Feedback Control Systems

(KON 313E)
Week 7

Prof.Dr. Volkan Sezer

Control and Automation Engineering Department

Root Locus Techniques
Root Locus Techniques
• Root locus, a graphical presentation of the closed-loop poles as a
system parameter is varied.
• Obtained from open-loop system.
• It is a powerful method of analysis and design for stability and
transient response

Closed loop transfer function changes by the change of K!

Root Locus Techniques
How do the closed loop poles change with K in the following unity feedback system?

𝑠 2 + 10𝑠 + 𝐾 = 0

−10 ± 100 − 4𝐾
𝑠1,2 =
2

𝐾
𝑠(𝑠 + 10) 𝐾
𝐶𝐿𝑇𝐹 = = 2
𝐾 𝑠 + 10𝑠 + 𝐾
1+
𝑠(𝑠 + 10)

Aim: To be able to draw this graph without calculating individual poles, even if the system is of high degree!
Root Locus Techniques– Preliminaries
Vector representation of complex numbers:

𝜎 + 𝑗𝑤 can be expressed as amplitude and angle.

𝑀= 𝜎 2 +𝑤 2
M‫𝜃ہ‬ 𝑤
θ = tan−1
𝜎

𝜎1 + 𝑗𝑤1 M1 උ𝜃1 M1
= උ𝜃 - 𝜃
𝜎2 + 𝑗𝑤2 M2 උ𝜃2 M2 1 2
Root Locus Techniques– Preliminaries
𝑁(𝑠)
The amplitude and angle of any transfer function TF = 𝐷(𝑠) at s --> +wj , can be found as:

If Find F(s) at

1 ς 𝐷𝑖𝑠𝑡𝑎𝑛𝑐𝑒 𝑜𝑓 𝑧𝑒𝑟𝑜𝑠 𝑡𝑜 "𝑠"

M= For S+1 :
1 ς 𝐷𝑖𝑠𝑡𝑎𝑛𝑐𝑒 𝑜𝑓 𝑝𝑜𝑙𝑒𝑠 𝑡𝑜 "𝑠"

For s :
θ = ෍ 𝐴𝑛𝑔𝑙𝑒𝑠 𝑜𝑓 𝑧𝑒𝑟𝑜𝑠 𝑡𝑜"𝑠" − ෍ 𝐴𝑛𝑔𝑙𝑒𝑠 𝑜𝑓 𝑝𝑜𝑙𝑒𝑠 𝑡𝑜"𝑠"

For S+2 :

Then:
Root Locus Techniques– Preliminaries
This expression contains both amplitude and angle conditions.

Amplitude Condition
This means that if the angle
Angle Condition condition is met:

1 ς 𝐷𝑖𝑠𝑡𝑎𝑛𝑐𝑒 𝑜𝑓 𝑧𝑒𝑟𝑜𝑠 𝑡𝑜 "𝑠"

The amplitude of at s=𝜎 + 𝑗𝑤 ′
1 ς 𝐷𝑖𝑠𝑡𝑎𝑛𝑐𝑒 𝑜𝑓 𝑝𝑜𝑙𝑒𝑠 𝑡𝑜 "𝑠"

1 ς 𝐷𝑖𝑠𝑡𝑎𝑛𝑐𝑒 𝑜𝑓𝑝𝑜𝑙𝑒𝑠 𝑡𝑜 "𝑠"

CLTF= Since 𝐾=
1 ς 𝐷𝑖𝑠𝑡𝑎𝑛𝑐𝑒 𝑜𝑓 𝑧𝑒𝑟𝑜𝑠 𝑡𝑜 "𝑠"

S values which make 1 + 𝐾𝐺 𝑠 𝐻(𝑠) = 0

Is S=-2+3j a closed loop pole?
are closed loop poles of the system.
Therefore, a point located in the s plane must meet
the following condition in order to be included in
the root-locus. (k integer)

𝐾𝐺 𝑠 𝐻 𝑠 = −1 Since it is not an odd multiple of 180, the given

= 1 ‫ ہ‬2𝑘 + 1 180 point cannot be a closed loop pole. This means it
cannot be located on the locus!!
Root Locus Techniques– Preliminaries
𝑗 2
Is the following point: −2 + a closed loop pole?
2

We can also find which K makes the given point a closed loop
pole, using amplitude condition.

Obtain by yourself if the angle condition is met!!

(Yes)
Root Locus Techniques– Preliminaries
Finite and Infinite Zeros and Poles

In a transfer function when s→ ∞ the result becomes ∞, this transfer function has pole(s) at infinity.

In a transfer function when s→ ∞ the result becomes 0, this transfer function has zero(s) at infinity.

𝐺 𝑠 =𝑠+7 It has pole at infinity.

𝐺 𝑠 = 1/𝑠 It has zero at infinity.

The total number of zeros and poles of a transfer function is equal, considering the number of zeros and poles at infinity.

1
𝐺 𝑠 = There are 3 zeros in infinity.
(𝑠 + 6)3
Sketching the Root Locus

Rule1:

Start by placing the open loop zeros and poles of KG(s)H(s) on the S plane.

The total number of branches will be equal to the number of poles.

Rule2:

Root locus will be symmetrical with respect to the real axis.

Sketching the Root Locus
Rule3:

On the real axis, for K > 0 the root locus exists to the left of an odd number of real axis,
finite open-loop poles and/or finite open-loop zeros.

R(s) 𝐾(𝑠 + 3)(𝑠 + 4) C(s)

+ (𝑠 + 1)(𝑠 + 2)
_
Sketching the Root Locus
Rule3:

On the real axis, for K > 0 the root locus exists to the left of an odd number of real axis,
finite open-loop poles and/or finite open-loop zeros.

R(s) 𝐾(𝑠 + 3)(𝑠 + 4) C(s)

+ (𝑠 + 1)2 (𝑠 + 2)
_
 𝑁𝐺 (𝑠)
𝐾
Sketching the Root Locus 𝑇 𝑠 =
𝐾𝐺(𝑠)
1 + 𝐾𝐺 𝑠 𝐻(𝑠)
=
𝐷𝐺 (𝑠)
𝑁 (𝑠) 𝑁𝐻 (𝑠)
1+𝐾 𝐺
𝐷𝐺 (𝑠) 𝐷𝐻 (𝑠)
Rule4:
The root locus begins at the finite and infinite poles of G(s)H(s) and 𝐾𝑁𝐺 (𝑠)𝐷𝐻 (𝑠)
=
ends at the finite and infinite zeros of G(s)H(s). 𝐷𝐺 (𝑠)𝐷𝐻 (𝑠) + 𝐾𝑁𝐺 (𝑠)𝑁𝐻 (𝑠)

𝐾𝑁𝐺 (𝑠)𝐷𝐻 (𝑠)

For the two previous same example 𝑓𝑜𝑟 𝐾 → 0 𝑇(𝑠) =
𝐷𝐺 (𝑠)𝐷𝐻 (𝑠) +∈
small
So when K goes to zero, the closed loop poles are the poles of G(s)H(s).

𝑁𝐺 (𝑠)𝑁𝐻 (𝑠)
Proof: 𝐺 𝑠 𝐻(𝑠) =
𝐷𝐺 (𝑠)𝐷𝐻 (𝑠)

𝐾𝑁𝐺 (𝑠)𝐷𝐻 (𝑠)

𝑓𝑜𝑟 𝐾 → ∞ 𝑇(𝑠) =
𝐷𝐺 (𝑠)𝐷𝐻 (𝑠) + 𝐾𝑁𝐺 (𝑠)𝑁𝐻 (𝑠)
Can be ignored for a big K

𝑁𝐺 (𝑠) 𝑁𝐻 (𝑠) For a big K, closed loop poles are zeros of G(s)H(s).
𝐺 𝑠 = 𝐻 𝑠 =
𝐷𝐺 (𝑠) 𝐷𝐻 (𝑠)
Sketching the Root Locus
Rule5: The root locus approaches straight lines as asymptotes as the locus approaches infinity.
Numberof Asymptotes = 𝑵𝒖𝒎𝒃𝒆𝒐𝒇𝑭𝒊𝒏𝒊𝒕𝒆𝑷𝒐𝒍𝒆𝒔 − 𝑵𝒖𝒎𝒃𝒆𝒓𝒐𝒇𝑭𝒊𝒏𝒊𝒕𝒆𝒁𝒆𝒓𝒐𝒔

The asymptotes intersect the real axis

and the point 𝜎𝑎 and the angle 𝜃𝑎 with
the real axis are calculated as follows:

Finite Poles: 0,-1,-2,-4

Finite Zeros: -3

Number of Asymptotes=4-1=3
With the information so far, the root-locus can be roughly skeched!
Sketching the Root Locus
Rule6: The points of departure from the real axis 𝜎1 and arrival at the real axis 𝜎2 can be found as follows:
For 𝜎1 , K is its highest value on real axis.

For 𝜎2 , K is its lowest value on real axis.

𝐾𝐺 −1
𝐶𝐿𝑇𝐹 = → 𝐾𝐺 𝑠 𝐻 𝑠 + 1 = 0 →K=
1+𝐾𝐺𝐻 𝐺 𝑠 𝐻(𝑠)

If the derivative of K with respect to 𝜎 is taken and set equal to zero, the minimum and maximum values are found!

−1
K= 𝐺 𝑠 𝐻(𝑠)

𝜎1 = −1.45 𝜎2 = 3.82
Along the real axis s= 𝜎
Sketching the Root Locus
Rule6: The points of departure from the real axis 𝜎1 and arrival at the real axis 𝜎2 can be found as follows:
For 𝜎1 , K is its highest value on real axis.

For 𝜎1 , K is its lowest value on real axis.

𝐾𝐺 −1
𝐶𝐿𝑇𝐹 = → 𝐾𝐺 𝑠 𝐻 𝑠 + 1 = 0 →K=
1+𝐾𝐺𝐻 𝐺 𝑠 𝐻(𝑠)

If the derivative of K with respect to 𝜎 is taken and set equal to zero, the minimum and maximum values are found!

𝜎1 = −1.45

𝜎2 = 3.82
Sketching the Root Locus
Rule7: The calculation of the point of intersection of the jw axis can be done in 2 ways:

Method A: The jw axis intersection point is the point at which the locus curve passes from the stable region to the
unstable region, and this can be found with the Routh Hurwitz table. According to the Routh-Hurwitz table, if a row
(except for the 𝑠 0 row) has the value '0', there is a pole on the jw axis. Cutting points are found with the polynomial
obtained from the previous line.

Method B: In the characteristic equation of the closed loop transfer function, making s=jw and 1+KG(jw)H(jw)=0+0j ;
K and W can be calculated.

W:Gives the places where it intersects the jw axis

K: Gives the values of K at these intersection points.
Sketching the Root Locus
Rule7 Method A: The jw axis intersection point is the point at which the locus curve passes from the stable region to the
unstable region, and this can be found with the Routh Hurwitz table. According to the Routh-Hurwitz table, if a row (except
for the 𝑠 0 row) has the value '0', there is a pole on the jw axis. Cutting points are found with the polynomial obtained from
the previous line.
Example:

By finding the CLTF of this system we have seen

before, the intersection point of the jw axis of the First, we find the K that makes a row zero. Then the polynomial of the
root curve can be found by using Routh Hurwitz. previous line is written and the roots are found.
→
(Other root of K which is -73,5, makes system unstable)
CLTF=
From previous row:
𝑠1,2 = ±1.59𝑗

jw intersection points
Sketching the Root Locus
Rule8: Determination of the angles of departure from the poles or arrival at zero.
We know that the sum of the angles from open-loop zero and poles to any closed-loop pole must be an odd
multiple of 180. (Angle Rule) (The angle of zeros was taken as positive and the angle of poles as negative)

Since the root-locus curve (closed-loop poles) necessarily passes through the open-loop zero and poles, the angle
condition can be applied to the zero departed from or to be reached as close as ∈ (∈⤏0)
Sketching the Root Locus
Rule8: Determination of the angles of departure from the poles or arrival at zero.
We know that the sum of the angles from open-loop zero and poles to any closed-loop pole must be an odd
multiple of 180. (Angle Rule) (The angle of zeros was taken as positive and the angle of poles as negative)

Since the root-locus curve (closed-loop poles) necessarily passes through the open-loop zero and poles, the angle
condition can be applied to the zero departed from or to be reached as close as ∈ (∈⤏0)
−𝜃1 −𝜃2 +𝜃3 −𝜃4 = 180(2𝑘 + 1)
𝜃1 : 𝐴𝑛𝑔𝑙𝑒 𝑜𝑓 𝐷𝑒𝑝𝑎𝑟𝑡𝑢𝑟𝑒
𝜃2 : 90
1
𝜃3 = tan−1 = 45
1
1
Zeros: -2 𝜃4 = tan−1 = 26.6
2
Poles: -3, -1+j, -1-j

−𝜃1 − 90 + 45 − 26.6 = 180

𝜽𝟏 = −𝟐𝟓𝟏. 𝟔 = 𝟏𝟎𝟖. 𝟒
Root Locus Example
X

-1,25
X >𝜎 < X

X
𝑑𝐺(𝑠) are given as: 𝑠1 = −0,467 𝑠2,3 = 1.64 ± 2.67𝑗
The poles which make =0
𝑑𝑠

Sketch the root locus of the given system. Departure from real axis (breaking point): 𝜎
𝐾𝐺 𝑠 𝐻(𝑠)
Open Loop Poles: 0 ; -1; −2 ± 3𝑗 𝐶𝐿𝑇𝐹 =
1 + 𝐾𝐺 𝑠 𝐻(𝑠)
Number of asymptotes: 4-0=4 −1
1 + 𝐾𝐺 𝑠 𝐻 𝑠 = 0 → 𝐾 = = -1[𝑠(𝑠 + 1)(𝑠 2 + 4𝑠 + 13)]
X 𝐺 𝑠 𝐻(𝑠)
Intersection point of asymptotes:
𝐾 = −1[𝜎 (𝜎 + 1)(𝜎 2 + 4𝜎 + 13)]=-1[𝜎 4 + 5𝜎 3 + 17𝜎 2 + 13𝜎]

𝑑𝐾
X > <X 𝑑𝜎
=0 −1(4𝜎 3 + 15𝜎 2 + 34𝜎+13)=0
𝜃𝑎1 = 45
(2𝑘 + 1)𝜋 𝜃𝑎2 = 135 𝜎1 = −0,467 𝜎2,3 = 1.64 ± 2.67𝑗
𝜃𝑎 =
X
4 𝜃𝑎3 = 225 Breaking point can not have
imaginary component!
𝜃𝑎4 = 315 We could have used the information given in the question
directly as the breaking point!
 jw intersection points: (METHOD B)
𝐾
Root Locus Example 𝐶𝐿𝑇𝐹 =
𝑠(𝑠 + 1)(𝑠 2 + 4𝑠 + 13)
𝐾
1+
𝑠(𝑠 + 1)(𝑠 2 + 4𝑠 + 13)

𝐾
𝐶ℎ𝑎𝑟𝑎𝑐𝑡𝑒𝑟𝑖𝑠𝑡𝑖𝑐 𝐸𝑞𝑢𝑎𝑡𝑖𝑜𝑛: 1 + 𝑠(𝑠+1)(𝑠2+4𝑠+13)=0

𝑑𝐺(𝑠) are given as: 𝑠1 = −0,467 𝑠2,3 = 1.64 ± 2.67𝑗

→ 𝑠 𝑠 + 1 𝑠 2 + 4𝑠 + 13 + 𝐾 = 0
The poles which make =0
𝑑𝑠

Sketch the root locus of the given system. 𝑠 4 + 5s 3 + 17𝑠 2 + 13𝑠 + 𝐾 = 0

(𝑗𝑤) 4 +5(𝑗𝑤)3 + 17(𝑗𝑤)2 +13 𝑗𝑤 + 𝐾 = 0 + 0𝑗
𝜃4 −𝜃1 − 𝜃2 − 𝜃3 − 𝜃4 = 180(2𝑘 + 1)
X 𝑤 4 − 5𝑤 3 𝑗 − 17𝑤 2 + 13𝑗𝑤 + 𝐾 = 0 + 0𝑗
𝜃1 =180-atan(3/2)=123,66
𝑤 4 − 17𝑤 2 + 𝐾 + 𝑗(13𝑤 − 5𝑤 3 ) = 0+0j
𝜃2 =180-atan(3/1)=108,4 w(13-5𝑤 2 )=0
𝜃2 𝜃1
-1,25
X >𝜎 < X 𝜃3 =90
𝑤1 = 0 → 𝐾 = 0
−123,66 − 108,4 − 90 − 𝜃4 =180
𝜃3 𝑤2 = 13/5 = 1,612 → 𝐾 = 37,42
X
𝜽𝟒 =-502,06≈-142,06 ≈217,94
13
𝑤3 = − = −1,612 → 𝐾 = 37,42
5
Root Locus Example Asymptotes:

𝑑𝐺(𝑠) are given as: 𝑠1 = −0,467 𝑠2,3 = 1.64 ± 2.67𝑗

The poles which make =0 Real Axis Intersection Points:
𝑑𝑠

Sketch the root locus of the given system.

217,04 Pole Departure Angle:

X
1,612j

1,25j

jw intersection points:

-1,25
X >-0,467< X

360-217,04 -1,25j
X
-1,612j
Generalized Root Locus

So far, we have obtained the variation of the closed

loop poles of the KG(s)H(s) with respect to K.

However, it may also be necessary to plot the root-locus curve

with respect to the variation of another parameter.

In this case, we can plot the root-locus by converting the

open-loop transfer function into the desired format.
Generalized Root Locus
In order to plot the root locus according to p1, firstly the closed
loop t.f. is found.

𝐶𝐿𝑇𝐹

We make a manipulation here: It is clear that the open-loop transfer

function of such a CLTF is as follows:

OLTF=
To make the denominator 1+p1G(s)H(s),
It can be written as follows; It is now in the format we
let us divide each side by 𝑠 2 + 2𝑠 + 10.
want. We can plot by
applying our rules (p=K).
Feed forward TF

Open loop TF
Generalized Root Locus

Number of Asymptotes: 2-1=1

−1 + 3𝑗 − 1 − 3𝑗 + 2
𝜎𝑎 = =0
2−1
(2𝑘 + 1)𝜋
𝜃𝑎 = = 2𝑘 + 1 𝜋 → 𝑘 = 0, 𝜃 = 180
2−1

𝑂𝐿𝑇𝐹

Zeros: -2
Poles: 𝑠1,2 = −1 ± 3𝑗

Let’s add details to the plot!

Generalized Root Locus
Departure angle from the pole
𝜃3 𝜃1 = atan 3 = 71,6
K X 𝜃2 = 90
Arrival point to real axis
𝐶𝐿𝑇𝐹 𝜃1 − 𝜃2 − 𝜃3 = 180(2𝑘 + 1)
𝜃1

𝐾𝐺𝐻 −1
𝐶𝐿𝑇𝐹 = 1+𝐾𝐺𝐻 → 1 + 𝐾𝐺𝐻 = 0 → 𝐾 = 𝐺𝐻 𝜃2
X 𝜃3 =71,6−90−180=−198,4 ≈161,6
−1 𝑠 2 +2𝑠+10 𝜎 2 + 2𝜎 + 10
𝐾= s+2 = − 𝑠+2 K=−
2
𝑠 +2𝑠+10
𝜎+2
161,6
𝑑𝐾 2𝜎 + 2 𝜎 + 2 − 𝜎 2 + 2𝜎 + 10 1
=−
𝑑𝜎 (𝜎 + 2)2
-5,16

2𝜎 2 + 4𝜎 + 2𝜎 + 4 −𝜎 2 −2𝜎 − 10 = 0

𝜎1 =-5,16, 𝜎2 =1,16