Andrew Hung A09995772 MAE143c HW1

1a.
180 degree root locus, za=5

Root Locus
20

15

Imaginary Axis (seconds-1)

10

-5

-10

-15

-20
-20

-15

-10

-5

10

15

20

25

30

Real Axis (seconds -1)

0 degree root locus (za=5)

Root Locus
0.8

0.6

Imaginary Axis (seconds-1)

0.4

0.2

-0.2

-0.4

-0.6

-0.8
-2

Real Axis (seconds -1)

10

12

Breakaway points
a(s)b(s) a(s)b(s) = 0.
(2S-10)(S^2-10S)-(S^2-10S+25)(2S-10)=0
S=5
We need a positive K to stabilize this system.
With Mp=15% and tr=18s
s= -5 +- 8.66i
Let G(s+)D(s+)=-1
Solve, and equate the constants of the imaginary and real components to each other
K1(-100)(-5-5+8.66i)(-5-za+8.66i)=-(-5+8.66i)(-15+8.66i)
Za=2.5
K1=.0114286
Root Locus
8
6

Imaginary Axis (seconds-1)

4
2
0
-2
-4
-6
-8
-15

-10

-5

10

Real Axis (seconds -1)

If we aim for the desired tr and Mp, a point on the root locus graph will hit the desired frequency,
-5+8.66i. At this point, we are barely achieving the performance specifications.

1b.
G(s)
Bode Diagram

Magnitude (dB)

100

50

Phase (deg)

-50
90

45

0
-2

-1

10

10

10

10

10

Frequency (rad/s)

G(s)D(s)

Bode Diagram
25

Magnitude (dB)

20
15
10
5
0

Phase (deg)

-5
270

225

180

135
-1

10

10

10

10

Frequency (rad/s)

After tuning the root locus plot to the appropriate approximate design guides, one would use the
design guides given for the bode plot, giving a more precise tuning of G(s)D(s), improving the
system.

These design guides are zeta=PM/100 and wc >~ 1.8/tr, so PM=50 degrees, and wc>~10.
Without additional poles/zeros added to G(s)D(s) by various filters, the magnitude at the desired
crossover frequency, wc=10 is -.03199, pretty close to 0, the desired magnitude at the crossover
frequency. However, there are at least one other frequency before wc=10 that crosses over.
The phase margin, however, is 4.3 degrees, well below the desired phase margin. To achieve a
higher phase margin, you could implement a lead filter with z>p, centered at the crossover
frequency. If p/z=10, the phase margin would be bumped up approximately 55 degrees, reaching
the desired phase margin of 50 degrees. However, the closer the pole and zero are to each other,
the greater change in magnitude at the desired crossover frequency, causing a higher rise time.
1c.
H(s)=-100(s+5)(s-5)(.0114)(s+10)(s+1)(s-2.5)/[s(s+10)(s+1)(s-10)(s+5)
-100(s+5)(s-5)(.0114)(s+10)(s+1)(s-2.5)]
1d.
L1(s) is a type 1, r=1 system. It has 1 pole at the origin.
Steady State Gain of H(s)= 1 (H(s) has unity steady state gain
No prefactor is needed, since H(s) has unity steady state gain.
1e.
Step Response
8

Amplitude

-2

-4

0.5

1.5

Time (seconds)

y(t)=
1f.
If D(s) is improper, the bode plot will have an increasing magnitude as the frequency increases.
When such a controller is implemented into the system, high frequency noise will plague the
system, rendering the system unstable as time goes on. To correct this problem, a lead

compensator can be implemented along with the D(s) in a closed loop with the plant to level off
the high frequency noise, thus making the system stable once more.
2a.
G2(s)= -s^2/[(s+10)(s-10)]
D2(s)=K2 (s+za)/(s+10)
s=-.5+.866i=> from design guides
G2(s)D2(s)=-1, keep za and K2 as variables
Solve, and equate the constants of the imaginary and real components to each other
K2=.10.08 za=-.908
180 degree
Root Locus

Imaginary Axis (seconds-1)

0.5

-0.5

-1
-10

-5

10

Real Axis (seconds -1)

0 degree
Root Locus
1.5

The 180 degree root locus will get the job

Imaginary Axis (seconds-1)

0.5

-0.5

-1

-1.5
-15

-10

-5

0
Real Axis (seconds -1)

10

15

2b
G2(s)
Bode Diagram
0

Magnitude (dB)

-10
-20
-30
-40
-50
180

Phase (deg)

180
180
180
180
180
0

10

10

10

Frequency (rad/s)

G2(s)D2(s)
Bode Diagram

Magnitude (dB)

50

-50

Phase (deg)

-100
180

150

120
-2

10

-1

10

10

10

10

10

Frequency (rad/s)

Recalling the design constraints of rise time and overshoot, my design guides are now getting the
magnitude of the system at wc=1 to equal zero, and the phase margin to be greater than 50
degrees. As of the G2(s)D2(s) found in 2a, the phase margin is about 40 degrees, a bit lower than
desired. However, the magnitude of the system at wc=1 is around -3, very close to the desired 0
magnitude at wc.
To reach the phase margin design requirement, I could put a phase lead compensator centered at
wc=1, with a zero at .316 and a pole at 3.16 to bump the phase down a bit increase the PM.
z=wc*sqrt(10)
p/z=10

Bode Diagram
50

Magnitude (dB)

Blue: G2(s)D2(s)
Green: G2(s)D2(s) with added lag
compensator

-50

Phase (deg)

-100
180

135

90

45
-2

10

-1

10

10

2c.

H2(s)= .948(-s^2)(s-19.8853)/[(s+10)(s-10)(s+10)+.948(-s^2)(s-19.8853)]
2d. L2(s) is a type 0, r=0 system. It has no poles at the origin.
The steady state gain of H2(s) is 0.
lim G2(s)D2(s)=0
s=>0
2e.
Step Response
1.2

0.8

Amplitude

0.6

0.4

0.2

-0.2

-0.4

5
Time (seconds)

y(t)=

10

10

Frequency (rad/s)

10

10

3. The blue dashed line and the red solid line are the step response of nonminimum phase system.

These responses occur when there are right half plane zeros in the system. The system oscillates
between the wrong way and the right way to finally reach the right value. I only expect this
behavior in 1e and not 2e, since 1e has right half plane zeros, whereas 1e does not.
4a.

The cutoff frequencies would be 100 for F1(s) and 10 for F2(s).
4b.

comes from the loop below.

5. k=. 0114286
Nyquist Diagram
2
1.5
1

Imaginary Axis

0.5
0
-0.5
-1
-1.5
-2
-2.5
-1.15

-1.1

-1.05

-1

-0.95

-0.9

-0.85

-0.8

-0.75

Real Axis

K=10
4

Nyquist Diagram

x 10

0.8
0.6

Imaginary Axis

0.4
0.2
0
-0.2
-0.4
-0.6
-0.8
-1
-1000

-900

-800

-700

-600

-500
Real Axis

-400

-300

-200

-100

Since G1(s)D1(s) has 1 RHP, you would need 1 encirclement of -1 for stability. The system is
stable at small Ks but unstable at large Ks.
6a.
>> PolyInertia([1 10 100])
Routh row 2:

Routh row 1:

10

Routh row 0:

100

100

inertia =
2

stable CT system
ans =
2

No sign changes, a stable system.

-1250k

750k-10

-1250k
750k-10<0
k<10/750==> k<.0133333
-100k+1<0==> k>.01
.01<k<.01333

Code mfile
%1a
gds=tf(-100*[1 -10 25],[1 -10 0])
rlocus(gds)
gdso=tf(100*[1 -10 25],[1 -10 0])
rlocus(gdso)
gd1a=tf(.0114286*[-100 750 -1250],[1 -10 0])
rlocus(gd1a)
K1=.0114286
%1b
g=tf(-100*[1 0 -25],[1 1 -100 -100 0])
bode(g)
bode(gd1a)
%1c
h1=gd1a/(gd1a+1)
%1e
h1=minreal(h1)
step(h1)
%2a
g2=tf(-[1 0 0],[1 0 -100])
K2=10.08
za=.9089
d2=tf(K2*[1 10],[1 za])
d2o=tf(-K2*[1 10],[1 za])
rlocus(g2*d2)
%rlocus(g2*d2o)
%2b
bode(g2)
bode(g2*d2)
lag=tf([1 -.316],[1 -3.16])
close all
bode(g2*d2)
bode(g2*d2*lag)
%close all
%2c
h2=(g2*d2)/(1+g2*d2)
%2e
h2=minreal(h2)
step(h2)
%5
nyquist(gd1a)
gd1a=tf(10*[-100 750 -1250],[1 -10 0])
nyquist(gd1a)
%6a
h1=gd1a/(gd1a+1)

h1=minreal(h1)
PolyInertia([1 10 100])