An active way linear synchronous motor with multiphase independent supply
ABSTRACT In this paper an active way linear synchronous motor with multiphase independent supply is presented. The paper is mainly focused on the power electronics and control structure. The whole system, motor, power electronics and control has been built and successfully tested. The obtained results show that the chosen multi DSP master-slave structure allows the control of the proposed linear motor with good performances and reasonable costs.
- [Show abstract] [Hide abstract]
ABSTRACT: In this paper, an active way linear synchronous motor with multiphase independent supply is presented. A proof-of-concept prototype of the proposed linear motor has been built, including power electronics and control. The control is based on a multi digital signal processor (multi DSP) master–slave structure. The controller design and the power and control boards have been described. The analytical model of the motor is presented. The analytical results are compared with a finite-element simulation, showing a good agreement. The analytical model has been also used to implement a dynamic model of the whole system in the Matlab/Simulink/Plecs simulation environment. The proof-of-concept prototype has been fully tested with one and two independent gliders. The obtained experimental results show that the chosen multi DSP master–slave structure allows the control of the proposed linear motor with good performances and reasonable costs.IEEE Transactions on Industrial Electronics 01/2012; 59(5):2178-2188. · 6.50 Impact Factor
Φ ΦAbstract – In this paper an active way linear synchronous
motor with multiphase independent supply is presented. The
paper is mainly focused on the power electronics and control
structure. The whole system, motor, power electronics and
control has been built and successfully tested. The obtained
results show that the chosen multi DSP master-slave structure
allows the control of the proposed linear motor with good
performances and reasonable costs.
Index Terms—Digital Signal Processor applications, Linear
Synchronous Motor, Long Stator, Motion Control, Power
Rs phase resistance
Ld phase synchronous inductance
us voltage induced by PM
ui total induced voltage
ub coil voltage
ib coil current
x glider position
v glider speed
KE b.e.m.f constant
KF force constant
HE demand of the industry for linear motors is multiple.
Machines tools, assembly machines, conveyors in
production lines, automatic doors - can be mentioned
among others. People transportation is another very
interesting application field for these systems 
In industrial applications, the permanent magnet (PM)
synchronous linear motor, while still more expensive, is
becoming a viable alternative to a rotative-to-linear
transmission (e.g. belts and pulleys, racks and pinions or
screw systems). Typically the mobile part is active. The
coils, installed on the glider, are supplied by a three-phase
electric drive. The permanent magnets are allocated on the
fixed part (magnetic way), associated to the mechanical
guidance. This configuration has some drawbacks, namely
the presence of the feeding cables which of course affect the
dynamics of the glider, the energetic concentration on the
ΦThis work was supported in part by HES-SO (Switzerland) under Grant
Sagex 20063 .
Tazio Beltrami is with the University of Applied Sciences of Western
Switzerland, Yverdon-les-Bains, 1400 CH, e-mail: tazio.beltrami@heig-
Mauro Carpita is with the University of Applied Sciences of Western
Switzerland, Yverdon-les-Bains, 1400 CH, phone: +41-24-5576305; fax:
+41-24-5576404; e-mail: firstname.lastname@example.org .
David Moser is with the University of Applied Sciences of Western
Switzerland, Yverdon-les-Bains, 1400 CH, e-mail: david.moser@heig-
Serge Gavin is with the University of Applied Sciences of Western
Switzerland, Yverdon-les-Bains, 1400 CH, e-mail: email@example.com.
glider and the presence of magnets on the way, which can
create problems with metallic parts present on the ambient.
In an active way linear motor, the concept is completely
reversed. The active part resides in the fixed part of the
motor and the mobile part is passive. For long distances, e.g.
long stator applications, the carriageway is arranged in
several electrically independent segments, also called
“sectors”, supplied by a three-phase inverter -. The
absence of cables on the moving part, the reduction of the
reactive power, the better cooling of the coils now contained
in the fixed part of the motor, allow an increase of the
overall system performances. Moreover, this structure allows
the control of several independent gliders.
In this paper a different approach to the active way linear
motor is discussed, making use of multiphase independent
supply and thus replacing the sector concept. This approach
can further increase the advantages listed above. The basic
principle of the proposed linear motor with multiphase
independent supply is represented in Fig. 1. Main target of
the paper is to illustrate the chosen power converter and
control structure, showing that a flexible and not expensive
system can be built based on a multi DSP master-slave
III. ACTIVE WAY LINEAR MOTOR WITH MULTIPHASE
A. “Proof of concept” prototype
The key concept of multiphase supply consists in
separately controlling the current supplied to each coil. The
instantaneous value to be imposed on the current of the
suitable coil is calculated from the measured position of each
glider. This structure allows managing at best the force to be
applied to the glider. The amplitude and the form of the
reference current can be optimized off-line in order to obtain
An Active Way Linear Synchronous Motor with
Multiphase Independent Supply
Tazio Beltrami, Mauro Carpita, Member, IEEE, Serge Gavin, David Moser
Fig. 1. Active way linear motor: basic principle
Fig. 2. Proof of concept prototype
XIX International Conference on Electrical Machines - ICEM 2010, Rome
978-1-4244-4175-4/10/$25.00 ©2010 IEEE
best performances and smoothed force.
In order to validate the power and control system, a “proof
of concept” prototype, manufactured by the enterprise Etel
SA (CH), has been made. The prototype, shown in Fig. 2, is
a synchronous ironless motor. The motor is made of 3
mechanical modular sections. On each section there are eight
coils, the distance between each coil is 45 mm. The
prototype sizes are 1080 mm x 200 mm x 50 mm. So, 24
coils must be separately supplied. The ironless choice has
been made in order to avoid problems related to the
saturation of the magnetic circuit of the coils and interaction
between magnets and stator. The reluctance effects are also
Two gliders have been designed for the prototype. On
each glider there are two permanent magnets. This choice,
even though is not the best one for creating a smooth force,
allows testing the system in a more demanding condition.
The glider positions are measured with optical linear
position transducers. The grating is placed on the stator,
while a read head is installed on each glider. Sensorless
techniques  to be applied to this structure are under
B. Equivalent model and Motor Parameters
In Fig. 3 the equivalent electrical schematic of a phase of
the motor is shown. The main parameters used for the
modeling of the motor are shown in Table 1. Those
parameters have been calculated and verified by
measurement on the prototype. The constant KE is the
rapport between the peak to peak value of the induced
voltage and the speed.
MAIN PARAMETERS OF THE MOTOR
coil resistance Rs 7.2 Ω
force constant (average)
C. Theoretical principles - basics
The force required to move the glider depends on the back
e.m.f. induced on the coil by the movement, from the current
injected in the active coils and from speed.
( ) ( )
( ) F x
u x i x
Fig. 4. Measured phase induced voltages (4 phases)
Fig. 5. “Constant” current profile: b.e.m.f and current in a coil, and force
produced on the glider as a function of the glider position.
Fig. 3. Equivalent schematic of a phase of the motor.
Fig. 6. “Reciprocal induced voltage” current profile:
(a) b.e.m.f and current in two coils as a function of the glider position,
(b) Force produced on the glider as a function of the position.
According to (1), by knowing the speed and the b.e.m.f. of
a coil, it is possible to obtain the required force acting on the
current taking into account the position of the glider.
Fig. 4 shows the voltage measured on four contiguous
coils by imposing a constant speed to the glider.
With the chosen distance between coils, and two PM on
the glider, the electromagnetic force is produced by
supplying two suitable coils at the same time. The
independent control of the current of each coil allows to
impose a desired shape to the current reference of the coil.
As a consequence, the force profile applied to the glider can
be optimised. To illustrate the flexibility of the system, two
different strategies are presented here: the simple “constant”
current profile and the “quasi constant force” current profile.
D. “Constant” current profile
This strategy is straigthforward, the current has a constant,
positive or negative value according to the sign of the
b.e.m.f (fig. 5). As it can be seen, the force profile produced
by two adjacent coils has an important ripple superposed to
the average value.
E. “Quasi constant force” current profile
The original idea was to impose a profile of the current
equal to the reciprocal profile of the b.e.m.f. In this way,
according to (1) and the induced voltage described in Fig. 4
the obtained force profile is theoretically constant. However,
in the zone where the voltage is near to zero the required
current should be too high, so the current in these ranges
should be limited at its maximum allowable value. Actually,
to take into account the action of the nearby coils, the current
has been shaped with lower values, as it can be seen on Fig.
6a. As already stated, two coils must be energized at each
glider position, and we can always identify a “first” (i.e. the
closer) and a “second” (i.e. the farther) coil. The current
profile in the second coil is symmetrical with respect to the
current symmetric axis. Only the current profiles in the
“exploited zone” (Fig. 6a) are used. With this strategy, the
behaviour of the force is the envelope of the forces that
would be generated by each couple of coils, as shown in fig.
POWER ELECTRONICS AND CONTROL SYSTEM
The whole assembled system is shown in fig. 8. The master
board and the four slave boards, together with the
corresponding four power boards, are shown. The 24 coils
linear motor prototype with two gliders is shown too.
A. General structure of the control
The control of the system has been entirely designed with
DSP processors. The architecture of the system is of Master-
Slave type, as shown in Fig. 7. The control structure is made
of two nested regulation loops, the external position loop
supplying the reference to the internal coil current loops. The
master processor manages the position loop. The sampling
frequency of the position loop is 8 kHz. Each slave
processor manages the current loops of six coils, with a
switching frequency of 24 kHz. The current of each
individual coil is then separately controlled.
The master deal with the measures of the gliders positions,
calculates the required forces and sends both information
(force and position for each glider) to all the slave processors
through an SPI (Serial Peripheral Interface) bus. The
communication is unidirectional; the slave processors do not
send information back to the master. From the glider
position received by the master, each slave is able to
evaluate if one or more of its controlled coils must be
energized, and from the force reference it can evaluate the
Fig. 7. Master-Slave configuration of the control
Fig. 8. Whole assembled system
current level to be imposed on the coil, according to the off-
line calculated profile.
The master-slave configuration of the control allowed to
successfully solve the problem of the passage of the glider
from a coil controlled by one slave DSP to a coil controlled
by the next slave DSP.
B. Control boards (DSP)
The control board layout is shown in Fig. 9a. The control
boards are based on the Texas Instruments DSP controller
TMS320F2812. The complete system is made of five control
boards, one for the master processor and four for the slave
processors. The master and the slave boards are identical.
The connectors for the optic encoder have been mounted on
the master card only. Each slave control board is connected
to a power board. In order to guarantee a good immunity a
galvanic isolation has been chosen for the SPI bus, making
use of the Analog Devices ADuM 1400 four-channel digital
C. Power boards
The power board layout is shown in fig. 9b. On each board
there are six inverters. The converter topology is a half
bridge with capacitor middle point. The switching
components are MOSFETs (Fairchild Semiconductor
FQD12N20, 200 V, 9 A). Each coil is then supplied by a
single inverter branch, connected to a symmetric ±25VDC
bus. This choice of the power structure has been made in
order to minimize the number of power semiconductors to
be used on the whole system. The maximum DC voltage is ±
60 V with a current of 2 A. The current measurements are
made with shunt resistors and Avago Technologies HCPL
D. Modeling of the system
Fig. 10 shows the block diagrams of the two control loops.
The current loop (Fig. 10a) includes the PI controller
Gci(s), the transfer function of the power converter Gcm(s),
the transfer function of the motor coil Gmot(s) and the
average force constant KFav. The effects of the b.e.m.f are
modeled by the average voltage constant KEav. The position
loop (Fig. 8b) includes the position PID controller Gcx(s), the
closed loop current transfer function Gi(s), the transfer
function of the mechanical part Gmec(s) and an integrator that
allows the calculation of the position from the speed. The
modeling of the mechanics takes into account the mass of
the cart and of the additional mechanical load and the
E. Control parameters calculation
The chosen current controller is a PI. The integral term has
been calculated by compensating the dominant pole of the
system, while the proportional term is calculated by
imposing as usual a phase margin of 60°. Fig. 11 shows the
Bode diagrams of the open loop transfer function of the
current. With reference to the form
Fig. 9. Electronics boards: a) Control board, master b) Power board, six coils
Fig. 10. System block diagram: (a) current loop, (b) position loop, (c) mechanics model.
The position controller is a PID. The derivative
component compensates for the integral behavior of the
system. The controller parameters have been calculated
according to the symmetrical criterion . The value of the
constants Tix and Tdx shown by eq. 6 is equal to the
mechanical time constant of the system. The proportional
term is calculated again in the frequency domain, imposing a
phase margin of 60°. Fig. 12 shows the Bode diagram of the
position open loop transfer function.
With reference to the form
Simulations have been performed using Matlab. The
model of the motor used for the simulation is the theoretical
one, but the actual system saturations due to the maximum
voltage and current available have been taken into account.
The voltage limitation is ± 25 V and the current limitation is
± 2.5 A. A simulated position step is shown in fig. 13
(together with measurements, see section V).
V. TEST RESULTS
A. Current loop test
The current loop has been tested with a nominal current
step response. The simulated and the measured results are
shown in Fig. 14, which shows the good agreement between
the two. The current loop dynamics allow to impose the
current profiles discussed in section I.
B. Position loop test, one glider
Transition from a coil controlled by one slave DSP to a
coil controlled by the next slave DSP can be easily and
smoothly achieved. The position regulation has been
exaustively tested. Fig. 13 shows the simulated and the
measured results with a 0.5 m step. The same behaviour is
shown by the two curves. The regulation time is about 400
ms. The position overshoot is lower than 3%. The maximum
speed of the system is around 1.4 m/s.
C. Position loop test, two gliders
The system has been tested with two gliders. In this case
four coils, two for each glider, are active at the same time.
The test results were successful.
In this paper an active way linear synchronous motor with
multiphase independent supply has been presented. The
whole system, motor, power elecronics and control has been
built and succesfully tested. In particular, the power and
control structure, based on a multi-DSP master-slave
approach, has been tested and validated.
Main advantages in comparison to the sector concept are
the possibilty to impose a suitable current profile on each
coil and a simpler management of several glider.
Next step will be the investigation on sensorless
The proposed control structure could be also used on other
kind of multiphase applications, such as high power
multiphase rotating motors for marine applications, of course
by completely redesigning and adequating the power section.
Fig. 12 Bode plots of the position loop
Fig. 13. Position step of 0.5 m
Fig. 11. Bode plots of the current loop