Proceedings of 2014 RAECS UIET Panjab University Chandigarh, 06 - 08 March, 2014

Modeling, simulation and control of single actuator

magnetic levitation system
Shimi S.L., Dnyaneshwar
Karanjkar

Tejinder Kumar
Physical Science Department
Indian Institute of Science Education
and Research
Mohali, India
tejinderdevgon@iisermohali.ac.in

I.

Hall

effect

sensor;

Magnetic

INTRODUCTION

Magnetic Levitation Systems are electromechanical

systems that are used to suspend an object in space without any
contact. This technology is becoming more popular due to its
frictionless and contactless properties as it removes energy
losses due to friction [1].
Magnetic Levitation System (MLS) have been widely used
in various fields such as frictionless bearings, high speed
maglev passenger trains, highly accurate positioning systems,
high speed computing [2]. Magnetic actuation has the potential
for numerous other applications. In addition to supporting
loads (levitation), it can dampen vibration, apply precision
force, and move objects precise distances all with no contact
between surfaces and essentially no friction. This type of
actuation can be used in harsh environments (corrosive,
vacuum, etc.) where traditional mechanical or hydraulic
actuators might not survive. A magnetic actuator can operate in
ultra clean environments without the hazard of producing
contaminants from its use [3-4]. Due to well-established
miniaturization trends in several sectors of industry e.g.,
electronics, materials, and bioengineering the manufacturing

978-1-4799-2291-8/14/$31.00 2014 IEEE

Electronics and Communication

Engineering Department
Gulzar Group of Institutes
Khanna, India
sakshi_rana@strobotix.com

Electrical Engineering Department

National Institute of Technical
Teachers Training and Research
Chandigarh, India
shimi.reji@gmail.com

Abstract This paper aims at Design, Fabrication and Control of

a single actuator magnetic levitation system. A physical design
model of Magnetic Levitation System have been presented at the
initial phase. Modeling and Simulation of this non-linear
magnetic levitation system is proposed with feedback
linearization where a non linear state space transformation is
used to linearize the system exactly. After this, experimental
setup and construction of actual magnetic levitation system is
presented. Experimental Levitation data is collected and
compared to the theory. Magnetic levitation system considered in
this study is taken as a neodymium magnet suspended in a
voltage controlled magnetic field. Dynamic behavior of the
system was modeled by the study of electromagnetic and
mechanical subsystems. State space model was derived from the
system equations. For high accuracy in position detection, Hall
effect sensor SS49E was utilized. The successful operation of this
system was obtained using relatively cheaper and simpler
magnetic levitation subsystems and components.
KeywordsElectromagnet;
levitation system; Non linear

Shakshi Rana

processes of certain components rely heavily on high precision

positioning systems[5-7]. Nanotechnology relates to devices
whose dimensions are in nanometers. This sector is at an early
stage of development but the number of its applications is
growing fast. Its growth is manifested by the increasing
demand for workers in the area [8-9]. The manufacture of
nanodevices employs processes like photolithography,
molding, and micromachining . In many cases, these processes
require positioning with nanometric accuracy. Magnetically
levitated stages are a better alternative for positioning systems
because of their more robust nature.
Most of the study have been made on the control of a MLS
using different techniques as its involves complexity and
requires the need for cheap and accurate control system. MLS
are generally highly non-linear and open loop unstable
systems. This unstable aspect of MLS and its inherent
nonlinearities make the modeling and control problems very
challenging [10-12]. Magnetic levitation system considered in
this study is taken as a cylindrical neodymium magnet
suspended in a voltage controlled magnetic field. A simple
model for non linear magnetic levitation system is introduced
for position control of a magnetically levitated permanent
magnet[13-15]. A state space model is derived from system
equations and feedback linearization for the transformation of
system to a linear MLS. The intent is to construct the system in
low cost and low precision components and still be able to
levitate the magnet with high precision.
II.

PAPER ORGANIZATION

In this paper, III will first present the physical design

aspects of the Magnetic Levitation System. This will involve
the details of various components and subsystems. IV
represents the system dynamics along with Electromagnetic
dynamics modeling and mechanical modeling. Simulation will
then be used to construct a controller based on feedback
linearization for stabilizing the maglev system and its open
loop unit step response is presented. V represents the
experimental setup and construction of actual magnetic
levitation system. Both simulation and real time control of
maglev system is done in matlab and results are presented.
Experimental data is collected and results are obtained.

III.

MAGNETIC LEVITAION SYSTEM

Fig. 1 represents the schematic diagram of magnetic

levitation system.

L: coil's inductance
B. Mechanical Modeling
Free body diagram of permanent magnet suspended by
balancing the electromagnetic force Fm and gravitational force
Fg is shown in Fig. 2.

Figure 1: Schematic Diagram of Magnetic Levitation System

Magnetic levitation system considered in this study is

taken as a neodymium magnet suspended in a voltage
controlled magnetic field. It consists of an electromagnet that
has been suspended with the help of a mechanical framework.
The electromagnet is made out of copper wire by making turns
around a hollow bobbin. A cylindrical iron rod of high
permeability is inserted inside the hollow bobbin to ensure the
flux transfer through its cross-section. Hall effect sensor is
placed at the bottom of this iron rod inside the electromagnet.
The main purpose of electromagnet here is to control the
position of the magnet levitated. This has been achieved by
controlling the current through the electromagnet in
proportion to attain a stable position for magnet being
levitated. A coil driver is used to control the current of this
electromagnet that is further driven from the control algorithm
designed in Simulink. Simulink gives the output signal to the
coil driver via interfacing achieved with the help of dSpace
DS1104 board. Feedback signal of magnetic position is
obtained with the help of Hall Effect sensor as input to the
dSpace board and signal is send to the Simulink workspace for
further calibrations. This acts as the feedback signal for the
control algorithm in Simulink that compares it with a
reference value and gives the output to the coil driver.
IV.

SYSTEM DYNAMICS AND MODELING

Dynamic behavior of the maglev system can be modeled by

the study of electromagnetic and mechanical models.
A. Electromagnetic Dynamics Modeling
Electromagnetic force produced by the current in the coil is
given by kirchoff's voltage law [2]
( )
( )
(1)
where,
v: applied voltage
i : current
R: coil's resistance

Figure 2: Free body diagram of Magnetic Levitation System

Net force Fa acting on magnet is given by Newton's third law

of motion while neglecting the damping, friction and drag
force of air, etc. [2],[7]
(2)
(3)
Where,
m: mass of magnet
x : position of ball
g: gravitational constant
K: magnetic force constant
C. Nonlinear Model
On the basis of electromagnetic dynamics modeling and
mechanical modeling, the system can be defined by following
set of differential equations [2];
(4)
( )

(5)
(6)

Eq. 5 indicates that L(x) is a nonlinear function of magnets

position x. Various approximations have been used for the
determining inductance of this system. If we take the
approximation that inductance varies with the inverse of
magnet position [4], that is
( )
(7)
where L is the constant inductance of the coil in the
absence of magnet and L0 is the additional inductance
contributed by the presence of the magnet, x0 is the equilibrium
position. Substituting Eq. 7 in Eq. 5 results in

( )

(8)
2 in Eq. 8, we get

substituting
( )

( )

(9)

D. Vector Form
Taking
,
,
, Eq. 4, Eq. 5 and Eq. 9 can
be represented in vector format where position of magnet is
taken as output variable as given below;
(

0
0

2
(11)

E. Linear Model
The system was linearized around the point
which results in state vector as:

0
0
108.7
100
The above parameters were added to the state space block in
SIMULINK and the model obtained for the unit step input
response is represented in Fig. 3;

(10)

100

In the state space form, we can write the linear model as

under;
0
1
0
813.8
0
26.04
0
113.22
130.43

(12)
At stable position of levitation, the time rate of position
must be zero, i.e. x02 = 0
Also current at the time of levitation of magnet can be
found from Eq. 7 and it must satisfy following condition:
(13)

Figure 3: Simulink model for magnetic levitation system using state space
block

The state space block includes the values of parameters A,

B, C and D as obtained above those are added with the help of
parameter selection represented in Fig. 4;

Thus linearized model in the state space form can be

written as under;
0
1
0
0
(14)
0
0
0

(15)

100

(16)

F. Simulation Model and Results

The theoretical non linear model with feedback linearization
controller was simulated using MATLAB with physical
system values as mentioned in table 1.
Table I. Physical parameters of Magnetic Levitation System

Parameter
m
g
R
L
K
x01
x02
x03

Value
0.04
9.8
1.2
0.0092
0.0001
0.012
0
0.75

Unit
Kg
m/s2
Ohms
H
-m
m/s
A

Figure 4: Parameter selection for the state space model in Simulink

The open loop response of the system is represented by Fig. 5

as under;

Table II. Electrical Design Specifications of Magnetic Levitation System

Component /
Subsystem
Power Supply

Output Voltage: 9V
Output Current: 1 A

Electromagnet

Material: Copper Coil

Thickness: 1mm
Internal Diameter: 17mm
External Diameter: 50mm
Coil Resistance: 1.2
Operating Voltage: 9V
Supply Voltage: 9V
Operating Temperature: 40~85C
Supply Current = 8mA
High-Voltage
Electrical
Isolation: 1.5-kV, 2.5-kV, or
3.55-kV Rating
High-Speed Switching:
tr = 7 ms, tf = 7 ms Typical
Collector current:500mA
Maximum output voltage: 50V
Operating Temperature:-20~85
C

Hall
Effect
Sensor SS49E
Figure 5: Open loop response of magnetic levitation system

V.

EXPERIMENTAL SETUP AND CONSTRUCTION

The hardware implementation of Magnetic Levitation

System is represented by the schematic diagram shown in Fig.
6. The position of the levitated object is sensed using SS49E
Hall-effect sensor. The analog output voltage thus generated is
fed to the dSpace board that further sends this signal to the PC
in the Simulink model. After processing of the signal using the
controller designed above, the control signal is send to the coil
driver IC 4N35 Optocoupler to further drive the amplifier
ULN2003 that controls the current in the electromagnet and
thus the field generated by it. This variable field in turn
controls the position of permanent magnet being levitated in
order to maintain a stable position.

Figure 6: Schematic diagram of magnetic levitation system hardware model

Technical Specifications

Optocoupler
IC 4N35

ULN2003

The design parameters of magnetic levitation have been

presented in table 2. This table presents the specifications of
various electronic components used in the hardware model of
the design. These components include Hall Effect sensor
SS49E, Electromagnet, Optocoupler IC 4N35, ULN2003
amplifier, and power supply. The complete control system
works on 9V.
In the model, the signal from hall effect sensor is first
converted to digital signal in discrete form and sampled at
different rates. Different sampling rates show different results.
Earlier, with lower sampling rates, the system was unstable
but as we increased the sampling rate, the system became
stable at a certain point. So, sampling rate of the ADC had a
great effect on the magnetic levitation system output. A
comparator is used next in the control system that compares
the analog output voltage to a reference voltage. The output is
sent through to the coil driver circuit from the PCI interface
used with the dSpace output port.
Procedure followed in dSpace
Select solver option in configuration parameters and
select start and stop time of about 0s to infinity.
The solver type that is selected is ode3 (bogacki
shampine) and fixed step size is selected.
The sampling time is selected to be 10us. Stability
was attained at this sampling time. When the time
was increased, system became unstable.
After that compiler optimization level is selected in
optimization option of dSpace.

Next, system target file is created in C language and its

name is given and its file is generated in .tlc extension under
code generation option. The experimental setup of magnetic
levitation system is represented in Fig. 8.

approx. 5.7 V as the voltage shown in the plotter is in divide

by 10 mode, so 0.57 represents 5.7 V.

Figure 10: Hall Effect Sensor Output when magnet is taken away from
equilibrium position

So, when the object is taken away from the Hall Effect
sensor, the voltage drops down to about 4.5 V.
Figure 8: Experimental Setup of Magnetic levitation system

VI.

EXPERIMENTAL RESULTS AND DISCUSSION

Fig. 9 shows the variation of hall effect sensor output

voltage in response to position of levitated object when the
Magnetic levitation system is in equilibrium state and Fig. 10
shows the variation of hall effect sensor output voltage in
response to position of levitated object when the magnet is
taken away from the equilibrium position respectively.
The maximum sampling frequency is limited by 500MHz
or 20 ns sampling time.

Figure 11: Variation of Output current in response to the position of

electromagnet

The above graph was obtained at the sampling frequency

of 100 kHz or more.

Figure 9: Hall Effect Sensor Output Voltage in Response to the Position of

Levitated Object at equilibrium position

Above graph shows the output voltage of Hall Effect sensor

when the magnet is in the near to the sensor, i.e. in the stable
position. It is estimated that the voltage of hall sensor is

When the sampling frequency was decreased, the response on

the system was not stable as the response time for the
electromagnet to switch on and off was increased. So, the
sampling frequency was set around 100 kHz for the model.
Finally, the system was stable and could levitate continuously
with the help of control system designed in MATLAB
Simulink and dSpace environment. Results are according to
the desired output. We were able to levitate neodymium
magnet with a fairly stable position with our magnetic
levitation system.

REFERENCES
[1] P. uster and A. Jadlovsk, Modeling and Control Design of
Magnetic Levitation System, SAMI 2012, 10th IEEE Jubilee
International Symposium on Applied Machine Intelligence and
Informatics, January 26-28, 2012 Slovakia.
[2] Ishtiaq Ahmad1, Muhammad Akram Javaid2 ,"Nonlinear Model &
Controller Design for Magnetic Levitation System", RECENT
ADVANCES in Signal Processing, Robotics And Automation,
ISBN:978-960-474-157-1.
[3] Stephen C. Paschall II, Design, "Fabrication and Control of

Single Actuator Magnetic Levitation System", Mechanical

Engineering Department, Texas A&M University, August
2002.
[4] Valer Dolga, Lia Dolga, Modeling and Simulation of a magnetic
levitation system, Annals of the Oradea University. Fascicle of
Management and Technological Engineering, Volume VI (XVI),
2007.
[5] Ugur Hasirci, Abdulkadir Balikci, Zivan Zabar, Senior Member,
IEEE, and Leo Birenbaum, Senior Member, IEEE, A Novel
Magnetic-Levitation System: Design, Implementation, and
Nonlinear Control, IEEE Transactions On Plasma Science,
Vol. 39, No. 1, January .
[6] Hong QIU, Guang-zhong CAO, J.F.PAN Lv-ming LIN, The
Development of Magnetic Levitation Ball Control System Based
on TMS320F2812, 3rd International Conference on Power
Electronics Systems and Applications, 2009.
[7] Norbert C. Cheung, Zhen Gang SUN, Shi Wei ZHAO, WaiChuen
GAN, Design and Simulation of a Testing Fixture for Planar
Magnetic Levitation System Control Using Switched Reluctance
Actuator, 2009 3rd International Conference on Power
Electronics Systems and Applications.
[8] J J. Raygoza-Panduro, Susana Ortega-Cisneros, J Rivera, A. de la
Mora, Design of a mathematical unit in fpgas for the
implementation of the control of a system of magnetic
levitation, 978-1-4244-1992-0/08/C2008 IEEE.
[9] Katie A. Lilienkamp and Kent Lundberg, Low-cost magnetic
levitation project kits for teaching feedback system design,
American Control Conference, 2004.

[10] F. Krach, B. Frackelton, J. Carletta, and R. Veillette, FPGABased Implementation of Digital Control for a Magnetic
Bearing, Proceedings of the American Control Conference,
Denver, Colorado June 4-6.2003.
[11] Chih-Min Lin, Fellow, IEEE, Ming-Hung Lin, Student Member,
IEEE, and Chun-Wen Chen, Student Member, IEEE, SoPCBased Adaptive PID Control System Design for Magnetic
Levitation System, IEEE SYSTEMS JOURNAL, VOL. 5, NO.
2, JUNE 2011.
[12] Do-Kwan Hong, Member IEEE, Ki-Chang Lee, Optimum
Design of Electromagnet in Magnetic Levitation System for
Contactless Delivery Application Using Response Surface
Methodology, Proceedings of the 2008 International
Conference on Electrical Machines.
[13] Jeng-Dao Lee1, Member, IEEE, Xuan-Ren Chen2 and RongYong Duan3, Mechatronics and Control System Design for

Magnetic-Levitation Platform, 2011 IEEE International

Conference on Control Applications (CCA), Denver, CO, USA.
September 28-30, 2011.
[14] Trent M. Guess and David G. Alciatore, "Model Development
and control Implementation for a magnetic levitation
Apparatus", Proceedings of ASME Computers in Engineering
Conference, Boston, September 1995, U.S.A.
[15] Wong T. "Design of a magnetic levitation system - an
undergraduate project", IEEE Transactions on Education, 29,
1986, 196-200.