Feedback Control Systems

(KON 313E)
Week 8

Prof.Dr. Volkan Sezer

Control and Automation Engineering Department

Root Locus Techniques
and Compansators
K Synthesis (P Controller Design) with Root Locus

By playing with the value of K in the root locus, we can change the location of the closed loop poles.

Another name of the gain K used here is the P controller.

K Synthesis (P Controller Design) with Root Locus
Reminder

For the following caharacteristic equation of 2nd order

transfer function, 𝑠 2 + 2𝜉𝑤𝑛 𝑠 + 𝑤𝑛 2 there are some
relationship between poles, 𝜉 and 𝑤𝑛 .

CL Poles: −𝜉𝑤𝑛 ± 𝑤𝑑 𝑗
𝜉=cos(𝛽)
𝑤𝑑 = 𝑤𝑛 1 − 𝜉 2

*In the higher order characteristic equations, the 2nd order dominant poles that provide the desired
behaviour are determined. In order to minimise the effect of other poles, the design can be made by
assigning the remaining poles to 5 times more to the left of the real axis.
K Synthesis (P Controller Design) with Root Locus

With the help of the amplitude condition, the gain value K can be calculated.

Here, the gain value K' of the open loop transfer function itself must also be taken into account.

𝐾′ 𝑠−𝑧 𝑠−𝑧 .. 1 ς 𝐷𝑖𝑠𝑡𝑎𝑛𝑐𝑒 𝑜𝑓𝑝𝑜𝑙𝑒𝑠 𝑡𝑜 "𝑠"

G(s)= 𝑠−𝑝 0 𝑠−𝑝 1.. 𝐾∗ = 𝐾 ′𝐾 =
0 1 1 ς 𝐷𝑖𝑠𝑡𝑎𝑛𝑐𝑒 𝑜𝑓𝑧𝑒𝑟𝑜𝑠 𝑡𝑜 "𝑠"
K Synthesis (P Controller Design) with Root Locus

By changing the value of K in root locus, we can change the location of the closed loop poles.

Another name of the gain K used here is: P controller.

K Synthesis (P Controller Design) with Root Locus
R(s) 4(𝑠 + 6) C(s)
K For a suitable value of K, the dominant poles are known
+ 𝑠(𝑠 + 2)(𝑠 + 4)
_ to pass through -0,725+ ഥ 2,3𝑗 . What is the K value that
ensures this?

𝐾 ′ = 4 (open loop system gain)

𝐴 =5,7546 𝐵 𝐶 𝐷
-0.725±2.3j 𝐵 = 4,002 𝐾∗ =
𝐴
𝐶 = 2,6298
𝐷 = 2,4116 𝐾 ∗ = 4,41

𝐾 ∗ = 𝐾 ′𝐾

4,41
𝐾= = 1,1
4
K Synthesis (P Controller Design) with Root Locus

R(s) (𝑠 + 1) C(s)
K It is known that one of the closed loop system poles
+ (𝑠 − 1)
_ is at -0.5. What is the K value that ensures this?

1
(𝑠 + 2)

(𝑠+1) 1
𝑂𝐿𝑇𝐹 = 𝐾 𝐵 𝐴
(𝑠−1) (𝑠+2) 𝐴 𝐵 1,5∗1,5
𝐾= = =4,5
-0,5 𝐶 0,5
x x
-2 -1 1
𝐶
K Synthesis (P Controller Design) with Root Locus
R(s) 1 C(s)
K
+ 𝑠(𝑠 + 5)(𝑠 + 8)
_ 𝑡𝑠 =4→ 4𝜉𝑤𝑛 = 4 → 𝜉𝑤𝑛 =1
ξ𝑤𝑛 = 1
Plot the root locus. Calculate the gain K It is seen on root locus
(P controller) which makes the settling that the required settling
time Ts=4sec. Show on the root locus. time can be satisfied.

O.L Poles: 0, -5, -8

Number of Asymptotes: 3-0=3

0−5−8 −13
𝜎𝑎 = =
3 3
(Calculate by yourself that the point of departure from the real axis is -2)
K Synthesis (P Controller Design) with Root Locus
𝐾
𝑠(𝑠 + 5)(𝑠 + 8)
𝐶𝐿𝑇𝐹 =
R(s) 1 C(s) 𝐾
1+
K 𝑠(𝑠 + 5)(𝑠 + 8)
+ 𝑠(𝑠 + 5)(𝑠 + 8)
_ 𝐾 𝑠 𝑠+5 𝑠+8 +𝐾 =0
𝑃𝑐 𝑠 = 1 + 𝑠(𝑠+5)(𝑠+8)=0
𝑠 3 +13𝑠 2 + 40𝑠 + 𝐾 = 0

Desired 2nd order behaviour Third Pole

𝑠 2 + 2𝜉𝑤𝑛 𝑠 + 𝑤𝑛 2 𝑠 + 𝑝 = 𝑠 3 + 13𝑠 2 + 40𝑠 + 𝐾

𝜉𝑤𝑛 =1

𝑠 2 + 2𝑠 + 𝑤𝑛 2 𝑠 + 𝑝 = 𝑠 3 + 13𝑠 2 + 40𝑠 + 𝐾

𝑠 3 + 𝑝 + 2 𝑠 2 + 2𝑝 +𝑤𝑛 2 𝑠+𝑝𝑤𝑛 2 = 𝑠 3 + 13𝑠 2 + 40𝑠 + 𝐾

𝑝 + 2 = 13 → 𝑝 = 11

2𝑝 +𝑤𝑛 2 =40 → 𝑤𝑛 2 = 18 → 𝑤𝑛 =3 2

𝑝𝑤𝑛 2 = 𝐾 →K=18*11=198
K=198
K Synthesis (P Controller Design) with Root Locus
R(s) 1
K
+ 𝑠(𝑠 + 5)(𝑠 + 8)
_

Or it can be obtained in a much simpler way with the

amplitude condition (calculate)

K= 𝐴 𝐵 𝐶
Compansators
Compansators

By changing only the value of K, we are limited to

behavior only on the root-locus.
By changing K, the point A which satisfies the ξ
value for a required %overshoot can be reached.
However the point B, which will satisfy both
overshoot and settling time requirements, cannot
be reached!
Therefore, situations may arise where we need to add
additional zeros and poles to the system. Structures that
need to add additional zero and (or) poles are called
compensators.
Compansators

There are two main purposes in using compensators:

• Improving (reducing/eliminating) steady-state errors
• Improving transient behavior

The main comparators used for this purpose are:

PI Controller → Improves steady state error. (This is the main purpose, it is also possible to improve
transient state behavior)
PD Controller → Improves transient behavior.
PID Controller → It can improve both steady-state error and transient behavior.
PI Controller (Ideal Integral Compensator)
By adding an open-loop pole at the origin, it increases the type of the system and
reduces the steady-state error. However, this added pole may undesirably change
the shape of the system's locus.

Let's say that the pole at point A provides the transient behavior we want.

Although adding a pole to the origin reduces the steady-state error, it prevents
us from reaching the poles that will provide the transient behavior we want.

To avoid this problem, the design can be realized by adding an open loop zero
very close to and to the left of the pole at the origin!

But this does not have to be the only method of PI controller design (You might
prefer to change the original root locus).
PI Controller (Ideal Integral Compensator)

In order to avoid this problem, the design can be realized by adding a zero very close
to and to the left of the origin pole.
PI Controller (Ideal Integral Compensator)
PI Controller

Ki

Kp
PI Controller (Ideal Integral Compensator)
In the given system, calculate the value of the gain K such that the
damping ratio of the closed-loop system is 0.174. Then, instead of
the K gain, design a PI controller that will provide the same
cos 𝛽 = 0.174
damping ratio and cancel the steady-state error for unit step
𝛽≈80
input.

In this case the steady state error for unit step input is:

𝑅(𝑠)
E(s)=1+𝐾𝐺(𝑠) 𝑒𝑠𝑠 = lim 𝑠𝐸(𝑠)
𝑠→0

1 1 1
𝑒𝑠𝑠 = lim 𝑠 = = 0.108
𝑠→0 𝑠1+ 164.4 1 + 8.23
(𝑠 + 1)(𝑠 + 2)(𝑠 + 10)
PI Controller (Ideal Integral Compensator)
In the given system, calculate the value of the gain K such that the
damping ratio of the closed-loop system is 0.174. Then, instead of
the K gain, design a PI controller that will provide the same
damping ratio and cancel the steady-state error for unit step
input.

This K value can be

calculated
geometrically (with
1 1 marking the closed
𝑒𝑠𝑠 = lim 𝑠
𝑠→0 𝑠 158.2(𝑠 + 0.1) 1 loop poles and
1+
𝑠 (𝑠 + 1)(𝑠 + 2)(𝑠 + 10) amplitude
condition) or
1
= =0 directly
∞
algebraically.
PI Controller (Ideal Integral Compensator)
Algebraic Method:
In the given system, calculate the value of the gain K such that the (𝑠 + 0.1) 1
damping ratio of the closed-loop system is 0.174. Then, instead of 𝐾 𝑠 (𝑠 + 1)(𝑠 + 2)(𝑠 + 10)
the K gain, design a PI controller that will provide the same 𝐾Ç𝑇𝐹 =
(𝑠 + 0.1) 1
damping ratio and cancel the steady-state error for unit step 1+𝐾
𝑠 (𝑠 + 1)(𝑠 + 2)(𝑠 + 10)
input.
𝐾(𝑠 + 1)
= 4
𝑠 + 13𝑠 3 + 32𝑠 2 + 𝑠 20 + 𝐾 + 0.1𝐾

Desired ξ=0,174

(𝑠 2 + 2 ∗ 0.174 ∗ 𝑤𝑛 ∗ 𝑠 + 𝑤𝑛 2 )(𝑠 + 𝑝1 )(𝑠 + 𝑝2 ) = 𝑠 4 + 13𝑠 3 + 32𝑠 2 + 𝑠 20 + 𝐾 + 0.1𝐾

Wn, p1, p2 and K values can be calculated by solving 4 equations with 4 unknowns !
PI Controller (Ideal Integral Compensator)
In the given system, calculate the value of the gain K such that the
damping ratio of the closed-loop system is 0.174. Then, instead of
the K gain, design a PI controller that will provide the same
damping ratio and cancel the steady-state error for unit step
input.
PI Controller (Ideal Integral Compensator)
• This is not the only design method of the PI controller.
• The ‘zero’ to be added can also be placed far from the ‘origin’
according to the closed loop requirements.
• Where to place the zero and the gain value can be calculated by
amplitude and angle condition; or it can be calculated by polynomial
equalization (solution examples with similar approaches are available
in the PD controller)
• The remaining closed-loop pole(s) that break the dominant pole
approximation must also be considered.
PD Controller (Ideal Derivative Compensator)
• It is used to improve the transient behavior of the system.

• It is designed by adding 'zero' to a suitable place on the real

axis in the open loop transfer function, and then calculating
the K gain.
Derivative(D)

𝐾𝐷 𝑠

𝐾𝑝
Kp
PD Controller (Ideal Derivative Compensator)
In the system below, which has an overshoot for the The dominant poles where the ξ=0.504 line intersects with the ground
unit step input by 16%, design an ideal derivative axis can be found geometrically by drawing a good root locus
compensator that provides the same amount of (1.205±2.064j).
overshoot and reduces the settling time to 1/3 times. cos 𝛽 = 0.504
𝛽≈59.74

−ln(0.16)
𝜁= =0.504
𝜋2 +𝑙𝑛2 (0.16)

The gain value for these poles can be obtained with the amplitude condition.

It is also possible to obtain it algebraically (polynomial equating) without using root locus.
PD Controller (Ideal Derivative Compensator)
In the system below, which has an overshoot for the
Algebraic method
𝐾
unit step input by 16%, design an ideal derivative 𝑠(𝑠 + 4)(𝑠 + 6)
compensator that provides the same amount of −ln(0.16) 𝐶𝐿𝑇𝐹 =
𝜁= =0.504 𝐾
overshoot and reduces the settling time to 1/3 times. 𝜋2 +𝑙𝑛2 (0.16) 1+
𝑠(𝑠 + 4)(𝑠 + 6)
𝐾 𝐾
= = 3
𝑠(𝑠 + 4)(𝑠 + 6) + 𝐾 𝑠 + 10𝑠 2 + 24𝑠 + 𝐾

Desired Reality

(𝑠 + 𝑝) 𝑠 2 + 2𝜉𝑤𝑛 𝑠 + 𝑤𝑛2 = 𝑠 3 + 10𝑠 2 + 24𝑠 + 𝐾

Dominant Poles: −𝜉𝑤𝑛 ± 𝑤𝑑 𝑗 0.504
−0.504 ∗ 2.391 ± 2.391 1 − 0.5042 j 𝑠 3 + 𝑠 2 𝑝 + 1.08𝑤𝑛 + 𝑠 𝑤𝑛2 + 1.08𝑤𝑛 𝑝 + 𝑤𝑛2 𝑝 = 𝑠 3 + 10𝑠 2 + 24𝑠 + 𝐾
−1.205 ± 2.064j
Settling time without 3 Unknowns, 3 Equations
compensation:
3.32 p + 1.008𝑤𝑛 = 10 𝑤𝑛 = 2.391
Desired settling time: 𝑇𝑆𝑁 = = 1.107
3 𝑤𝑛2 + 1.008𝑤𝑛 𝑝 = 24 K=43.35
4 𝑤𝑛2 𝑝 = 𝐾 P=7.59 (Place of 3rd Pole)
𝑇𝑆𝑁 = = 1.107 𝜉𝑤𝑛 = 3.613
𝜉𝑤𝑛
Desired with Compensation
PD Controller (Ideal Derivative Compensator)
What is required:
In the system below, which has an overshoot for the
unit step input by 16%, design an ideal derivative
𝜉=0.504 (The same overshoot) 𝜉𝑤𝑛 = 3.613 (1.107sec settling time)
compensator that provides the same amount of 4
overshoot and reduces the settling time to 1/3 times. 𝑤𝑛 = = 7.169
0.504 ∗ 1.107
In this case, the locations of the new dominant poles are:
𝑠1,2 = −𝜉𝑤𝑛 ± 𝑤𝑛 1 − 𝜉 2 j=-3.613 ±6.19j
Once the desired dominant poles are determined, the design can be
accomplished by angle&amplitude condition or algebraically!
Where to place the open loop zero? (Angle Condition)

𝜃1 = 68.9 𝜃2 = 86.4 𝜃3 = 120.3

−𝜃1 −𝜃2 −𝜃3 + 𝜃𝑥 = (2𝑘 + 1)180

For k=-1 ,
𝜃𝑥 = 68.9 + 86.4 + 120.3 − 180
𝜃𝑥 =95.7
6.19
𝑡𝑎𝑛 180 − 95.7 = a≈3
3.613 − 𝑎 a
PD Controller (Ideal Derivative Compensator)
In the system below, which has an overshoot for the
unit step input by 16%, design an ideal derivative
compensator that provides the same amount of
overshoot and reduces the settling time to 1/3 times.

The gain value at this point can also be calculated from the amplitude condition.
𝐴 = 6.63
𝐴 𝐵 𝐷
B=6.20 𝐾= =47.4
𝐶
PD Controller C=6.22
D=7.17

PD Controller: K(s+a) = 47.4(s+3)

PD Controller (Ideal Derivative Compensator)
In the system below, which has an overshoot for the
unit step input by 16%, design an ideal derivative
compensator that provides the same amount of
overshoot and reduces the settling time to 1/3 times.

PD Controller

3. Pole is at -2.775. Its effect is not felt much because

of the zero right next to it!
PD Controller (Ideal Derivative Compensator)
Algebraic Method:
In the system below, which has an overshoot for the
unit step input by 16%, design an ideal derivative Requirements:
compensator that provides the same amount of 𝜉=0.504
overshoot and reduces the settling time to 1/3 times.
𝑤𝑛 = 7.169

𝐾(𝑠 + 𝑎)
𝑠(𝑠 + 4)(𝑠 + 6) 𝐾(𝑠 + 𝑎) 𝐾(𝑠 + 𝑎)
𝐶𝐿𝑇𝐹 = = = 3
𝐾(𝑠 + 𝑎) 𝑠(𝑠 + 4)(𝑠 + 6) + 𝐾(𝑠 + 𝑎) 𝑠 + 10𝑠 2 + (24 + K)𝑠 + 𝐾𝑎
1+
𝑠(𝑠 + 4)(𝑠 + 6)
3 Unknowns, 3 Equations
(𝑠 + 𝑝) 𝑠 2 + 2𝜉𝑤𝑛 𝑠 + 𝑤𝑛2 = 𝑠 3 + 10𝑠 2 + (24 + K)𝑠 + 𝐾𝑎
7.22+p=10 P=2.78 (𝑠 + 𝑝) 𝑠 2 + 7.22𝑠 + 51.26 = 𝑠 3 + 10𝑠 2 + (24 + K)𝑠 + 𝐾𝑎
51.26+7.22p=24+K K=47.4
𝑠 3 + 7.22 + 𝑝 𝑠 2 + 51.26 + 7.22p s + 51.26p = 𝑠 3 + 10𝑠 2 + (24 + K)𝑠 + 𝐾𝑎
51.26p=Ka a=3

PD Controller: K(s+a) = 47.4(s+3)

PD Controller (Ideal Derivative Compensator)
In the system below, which has an overshoot for the
unit step input by 16%, design an ideal derivative
compensator that provides the same amount of
overshoot and reduces the settling time to 1/3 times.

PD Controller
PID Controller
• It is used to improve both steady-state error and transient behavior
simultaneously.
• It is one of the most popular controller types in real applications.
PID Controller
• It is used to improve both steady-state error and transient behavior
simultaneously.
• It is one of the most populer controller types in real applications.
PID Controller
Different design methods are available. One of the most used ones has 2 stages,
as expressed below.

1) A PI controller (without gain) is created that improves the steady-state error,

adding 1 pole to the origin and 1 zero near it: (s+b)/s

2) The position of the other zero is found. Total gain of controller is calculated with
the amplitude condition:K(s+a)

Apart from this, solution is also possible by algebraic method by equalizing polynomials!
Note: Instead of placing a zero right near the pole at the origin, the location of that zero can also be obtained by designing.
In this case, the controller's degrees of freedom will increase. Even as the design becomes more complex, more
requirements can be achieved. (For example: in addition to time response requirements, steady state error requirements
can also be provided (Ex: A specific error value for ramp input)
PID Controller
The system shown below, whose root locus also given, makes 20% overshoot. Design a PID controller that
makes the steady-state error zero for unit step input, continues the 20% overshoot as in the current system,
and reduces the peak time to 2/3 times. Overshoot: 0.2
ln 0,2
𝜉= = 0.456
𝜋2 + 𝑙𝑛2 0,2

−3 − 6 − 10 − −8
𝜎𝑎 = = −5.5
3−1
(2k + 1)180 𝜋
Actual System: 𝜃𝑎 = 𝜃𝑎1 =
3−1 2
𝜉=0.456 𝜉𝜔𝑛 ∓ 𝜔𝑛 1 − 𝜉 2 = 5,445 ∓ 10,57𝑗 3𝜋
𝜃𝑎2 =
2
𝜉𝜔𝑛 =5.445 Departure(breaking) point can
be calculated
𝜔𝑛 =11.94 𝜎 = −4.6
𝜋 3,14 Note: Even if the root locus is not given, the K value for 20% overshoot and the
𝑇𝜌 = ⇒ 𝑇𝑝 = = 0.296 corresponding closed loop poles can be calculated with an algebraic approach
𝑤𝑛 1 − 𝜉2 11.94 1 − 0, 4562 (polynomial equalization) as we did in the previous examples.
PID Controller
The system shown below, whose root locus also given, makes 20% overshoot. Design a PID controller that
makes the steady-state error zero for unit step input, continues the 20% overshoot as in the current system,
and reduces the peak time to 2/3 times.
2
What is required: 𝑇𝑝𝑁 = 0.296 ⋅ = 0,197
3
(same overshoot) 𝜋
𝑇pN = ⇒ 𝜔𝑛 = 17.91
𝜉=0.456 𝜔𝑛 1 − 𝜉2

So, we can calculate the required closed loop poles:

Actual System:
−𝜉𝜔𝑛 ∓ 𝑤𝑛 1 − 𝜉 2 𝑗 = −8.17 ∓ 15.94𝑗
𝜉=0.456 𝜉𝜔𝑛 ∓ 𝜔𝑛 1 − 𝜉 2 = 5,445 ∓ 10,57𝑗

𝜉𝜔𝑛 =5.445
From this point on, we can achieve the solution in 2 ways:
𝜔𝑛 =11.94 • Classical method using angle&amplitude condition
𝜋 3,14 • Algebraic approach
𝑇𝜌 = ⇒ 𝑇𝑝 = = 0.296
𝑤𝑛 1 − 𝜉2 11.94 1 − 0, 4562
PID Controller
The system shown below, whose root locus also given, makes 20% overshoot. Design a PID controller that
makes the steady-state error zero for unit step input, continues the 20% overshoot as in the current system,
and reduces the peak time to 2/3 times.
Classical method using angle&amplitude condition:
Let's place the zero of the PI controller at -0.5 (There must be a pole
at the origin to remove the steady-state error )
(s+0.5)/s
the required closed loop poles

−𝜉𝜔𝑛 ∓ 𝑤𝑛 1 − 𝜉 2 𝑗 = −8.17 ∓ 15.94𝑗

x x x -0.5
x
-10 -8 -6 -3
PID Controller
The system shown below, whose root locus also given, makes 20% overshoot. Design a PID controller that
makes the steady-state error zero for unit step input, continues the 20% overshoot as in the current system,
and reduces the peak time to 2/3 times.
Classical method using angle&amplitude condition:
To obtain the dominant poles determined for the transient behavior,
we can locate the zero of the PD controller with the angle condition:
−𝜃1 +𝜃2 −𝜃3 −𝜃4 +𝜃5 −𝜃6 +𝜃𝑥 = 2𝑘 + 1 180
15.94

the required closed loop poles

−𝜉𝜔𝑛 ∓ 𝑤𝑛 1 − 𝜉 2 𝑗 = −8.17 ∓ 15.94𝑗

𝜃2 𝜃3 𝜃4 𝜃5 𝜃6
𝜃1
x -8.17 x x -0.5
x
-10 -8 -6 -3

𝑓𝑜𝑟 𝑘 = −1 , 𝜃𝑥 ≈ 20°
contribution of zero
PID Controller
The system shown below, whose root locus also given, makes 20% overshoot. Design a PID controller that
makes the steady-state error zero for unit step input, continues the 20% overshoot as in the current system,
and reduces the peak time to 2/3 times.
Classical method using angle&amplitude condition:
To obtain the dominant poles determined for the transient behavior,
we can locate the zero of the PD controller with the angle condition:
15.94

the required closed loop poles

−𝜉𝜔𝑛 ∓ 𝑤𝑛 1 − 𝜉 2 𝑗 = −8.17 ∓ 15.94𝑗

𝜃𝑥 ≈ 20°
𝜃2 𝜃3 𝜃4 𝜃5 𝜃6
𝜃𝑥 = 20 𝜃1
contribution of zero x
-10
-8.17
-8
x-6 x-3 -0.5
x

a
PID controller without gain:
PID Controller
The system shown below, whose root locus also given, makes 20% overshoot. Design a PID controller that
makes the steady-state error zero for unit step input, continues the 20% overshoot as in the current system,
and reduces the peak time to 2/3 times.
Classical method using angle&amplitude condition:
We can use the amplitude condition to calculate the K gain.

15.94

the required closed loop poles

−𝜉𝜔𝑛 ∓ 𝑤𝑛 1 − 𝜉 2 𝑗 = −8.17 ∓ 15.94𝑗

A B G
PI controller without gain: C D F
E
𝜃2 𝜃3 𝜃4 𝜃5 𝜃6
𝜃𝑥 = 20 𝜃1
x
-10
-8.17
-8
x-6 x-3 -0.5
x

a
PID Controller
The system shown below, whose root locus also given, makes 20% overshoot. Design a PID controller that
makes the steady-state error zero for unit step input, continues the 20% overshoot as in the current system,
and reduces the peak time to 2/3 times.

5,91(𝑠 + 0,5)(𝑠 + 51,77)

𝑃𝐼𝐷 𝐶𝑜𝑛𝑡𝑟𝑜𝑙𝑙𝑒𝑟:
𝑠
PID Controller
The system shown below, whose root locus also given, makes 20% overshoot. Design a PID controller that
makes the steady-state error zero for unit step input, continues the 20% overshoot as in the current system,
and reduces the peak time to 2/3 times.
Algebraic Method

Let's choose the zero of the PI controller at -0.5 again!

𝐾(𝑠 + 0.5)(𝑠 + 𝑎) (𝑠 + 8)
𝑠 (𝑠 + 3)(𝑠 + 6)(𝑠 + 10) 𝐾(𝑠 + 0.5)(𝑠 + 𝑎)(𝑠 + 8)
𝐶𝐿𝑇𝐹 = 𝐶𝐿𝑇𝐹 =
𝐾(𝑠 + 0.5)(𝑠 + 𝑎) (𝑠 + 8) 𝑠 𝑠 + 3 𝑠 + 6 𝑠 + 10 + 𝐾(𝑠 + 0.5)(𝑠 + 𝑎)(𝑠 + 8)
1+ 𝑠 (𝑠 + 3)(𝑠 + 6)(𝑠 + 10)

Pc: 4th degree

Required: 𝜉 = 0.456 𝑤𝑛 = 17.91
PID Controller
The system shown below, whose root locus also given, makes 20% overshoot. Design a PID controller that
makes the steady-state error zero for unit step input, continues the 20% overshoot as in the current system,
and reduces the peak time to 2/3 times.
Algebraic Method

Let's choose the zero of the PI controller at -0.5 again!

order equalizers

(𝑠 2 + 2 ∗ 0.456 ∗ 17.91𝑠 + 17.912 )(𝑠 + 𝑝1 )(𝑠 + 𝑝2 ) = 𝑠 𝑠 + 3 𝑠 + 6 𝑠 + 10 + 𝐾(𝑠 + 0.5)(𝑠 + 𝑎)(𝑠 + 8)

desired reality

• By equalizing the polynomial coefficients, 4 equations with 4 unknowns are obtained.

• From these equations, p1, p2, a,K values can be calculated.
• Once K and a are obtained, the controller design is completed.
• With the help of p1 and p2, positions of 3rd and 4th poles can be examined