Solutions to Problems 5-1

Chapter 5 – Solutions to Instructor Reserve

Problems

I-1 Instructor
Push G1(s) to the left past the pickoff point.

H1

-
+
+
G1 G2 G3

+
1
G1

( )( )
( G1 G 2 +1 ) G3
G1 1
T ( s )= G 2+ G3 =
Thus,
1+ G 1 H 1 G1 ( 1+G1 H 1 )

I-2 INSTRUCTOR

a. Split G3 and combine with G2 and G4. Also use feedback formula on G6 loop.

Push G2 +G3 to the left past the pickoff point.

Using the feedback formula and combining parallel blocks,

Multiplying the blocks of the forward path and applying the feedback formula,
 Solutions to Problems 5-3

I-3 Instructor
a. Push G7 to the left past the pickoff point. Add the parallel blocks, G3+G4.

Push G3+G4 to the right past the summing junction.

Collapse the minor loop feedback.

Push to the left past the pickoff point.

Push G1 to the right past the summing junction.

 Solutions to Problems 5-5

Add the parallel feedback paths to get the single negative feedback,

H(s) = + - . Thus,

T(s) =

b.
Program:
G1=tf([0 1],[1 7]); %G1=1/s+7 input transducer
G2=tf([0 0 1],[1 2 3]); %G2=1/s^2+2s+3
G3=tf([0 1],[1 4]); %G3=1/s+4
G4=tf([0 1],[1 0]); %G4=1/s
G5=tf([0 5],[1 7]); %G5=5/s+7
G6=tf([0 0 1],[1 5 10]); %G6=1/s^2+5s+10
G7=tf([0 3],[1 2]); %G7=3/s+2
G8=tf([0 1],[1 6]); %G8=1/s+6
G9=tf([1],[1]); %Add G9=1 transducer at the input
T1=append(G1,G2,G3,G4,G5,G6,G7,G8,G9);
Q=[1 -2 -5 9
2180
3180
4180
5 3 4 -6
6700
7 3 4 -6
8 7 0 0];
inputs=9;
outputs=7;
Ts=connect(T1,Q,inputs,outputs);
T=tf(Ts)
 Computer response:

Transfer function:

6 s^7 + 132 s^6 + 1176 s^5 + 5640 s^4 + 1.624e004 s^3

+ 2.857e004 s^2 + 2.988e004 s + 1.512e004

-----------------------------------------------------------

s^10 + 33 s^9 + 466 s^8 + 3720 s^7 + 1.867e004 s^6

+ 6.182e004 s^5 + 1.369e005 s^4 + 1.981e005 s^3

+ 1.729e005 s^2 + 6.737e004 s - 1.044e004

I-4 Instructor
Push G3 to the left past the pickoff point. Push G6 to the left past the pickoff point.
 Solutions to Problems 5-7

Hence,

Thus the transfer function is the product of the functions, or

=
I-5 Instructor
Push G3(s) to the left past the pickoff point.

Push G2(s)G3(s) to the left past the pickoff point.

Push G1(s) to the right past the summing junction.

 Solutions to Problems 5-9

Collapsing the summing junctions and adding the feedback transfer functions,

G1 (s)G2 (s)G3 (s)

T (s) 
1  G1 (s)G2 (s)G3 (s)Heq (s)
where

H3(s) H 1( s ) H2(s) H4(s)

H eq ( s )= + + + +1
G 3 ( s ) G 2 ( s ) G 3 ( s ) G 1 ( s )G 3 ( s ) G 1 ( s )
I-6 Instructor
The equivalent mechanical system is found by reflecting all mechanical impedances
to the spring.

Writing the equations of motion:

( 4s 2+2 s+5 ) θ 1( s)−5θ2( s)=4T ( s)
−5 θ1 (s)+ ( 2 s2 +5 ) θ2 ( s)=0
Solving for 2(s),
( 2
| 4 s +2 s+5
) 4 T (s ) |
−5 0 20T ( s )
θ2 (s )= = 4 3 2
( 4 s +2 s+ 5 )
2
−5 8 s + 4 s + 30 s +10 s
| |
−5 ( 2 s 2 +5 )
The angular rotation of the pot is 0.25 that of 2, or
θ p (s ) 2.5
=
T ( s ) s ( 4 s +2 s 2 +15 s+5 )
3

For the pot:

E p (s) 25 2.5
= =
θ p ( s) 5(2 π ) π
For the electrical network: Using voltage division,
E o( s ) 200 , 000 s
= =
Ep (s ) 1 1
+200 ,000 s+
−5
10 s 2

Substituting the previously obtained values,

6 .25
s
Eo (s )
T (s )
= ( )(
θ p (s )
T( s)
Ep ( s )
θ p (s ) )( ) ( )(
E o( s )
E p (s )
=
s s+
1
π

4 s3 +2 s 2 +15 s +5 )
2
 Solutions to Problems 5-11

I-7 Instructor
a.
·
x 1=x 2
·
x 2 =x3
·
x 3 =- 2 x 1−4 x 2−6 x3 +r
y=x 1 + x2

1 1 1 1
r 1 x x x
s 3 s 2 s 1 1 y

-6

-4
-2

b.
·
x 1=x 2
·
x 2 =- 3 x 2 + x 3 +r
·
x 3 =- 3 x 1−4 x 2 −5 x 3 +r
y=x 1 + 2 x 2
1

c.
 ·
x 1=7 x 1 + x2 +r
·
x 2 =- 3 x 1 +2 x2 −x 3 + 2r
·
x 3 =- x 1 +2 x 3 +r
y=x 1 + 3 x 2 + 2 x 3

-3

I-8 Instructor
Closed-loop gains: G2G4G6G7H3; G2G5G6G7H3; G3G4G6G7H3; G3G5G6G7H3; G6H1; G7H2

Forward-path gains: T1 = G1G2G4G6G7; T2 = G1G2G5G6G7; T3 = G1G3G4G6G7; T4 =

G1G3G5G6G7

Nontouching loops 2 at a time: G6H1G7H2

 = 1 - [H3G6G7(G2G4 + G2G5 + G3G4 + G3G5) + G6H1 + G7H2] + [G6H1G7H2]

1 = 2 = 3 = 4 = 1

T(s) =

=
 Solutions to Problems 5-13

I-9 Instructor

T(s) = =

I-10 Instructor
Program:
numg=8;
deng=poly([0 -8 -10]);
G=tf(numg,deng);
T=feedback(G,1);
[numt,dent]=tfdata(T,'v');
[A,B,C,D]=tf2ss(numt,dent); %Obtain controller canonical form
'(a)'; %Display label
A=rot90(A,2); %Convert to phase-variable form
A=fliplr(A); %Convert to phase-variable form
B=rot90(B,2); %Convert to phase-variable form
C=fliplr(C); %Convert to phase-variable form
'(b)'; %Display label
[a,b,c,d]=canon(A,B,C,D); %Convert to parallel form

Computer response:

T=
8

-----------------------

s^3 + 18 s^2 + 80 s + 8

Continuous-time transfer function.

numt =

0 0 0 8

dent =

1 18 80 8

ans =

(a)
A=
 0 1 0
1 0 0
-18 -80 -8

B=
0
0
1

C=
8 0 0

ans =
(b)
a=
-8 0 0
0 1 0
0 0 -1

b=
0.0625
0
0

c=
0 0.0442 0.0442
d=
0
 Solutions to Problems 5-15

I-11 Instructor

x1 = A1x1 + B1r (1)

y1 = C1x1 (2)

x2 = A2x2 + B2r (3)

y2 = C2x2 (4)
In vector-matrix form,

x1 A1 O x 1
= - + B1 r
x2 O A2 x 2 B2
x1
y = y1 + y 2 = C1 C2
x2
I-12 Instructor
 I-13 Instructor
Eigenvalues are -1, -2, and -3 since,

| - A | = (+ 3) (+ 2) (+ 1)

Solving for the eigenvectors, Ax = x

or,

For  = -1, x2 = 0, x1 = x3 . For  = -2, x1 = x2 = . For  = -3, x1 = - , x2 = x3 . Thus,

z= P-1APz + P-1Bu ; y = CPz, where

I-14 Instructor
Program:
numg1=-0.125*[1 0.435]
deng1=conv([1 1.23],[1 0.226 0.0169])
'G1'
G1=tf(numg1,deng1)
'G2'
G2=tf(2,[1 2])
G3=-1
'H1'
H1=tf([-1 0],1)
'Inner Loop'
Ge=feedback(G1*G2,H1)
'Closed-Loop'
T=feedback(G3*Ge,1)

Computer response:
numg1 =

-0.1250 -0.0544

deng1 =

1.0000 1.4560 0.2949 0.0208

ans =

G1

Transfer function:
-0.125 s - 0.05438
 Solutions to Problems 5-17

------------------------------------
s^3 + 1.456 s^2 + 0.2949 s + 0.02079

ans =

G2

Transfer function:
2
-----
s+2

G3 =

-1

ans =

H1

Transfer function:
-s

ans =

Inner Loop

Transfer function:
-0.25 s - 0.1088
------------------------------------------------
s^4 + 3.456 s^3 + 3.457 s^2 + 0.7193 s + 0.04157

ans =

Closed-Loop

Transfer function:
0.25 s + 0.1088
-----------------------------------------------
s^4 + 3.456 s^3 + 3.457 s^2 + 0.9693 s + 0.1503
I-15 Instructor
a. Since VL(s) = Vg(s) – VR(s), the summing junction has Vg(s) as the positive input and VR(s) as the
negative input, and VL(s) as the error. Since I(s) = VL(s) (1/(Ls)), G(s) = 1/(Ls). Also, since VR(s) =
I(s)R, the feedback is H(s) = R. Summarizing, the circuit can be modeled as a negative feedback
system, where G(s) = 1/(Ls), H(s) = R, input = Vg(s), output = I(s), and error = VL(s), where the
negative input to the summing junction is VR(s).
1
I ( s) G( s ) Ls 1
T ( s )= = = =
V g ( s ) 1+G( s ) H ( s ) 1 Ls + R 1
1+ R I (s )=V g ( s )
b. Ls . Hence, Ls +R .
V g( s )
I ( s )=
c. Using circuit analysis, Ls+ R .

I-16 Instructor
a. Adding currents at the op-amp’s inverting terminal, under ideal condition we get
vi  v1 v1  vo

Ri Rf
which after some algebraic manipulations gives
Rf Ri
v 1= vi+ vo
Rf + Ri Rf + Ri
v =−Av 1
Also from the circuits diagram o

b.These equations can be represented by the following block diagram

+
vi Vo
vi
Rf +
-A
vo
++ -
R f  Ri
v1
Ri
R f  Ri

Rf Ri Ri
T 1=− A L=− A L=− A
We have that
Rf + Ri ; Ri + R f ;
Ri + R f ;  1 1
Rf
−A
v o T 1 Δ1 Ri + R f
= =
vi Δ Ri
1+ A
Ri + R f
 Solutions to Problems 5-19

Rf Rf
−A −
vo R i + Rf R i + Rf R f
=Lim = =
v i A →∞ Ri Ri Ri
1+ A
c.
Ri + R f R i + Rf
 I-17 Instructor

Block Diagram:
 Solutions to Problems 5-21

Case
Structure Details:
 Solutions to Problems 5-23

I-18 Instructor
a. There are three forward paths:

(
M 1=K p 2 K p 1 + )
Ki1 1
^ s3
s Lf C f C f

M =( K )
K i11
2 p1 + 2
s LC s f f
1
M 3=( L^ f s + ^
Rf ) 2
Lf C f s
The loops are:

(
L1=−K p 2 K p1 +
Ki1
) 1
s Lf C f s2

L =−( K )
K 1 i1
2 p1 +
s L s f

−Rf
L3 =
Lf s
There are no non-touching loops.

(
∆=1+ K p 2 K p 1 +
Ki1
)
1
s Lf C f s 2
+ K p1+
K i1 1
+ (
Rf
s Lf s Lf s )
and ∆ 1=∆ 2=∆ 3=1

V Load M 1 ∆1 + M 2 ∆2+ M 3 ∆ 3
= =
(
K p2 K p1+ )
Ki1 1
^
s Lf C f C f s 3
+ K p 1 +
K i1 1
s Lf C f s (
+ ^L s + ^
2 ( f )
Rf )
1
Lf C f s
2

1+ K ( K ) ( )
I Cf ∆ Ki1 1 K 1 R
p2 p1 + 2
+ K p 1+ i 1 + f
s Lf Cf s s Lf s L f s
b. There are three forward paths:

(
M 1=K p 2 K p 1 + )
Ki1 1
^ s3
s Lf C f C f

M =( K )
K 1i1
2 p1 + 2
s LC s f f
1
M 3=( L^ f s + ^
Rf ) 2
Lf C f s
The loops are:
 (
L1=−K p 2 K p1 + )
Ki1 1
s Lf C f s2

L =−( K )
K 1
i1
2 p1 +
s L s f

−Rf
L3 =
Lf s
There are no non-touching loops.

(
∆=1+ K p 2 K p 1 +
Ki1 1
)
s Lf C f s 2
+ K p1+ (
K i1 1
+
Rf
s Lf s Lf s )
and ∆ 1=∆ 2=∆ 3=1

V Load M 1 ∆1 + M 2 ∆2+ M 3 ∆ 3
= =
(
K p2 K p1+ )
Ki1 1
^ s
s Lf C f C f
3
+ K p1+
K i1
(1
s Lf C f s 2 )
+ ( ^Lf s + ^
Rf )
1
Lf C f s
2

1+ K ( K ) ( )
I Cf ∆ Ki1 1 K 1 R
p2 p1 + 2
+ K p 1+ i 1 + f
s Lf Cf s s Lf s L f s
The loops and are the same as in part a. There is only one forward path
−1
M 1= and ∆ 1=1
Cf s
1
V Load M 1 ∆1 Cf s
= =-
I Dist ∆
(
1+ K p 2 K p 1+
Ki1
) 1
s Lf Cf s 2 (
K
+ K p 1+ i 1
1 R
+ f
s Lf s Lf s )
I-19 Instructor
X 1 aV 1
=
δ bh g
s2 −
a. Following the procedure described in Chapter 3 we define h and
V g aV V
ϕ=(s + ) X 1( s ) ẍ 1− x 1 = δ ϕ= ẋ 1 + x 1
A . In time domain h bh . a and we also define
ẋ 1=x 2 . These equations give

[ ][ ] [ ]
0 1 0
[ ]
ẋ1
ẋ 2
= g
h
0
x1
x2
+ aV δ
bh
 Solutions to Problems 5-25

ϕ=
[ V
a
1
][ ]
x1
x2

b.The eigenvalues can be obtained directly from the transfer function poles. Thus
λ 12=±
√ g
h

Consider
λ 1=
√ g
h , the first eigenvector is found from the solution of Ax1 =λ1 x 1 or

[ ][
0 1
g
h
0
x1
x2
=
]√[ ]
g x1
h x2
. This results in
x 2=
√ g
x
h 1 . Arbitrarily let x 1=1 so the first

[√ ]
1
v 1= g
eigenvector is h .

[ ][
0 1

Similarly for
λ 2=−
g
√ g
h ; Ax2 =λ1 x 2 or h
0
x1
x2 ] √[ ]
=−
g x1
h x2
resulting in

[√ ]
1

x 2 =−
√ g
x
h 1 . Letting arbitrarily x 1=1 the second eigenvector is
v 2= g
h .

[√ √ ]
1 1
P= [ v 1 v 2 ]= g g
−
c. The similarity transformation matrix is h h

[ ] √
g
− −1
h

√
[ ] √
g 1 1 h
− 1
−1
h 2 2 g
P = =
−2
√ g
h
1
2
−
1
2 √ h
g

The matrices for the diagonalized are calculated as follows:

[ √
][ ][ √ ][ √
]
1 1 h g
0 1 1 1 0
2 2 g h
A d =P−1 AP=
1
2
−
1
2 √ h
g
g
h
0 g
h
−
√ g =
h 0 −
√ g
h
 [ ][ ] [ ] √
1 1 h 1 aV
0
2 2 g 2 b √ gh
Bd =P−1 B= aV =

√
1 1 h 1 aV
− bh −
2 2 g 2 b √ gh

[ √][ √
1 1
C d =CP=
V
a [ 1 ]√ g
h
−
h
V
g = +
a
g
h
V
a
−
g
h √]
The diagonalized representation is:

√
[ ] [ ]
g 1 aV
0

[] ż1

√ [
ż 2 ]
=
h
0 −
g
h
z1
+
2 b √ gh
z 2 − 1 aV
2 b √ gh
δ

[ √ √ ][ ]
ϕ=
V
a
+
g
h
V
a
−
g
h
z1
z2

SOLUTIONS TO DESIGN PROBLEMS

I-20 Instructor
a.
T(s) = ; from which, 2n = 1 and n = 5. Hence,  = 0.1. Therefore,

%OS=e−ζπ / √ 1−ζ x 100=72 . 92% ; Ts = = 8.

2

b.
T(s) = ; from which, 2n = 1+25K2 and n = 5.

% OS
−ln ( )
100 −ln 0 .2
ζ= = =0 . 456

Hence, √ 2
π + ln (
% OS
100
2
)
√ π 2 +ln 2 0. 2

32
=1. 28
Also, Ts = = 0.25, Thus, n = 16; from which K2 = 25 and

n = 35.09. Hence, K1 = 49.25.

 Solutions to Problems 5-27

I-21 Instructor

a. The transfer function derived for that system was:

θ1 (s) 1
G( s) = =
T ( s) (J eq s + Deq s + K eq )
2

, where:

N1 N1 N1
J eq = 1+ ( )2 D eq = ( )2 K eq = ( )2
N2 N2 N2
; ; .

or

1
θ 1( s ) 1+ n2
G( s ) = =
T ( s) n2 n2
s2 + s +
1+ n2 1+ n2

n2 n2
∴ 2 ςωn= ω 2n =
1+n2 1+n 2
and . From Ts = 10 sec = 4/n, we find n = 2.

0.8 0.8
2 2
s + 0.8 s + 0.8 s + 0.8 s + 1.6
b. In Figure P5.58, KG(s) = ; and T(s) = ,

where 2n = 0.8; 2n = 1.6

This results in: n = 1.265 rad/sec &  = 0.32; % O.S. = 33% and Ts = 10 sec.

I-22 Instructor

a. The block diagram of this system is shown below.

b. The transfer function of GP(s) of the circuit representing the process may be derived as
shown in example 2.6. That yields (see equation 2.66):
 V o(s) 1/ LC 106
G P (s )= = = 2
V f (s) R 1 s +10 R p s+ 106
s 2 + p s+
L LC

Hence:
ω n =103 and for a critically damped process (ζ =1), we have:

2 ςωn =10 R p ⇒ R p =2×10 3 /10=200 Ω

c. In the circuit representing the final control element:

10 8
104 ×
1 .8 s 5 .556×107
Z f =10 K Ω // 0. 022 μF ⇒ Z f (s )= =
4 10
8 ( s+5556 )
10 + Z (s)=10 4
1. 8 s ; i

V f ( s ) Z f ( s ) 5 .556×103
∴ GF (s )= = =
V c (s ) Z i ( s ) s+5556

The overall system transfer function is given by:

C (s) G ( s )⋅GF ( s )⋅G P ( s ) 5 .556×109 K P

T ( s )= = C =
R ( s ) 1+ GC ( s )⋅G F ( s )⋅GP ( s ) ( s+5556 )( s2 +2000 s +106 )+5 .556×109 K P

4
T s= =0 . 004 ⇒ ωn =2000
Given that ς=0 . 5 and
ςωn , the dominant poles

should be:

s1 , 2=− ςωn ± jωn √1− ς2 =−1000± j 1732 , which are the roots of
s2 +2000 s+4×106 .

Thus, the third closed-loop pole may be found by dividing the characteristic polynomial

by s2 +2000 s+4×106 , which gives s+1389 (1+K P ) .

This third pole is non-dominant for KP > 2.6.

Given that its tolerance is ± 10%, I would set the controller potentiometer, RF, to:

R F=10 kΩ×K p ×1 .1=28 .6 k Ω

Solutions to Problems 5-29