∴ Magnitude response will be zero at θ = π and large at θ0= ± π /4 because the poles are close to the unit circle.

Profes

sor E

. Ambik

airaja

h

UNSW, A

ustra

lia

Page 58: Part B Chapter1 Z-Transform-modeemedia.ee.unsw.edu.au/contents/elec3104/LectureNotes/DSP3.pdf · Clearly, the z-transform is a power series with an infinite number of terms and so

-π π-π/4 π/4

|H(θ)|

θ

Magnitude Response π/40.9

θ=0θ = -π

θ = π

θ

Profes

sor E

. Ambik

airaja

h

UNSW, A

ustra

lia

Page 59: Part B Chapter1 Z-Transform-modeemedia.ee.unsw.edu.au/contents/elec3104/LectureNotes/DSP3.pdf · Clearly, the z-transform is a power series with an infinite number of terms and so

Example :Sketch the approximate magnitude response from the pole-zero map given below:

|z|= 1

+ 0.8

- 0.8Re(z)

Im(z)dB

-π -π/2 0 π/2 π

-fs/2fs/2

θ

Profes

sor E

. Ambik

airaja

h

UNSW, A

ustra

lia

Page 60: Part B Chapter1 Z-Transform-modeemedia.ee.unsw.edu.au/contents/elec3104/LectureNotes/DSP3.pdf · Clearly, the z-transform is a power series with an infinite number of terms and so

Example :

Sketch an approximate magnitude response from the pole-zero map given below:

+ 0.5

- 0.5 Re(z)1

|z|=1

Im(z)

-π -π/2 0 π/2 π θ

|H(θ)| in dB

Profes

sor E

. Ambik

airaja

h

UNSW, A

ustra

lia

Page 61: Part B Chapter1 Z-Transform-modeemedia.ee.unsw.edu.au/contents/elec3104/LectureNotes/DSP3.pdf · Clearly, the z-transform is a power series with an infinite number of terms and so

Summary of Part B Chapter 1At the end of this chapter, it is expected that you should know:

The properties of z transforms and their application.

Discrete convolution in the time and z domains.

How to find the inverse z transform, given a transfer function.

The difference between a z transform and a Laplace transform and when to use each.

Estimation of frequency response from a transfer function

Hand-calculate magnitude and phase responses for simple transfer functions and plot them.Prof

esso

r E. A

mbikair

ajah

UNSW, A

ustra

lia

Page 62: Part B Chapter1 Z-Transform-modeemedia.ee.unsw.edu.au/contents/elec3104/LectureNotes/DSP3.pdf · Clearly, the z-transform is a power series with an infinite number of terms and so

The pole-zero description of a discrete time system.

Given a pole-zero diagram, transfer function or difference equation, how to find any of the other representations and discuss the system’s stability with reference to the pole-zero diagram.

How to derive the resonance frequency equation for a second-order resonance system (a complex pole pair).

Profes

sor E

. Ambik

airaja

h

UNSW, A

ustra

lia

Page 63: Part B Chapter1 Z-Transform-modeemedia.ee.unsw.edu.au/contents/elec3104/LectureNotes/DSP3.pdf · Clearly, the z-transform is a power series with an infinite number of terms and so

Examples

Profes

sor E

. Ambik

airaja

h

UNSW, A

ustra

lia

Page 64: Part B Chapter1 Z-Transform-modeemedia.ee.unsw.edu.au/contents/elec3104/LectureNotes/DSP3.pdf · Clearly, the z-transform is a power series with an infinite number of terms and so

A digital filter structure is shown below. Determine the Transfer function H(z).

Z-1

Z-1

X(z) a0

a0

+

-b1

b2

Y(z)

+

+

;1

)()()(

22

11

100

−−

−++

==zbzb

zaazHzXzY

Profes

sor E

. Ambik

airaja

h

UNSW, A

ustra

lia

Page 65: Part B Chapter1 Z-Transform-modeemedia.ee.unsw.edu.au/contents/elec3104/LectureNotes/DSP3.pdf · Clearly, the z-transform is a power series with an infinite number of terms and so

Sketch an approximate magnitude response from the pole-zero map given below:

Im(z)

Re(z)1

-0.5

-1

0.5 x

x

-π -π/2 0 π/2 π θ

|H(θ)|

Profes

sor E

. Ambik

airaja

h

UNSW, A

ustra

lia

Page 66: Part B Chapter1 Z-Transform-modeemedia.ee.unsw.edu.au/contents/elec3104/LectureNotes/DSP3.pdf · Clearly, the z-transform is a power series with an infinite number of terms and so

Question 5(b)Determine the magnitude response of the following filter and show that it has an all-pass

characteristic.

[4 marks]1)1(

)( 1

1

<−−

= −

azazazH

filter Pass -All

=−−+−−+

==

−−

=→−−

=+=

−−

11

1)()().(

1)(

1)()1(

21)(

2

22*

*2

θθ

θθ

θ

θ

θ

θ

θθθ

θθ

jj

jj

j

j

j

j

aeaeaaeaeaHHH

aeeaH

aeeaHzzH

)(θH

θ-π π

1

Profes

sor E

. Ambik

airaja

h

UNSW, A

ustra

lia

Page 67: Part B Chapter1 Z-Transform-modeemedia.ee.unsw.edu.au/contents/elec3104/LectureNotes/DSP3.pdf · Clearly, the z-transform is a power series with an infinite number of terms and so

A first-order digital filter is described by the system function :

11

1111)( −−

−++−

=zazb

bazH

Draw a canonic realisation of the transfer function H(z).

Z-1

X(z)

b

+

a

+ Y(z)

1-a/1-b

Profes

sor E

. Ambik

airaja

h

UNSW, A

ustra

lia

Page 68: Part B Chapter1 Z-Transform-modeemedia.ee.unsw.edu.au/contents/elec3104/LectureNotes/DSP3.pdf · Clearly, the z-transform is a power series with an infinite number of terms and so

Sketch roughly the magnitude response corresponding to the pole-zero pattern given below:

Question 3

Profes

sor E

. Ambik

airaja

h

UNSW, A

ustra

lia

Page 69: Part B Chapter1 Z-Transform-modeemedia.ee.unsw.edu.au/contents/elec3104/LectureNotes/DSP3.pdf · Clearly, the z-transform is a power series with an infinite number of terms and so

Sketch an approximate magnitude response from the pole-zero map given below:

Im(z)

Re(z)1

-0.5

-1

0.5 x

x0.5

θ

Mag

Profes

sor E

. Ambik

airaja

h

UNSW, A

ustra

lia

Page 70: Part B Chapter1 Z-Transform-modeemedia.ee.unsw.edu.au/contents/elec3104/LectureNotes/DSP3.pdf · Clearly, the z-transform is a power series with an infinite number of terms and so

The difference equation of a digtal filter is given by y(n) = x(n)-x(n-8)- y(n-2 )

Find the transfer function for the above filter.

28

28

11).1(

)()(

)()()()(

−−

−−

+−=

−−=

zz

zXzY

zzYzzXzXzY

Profes

sor E

. Ambik

airaja

h

UNSW, A

ustra

lia

Page 71: Part B Chapter1 Z-Transform-modeemedia.ee.unsw.edu.au/contents/elec3104/LectureNotes/DSP3.pdf · Clearly, the z-transform is a power series with an infinite number of terms and so

A first-order digital filter has a transfer function given by

Determine the impulse response of the above digital filter H(z) and show that the filter is stable if a < 1.

11

11)( −−

−+=

zazkzH

⎥⎦

⎤⎢⎣

⎡−

+−

= −

− 1

1

1 111)(

azz

azkzH

[ ])1()()( 1 −+= − nuanuaknh nn

if a< 1, then h(n) decays to zero.Therefore, stable system.

Profes

sor E

. Ambik

airaja

h

UNSW, A

ustra

lia