A DSP Laboratory Platform

for Teaching Power Electronics and Drives

M. F. Rahman, K V. Baburaj and L. Zhong
School of Electrical Engineering
University of New South Wales
Sydney, NSW 2052
email: [email protected]

ABSTRACT: with virtual instrument panels for quantities such as

voltage, current, speed, position, torque, % voltage
Digital signal processors (DSP) have recently been boost, fft and so on. Numerical values of some
incorporated in the experiments for the professional useful quantities such various controller
elective subjects in power electronics and electrical parameters, amplitude and rate limits, filter time
drives at UNSW. These experiments are supported constants, etc. are also displayed in the panel.
with virtual instrument panels for quantities such as These DSPs also produce the control and switching
voltage, current, speed, position, torque, 3'% voltage signals which are used in implementing certain
boost, fft and so on. Numerical values of some types of power electronic converter or drive
useful quantities such various controller parameters, systems. These facilities have been added
limits, etc are also displayed in the panel. These progressively over the last five years and their
DSPs also produce the control and switching signals effectiveness in teaching the subjects have been
which are used in implementing certain types of watched closely.
power electronic converter or drive systems. These
facilities have been added progressively over the last The incorporation of the dsp in the laboratory
five years and their effectiveness in teaching the instruction has allowed the experiments to include
subjects have been watched closely. This paper many complex control and computations to be
describes these facilities at the UNSW and their carried out in real time, The results of these are also
usefulness. shown in graphical and numerical form in real time,
so that students have a more hands-on feel for the
1. INTRODUCTION responses of the power electronic circuit and the
drive system at hand, The overall results is a more
The School of Electrical Engineering at the informative, interesting and satisfying laboratory
University of New South Wales teaches power for the two subjects. The students can now
electronics and electrical drives perhaps to the concentrate more on studying the performance,
largest group of undergraduate students in Australia. while the rigorous, time consuming and quite often
Over the last 5 years, various drive simulation boring tasks of waveform analysis, plotting and
programs have been extensively used in the teaching hardware configuration and set-up are now handled
of power electronics and drives. Most important by the computer. The waveforms are easily plotted
aspects of the circuits/systems under experimental on a printer so that more time can be spent on the
study are modeled in the same way as they are critical examination and evaluation of the circuits
experimented with and some of the results, such as and systems being tested. A few of the dsp based
the fft of some important waveforms and dynamic experiments in power electronics and drives are
and steady-state drive responses are correlated. The described in the following sections.
students are encouraged to use these program to
study the circuidsystem under ' study more 2. SINGLE- AND THREE-PHASE INVERTER
extensively than is possible within a three-hour time CIRCUITS
period of the experiment.
A TMS320C31 dsp board with several analogue
Recently, digital signal processors (DSP) have been and digital input and output channels is interfaced
incorporated in the experiments for the professional to a six-mosfet inverter as indicated in figure 1.
elective subjects in power electronics and electrical This set-up allows single- and three-phase inverter
drives at UNSW. These experiments are supported circuits with different modulation strategies

0-7803-4879-6EJ8/$10.00 0 1998 IEEE 751

 Figure 1. The dsp based inverter experiment.

implemented on the dsp. Switching frequencies, the above variables are adjusted can be observed on
dead times, unipolar or bipolar switching strategies, an oscilloscope and printed out. Alternatively, data
depth of modulation, modulating waveforms,etc, can on these can be analysed on the PC using Matlab. A
be easily adjusted. The output waveforms. .of few sample waveforms and fft data are indicated in
voltage, current and their harmonic content as any of figure 2.
m 2z.mv T2.01. 5.m fz i
,. . .. ... ... ? : . .

Figure 2(a) 1-$ SPWM inverter output voltage (b) fft of the voltage waveform of figure 2(a)
&current, Bipolar switched, & = 1 kHz,
m = 0.8, Vk =200v,Td = 1opseC
xm 2 5 . m v 2.-
; ; , : . P.W . , . . .
~

. ~

. . . ..........,.........
.
I
.
:
:
:
:
:
:
:
:
......... / . . . . . I . . . / . ....... ....... .... ..:.
~ ~ . .......................
... ~

... ... I : : : :
. . . .....
... ......,.
.. ,.. .. ......,....
.
..
.

. . . .. .. ..

. . .
... .. .. .. . . .. ... ...
. .. .. .. 7 .
. . .
I
*,<,a= arc .
,
e - '"re
I

e<+
. .

b,<FZ> i f"r.2 0,.

Figure 2(c) I-$ SPWM inverter output voltage Figure 2 (d) 1-@SPWM inverter output voltage
& CUtTent, Td = 10 WeC & CUITent, Td = 30 PE.

752
 .. .. .. .. .. .. .. ..

,. .+..........

Figure 2(e) Voltage and current waveform Figure 2(fl fft of the voltage waveform

.. .. ..

. . . . ..., .... ..,...... .,............,.. . .

. .. .. ..

.. .. ..
. . .

Figure 2(g) Voltage & current waveforms of Figure 2(h) fft of the voltage waveform of
a 3-41 sine-wave inverter,& = 1 IdIz,m = 0.8 figure 2(g)
Td= 10 usec, Vdc= 200V

DSP BOARD INTERFACE

U1 TMS32OC3i DSP B A S E 0 CONTROLLER

Figure 3. Experimental set-up for the dc drive.

753
3. A FOUR QUADRANT DC DRIVE (i) Setting up the cyrrent loop with current
feedback from an hall effect transducer
This experiment consists of a separately excited compared against a dc reference. The motor is
brushed dc motor driven from a Mosfet, four quadrant kept unexcited and the dc voltage to the H-
H-bridge. The drive is loaded with a permanent bridge is gradually increased to the full rated
magnet brushed dc generator. Both machines are rated voltage of the motor (=200V)
for 180V. 6A operation. The drive is fitted with (ii) A study of the step responses of the current
isolated voltage and current sensors and an loop with various types of controllers and their
incremental encoder of a resolution of 2048 pulsedrev proportional gains and integration time
on the shaft. These are interfaced to the DSP board constants. The dc input voltage to the H-
mentioned in section 2. The encoder pulses are used bridge, the pwm switching frequency and the
in calculating the speed and absolute position of the filter network parameters are also adjusted.
rotor. The motor is driven from a pwm current source, The various current controllers, filter networks
the reference for which is derived from the speed are all implemented in the dsp. Their
controller as indicated in figure 4. The dsp is also parameters, such as the gain, integration times
capable of producing the switching signals for the H- etc are adjusted through cursors in the cockpit
bridge switches. A variable switching frequency which is displayed on the PC monitor. The
modulator followed by a dead-time protector is also reference signals for the step and frequency
available in hardware. responses are also obtained from the dsp.
(iii) A frequency response test on the best current-
This experiment is designed to familiarise students controller found in (ii).
with control loops typically encountered in industrial
drives. The DSP board implements the current and In the second stage, the speed loop is implemented
speed controllers. These are PID type controllers, with with the current loop designed previously retained
the P-type being studied first. The experiment consists as the inner loop of the drive. This stage consists of
of the design of the current loop with the motor field the following:
unexcited, followed by the speed loop and
,

optimisation of each controller through, e x a h a t i o n (iv) Setting up of the speed loop with an amplitude
of the drive dynamic and steady-state responses. Full limit on the output of the speed controller.
frequency response tests are also carried out on the This limit acts as the current limiter.
current loop in order to determine the currendtorque
control bandwidth of the drive. The controller gain (v) Study of step responses of the drive for various
and integration parameters, filter network parameters, controller parameters. Steady-state offest with
supply voltage, load, etc, are easily adjusted through load and PI type controllers are investigated.
the DSP instrument panel (the cockpit) and the Motor accelerating and braking currents are
hardware facilities. identified and correlated with controller action.
The inner current loop design studies are followed by (vi) Removal of regenerating braking current by
the speed loop, where the speed signal is derived from disabling half of the H-bridge and
the pulses of the incremental encoder. An analogue demonstrating its effect on the drive response.
tachc-generator feedback signal is available which
helps to demonstrate the influence of filtering of the
noisy tacho-generator signal on the performance of
the drive. The speed loop controller gain and
integration parameters are easily adjusted through the
dsp cockpit, as in the case of the inner current loop.

3.1 EXPERIMENTAL WORKS ON THE DC

DRIVE
I I

The experimental work consists of two stages; the Figure 4. Qnamc response of a
first stage is for the inner current loop which consists non o p t i m i x d c m t l q
of the following:

154
 4. A THREE-PHASE V-f INDUCTION
MOTOR DRIVE
In this experiment a three-phase induction motor
is driven from a voltage source variable-voltage,
variable frequency inverter as indicated in figure
6.. The motor is rated for 415V, 7.8A and it’s
loading dc generator has equivalent rating. The
dsp produces the switching signals for the inverter
and its output rms voltage and frequency are
determined by the dsp according to the
programmable % voltage boost at low frequency.
figure5. Cbwd Imp speedadomm1wl
s n The motor is fitted with an encoder using which
speed is calculated for display and for control.

Figure 6. Experimental set-up for a V-f induction motor drive

4.1 EXPERIMENTAL WORKS ON THE AC of 2048 counts/rev. The switching frequency

DRIVE can be varied to indicate the variation of motor
current ripples. The input voltage to the
The experimental work consists of the following: inverter is set at 570Vdc.
(ii) The induction motor is driven at a few speeds
(i) Setting up of a sinusoidal pwm IGBT inverter (frequencies) and for each speed, the motor is
with variable voltage and variable frequency loaded through a dc generator in order to.
output. The output frequency and %boost obtain the motor torque-speed characteristic at
voltage are adjustable by cursors in the dsp each speed. Both stator resistance and low
cockpit. The induction motor position and frequency %voltage boost are adjusted in order
speed are obtained from an incremental encoder to demonstrate the need for the boost.

755
(i)The airgap and rotor fluxes and the stator 4. LABORATORY EVALUATION
current amplitudes are noted when the motor is
accelerated and decelerated. This is used as a The introduction of the dsp in the set of
prelude to other control methods such as vwtor experiments described above have very
control. significantly improved the quality of the laboratory
time for students of the professional elective
The above procedures familiarise the students with subjects on power electronics and electrical drive.
the steady-state torque-speed performance of an This is evidenced by rather elaborate and elegant
induction motor drive under a v-f drive. The reports the students submit, Also, the students are
dynamic performance under such a drive is also able to interact with the experiments more and
investigated briefly. Figures 7 show some of the analyse the experimental results in much greater
experimental results. detail. The dsp in each case minimises the hardware
to a minimum, even dispensing with many of the
Toque Sped CharectefiW
meters. Laboratory set-up time and hardware
troubleshooting are quickly becoming a rarity.
Iok
5. CONCLUSION

The paper has briefly presented the outlines of

several dsp based experimental facilities at the
UNSW. These have already proved to be very
popular and effective teaching tools for the
teaching the complexities of subjects in power
electronics and drives. Complicated waveform
analysis, control and hardcopy output for each case
are very easily obtained.

REFERENCES:

1. Matlab/Simulink User Guide 1997

5w 1wO 2. Laboratory Manuals, EE. UNSW, 1998
Sped in RPM

756