Design via Root Locus

CEN455: Dr. Nassim Ammour 1

 Root Locus Techniques
• Root locus is a graphical presentation of the closed-loop poles as a system parameter k is varied.

• The graph of all possible roots of this equation (K is the variable parameter) is called the root locus.

• The root locus gives information about the stability and transient response of feedback control systems.

Rout-locus (poles motion graph)

Zeros of T(s) are zeros of G(s) and poles of H(s).

Poles of T(s) depends on gain K

Closed-loop system CLCF is a function of K.

Equivalent Transfer Function
Root Locus graphically shows poles of T(s) as K varies
CEN455: Dr. Nassim Ammour 2
 Evaluation of a Complex Function via Vectors
Any complex number, 𝜎 + 𝑗𝜔, described in Cartesian coordinates can be graphically represented by a vector,

Magnitude and phase Of F(s) at S

ς 𝑧𝑒𝑟𝑜 𝑙𝑒𝑛𝑔𝑡ℎ𝑠 ς𝑚
𝑖=1 (𝑠 + 𝑧𝑖 )
𝑀= = 𝑛
ς 𝑝𝑜𝑙𝑒 𝑙𝑒𝑛𝑔𝑡ℎ𝑠 ς𝑗=1 (𝑠 + 𝑝𝑗 )
ς𝑚
𝑖=1(𝑠+𝑧𝑖 )
If 𝐹 𝑠 = ς𝑛
𝑚 𝑛
𝑗=1(𝑠+𝑝𝑗 )
𝜃 = ෍ 𝑧𝑒𝑟𝑜 𝑎𝑛𝑔𝑙𝑒𝑠 − ෍ 𝑝𝑜𝑙𝑒 𝑎𝑛𝑔𝑙𝑒𝑠 = ෍ ≺ 𝑠 + 𝑧𝑖 − ෍ ≺ 𝑠 + 𝑝𝑗
𝑖=1 𝑗=1

Problem: Given Find F(s) at the point 𝑠 = −3 + 𝑗4

Solution: Any complex number can be represented by a vector 𝑠 = −3 + 𝑗4

For zero ( point 𝑠1 = −1) the vector is:

4
𝑠 − 𝑠1 = 𝑠 − (−1) = −3 + 𝑗4 − −1 = −2 + 𝑗4 = −2 2 + 4 2 𝑡𝑎𝑛−1 =
−2

For pole at 0: For pole at -2: Magnitude and phase

(polar form)

Vector magnitude,

Vector angle,

Vector CEN455: Dr. Nassim Ammour 3

 Defining the Root Locus
Gain less than 25, over-damped.
Gain = 25, critically damped.

Gain over 25, under-damped.

Stable system, as no pole on right-hand plane.
Poles location for
different values of K

During underdamped, real parts are same; so settling

time (which is related to real part) remains the same.

Damping frequency (imaginary part) increases with gain,

resulting in reduction of peak time.

CEN455: Dr. Nassim Ammour 4

 Properties of Root Locus
𝐾𝐺(𝑠) Module condition Angle condition
The closed-loop transfer function 𝑇 𝑠 = 1+𝐾𝐺 𝑠 𝐻(𝑠)

𝑠0 is a pole if 1 + 𝐾𝐺 𝑠0 𝐻 𝑠0 = 0 ⟹ 𝐾𝐺 𝑠0 𝐻 𝑠0 = −1 = 1 ≺ 2𝑘 + 1 1800 𝑘 = 0, ±1, ±2, …

1 Π 𝑝𝑜𝑙𝑒 𝑙𝑒𝑛𝑔𝑡ℎ𝑠
𝐾𝐺 𝑠 𝐻 𝑠 = −1 ⟹ 𝐾𝐺 𝑠 𝐻 𝑠 =1⟹𝐾= =
𝐺 𝑠 𝐻 𝑠 Π 𝑧𝑒𝑟𝑜 𝑙𝑒𝑛𝑔𝑡ℎ𝑠

Find if the point -2+j3 is on root locus for some value of gain, K:
From the angle condition
For the point −2 + j( 2ൗ2) which is on root locus, the gain K is:
Σ zero angle - Σ pole angle
2
ς 𝑝𝑜𝑙𝑒 𝑙𝑒𝑛𝑔𝑡ℎ𝑠 𝐿3 𝐿4 (1.22)
𝐾= = = 2 = 0.33
Not a multiple of 1800 . So, −2 + j3 is not in the root locus ς 𝑧𝑒𝑟𝑜 𝑙𝑒𝑛𝑔𝑡ℎ𝑠 𝐿1 𝐿2 (2.12)(1.22)
(can not be a pole for some value of K).
CEN455: Dr. Nassim Ammour 5
 Sketching the Root Locus1
1.Number of branches: Equals the number of closed loop poles.
2.Symmetry: Symmetrical about the real axis (conjugate pairs of poles, real coefficients
of the characteristic equation polynomial).
3.Real axis segments: For K > 0, root locus exists to the left of an odd number real Real-axis segments of the root locus
axis poles and/or zeros (angle condition).
4.Start and end points: The root locus begins at finite and infinite poles of 𝐺 𝑠 𝐻 𝑠 and ends
at finite and infinite zeros of 𝐺 𝑠 𝐻 𝑠 .
5.Asymptotes: The root locus approaches straight lines as asymptotes as the locus
approaches infinity. the equation of the asymptotes is given by:
σ 𝑓𝑖𝑛𝑖𝑡𝑒 𝑝𝑜𝑙𝑒𝑠 − σ 𝑓𝑖𝑛𝑖𝑡𝑒 𝑧𝑒𝑟𝑜𝑠 Intersection with
𝜎𝑎 = Real axis
≠ 𝑓𝑖𝑛𝑖𝑡𝑒 𝑝𝑜𝑙𝑒𝑠 −≠ 𝑓𝑖𝑛𝑖𝑡𝑒 𝑧𝑒𝑟𝑜𝑠 Angle in radian of the Complete root locus for the system
(2𝑘 + 1)𝜋 asymptote with real axis
𝜃𝑎 = 𝑓𝑜𝑟 𝑘 = 0, ±1, ±2, …
≠ 𝑓𝑖𝑛𝑖𝑡𝑒 𝑝𝑜𝑙𝑒𝑠 −≠ 𝑓𝑖𝑛𝑖𝑡𝑒 𝑧𝑒𝑟𝑜𝑠
6.Real-Axis Breakaway and Break-in Points:
σ1: Breakaway point (leave the real axis);
σ2: Break-in point (return to the real axis ); .
Breakaway point: at maximum gain on the real axis between -2 and -1.

Break-in point: at minimum gain on real axis (increases

when moving towards a zero) between +3 and +5.
CEN455: Dr. Nassim Ammour 6
 Sketching the Root Locus2
Problem1: Sketch the Root Locus for the system shown in the following figure.
Solution1: Calculate asymptotes to find real axis intercept:

𝑎𝑛𝑑 Asymptote

• The angles of the lines that intersect at −4Τ3 is given by 𝜃𝑎 :

• For higher values of k, the angles would begin to repeat.
• There are four poles and one finite zero. Root locus begins at poles
and ends at zeros.(Three zeros at infinity are at the ends of the
asymptotes.)
Problem2: From root-locus graph on figure find break-in and break-away points
Method 1:(transition method) From figure, we get
𝑧𝑖 and 𝑝𝑖 are negative of zero and
pole values, respectively, of G(s)H(s).
Method 2: (Differentiation method)
𝐾(𝑠 − 3)(𝑠 − 5) 𝐾(𝑠 2 − 8𝑠 + 15) Differentiating K with respect to 𝜎 (max and min)
𝐾𝐺 𝑠 𝐻 𝑠 = =
(𝑠 + 1)(𝑠 + 2) (𝑠 2 + 3𝑠 + 2) 𝑑𝐾 (11𝜎 2 − 26𝜎 − 61)
= =0
𝑑𝜎 (𝜎 2 − 8𝜎 + 15)2
Along the real axis ( 𝑠 = 𝜎) 𝑎𝑛𝑑 𝐾𝐺 𝑠 𝐻 𝑠 = −1
Solving for 𝜎,
Solving for K
2
𝐾(𝑠 − 8𝑠 + 15) −(𝜎 2 + 3𝜎 + 2)
= −1 𝐾= 𝜎 = −1.45 𝑎𝑛𝑑 𝜎 = 3.82 -1.45 3.82
(𝑠 2 + 3𝑠 + 2) (𝜎 2 − 8𝜎 + 15)
7
CEN455: Dr. Nassim Ammour
7.Imaginary-Axis Crossing
Sketching the Root Locus3
Stability: the system's poles are in the left half-plane up to a particular value of gain K.
PROBLEM: For the system, find the frequency and gain, K, for which the root locus
crosses the imaginary axis. For what range of K is the system stable?
SOLUTION: The closed-loop transfer function 8.Angles of Departure and Arrival
Departure: from complex poles. Arrival: to complex zeros.
Characteristic Eq.:
= =0
We get Routh
table as follows: Poles:-3, -1+j, -1- j
Angle of departure: Zero: -2

we calculate the sum of angles drawn to a point

𝜀 close to the complex pole, -1 + j

A complete row of zeros yields the possibility for imaginary-axis roots.

For K > 0, only 𝑠1 row can be zero.

Forming the even polynomial by using the 𝑠 2 row (above) with K= 9.65,
𝜃3 𝜃4
Gives
Thus, the root locus crosses the imaginary-axis at 𝝎𝒅 = ±𝒋𝟏. 𝟓𝟗 at
a gain of K= 9.65 So, the system is stable for 0 ≤ K < 9.65
CEN455: Dr. Nassim Ammour 8
 Improving System Response Desired transient response

Speed up the response : move pole from A to B without affecting the percent overshoot
Solution: move the root locus to put the desired pole on it for some value of gain k Obtained transient
(compensation by adding poles and zeros). response
• Dynamic compensator is used if a satisfactory design cannot be
obtained by adjustment of gain k alone.
Compensators
Dynamic compensators (function of s) are designed to improve:
Transient response by adding an ideal compensator PD (pure differentiation
using active amplifiers) or a Lead compensator (implemented with passive a. Sample root locus, showing possible design point via gain adjustment (A)
and desired design point that cannot be met via simple gain adjustment (B);
elements) in the forward path or feedback path.

Steady-state error by adding an ideal compensator PI (pure integration

using active amplifiers) or a Lag compensator (implemented with passive
elements) in the forward path or feedback path.
𝑠+𝑧
• Compensator transfer function : 𝐶 𝑠 = 𝐾
𝑠+𝑝
b. Responses from poles at A and B
lead compensation if 𝑧 < 𝑝 and lag compensation if 𝑧 > 𝑝 .

CEN455: Dr. Nassim Ammour 9

 Ideal Integral Compensation (PI)
Improving Steady-State Error
• Steady-state error can be improved (without appreciably affecting the transient response) by placing an open-loop pole at the
origin, because this increases the system type by one.
If we add a Changes root-locus Solution: add a zero close to the pole at the origin
System operating with closed-loop poles at A to pole cancel out the effect of the added pole on
pole at the (Point A not on
(desirabled transient response ) the root-locus.
origin root locus.)

Pole at A is: we have improved the steady-state error without appreciably

a. on the root locus without compensator;
b. not on the root locus with compensator pole added;
affecting the transient response
c. approximately on the root locus with compensator pole and zero added CEN455: Dr. Nassim Ammour 10
 Example1
Given the system of Figure (a), operating with a damping ratio of 0.174, show that the
addition of the ideal integral compensator shown in Figure (b) reduces the steady-state
error to zero for a step input without appreciably affecting transient response.
SOLUTION Figure

For gain 𝐾 = 164.6, searching along the line of 𝜁 = 0.174 for the uncompensated
system : dominant poles are 0.694 ∓ 𝑗3.926 (third pole at −11.61) Figure (c).
164.6
This gain yields Position constant 𝐾𝑝 = lim 𝐺(𝑠) = = 8.23.
𝑠→0 20

Hence, the steady-state error is:

Closed-loop system Figure
We add an ideal integral compensator with a zero at −0.1. a. before compensation;
For gain 𝐾 = 158.2, searching along the line of 𝜁 = 0.174 for the compensated b. after ideal integral
compensation
system : dominant poles are 0.678 ∓ 𝑗3.837 (forth pole at −0.0902) Figure (e).
Poles and gain are approximately the same Same transient response
Damping Ratio unchanged (with K = 158:2). Steady State Error is ZERO!.

Desired response Figure (c)

Figure (e)
Figure (d)

CEN455: Dr. Nassim Ammour 11

 Lag Compensation
Improving Steady-State Error
• Similar to the Ideal Integrator, however it has a pole not on the origin but close to the origin (fig c) due to the passive networks.

• Steady State Improvement:

Before compensation: The static error constant,
KVo, for the system is:

After compensation:

• The effect on the transient response is negligible:

If the lag compensator pole and zero are close together, the angular contribution
of the compensator to point P is approximately zero degrees. point P is still at
approximately the same location on the compensated root locus.

Root locus: a. before lag compensation; b. after lag compensation CEN455: Dr. Nassim Ammour 12
 Example2
Compensate the system of Figure (a), whose root locus is shown in Figure (b), to improve the
steady-state error by a factor of 10 if the system is operating with a damping ratio of 0.174.
SOLUTION
• From example 1: uncompensated system error was 0.108 with 𝐾𝑃𝑜𝑙𝑑 = 8.230. A tenfold improvement means a
steady-state error of:
𝑛𝑒𝑤
𝑒 𝑜𝑙𝑑 ∞ 0.108 1 𝑛𝑒𝑤
𝑒 ∞ = = = 0.0108, 𝑠𝑖𝑛𝑐𝑒 𝑒 ∞ = 𝑛𝑒𝑤 ⇒ 𝐾𝑃 = 91.59
10 10 1 + 𝐾𝑃

𝑧𝑐 𝐾𝑃𝑛𝑒𝑤 91.59
• For the compensated system = = = 11.13
𝑝𝑐 𝐾𝑃𝑜𝑙𝑑 8.23

Arbitrarily selecting 𝑝𝑐 = 0.01 𝑧𝑐 = 11.13 ∙ 𝑝𝑐 = 11.13 0.01

𝑧𝑐 ≈ 0.111

• The compensated system

Root locus for uncompensated system

13
CEN455: Dr. Nassim Ammour
 Example2-Conted
• The transient response of both systems is approximately the same with reduced steady state error

Less Steady State Error 0.0108

the transient responses of the uncompensated

and lag-compensated systems are the same

Root locus for compensated system

• Comparison of the Lag-Compensated and the Uncompensated Systems

On the ζ= 0.174 line: (compensated system):
The second-order dominant poles are at
- 0.678 ±j3.836 (K=158.1)
The third and fourth closed-loop poles
are at -11.55 and - 0.101.
The fourth pole of the compensated
system cancels its zero.

14
CEN455: Dr. Nassim Ammour
 Ideal Derivative Compensation (PD)
Improving Transient Response

• The objective is to design a response that has a desirable percent overshoot and a shorter settling time than the
uncompensated system. (two approaches).
1. Ideal derivative compensation (Proportional-plus-Derivative (PD) active elements ): a pure differentiator is added
to the forward path of the feedback control system. sensitive to high frequency noise.

2. Lead Compensation: (not pure differentiation) approximates differentiation with a passive network by adding
a zero and a more distant pole to the forward-path transfer function.

Distant Less sensitive to high frequency noise.

pole

• The transient response of a system can be selected by choosing an appropriate closed-loop pole location on the s-plane.
• If this point is on the root locus, then a simple gain adjustment is all that is required in order to meet the transient response
specification.
• If the closed loop root locus doesn’t go through the desired point, it needs to be reshaped (add poles and zeros in the forward
path).
• One way to speed up the original system is to add a single zero to the forward path. 𝐺𝑐 𝑠 = 𝑠 + 𝑧𝑐

CEN455: Dr. Nassim Ammour 15

 Ideal Derivative Compensation (PD)
Improving Transient Response
• See how it affects by an example of a system operating with a damping ratio of 0.4:
(a) uncompensated (b) compensated, zero at -2 (c) zero at -3 (d) zero at -4

Observations and facts:

- In each case gain K is chosen such that percent overshoot is same.

- Compensated poles have more negative real part (smaller settling time)
and larger imaginary part (smaller peak time).

- Zero placed farther from the dominant poles, compensated dominant

poles move closer to the origin.
CEN455: Dr. Nassim Ammour 16
 Example31
Given the system of Figure (a), design an ideal derivative compensator
to yield a 16% overshoot, with a threefold reduction in settling time. Fig (a) Feedback
control system
SOLUTION
The performance of the uncompensated system operating with 16% overshoot fig (b).
along damping ratio line
16% Overshoot ζ = 0.504, -1,205 ±j2.064.
Dominant second-order poles
(𝑤𝑖𝑡ℎ 𝑘 = 43.35 and third pole at -7.59.)
Settling time 4 4
𝑇𝑠 = = = 3.320
ζ𝜔𝑛 1.205
Location of the compensated system's dominant poles.(Desired poles)
𝑇𝑠𝑜𝑙𝑑
𝑇𝑠𝑛𝑒𝑤 = = 1.107 threefold reduction in the settling time
3
4 4 real part of the compensated system's
𝜎= = = 3.613 dominant, second-order pole
𝑇𝑠𝑛𝑒𝑤 1.107
Fig (b) Compensated
dominant pole
Imaginary part of the compensated
𝜔𝑑 = 3.613 𝑡𝑎𝑛 180𝑜 − 120.26 0 = 6.193 system's dominant pole on line ζ = 0.504

17
CEN455: Dr. Nassim Ammour
 Example32 Zero contribution angle >
900 → zero position less than
Design the location of the compensator zero desired pole real part.

- The angle contribution of poles for the desired pole location:−275.60 .

- To achieve −1800 the angle contribution of the placed zero should be: −275.60 + 𝑥 = −1800 → 𝑥 = 95.60
- From the fig (c):
6.193
= tan(180 − 95.6) 𝜎 = 3.006
3.613 − 𝜎

Adding zero

Fig (c) Evaluating the location

of the compensating zero
Fig (c) Root locus for the compensated system

Fig (d) Uncompensated and compensated system step responses

18
CEN455: Dr. Nassim Ammour
 Lead Compensation
Basic Idea: The difference between 180° and the sum of the angles must be the
angular contribution required of the compensator.
Example: looking at the Figure, we see that:
𝜃2 − 𝜃1 − 𝜃3 − 𝜃4 + 𝜃5 = 2𝑘 + 1 1800
𝑤ℎ𝑒𝑟𝑒 𝜃1 − 𝜃2 = 𝜃𝑐 𝑖𝑠 𝑡ℎ𝑒 𝑎𝑛𝑔𝑢𝑙𝑎𝑟 𝑐𝑜𝑛𝑡𝑟𝑖𝑏𝑢𝑡𝑖𝑜𝑛 𝑜𝑓 𝑡ℎ𝑒 𝑐𝑜𝑚𝑝𝑒𝑛𝑠𝑎𝑡𝑜𝑟
• There are infinitely many choices of zc and pc providing same 𝜃𝑐
Example41
Design three lead compensators for the system in Figure to reduce the settling
time by a factor of 2 while maintaining 30% overshoot.
SOLUTION
• Characteristics of the uncompensated system operating at 30% overshoot
damping Dominant second-order
30% ratio pair of poles settling time
Overshoot ζ = 0.358, -1,007 ±j2.627. 𝑇𝑠 = 4ൗ1.007 = 3.972 s
along damping ratio line From pole's real part

• Design point (Desired Poles location)

real part of the desired
twofold reduction
pole location
in settling time
𝑇𝑠 = 3.972ൗ2 = 1.986 s −ζ𝜔𝑛 = − 4ൗ𝑇 = −2.014
𝑠
Imaginary part of the
desired pole location
𝜔𝑑 = −2.014 tan 110.980 = 5.252 19
CEN455: Dr. Nassim Ammour
 Example42
• Lead compensator Design.
Place the zero on real axis at -5 (arbitrarily as possible solution).
sum the angles (this zero and uncompensated system's poles and zeros),
the angular contribution required
resulting angle
from the compensator pole
𝜃0 = −172.690 𝜃𝑐 = −1800 + 172.690 = −7.310
location of the
compensator pole compensator pole
5.252
= tan(7.310 ) 𝑝𝑐 = 42.96
From geometry 𝑝𝑐 − 2.014
Fig (a) 5-plane picture used to calculate the location
in fig(a)
of the compensator pole

approximation is not
valid for case C

Fig (c) Uncompensated system and lead compensation responses (zeros at a:-5, b:-4 c: -2)
Fig (b) Compensated system root locus 20
CEN455: Dr. Nassim Ammour
 Improving Steady-State Error
and Transient Response

• Combine the design techniques to obtain improvement in steady-state error and transient response independently.

- First improve the transient response.(PD or lead compensation).

- Then improve the steady-state response. (PI or lag compensation).

• Two Alternatives

- PID (Proportional-plus-Integral-plus-Derivative) (with Active Elements).

- Lag-Lead Compensator. (with Passive Elements).

CEN455: Dr. Nassim Ammour 21

 PID Controller Design
• Transfer Function of the compensator (two zeros and one pole):
𝑘1 𝑘
𝑘2 𝑘1 𝑠 + 𝑘2 + 𝑘3 𝑠 2 𝑘3 ( 𝑠 2 + 𝑠 + 2)
𝑘3 𝑘3
𝐺𝑐 𝑠 = 𝑘1 + + 𝑘3 𝑠 = =
𝑠 𝑠 𝑠

• Design Procedure (Fig (a) ) Fig (a) PID controller implementation

1. From the requirements figure out the desired pole location to meet transient response specifications.

2. Design the PD controller to meet transient response specifications.

3. Check validity (all requirements have been met) of the design by simulation.

4. Design the PI controller to yield the required steady-state error.

5. Determine the gains, 𝑘1 , 𝑘2 𝑎𝑛𝑑 𝑘3 (Combine PD and PI).

6. Simulate the system to be sure all requirements have been met.

7. Redesign if simulation shows that requirements have not been met.

CEN455: Dr. Nassim Ammour 22
 Example51
Given the system of Figure (a), design a PID controller so that the
system can operate with a peak time that is two-thirds that of the
uncompensated system at 20% overshoot and with zero steady-
state error for a step input.
Fig (a) Uncompensated feedback control system
SOLUTION
• Evaluation of the uncompensated system
Dominant second-order
20% damping ratio pair of poles
Overshoot ζ = 0.456, -5,415 ±j10.57 with gain of 121.5.
along damping ratio line
Peak time 𝜋
𝑇𝑝 = = 0.297 𝑠𝑒𝑐𝑜𝑛𝑑𝑠
𝜔𝑑
between - 8 and -10 for a gain
equivalent to that at the dominant poles
A third pole at -8.169

• To reduce the peak time to two-thirds. (find the compensated system's dominant pole location)
The imaginary part
𝜋 𝜋
𝜔𝑑 = = = 15.87
𝑇𝑝 (2ൗ )(0.297)
3

The real part 𝜔𝑑 tan 117.130 = - tan(180 − 117.130 )

𝜎= = −8.13
tan(117.130 )
CEN455: Dr. Nassim Ammour 23
 Example52
• Design of the compensator
(sum of angles from the uncompensated system's poles and zeros to the
desired compensated dominant pole is −198.370 )

the required contribution

from the compensator zero 𝑧𝑐
−198.370 + 𝜃𝑐 = −1800 𝜃𝑐 = 18.370
Fig (a) Calculating the PD compensator zero

compensating
zero's location. the PD controller.
15.87
From geometry = 𝑡𝑎𝑛18.370 𝑧𝑐 = 55.92 𝐺𝑃𝐷 (𝑠) = (𝑠 + 55.92)
in Fig(a) 𝑧𝑐 − 8.13

gain at the design point

𝑘 = 5.34

• The PD-compensated system is simulated. Fig (b) (next slide) shows the reduction
in peak time and the improvement in steady-state error over the uncompensated Fig (b) Root locus for PD-compensated system
system. (step 3 and 4)

CEN455: Dr. Nassim Ammour 24

 Example53
• A PI controller is used to reduce the steady-state error to zero
(for PI controller the zero is placed at -0.5 close to the origin)

PI controller is used as 𝑠 + 0.5

𝐺𝑃𝐼 (𝑠) =
𝑠
Searching the 0.456
damping ratio line
−7.516 ± 𝑗 14.67 𝑤𝑖𝑡ℎ 𝑎𝑠𝑠𝑜𝑐𝑖𝑎𝑡𝑒𝑑 𝑔𝑎𝑖𝑛 𝑘 = 4.6
we find the dominant,
second-order poles
Fig (a) Root locus for PID-compensated system
• Now we determine the gains (the PID parameters),
𝑘 𝑠 + 55.92 𝑠 + 0.5 4.6 𝑠 + 55.92 𝑠 + 0.5
𝐺𝑃𝐼𝐷 𝑠 = =
𝑠 𝑠

1 1
= 256.5 + 128.6 + 4.6 𝑠 = 𝑘1 + 𝑘2 + 𝑘3 𝑠
𝑠 𝑠

Matching: 𝑘1 = 256.5, 𝑘2 = 128.6, 𝑘3 = 4.6

CEN455: Dr. Nassim Ammour 25

Fig (b) Step responses for uncompensated, PD compensated, and PID compensated systems
 Lag-Lead Compensator Design
(Cheaper solution then PID)
• First design the lead compensator to improve the transient response. Next we design the lag compensator to meet the steady-
state error requirement.

• Design procedure:

1. Determine the desired pole location based on specifications. (Evaluate the performance of the uncompensated system).

2. Design the lead compensator to meet the transient response specifications.(zero location, pole location, and the loop gain).

3. Evaluate the steady state performance of the lead compensated system to figure out required improvement.(simulation).

4. Design the lag compensator to satisfy the improvement in steady state performance.

5. Simulate the system to be sure all requirements have been met. (If not met redesign)

CEN455: Dr. Nassim Ammour 26

 Example61
Design a lag-lead compensator for the system of Figure so that the
system will operate with 20% overshoot and a twofold reduction in
settling time. Further, the compensated system will exhibit a tenfold
improvement in steady-state error for a ramp input.
SOLUTION Fig (a) Uncompensated system

• Step 1: Evaluation of the uncompensated system

Dominant second-order
20% damping ratio pair of poles
Overshoot ζ = 0.456, -1,794 ±j3.501 with gain of 192.1.
along damping ratio line

• Step 2 : Lead compensator design (selection of the location of the compensated system's dominant poles).
Twofold reduction the imaginary part of
of settling time the dominant pole
−ζ 𝜔𝑛 = −2 1.794 = −3.588 𝜔𝑑 = ζ 𝜔𝑛 tan(117.130 ) = 7.003
the real part of
the dominant pole

lead compensator design.

𝑧𝑐 = −6
Arbitrarily select a location
for the lead compensator zero.
- compensator zero coincident with the open-loop pole to eliminate a zero and leave the lead-compensated system
with three poles. (same number that the uncompensated system has)
27
CEN455: Dr. Nassim Ammour
 Example62
- Finding the location of the compensator pole.
- Sum the angles to the design point from the uncompensated system's poles
and zeros and the compensator zero and get -164.65°.
- The difference between 180° and this quantity is the angular contribution
required from the compensator pole (—15.35°).
- Using the geometry shown in Figure (b)

Compensator pole.
7.003 𝑝𝑐 = −29.1
= tan(15.35° )
𝑝𝑐−3.588
Fig (a) Root locus for uncompensated system

- The complete root locus for the lead-compensated system is sketched in Figure (c)

Fig (b) Evaluating the compensator pole

Fig (c) Root locus for lead-compensated systemCEN455: Dr. Nassim Ammour 28
 Example63
• Steps 3 and 4: Check the design with a simulation. (The result for the lead compensated system is shown in Figure(a) and is
satisfactory.)

• Step 5: design the lag compensator to improve the steady-state error. inversely the addition of lead
proportional to compensation has improved
uncompensated system's the steady- the steady-state error by a
open-loop transfer function state error factor of 2.122
static error
192.1 constant
𝐺 𝑠 = 𝑘𝑣𝑂 = 3.201
𝑠 𝑠 + 6 (𝑠 + 10) Need of tenfold
lead-compensated system's 6.794 improvement 10
open-loop transfer function static error = 2.122 𝑘𝑣𝑁 = = 4.713
constant 3.201 lag compensator factor 2.122
1977 for steady-state error
𝐺𝐿𝐶 𝑠 = 𝑘𝑣 = 6.794 improvement
𝑠 𝑠 + 10 (𝑠 + 29.1)

Step 6: We arbitrarily choose the lag compensator pole at 0.01, lag-lead-compensated

lag compensator lag Open loop TF
zero 𝑘𝑣𝑁 4.713
compensator (𝑠 + 0.04713) 𝐾 (𝑠 + 0.04713)
𝑧𝑐 = 𝑝𝑐 = 0.01 = 0.04713 𝐺𝐿𝑎𝑔 𝑠 = 𝐺𝐿𝐿𝐶 𝑠 =
𝑘𝑣𝑂 3.201 𝑠 + 0.01 𝑠 𝑠 + 10 (𝑠 + 29.1) 𝑠 + 0.01

- The uncompensated system pole at - 6 canceled the lead compensator zero at -6.
- Drawing the complete root locus for the lag-lead-compensated system and by searching along the 0.456 damping ratio line
closed-loop
dominant poles
𝑝𝑐 = −3.574 ± 𝑗 6.976 with a gain of 1971. 29
CEN455: Dr. Nassim Ammour
 Example64

- Fig (b) shows the complete draw of the lag-lead-compensated root locus.
- The lag-lead compensation has indeed increased the speed of the system (settling time or the peak time).

Step 7: The final proof of our designs is shown by the simulations of Figure (b)

Fig (a) Root locus for lag-lead-compensated system Fig (b) Improvement in step response for lag-lead-compensated system
CEN455: Dr. Nassim Ammour 30
 Feedback Compensation
Compensator 𝐻𝑐 𝑠 is used at the minor feedback to improve
transient response and steady-state response independently
• More complicated than cascade.
• Generally provide faster response.
• Can be used in cases where noise is a concern if we use
cascade compensators.
• May not require additional gain.

The design of feedback compensation consists of finding the gains, such as 𝐾, 𝐾1 𝑎𝑛𝑑 𝐾𝑓 .
Similar to cascade compensation. Consider compensation as adding poles and zeros to feedback section for the equivalent
system. Moving a
Blocks in cascade
(product of blocks)
summing Moving a
point ahead pickoff point
of a block 𝐾 𝐾1 𝐺1 (𝑠) 𝐺2 (𝑠) behind a block

1 1
𝐾 𝐾𝑓 𝐻𝑐 (𝑠)
𝐺2 (𝑠)

𝐾𝐾1 𝐺1 (𝑠)𝐺2 (𝑠)

𝐾𝐾1 𝐺1 (𝑠)𝐺2 (𝑠)
𝐾𝑓 𝐻𝑐 (𝑠)
𝐾𝐺2 (𝑠) 𝐾𝑓 𝐻𝑐 (𝑠)
+1
𝐾𝐺2 (𝑠)
1 Two blocks in parallel
(sum of blocks)
CEN455: Dr. Nassim Ammour 31
 Example71
Given the system of Figure (a), design rate feedback compensation, as shown in
Figure (b), to reduce the settling time by a factor of 4 while continuing to operate
the system with 20% overshoot.
System
SOLUTION
• First design a PD compensator.
- For the uncompensated system, Search along the 20% overshoot line (𝜁 = 0.456)
The angle of the 20%
overshoot line 180° − arccos(𝜁) = 117.13°
the dominant −1.809 ± 𝑗 3.531 (𝑠𝑒𝑒 𝑓𝑖𝑔 𝑒 )
poles system with rate feedback compensation
- The settling time is 2.21 seconds and must be reduced by
a factor of 4 to 0.55 second.
𝐾𝑓 𝑠 + 1 feedback,
1
zero at
𝐾𝑓

• Next determine the location of the dominant

poles for the compensated system.
Equivalent BD from fig (b) equivalent compensated system;
- To achieve a fourfold decrease in the settling time, the real unity feedback

part of the desired pole must be increased by a factor of 4.

Real part of
Compensated pole 4 −1.809 = −7.236
Imaginary part of
Compensated pole
𝑤𝑑 = −7.236 tan 117.13° = 14.12 (e) Root locus for uncompensated system
equivalent compensated system showing unity feedback
CEN455: Dr. Nassim Ammour 32
 Compensated dominant Example72
pole position

𝑝𝑐 = −7.236 ± 𝑗 14.12

• Sum of the angles from the uncompensated system's poles (add zero to yields 180° )
compensator
zero contribution

𝜃 = −277.33° 𝜃𝑧 − 277.33° = −180° → 𝜃𝑧 = +97.33°

• Using the geometry shown in Figure (a)

Compensator's 14.12
zero location = tan(180° − 97.33° ) 𝑧𝑐 =5.42 (a) Finding the compensator zero
7.236 − 𝑧𝑐
• The root locus for the equivalent compensated system (fig (c) previous slide) is shown in Figure (b)

The gain at
the design point, 𝐾1 = 1388

Since 𝐾𝑓 is the reciprocal

1
𝑧𝑐 = 1 1
of the compensator zero, 𝐾𝑓 𝐾𝑓 = 𝑧 = 5.42 = 0.185 𝐾1 𝐾𝑓 = 256.7
𝑐

• steady-state error characteristic (fig (d) slide 32 ) (b) Root locus for the
𝐾1 compensated system
𝐾𝑣 = lim 𝑠𝐺(𝑠) = = 4.18
𝑠→0 75 + 𝐾1 𝐾𝑓
CEN455: Dr. Nassim Ammour 33
 Example73

• The closed-loop transfer function is (fig (d) slide 32)

𝐺(𝑠) 𝐾1
𝑇 𝑠 = = 3
1 + 𝐺 𝑠 𝐻(𝑠) 𝑠 + 20 𝑠 2 + 75 + 𝐾1 𝐾𝑓 𝑠 + 𝐾1

• The results of the simulation are shown in Figure (a) and (b)

over-damped response
The settling time is 2.21 seconds with a settling time of 0.75 second

(a) Step response for uncompensated system CEN455: Dr. Nassim Ammour
(b) Step response for the compensated system 34
 Physical Realization of Compensation
Active-Circuit Realization
• 𝑍1 (𝑠)and 𝑍2 (𝑠)are used as a building block to implement the compensators and controllers, such as PID controllers.

• The transfer function of an inverting operational amplifier

𝑉0 (𝑠) 𝑍2 (𝑠)
=−
𝑉𝑖 (𝑠) 𝑍1 (𝑠)

Fig (a) Operational amplifier for transfer function realization

CEN455: Dr. Nassim Ammour 35

• Table1 summarizes the realization of PI, PD, and
PID controllers as well as lag, lead, and lag-lead
compensators using Operational amplifiers.

• Fig (a) : lag-lead compensator can be formed by

cascading the lag compensator with the lead
compensator.

Fig (a) Lag-lead compensator implemented with operational amplifiers

Table 1 Active realization of controllers and compensators, using an operational amplifier

CEN455: Dr. Nassim Ammour 36
 Example8
Implement the PID controller of Example 5

SOLUTION
4.6 𝑠 + 55.92 𝑠 + 0.5
• The transfer function of the PID controller is 𝐺𝑐 𝑠 =
𝑠
27.96
• which can be put in the form 𝐺𝑐 𝑠 = 𝑠 + 56.42 +
𝑠
• Comparing the PID controller in Table 1 with this equation we obtain
the following three relationships:
𝑅2 𝐶1 1
+ = 56.42 𝑅2 𝐶1 = 1 = 27.96
𝑅1 𝐶2 𝑅1 𝐶2
• Since there a re fou ru nk now ns a nd three equ a tions
we arbitrarily select a practical value:
Fig (a) PID controller

𝐶2 = 0.1 𝜇𝐹 𝑅1 = 357.65 𝑘Ω, 𝑅2 = 178.891 𝑘Ω 𝑎𝑛𝑑 𝐶1 = 5.59𝜇𝐹

• The complete circuit is shown in Figure (a) where the circuit element values have been rounded off.

CEN455: Dr. Nassim Ammour 37

 Passive-Circuit Realization
• Lag, lead, and lag-lead compensators can also be implemented with passive networks (Table 2) .

TABLE 2 Passive realization of compensators

CEN455: Dr. Nassim Ammour 38
 Example9
Realize the lead compensator designed in Example 4 (Compensator b zero at -4).

SOLUTION
𝑠+4
• The transfer function of the lead compensator is 𝐺𝑐 𝑠 =
𝑠 + 20.09

• Comparing the transfer function of a lead network shown in Table 2 with The equation, we obtain the
following two relationships:
1 1 1
=4 𝑎𝑛𝑑 + = 20.09
𝑅1 𝐶 𝑅1 𝐶 𝑅2 𝐶

• Since there are three network elements and two equations, we may select one of the element values
arbitrarily
𝐶 = 1 𝜇𝐹 𝑅1 = 250 𝑘Ω 𝑎𝑛𝑑 𝑅2 = 62.2 𝑘Ω

CEN455: Dr. Nassim Ammour 39