1 1 0.5z 1 1 0.3z 1 1 0.7z 1 (1 0.5z) (1 0.3z) 1 (1 0.5z) (1 0.7z) 1 1 0.5z 1 1 0.3z 1 1 0.7z

Download as pdf or txt
Download as pdf or txt
You are on page 1of 4

P. C.

Yeh Application of Z-Transform to System Analysis 1

I. Transfer function

Given the impulse response of h(n) of a system, the transfer function of the system is defined as

Z
h(n) −
→ H(z).

For such a system, we often denote it by h(n) or H(z) .

Since we know x(n)→ h(n) →y(n) = h(n) ∗ x(n) ⇒ Y (z) = H(z) · X(z). As long as we know the Z-tx
of the input signal, we can obtain Y (z) and thus y(n) through the transfer function and inverse Z-tx.
On the other hand, since Y (z) = H(z) · X(z) ⇒ H(z) = Y (z)/X(z), which implies that the transfer
function can be computed if we know the Z-tx of any input-output signal pair.

Example 1: h(n) = 0.5n u(n), x(n) = 0.3n u(n) + 0.7n u(−n − 1) ⇒ y(n) = ?
1 1 1
H(z) = 1−0.5z −1 , X(z) = 1−0.3z −1 − 1−0.7z −1

⇒ Y (z) = (1−0.5z −1 )1(1−0.3z −1 ) − (1−0.5z −1 )1(1−0.7z −1 ) = 5 · 1−0.5z


1 1
−1 − 1.5 · 1−0.3z −1 − 3.5 ·
1
1−0.7z −1
,
with ROCY : {z : |z| > 0.5} ∩ {z : 0.3 < |z| < 0.7} = {z : 0.5 < |z| < 0.7}
⇒ y(n) = 5 · 0.5n u(n) − 1.5 · 0.3n u(n) + 3.5 · 0.7n u(−n − 1).

Example 2: Given that an input-output pair of a causal system: x(n) = 0.5n u(n), y(n) = 0.7n u(n).
What is h(n)?
To find h(n), we first find the transfer function H(z):
1
1−0.7z −1 1−0.5z −1 5 2 1
H(z) = Y (z)/X(z) = 1 = 1−0.7z −1
= 7 − 7 · 1−0.5z −1
.
1−0.5z −1
5 2
Since h(n) is causal ⇒ h(n) = Z −1 (H(z)) = 7 · δ(n) − 7 · 0.5n u(n).

II. How to get transfer function for the system block diagram?

Example: Given the following block diagram of the system

0.2

-0.5
x(n)
w(n)
z -1 z-1 y(n)

0.3

What is the transfer function H(z)?


P. C. Yeh Application of Z-Transform to System Analysis 2

Let the Z-tx of x(n), y(n) and w(n) be X(z), Y (z) and W (z) respectively.
From the node closest to y(n), we have
£ ¤ ¡ ¢
Y (z) = z −2 W (z) + 0.2 W (z) − 0.5 z −1 W (z) = W (z) 0.2 − 0.5z −1 + z −2 . (1)

From the node closest to x(n), we have

X(z)
W (z) = X(z) + 0.3 z −1 W (z) ⇒ W (z) = (2)
1 − 0.3z −1
X(z) ¡ ¢ 0.2−0.5z −1 +z −2 0.2−0.5z −1 +z −2
By (1), (2) ⇒ Y (z) = 1−0.3z −1
· 0.2 − 0.5z −1 + z −2 = 1−0.3z −1
· X(z) ⇒ H(z) = 1−0.3z −1
.

III. How to get transfer function from the system difference equation?

Given a system is descried by the difference equation

y(n) + a1 y(n − 1) + · · · + aM y(n − M ) = b0 x(n) + b1 x(n − 1) + · · · + bK x(n − K),

what is the transfer function H(z)?


If we take the Z-tx of both sides of the difference equation, we have

Y (z) + a1 z −1 Y (z) + · · · + aM z −M Y (z) = b0 X(z) + b1 z −1 X(z) + · · · + bK z −K X(z)


¡ ¢ ¡ ¢
⇒ Y (z) 1 + a1 z −1 + · · · + aM z −M = X(z) b0 + b1 z −1 + · · · + bK z −K
b0 +b1 z −1 +···+bK z −K
⇒ H(z) = Y (z)/X(z) = 1+a1 z −1 +···+aM z −M
.

Let A(z) be the Z-tx of the coefficients of y(·) and B(z) be the Z-tx of the coefficients of x(·), i.e.
A(z) , 1 + a1 z −1 + · · · + aM z −M ) and B(z) , b0 + b1 z −1 + · · · + bK z −K ⇒ H(z) = B(z)
A(z) .

> Modes and Poles : For each system difference equation, modes are defined as the roots of A(z) = 0.
Since A(z) and B(z) might have some common factors that cancel out each other
⇒ some modes might be cancelled out when we try to find the poles of H(z)
⇒ {poles} ⊆ {modes}.
The equality holds only when there is no cancellation between A(z) and B(z).

Example: Given the system difference equation: y(n) − 0.5 y(n − 1) + 0.06 y(n − 2) = x(n) − 0.3 x(n −
1) + 0.03 x(n − 2), H(z) = ? Modes? Poles?
First find A(z) and B(z):
A(z) = Z{{1, −0.5, 0.06}} = 1 − 0.5z −1 + 0.06z −2 = (1 − 0.2z −1 ) (1 − 0.3z −1 )
B(z) = Z{{1, −0.3, 0.03}} = 1 − 0.3z −1 + 0.03z −2 = (1 − 0.1z −1 ) (1 − 0.3z −1 ).
⇒ modes: 0.2, 0.3.
P. C. Yeh Application of Z-Transform to System Analysis 3

»
»
Since H(z) = B(z)
= (1−0.1z −1 ) » »» −1 )
(1−0.3z
= 1−0.1z −1
⇒ poles: 0.2.
»
A(z) −1
(1−0.2z ) » »»»
(1−0.3z −1 ) 1−0.2z −1

IV. How to get the system difference equation from the transfer function?

0.2−0.5z −1 +z −2
Example: Given H(z) = 1−0.3z −1
, what is the system difference equation?
Y (z) 0.2−0.5z −1 +z −2
H(z) = X(z) = 1−0.3z −1

⇒ Y (z) (1 − 0.3z −1 ) = X(z) (0.2 − 0.5z −1 + z −2 )

⇒ Y (z) − 0.3z −1 Y (z) = 0.2 X(z) − 0.5 z −1 X(z) + z −2 X(z).

Take the inverse Z-tx of both side

⇒ y(n) − 0.3 y(n − 1) = 0.2x(n) − 0.5x(n − 1) + x(n − 2).

V. How to represent a system?

As we have seen,

M
X K
X
h(n) ⇐⇒
|{z} H(z) ⇐⇒
|{z} System D.E. ai y(n − i) = bj x(n − j).
Z{h(n)} B(z)/A(z) i=0 j=0

The system can be represented either by h(n), H(z), or the system difference equation. They are all
equivalent.

VI. How to solve difference equation with all-zero initial condition?

Given the system difference equation and the input x(n). Assume that y(n) = 0 ∀ n < 0 (all-zero
initial condition), we can solve y(n) by two-sided Z-tx easily. The y(n) correspond to all-zero condition
is called the ZSR (Zero State Response) of the system.

Example: Given the system difference equation: y(n) − 0.5 y(n − 1) + 0.06 = x(n) − 0.3 x(n − 1) + 0.03
and x(n) = 0.3n u(n). The system is causal. What is the ZSR of the system?
1−0.2z −1
In the example of III., we have the transfer function of the system: H(z) = 1−0.1z −1
1−0.2z −1 1 1−0.2z −1 1 1
⇒ Y (z) = H(z) X(z) = 1−0.1z −1 · 1−0.3z −1 = (1−0.1z −1 )(1−0.3z −1 ) = 0.5 · 1−0.1z −1 + 0.5 · 1−0.3z −1 .

Since the system is causal


⇒ ROCH : |z| > 0.1 ⇒ ROCY = ROCH ∩ ROCX = {z : |z| > 0.3}
⇒ ZSR: y(n) = 0.5 · 0.1n u(n) + 0.5 · 0.3n u(n).
P. C. Yeh Application of Z-Transform to System Analysis 4

VII. Causality and stability of a system

Recall that we can tell the causality and the stability of a system from ROCH , the ROC of H(z).

• If the shape of ROCH is the exterior of a disc ⇒ the system is causal .

• If the shape of ROCH is the interior of a disc ⇒ the system is anti-causal .

• If the shape of ROCH is donut-shaped ⇒ impuls response h(n) is two-sided .

• Unit circle: |z| = 1 ⊆ ROCH ⇔ the system is BIBO stable .

• If the system is stable and causal ⇒ all poles inside the unit circle .

• If the system is stable and anti-causal ⇒ all poles outside the unit circle .

1−3z −1
Example: Given the transfer function H(z) = (1−0.2z −1 ) (1−1.5z −1 ) (1−2z −1 )
.

There are four possible ROC’s: {|z| < 0.2}, {0.2 < |z| < 1.5}, {1.5, < |z| < 2}, {|z| > 2}

• If the system is causal ⇒ ROCH : |z| > 2.

• If the system is anti-causal ⇒ ROCH : |z| < 0.2.

• If h(n) is two-sided ⇒ ROCH is either 0.2 < |z| < 1.5 or 1.5 < |z| < 2.

• If the system is stable ⇒ ROCH : 0.2 < |z| < 1.5.

>If the system is causal or anti-causal or stable ⇒ the ROC is uniquely determined.

You might also like