Problem 1 (10 marks, 3.

1 Textbook)
Determine the 75%, 90%, and 95% response time for each of the systems given
(assume zero initial conditions):

(a)
5 ẏ +5 y=U (t )
(b)
2 P̈+ 8 Ṗ+ 8 P=2 U (t)

For part b) determine whether the system is overdamped, underdamped or critically

damped. And solve b) by two methods:

(i) MATLAB plot,

(ii) MATLAB’s fzero function.

Note: for part (b), need to find t where y(t) = 0.75KA, et cet

Solution:
2 P̈+ 8 Ṗ+ 8 P=2 U (t)
Comparing with,
a 2 P̈+a1 Ṗ+a0 =U (t)
We have,
a2=1, a1=4, a0=4
natural frequency:
a 4
√ √
w n= 0 = =2
a2
Damping ratio:
1

a1 4
ξ= = =1
2 √ a0 a2 2 √ 4∗1

Since, the value of damping ratio is 1. The given system is critically damped.
For critically damped system, we have

y (t)=KA−KA ( 1+ wn t ) ⅇ−w t
n

y (t)
=1−( 1+2 t ) ⅇ−2 t
KA

This is the system function that can be used to find the necessary time for given
percentage of response.

(i) Using MATLAB Plot,

1
 Figure 1: Plot of given system with time

The function was plotted using the MATLAB. The necessary value of time for 75%,
90%, and 95% response was calculated.

For 75% response, y(t)=0.75KA t=1.35

For 90% response, y(t)=0.90KA t=1.94
For 95% response, y(t)=0.95KA t=2.37

Code
clc
clear all
close all
t=[0:0.01:10];
fun=@(t) (1-((1+2*t).*exp(-2*t)));
y=fun(t);
hold on

2
plot(t,y);
ylim([0,2]);
xlabel('t');
ylabel('y(t)/KA');
% to determine 75% response time
[difference,index1]=min(abs(y-0.75));
x75=t(index1);
stem(x75,0.75,'-.or');
text(x75,0.75,'(1.35,0.75)75% response time');

% to determine 90% response time

[difference,index2]=min(abs(y-0.9));
x90=t(index2);
stem(x90,0.90,'-.om');
text(x90,0.90,'(1.94,0.90) 90% response time');

% to determine 95% response time

[difference,index3]=min(abs(y-0.95));
x95=t(index3);
stem(x95,0.95,'-.og');
text(x95,0.95,'(2.37,0.95) 95% response time');

(ii) Using Function fzero()

For 75% response, y(t)=0.75KA t=1.3463
For 90% response, y(t)=0.90KA t=1.9449
For 95% response, y(t)=0.95KA t=2.3719

Code
clc
clear all
close all
t=[0:0.01:10];
fun=@(t) (1-((1+2*t).*exp(-2*t)));

%using fzero function

% to determine 75% response time

xx75t=fzero(@(t) fun(t)-0.75,[0,20]);

% to determine 90% response time

xx90t=fzero(@(t) fun(t)-0.90,[0,20]);

% to determine 95% response time

xx95t=fzero(@(t) fun(t)-0.95,[0,20]);

3
Problem 3 (10 marks, 3.32 text)
A measuring system has a known natural frequency of 2000 Hz and damping ratio of
0.8. Would it be satisfactory to use it to measure a signal expected to have frequency
content up to 500 Hz? Demonstrate your answer by finding the expected dynamic error
and phase shift at 500 Hz. What is the system frequency bandwidth? To solve the
equation for the frequency band with use the MATLAB function roots and show your
MATLAB output

Solution:

ω n=2000 Hz , ζ=0.8 ,ω=500 Hz

For the 2nd order system, the magnitude ratio is given by,
1 1
M ( ω )= 1 /2
= 1/ 2
2 2 2 2 2

{[ 1−
ω
ωn( ) ] +(2 ζω/ω ) } {[ ( ) ] (
n
2
1−
500
2000
+ 2× 0.8 ×
500
2000 )}
¿ 0.981
Hence the dynamic error is given by,
δ ( ω )=M ( ω ) −1=0.981−1=−0.019
Hence for frequency upto 500 Hz, δ ( ω ) ≤1.9 % . Hence, it can be satisfactory to measure
frequency upto 500 Hz.
Now phase shift is given by,
ω 500
−2 ζ

( ( )) (
−2× 0.8 ×
∅ ( ω ) =tan −1
1−
ω
ωn

ωn
2
=tan−1
1−
2000
500 2
2000
=−23.106 °
( ) )
For 2nd order system, frequency bandwidth is given by:
−3 dB ≤ 20 log M ( ω)≤ 3 dB
¿ , 0.707 ≤ M (ω)≤ 1.412
1
¿ , 0.707 ≤ 1 /2
≤ 1.412
2 2

{[ 1−
ω
( )]
ωn
+ ( 2 ζω/ω n )
2

4
 1
¿ , 0.707 ≤ 1 /2
≤ 1.412
2 2 2

{[ 1−
ω
( 2000 ) ] +(2 × 0.8× 2000
ω
)}
This range of ω gives the require frequency bandwidth of the system.
Taking reciprocal and squaring both sides,
We get,
2 2 2
¿ , 0.4998 ≥ 1−
{[ω
2000 ( )] ( + 2× 0.8 ×
ω
2000 ) }≥ 1.9937
ω 2 ω 4 2
ω
{ ( )( )(
¿ , 0.4998 ≥ 1−2 × )} 2000
+
2000
+ 1.6 ×
2000
≥1.9937

ω 2 ω 4 2
ω
{ ( ) ( ) ( )}
¿ , 0.4998 ≥ 1−2 ×
2000
+
2000
+2.56 ×
2000
≥1.9937

ω 2 4
ω
{ ( ) ( )}
¿ , 2≥ 1+ 0.56 ×
2000
+
2000
≥0.501

ω
Considering, x= , the equation becomes,
2000
2 ≥ {1+ 0.56 x 2 + x 4 } ≥ 0.501
So, the polynomial 1+0.56 x 2+ x 4 is a parabola symmetric about Y-axis. On plotting the
polynomial in MATLAB, the curve as shown in Figure 2 is obtained.

5
 ω
Figure 2: Plot of 1+0.56 x 2+ x 4 where x =
2000
As seen from the plot, parabola represented by the polynomial turns in upward direction
with its minimum value at 1. So, at the lower limits of the inequality obtained i.e. 0.501,
the roots x will be imaginary which can be verified with the help of roots function in
MATLAB (Figure 3).

Figure3: Output of 1+0.56 x 2+ x 4=0.501

Also, the roots when the polynomial 1+0.56 x 2+ x 4 is equal to 2 is obtained applying the

6
roots function of MATLAB. As seen in Figure 4, the two real roots are equal to -0.87 and
0.87.

Figure 4: Output of 1+0.56 x 2+ x 4=2

Since the polynomial obtained for the frequency bandwidth is constrained between
0.501 to 2, it can be seen that the value of x lies between -0.87 to 0.87 to satisfy the
constraints. Furthermore, since the frequency can’t be negative, the value of x is taken
between 0 to 0.87.
Hence, we obtain;
0 ≤ x ≤ 0.87
ω
0≤ ≤0.87
2000
0 ≤ ω ≤1740
Hence, the system’s frequency bandwidth is0 ≤ ω ≤1740 Hz .

7
8