CHEE 319 – Tutorial 4 Solutions

[This version: 09:16 Tuesday 15th March, 2011]

1. (a) Solution: We must find the transfer function for the block diagram shown in Figure 1,
using the ideas of “block diagram simplification”. We begin with some terminology. The block

R+ + Y
- h - G1 q + h -
- G2 q h
+ ?
- q -
−6 +6

H2 

q G3  h
+6+

- H3

Figure 1: Block diagram

diagram arrangements in Figures 2 and 3 are called negative feedback and positive feedback,
respectively. For negative feedback, the gain is given by
G1 R
Y = ,
1 + G1 G2
whereas for positive feedback, the gain is given by
G1 R
Y = .
1 − G1 G2
It is clear that there are two positive feedbacks in Figure 1, involving G2 , H2 and G3 , H3 .
The block diagram in Figure 1 thus reduces to the block diagram shown in Figure 4. The top
“row” in Figure 4 can be simplified further, so Figure 4 reduces to the block diagram shown
in Figure 5. Figure 5 is a negative feedback. Hence, the transfer function for the given block

R+ Y
- h - G1 q -
−6

G2 

Figure 2: Negative feedback

1
 R+ Y
- h - G1 q -
+6

G2 

Figure 3: Positive feedback

R+ + Y
- h - G1 q - G2 + h
- ? q -
1−G2 H2
−6

G3 
1−G3 H3

Figure 4: Block diagram Figure 1 – After first reduction

R+   Y
- h - G1 1 + G 2 q -
1−G2 H2
−6

G3 
1−G3 H3

Figure 5: Block diagram Figure 1 – After second reduction

2
 R
q q

? ? ?
b3 b2 b1

+ Y3+ ?
+ Y2+ ?
+ Y
h -
? 1 - h - 1 - h - 1 q -
−6 s −6 s −6 s

a3 a2 a1

6 6 6
q q

Figure 6: Block diagram

diagram is (after some simplification)

Y G1 (1 − G2 H2 )(1 − G3 H3 ) + G1 G2 (1 − G3 H3 )
= .
R (1 − G2 H2 )(1 − G3 H3 ) + G1 G3 (1 − G2 H2 ) + G1 G2 G3

(b) Solution: We must find the transfer function for the block diagram shown in Figure 6.
There is not much we can do here in the way of block diagram simplification. Observe that
1
Y = (b1 R + Y2 − a1 Y ),
s
1
Y2 = (b2 R + Y3 − a2 Y ), and
s
1
Y3 = (b3 R − a3 Y ).
s
Plugging the expressions for Y2 and Y3 into the expression for Y , after some simplification we
obtain
1
Y = (b1 Rs2 + b2 Rs + b3 R − a1 Y s2 − a2 Y s − a3 Y ),
s3
or equivalently

(s3 + a1 s2 + a2 s + a3 )Y = (b1 s2 + b2 s + b3 )R.

It follows that the transfer function for the given block diagram is

Y b1 s2 + b2 s + b3
= 3 .
R s + a1 s2 + a2 s + a3

(c) Solution: We must find the transfer function for the block diagram shown in Figure 7.
Let’s first deal with the negative feedback involving 1/s and a1 . Figure 7 reduces to the block

3
R
q q

? ? ?
b3 b2 b1

+ + + + Y
h -
? h -
+ h - 1 q + h
- ? - 1 q - h -
+ ? 1 q -
−6 −6 −6 s s s

a1 

a2 

a3 

Figure 7: Block diagram

R
q q

? ? ?
b3 b2 b1

+ + + + Y
h -
? h - 1 + h
- ? - 1 q + h
- ? - 1 q -
s+a1 s s
−6 −6

a2 

a3 

Figure 8: Block diagram Figure 7 – After first reduction

4
 R
q

? ?
b3 + b2 (s + a1 ) b1

+ + + Y
h -
? h - 1 q h -
+ ?
- 1 q -
s(s+a1 ) s
−6 −6

a2 

a3 

Figure 9: Block diagram Figure 7 – After rearrangement

R
q

? ?
b3 + b2 (s + a1 ) b1

+ + Y
h
? - 1 - h -
+ ? 1 q -
s(s+a1 )+a2 s
−6

a3 

Figure 10: Block diagram Figure 7 – After third reduction

diagram shown in Figure 8. In order to put Figure 8 in a more “recognizable” feedback form,
we must move block b2 out of the inner “loop.” Since the block b2 must multiply R, the most
obvious choice is to move block b2 into block b3 in some manner. It is not hard to verify
directly that Figure 8 is equivalent to the block diagram shown in Figure 9. Reducing the
negative feedback involving 1/(s(s+a1 )) and a2 , Figure 9 reduces to the block diagram shown
in Figure 10. As we did for b2 , we now move block b1 out of the inner “loop.” Figure 10 is
thus equivalent to the block diagram shown in Figure 11. Reducing the negative feedback
involving 1/(s3 + a1 s2 + a2 s) and a3 , Figure 11 reduces to the block diagram shown in Figure
12. From Figure 12, we can easily read off the transfer function for the given block diagram:

Y b3 + b2 (s + a1 ) + b1 (s2 + a1 s + a2 )
= .
R s3 + a1 s2 + a2 s + a3

(d) Solution: We must find the transfer function for the block diagram shown in Figure 13.
Let’s first deal with the negative feedback involving B and H. Figure 13 reduces to the block

5
 R

?
b3 + b2 (s + a1 ) + b1 (s2 + a1 s + a2 )

+ Y
h
? - 1 q -
s3 +a1 s2 +a2 s
−6

a3 

Figure 11: Block diagram Figure 7 – After further rearrangement

R 1 Y
- b3 + b2 (s + a1 ) + b1 (s2 + a1 s + a2 ) -
s3 +a1 s2 +a2 s+a3
-

Figure 12: Block diagram Figure 7 – After fourth reduction

- D

R+ + Y
- h q - A
+ h -
- B q h
+ ?
- q -
−6 −6

H 

G 

Figure 13: Block diagram

- D

R+ + Y
- h q - A - B + h
- ? q -
1+BH
−6

G 

Figure 14: Block diagram Figure 13 – After first reduction

6
 R+ Y
- h - D + BA q -
1+BH
−6

G 

Figure 15: Block diagram Figure 13 – After second reduction

H3 

R − + Y
q h -
+ ?
- h - G1 q - G2 + h
- q -
−6 −6 +6

H1  q

H2 

- G3

Figure 16: Block diagram

diagram shown in Figure 14. Clearly, Figure 14 reduces to the block diagram shown in Figure
15. Figure 15 is a negative feedback. Hence, the transfer function for the given block diagram
is (after some simplification)

Y D + DBH + AB
= .
R 1 + BH + GD + GDBH + GAB

2. Solution: We must find the transfer function for the block diagram shown in Figure 16.
There is an obvious negative feedback involving G1 and H1 , so Figure 16 reduces to the
block diagram shown in Figure 17. Again, there is an obvious negative feedback involving
G1 /(1 + G1 H1 ) and H2 . Since
G1
1+G1 H1 G1
G1
= ,
1+ 1+G1 H1 · H2 1 + G1 H1 + G1 H2

Figure 17 reduces to the block diagram shown in Figure 18. Figure 18 obviously reduces to
the block diagram shown in Figure 19. With YM and Y3 as in Figure 19, it is clear that
G1 G2 G 1 G2 G1 G2
YM = (R − Y3 ) = ·R− · Y3 .
1 + G1 H1 + G1 H2 1 + G1 H1 + G1 H2 1 + G1 H1 + G1 H2

7
 H3 

R − Y
q - h
+ ? - G1 q - G2 + h
- q -
1+G1 H1
−6 +6

H2 

- G3

Figure 17: Block diagram Figure 16 – After first reduction

H3 

R − Y
q - h
+ ? - G1 - G2 + h
- q -
1+G1 H1 +G1 H2
+6

- G3

Figure 18: Block diagram Figure 16 – After second reduction

H3 
Y3

R − YM Y
q - h
+ ? - G1 G2 + h
- q -
1+G1 H1 +G1 H2
+6

- G3

Figure 19: Block diagram Figure 16 – After third reduction

8
 G1 G2 H3
1+G1 H1 +G1 H2
Y2
R Y1 − Y4 Y
q - G1 G2 - h
+ ? + h
- q -
1+G1 H1 +G1 H2
+6
Y3
- G3

Figure 20: Block diagram Figure 16 – After rearrangement

G1 G2 H3
1+G1 H1 +G1 H2

R − Y
q - G1 G2 + h -
- h
+ ? q -
1+G1 H1 +G1 H2
+6

- G3

Figure 21: Block diagram Figure 16 – After further rearrangement

It follows that Figure 19 can be written equivalently as the block diagram shown in Figure 20.
Obviously, since Y4 = Y3 + (Y1 − Y2 ) = (Y3 + Y1 ) − Y2 , Figure 20 can be written equivalently as
the block diagram shown in Figure 21. From Figure 21 we can read off the transfer function.
Writing S , G1 G2 /(1 + G1 H1 + G1 H2 ), we have

Y 1
= (S + G3 ) ·
R 1 + SH3
which after some simplification gives

Y G1 G2 + G3 + G3 G1 H1 + G3 G1 H2
= .
R 1 + G1 H1 + G1 H2 + G1 G2 H3

3. (a) Write the Laplace transform of the signal in Figure 22.

Solution: After some trial and error, one can see that the signal (let’s call it u) in Figure 22
can be written as

u(t) = t11(t) − (t − 1)11(t − 1) − (t − 2)11(t − 2) + (t − 3)11(t − 3).

Then the Laplace transform of u is given by

L (u)(s) = L (t11(t))(s) − L ((t − 1)11(t − 1))(s) − L ((t − 2)11(t − 2))(s) + L ((t − 3)11(t − 3))(s).

9
 Figure 22: The input signal u of Problem 3.11

Using the “time delay” property of Laplace transforms, we obtain

1 e−s e−2s e−3s

L (u)(s) = − − 2 + 2
s2 s2 s s
1 − e−s − e−2s + e−3s
= .
s2

(b) Determine the Laplace transform of the output if u is applied to G, where

ωn2
G(s) , , ζ, ωn ∈ R.
s2 + 2ζωn s + ωn2

Solution: By definition, we have

Y (s)
= G(s), or Y (s) = G(s)U (s).
U (s)

In part (a), we computed U (s) ≡ L (u)(s). Plugging it in, we see that Y (the Laplace
transform of the output) is given by

ωn2 1 − e−s − e−2s + e−3s

Y (s) = G(s)U (s) = · .
s2 + 2ζωn s + ωn2 s2

(c) Assuming that ωn 6= 0, find the output of the system if u is applied to G.

Solution: We will first compute the output y(t) = L −1 (Y )(t) of the system if a ramp input
is applied, i.e., when U (s) = 1/s2 . In this case, we have

ωn2
Y (s) = .
s2 (s2 + 2ζωn s + ωn2 )

10
Without further assumptions on ζ, the partial fraction expansion of Y is given by

1 2ζ 1 1 2ζs + 4ωn ζ 2 − ωn
Y (s) = − · + · ,
s2 ωn s ωn s2 + 2ζωn s + ωn2

which can be rewritten as

1 2ζ 1 1 2ζs + 4ωn ζ 2 − ωn
Y (s) = 2
− · + · p .
s ωn s ωn (s + ωn ζ)2 + (ωn 1 − ζ 2 )2

To compute this partial fraction expansion in Maple, one issues the commands

> u := omega^2 / (s^2 * (s^2 + 2 * zeta * omega * s + omega^2));

2
omega
u := ---------------------------------
2 2 2
s (s + 2 zeta omega s + omega )
> convert(u,parfrac,s);
2
2 zeta s + 4 omega zeta - omega 2 zeta 1
------------------------------------ - ------- + ----.
2 2 omega s 2
omega (s + 2 zeta omega s + omega ) s

We must make use of the fact that for γ 6= 0,

e−αt (β − α)
 
−1 s+β −αt
L (t) = e cos(tγ) + sin(tγ).
(s + α)2 + γ 2 γ

Let us first see what happens when ζ 6= 0.

p
Case 1: Suppose that ζ 2 6= 1 so that ωn 1 − ζ 2 6= 0. We then have

s + 2ωn ζ − ω2ζn
!
−1 2ζ 1 −1
L (Y )(t) = t − + ·L p (t)
ωn 2ζωn (s + ωn ζ)2 + (ωn 1 − ζ 2 )2
" #
−ωn ζt
 
2ζ 1 −ωn ζt
p e ωn
p
= t− + · e cos(tωn 1 − ζ 2 ) + p · ωn ζ − sin(tωn 1 − ζ 2 )
ωn 2ζωn ωn 1 − ζ 2 2ζ

for t ≥ 0.
p
Case 2: Suppose that ζ 2 = 1, so that ωn 1 − ζ 2 = 0. The partial fraction expansion of Y
becomes

1 2ζ 1 1 s + 2ωn ζ − ω2ζn
Y (s) = 2 − · + ·
s ωn s 2ζωn (s + ωn ζ)2

11
 We then have
     
−1 2ζ 1 s ωn 1
L (Y )(t) = t − + + L −1 (t) + 2ωn ζ − ·L −1
(t)
ωn 2ζωn (s + ωn ζ)2 2ζ (s + ωn ζ)2
 
2ζ 1 −ωn ζt ωn
= t− + + (1 − ωn ζ)e + 2ωn ζ − · te−ωn ζt
ωn 2ζωn 2ζ

for t ≥ 0.
On the other hand, if ζ = 0, then the partial fraction expansion of Y is given by
1 1
Y (s) = − ,
s2 s2 + ωn2

so that

1
L −1 (Y )(t) = t − · sin(ωn t)
ωn

for t ≥ 0.
Finally, using the “time delay” property of Laplace transforms, we obtain

y(t) = y(t) − y(t − 1) + y(t − 2) + y(t − 3).

4. Solution: For a second-order system with transfer function

3 3
G(s) = = ,
s2 + 2s − 3 (s − 1)(s + 3)

we must determine its DC gain, and its final value given a unit step input. For the former,
recall that the DC gain of a system with transfer function G is given by

lim G(s).
s→0

Since G is continuous at s = 0, its DC gain is given by

3
G(s)|s=0 = = −1.
−3

To determine the system’s final value given a unit step input, we use the final value theorem.
We have
Y (s) 3
G(s) = = .
U (s) (s − 1)(s + 3)

When U (s) = 1/s (i.e., given a step input), we have

1 3
Y (s) = · .
s (s − 1)(s + 3)

12
By the final value theorem, if all the poles pi of sY (s) satisfy Re(pi ) ≤ 0, then

lim y(t) = lim sY (s),

t→∞ s→0

where y denotes the inverse Laplace transform of Y . However, it is clear that sY (s) has poles
at s = 1 and s = −3, so the conditions of the final value theorem are not satisfied. Indeed,
using a partial fraction expansion (left as an exercise), one can easily show that
3
y(t) = (et − e−3t ),
4
and so

lim y(t) = ∞.
t→∞

13