INTERNATIONAL SCIENTIFIC CONFERENCE

22 23 November 2013, GABROVO

DESIGN AND CONTROL OF PID-CONTROLLED

BALL AND BEAM SYSTEM
Stk KOCAOLU
Krklareli University / TURKEY
sitki.kocaoglu@kirklareli.edu.tr

Hilmi KUU
Trakya University / TURKEY
hilmi@trakya.edu.tr

Abstract
The control of The Ball and Beam system is very important in the aviation and space fields because of its similarity
to the control of aircraft during the flight, landing and turbulent,. Since real experiments are not possible in these
areas, using Ball and Beam systems in the control laboratories has become necessary. The system aims to keep the ball
on the beam in desired position. Current position of the ball is measured by using various sensors and the location of
the ball is changed by changing the angle of beam by using various motors. The mathematical model of this system is
not linear. By using several control algorithms, system tries to move the ball to the desired position appropriately. Here
system tries to set the rising time, the overshoot and offset in optimum level. In this study the PID control method is
applied and system is fixed as an experimental set. The aim is lecturing the effects of PID constants to students in
control laboratories.
Keywords: Ball and Beam, PID, Control, Microprocessor

INTRODUCTION
This study aims to control the position of
the ball moving freely on a beam. The angle
between the horizontal and the beam depends
on the movement of the motor that controls the
beam. The system brings the ball to the desired
area on the beam by using closed loop control.
Designed as an experimental set, this system
suggests a new control method and material
selection way for one of the most popular
models of automatic control.
The control of unstable systems is very
important for many control problems. Since
such systems prove dangerous to test in
vertical position control of aerospace and
airplanes, we can only study them in
laboratories by modeling the system.
In this project we used an aluminum beam,
a stepper motor, a ping pong ball, and two
infrared sensors to measure the position of the
ball. The system attempts to measure the angle
of beam and the position of the ball. We can
calculate the angle between beam and
horizontal because we know the motors angle
per step. But we cant calculate the linear
speed of the ball which is under outer effects.
We designed P, PD, PI and PID
controllers for this system. The user can
choose Kp, Kd and Ki coefficients freely. So

the user can see the effects of coefficients and

compare the controller structures.
The ball and beam system has been studied
and used in the tests of new methods by many
researchers. In particular, I. Petkovi, M.
Brezak, and R. Cupec have conducted artificial
vision based tests of the system [1]. P. Dadios
and R. Baylon did a similar study on fuzzy
logic [2]. J. Whelan and J. V. Ringwood, on
the other hand, used the same system for
position tracking through a camera instead of
sensors, using PID controllers [3]. S. Sridharan
and G. Sridharan developed a new control lab
device named beam and ball system sliding on
a cylinder in which they used PID controllers [4].
Using PID technique on unstable systems
control is very common and this experimental
set is not dependent on computer, which
makes the system a usable set for automatic
control laboratories.
EXPOSITION
As seen in Figure 1, the system includes
three mechanic components. These are the
ball, the beam and the stepper motor that
changes the angle of the beam. We supposed
that the shaft of the motor is rigid and we took
the centrifugal force as zero to simplify the
function.

2013

III-41

(5)
The angle is variable between 00 and 300.
Also
is variable between 0 and 0,5. The
force on the ball is also variable between 0N
and 0,0095 N.
The moment of the motor is
(6)

Fig. 1. The drawing of the system components

We supposed that beams axis is too close

to balls contact plane and that the ball is not
sliding on the beam. We must calculate the
forces on the contact points of the components
to find out movement equations.
Initially, we can say there are two forces
that effect the balls position on the beam. The
first one is the gravity that affects the ball
according to angle of the beam and second is
the friction force between ball and beam
according to the friction coefficient.

And we can find out the transfer function of

the system by using the angle. We know that
the step angle of the motor is equal to 1,8o.
(7)
(8)
Formula 8 is not a complete transfer
function, because this involves many
unverified assumptions. Also, since this study
is an experimental set, we are abstracting away
from putting restrictions on controller type and
variables like settling time or overshoot.

(1)
The position of the ball on the beam is
(2)
where is angular turning and r is the distance
between turning axis of ball and contact
position between ball and beam. r is nearly
equal to radius of the ball. The momentum of
the ball is
(3)
where

is the moment of inertia of the ball.

(4)

The radius of the ball is equal to 0,63 cm

and the mass of the ball is 2 g. By calculation,
the moment of inertia is equal to 0,3.107
kg/m2, which is considered approximately
zero. Thus, for formula 3, the frictional force
is equals to zero.

III-42

Fig. 2. Photo of experimental set

Signals from sensors have been filtered by

using suitable capacitors due to high
distortion. The microcontroller takes the
signals through built-in 10-bit ADC ports. The
microcontroller compares the signals to the
reference value. After that, it calculates the
number of steps it will feed into the motor.

2013

START

LCD1
LM044L

OK

b1

b2

SOL

SAG

D0
D1
D2
D3
D4
D5
D6
D7

RS
RW
E

b3

R1

R2

R3

R4

1k

1k

1k

1k

7
8
9
10
11
12
13
14

4
5
6

1
2
3

VSS
VDD
VEE

b4

+6V

d4
d5
d6
d7

rs
rw
e

+6V

U3

sharp_digital
rw
d4
d5
d6
d7

33
34
35
36
37
38
39
40

18

RA0/AN0
RC0/T1OSO/T1CKI
RA1/AN1
RC1/T1OSI/CCP2/UOE
RA2/AN2/VREF-/CVREF
RC2/CCP1/P1A
RA3/AN3/VREF+
RC4/D-/VM
RA4/T0CKI/C1OUT/RCV
RC5/D+/VP
RA5/AN4/SS/LVDIN/C2OUT
RC6/TX/CK
RA6/OSC2/CLKO
RC7/RX/DT/SDO
OSC1/CLKI
RB0/AN12/INT0/FLT0/SDI/SDA
RB1/AN10/INT1/SCK/SCL
RB2/AN8/INT2/VMO
RB3/AN9/CCP2/VPO
RB4/AN11/KBI0/CSSPP
RB5/KBI1/PGM
RB6/KBI2/PGC
RB7/KBI3/PGD

VUSB

RD0/SPP0
RD1/SPP1
RD2/SPP2
RD3/SPP3
RD4/SPP4
RD5/SPP5/P1B
RD6/SPP6/P1C
RD7/SPP7/P1D
RE0/AN5/CK1SPP
RE1/AN6/CK2SPP
RE2/AN7/OESPP
RE3/MCLR/VPP

15
16
17
23
24
25
26

Q1

R5

motor_puls
b1

BC556AP

sens1
1K

motor_enable

R6
sens2

RV1

RV2

100

100

Q2
BC556AP

1k

+3
48%

53%

+5

19
20
21
22
27
28
29
30

U2

b3
sens1
sens2
motor_ileri_geri
b2
e
rs

8
9
10
1

D1
LED-BLUE

8
1
5
7
6

VCC1
VCC2

X1

X1
RST
SCLK
I/O

CRYSTAL
X2

sharp2

b4

2
3
4
5
6
7
14
13

sharp1

sharp1
sharp2
pot

DS1302

RV3

+5V

100k

PIC18F4550

R7

pot

1k

56%

+5V

Fig. 5. Second Sensors Graph

Fig. 3. Proteus Circuit Model

transformation

(10)

Voltage
(V)

sensors

Position
(cm)

and the second

equation is

Table 1 Position- Voltage Values

Voltage
(V)

(9)

The graph of the two Sharp sensors

provided in the catalogue is shown in Figure 6.
As seen therein, these sensors cant measure
the positions between 0 to 15cm. Table 1
below gives the sensors position-voltage
values.

Position
(cm)

This is the longest range sensor of its type

and has 1,5m distance for catalogue
information. But, because of surface curvature
of the ball, our sensor could measure up to
30cm only in this project. So we used two
analogue sensors, a 60cm beam, and a third
digital sensor to find out which half of the
beam the ball is on.
The values measured by sensors were
calculated on MS Excel graph application as
position-voltage variations, which gave us
fourth degree transformation equations shown
below. We can see that the two sensor
equations are not identical. The first sensors
transformation equation is

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30

2,14
2,07
2,05
2,01
1,98
1,94
1,92
1,89
1,86
1,83
1,80
1,77
1,74
1,70
1,67
1,63
1,61
1,57
1,56
1,54
1,52
1,50
1,48
1,46
1,45
1,44
1,43
1,42
1,41
1,40

31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60

1,33
1,34
1,35
1,36
1,365
1,37
1,385
1,40
1,405
1,41
1,44
1,47
1,49
1,51
1,55
1,58
1,64
1,68
1,74
1,79
1,81
1,88
1,93
1,98
2,02
2,05
2,07
2,10
2,18
2,25

Fig. 4. First Sensors Graph

2013

III-43

Fig. 6. GP2Y0A0 Voltage-Distance Graph

Fig. 7. Control Panel

By stabilizing rotation speed and input

pulse rate in open-loop control systems,
stepper motors ensure exact position
correctness by stabilizing rotation speed and
pulse frequency, i.e. input pulse rate, and by
minimizing angular fault to negligible levels
[5].
In this project we used Nema23 model
hybrid stepper motor. Hybrid stepper motors
are used for precise position control. Such
motors can produce high torque particularly
when they are used with bipolar drivers. [6].
This motor, has a step angle of 1,8o and is
appropriate for frequencies up to 20000Hz.
Our square wave frequency is 2000Hz and the
driver switch is at 25600 puls/rev. position.
This provides us with a resolution 128 times
higher than normal in using micro stepping,
thus reducing the stepping angle all the way to
0,014o.
Steps fed into the motor are calculated as a
function of the difference between reference
and desired positions of the ball and the
coefficients Kp, Kd and Ki. Steps obtained
through PID calculation are fed to the motor
driver approximately every 300ms.
As shown in Figure 7, an LCD display with
4 lines and 20 characters has been preferred as
the user interface. The reference position
selected by the user, the actual position, and
the coefficients Kp, Kd, Ki are displayed on
the LCD. The 3rd line of LCD is reserved to
display the date and time. Also integrated into
the system are 4 buttons for menu operations.

III-44

The block diagram of the system is

shown in Figure 8. Here, we have the sensors
for the position of the ball and a potentiometer
for beams angle, providing a control system
of two closed loop. Figure 9 gives the signal
flowchart of the system.

Fig. 8. Block Diagram of the System

By comparing the actual position to the

reference, the system determines the direction
of rotation for stepper motor and calculates the
number of steps. For this, the microcontroller
uses digital PID calculation. Every time the
ball passes by the middle point of the beam,
one sensor is turned off and the other on by the
third infrared sensor integrated for that
purpose specifically. The potentiometer
constantly measures beams angle. The shaft
of potentiometer is coupled with the shaft of
stepper motor, thus moving together. The
system stops when ball arrives at the reference
position. The system starts working again in
case the position of the ball changes because
of outer effects.

2013

Table 4 PI Controller Experimental Results

Kp

Ki

0,2
0,2
0,2
0,4
0,4
0,4
0,6
0,6
0,6
0,8
0,8
0,8

0,3
0,6
0,9
0,3
0,6
0,9
0,3
0,6
0,9
0,3
0,6
0,9

Time
(s)
30
47
91
-

Offset
(cm)
+2
-1
-0,5
-

Result
Success
Success
Unstable
Unstable
Unstable
Success
Unstable
Unstable
Unstable
Unstable
Unstable
Unstable

Table 5 PID Controller Experimental Results

Fig. 9. Flowchart of the System

Table 2 System Parameters

Kp

Time (s)

0,2
0,4
0,6
0,8

36
74
11
-

Offset
(cm)
+2
+1,5
+0,5
-

Result
Success
Success
Success
Unstable

Table 3 PD Controller Experimental Results

Kp

Kd
0,3

Time
(s)
9

Offset
(cm)
+9

0,2
0,2

0,6

22

+9

0,2
0,4
0,4
0,4
0,6
0,6
0,6
0,8
0,8
0,8

0,9
0,3
0,6
0,9
0,3
0,6
0,9
0,3
0,6
0,9

30
15
5
51
118
77
107

-1,5
0
-0,5
-2
-1
+0,5
+0,5

Result
Big Steadystate Error
Big Steadystate Error
Success
Success
Success
Unstable
Success
Delayed
Unstable
Unstable
Success
Delayed

Kp

Kd

Ki

Time (s)

0,2
0,2
0,2
0,2
0,2
0,2
0,2
0,2
0,2
0,4
0,4
0,4
0,4
0,4
0,4
0,4
0,4
0,4
0,6
0,6
0,6
0,6
0,6
0,6
0,6
0,6
0,6
0,8
0,8

0,3
0,3
0,3
0,6
0,6
0,6
0,9
0,9
0,9
0,3
0,3
0,3
0,6
0,6
0,6
0,9
0,9
0,9
0,3
0,3
0,3
0,6
0,6
0,6
0,9
0,9
0,9
0,3
0,3

0,3
0,6
0,9
0,3
0,6
0,9
0,3
0,6
0,9
0,3
0,6
0,9
0,3
0,6
0,9
0,3
0,6
0,9
0,3
0,6
0,9
0,3
0,6
0,9
0,3
0,6
0,9
0,3
0,6

11
44
9
27
37
20
15
25
29
31
59
88
40
52
30
-

2013

Offset
(cm)
+1
0
-2
+1
-0,5
0
0
0
+1
-0,5
-0,5
+0,5
-0,5
+0,5
+1,5
-

Result
Success
Success
Unstable
Success
Success
Unstable
Success
Unstable
Success
Unstable
Unstable
Unstable
Success
Success
Unstable
Unstable
Success
Unstable
Success
Success
Success
Unstable
Success
Unstable
Success
Success
Unstable
Unstable
Unstable
III-45

0,8
0,8
0,8
0,8
0,8
0,8
0,8

0,3
0,6
0,6
0,6
0,9
0,9
0,9

0,9
0,3
0,6
0,9
0,3
0,6
0,9

131
-

+0,5
-

Unstable
Delayed
Unstable
Unstable
Unstable
Unstable
Unstable

CONCLUSION
This study presents a new model and
design of Ball and Beam system for one of the
most popular models of automatic control. The
set developed can be used in automatic control
labs. The goal of in this study is not to bring
the ball to the desired position in the most
optimal way, but rather to observe the effects
of coefficients Kp, Kd and Ki clearly. To that
end, we kept working independently from the
computer and created a control panel to
conduct experiments with the system easily
and rapidly.
Upon getting the system to work as
desired, we conducted a number of
experiments to observe the effects of the
coefficients by increasing and decreasing
them. Approximately 52% of the trials were
successful, and the ball settled in a reasonable
period by an appropriate steady-state error
(Tables 2, 3, 4, 5).
During this experiment P, PD, PI and
PID controllers were tested separately. P
controller was successful in 75% of the trials,
but if a very large coefficient Kp is selected,
the oscillations continued (Table 2). ..PD
controllers succeeded in 83% of the cases, but
when the proportional coefficient is increased,
it failed too (Table 3).
Proportional coefficient is significant
for PI-controlled first order systems. This is
because, when ramp input is applied, steadystate error is proportional to the amplitude of
the input, but high
may render the system
unstable [7]. PI controller was successful in
25% of the cases, with no attempt with zero
offset, and all the attempts with higher values
of the proportional coefficient failed (Table 4).
PID controller attempts had a success rate of
48%, 25% of which had zero offset (Table 5).
PD controller structure came out as the most

III-46

successful in practice. Given that the control

signal continuous in integral controllers even
with zero offset, the success of PD controllers
can be attributed to the lack of an integral
controller in its structure.
In future studies, different controller
structures should be tested on this experiment
set. Due to the close overlap between the
model developed here and the conditions in
aviation and spacecraft trajectory tracking
steps, further research on this system could
contribute to the future developments in the
field.

REFERENCE
[1] E. P. Dadios, R. Baylon, R. D. Guzman, A.
Florentino, R. M. Lee ve Z. Zulueta, Vision
Guided Ball-Beam Balancing System Using
Fuzzy Logic, 26th Annual Conference of the
IEEE Industrial Electronics Society, Cilt 3,
pp. 1973-1978, 2000.
[2] J. Whelan ve J. W. Ringwood, A
Demonstration Rig for Control Systems Based
on the Ball and Beam with Vision Feedback,
Proc.3rd IFAC Symposium on Control
Education, Tokyo, 1994.
[3] S. Sridharan ve G. Sridharan, Ball and Beam
on Roller: A New Control Labrotary Device,
Proceedings of the 2002 IEEE International
Symposium on Industrial Electronics, Cilt 4,
pp. 1318-1321, 2002.
[4] T. Tez, Manyetik Ask Sisteminde Kullanlan
Kontrol Algoritmalarnn Deneysel ve Teorik
Aratrlmas, Master Thesis, Trakya Uni.
Inst. of Science, p. 64., Edirne, 2011,
[5] Quin Tep;Tamagawa Seiki Co., Ltd., 5 Phase
Step Motor and Driver, Motro Compo, No.
T12-1596N2, 1989.
[6] D. Uygun, Hibrit Adm Motorunun Saysal
Kontrol, Master Thesis, Marmara Uni.
Inst.of Science, stanbul, 2006.
[7] H. N. Hurma, PID Kontrolr ve PLC
Uygulamas,
Master
Thesis,
stanbul
Technical Uni. Inst. of Science, pp. 1130,istanbul, 1998.

2013