Вы находитесь на странице: 1из 384

The Control Techniques

Drives and Controls


Handbook

Bill Drury

The Institution of Electrical Engineers


Published by: The Institution of Electrical Engineers, London,
United Kingdom

2001: The Institution of Electrical Engineers

This publication is copyright under the Berne Convention and the


Universal Copyright Convention. All rights reserved. Apart from any fair
dealing for the purposes of research or private study, or criticism or
review, as permitted under the Copyright, Designs and Patents Act, 1988,
this publication may be reproduced, stored or transmitted, in any forms or
by any means, only with the prior permission in writing of the publishers,
or in the case of reprographic reproduction in accordance with the terms
of licences issued by the Copyright Licensing Agency. Inquiries
concerning reproduction outside those terms should be sent to the
publishers at the undermentioned address:

The Institution of Electrical Engineers,


Michael Faraday House,
Six Hills Way, Stevenage,
Herts. SG1 2AY, United Kingdom

While the author and the publishers believe that the information and
guidance given in this work are correct, all parties must rely upon their
own skill and judgment when making use of them. Neither the author
nor the publishers assume any liability to anyone for any loss or damage
caused by any error or omission in the work, whether such error or
omission is the result of negligence or any other cause. Any and all such
liability is disclaimed.

The moral right of the author to be identified as author of this work has
been asserted by him in accordance with the Copyright, Designs and
Patents Act 1988.

British Library Cataloguing in Publication Data

Control Techniques drives and controls handbook.


(lEE power series; no. 35)
1. Electric motors 2. Electric controllers 3. Electric driving
I. Drury, W. I1. Control Techniques Drives PIc
621.4' 6

ISBN 0 85296 793 4

Typeset by Newgen Imaging Systems, India


Printed in England by Cambridge University Press, Cambridge
Preface

Variable-speed drives remain a key component of the boom


in all aspects of automation and energy saving which is
becoming of ever greater importance throughout the world.
The words of Harry Ward Leonard first uttered on
18 November 1896 in his paper entitled 'Volts versus
o h m s - s p e e d regulation of electric motors' still hold true:
'The operation by means of electric motors of elevators,
locomotives, printing presses, travelling cranes, turrets on
men-of-war, pumps, ventilating fans, air compressors,
horseless vehicles and many other electric motor applica-
tions too numerous to mention in detail, all involve the des-
irability of operating an electric motor under perfect and
economical control at any desired speed from rest to full
speed'.
It can, and should, be argued that electrical variable-speed
drives have facilitated the automation revolution. They have,
like so many enabling technologies, developed rapidly, fuel-
led by their success, stretched by demands never dreamed
possible a generation earlier. The development cycle of
drives products is now such that product ranges have
expected lifetimes of only three to five years - a problem in Figure P. 1 Michael Faraday (1791-1867)
itself to many OEM customers whose own products have a
much longer design life.
The world of variable-speed drives is an exciting and rapidly
moving one. To predict the future and the pace of devel-
1824 Arago discovered that if a copper disc is rotated
opment is difficult. A historical perspective is helpful and,
rapidly beneath a suspended magnet, then the magnet
for those who need any convincing, shows how quickly
also rotates in the same direction as the disc.
things are moving:
1825 Babbage and Herschel demonstrated the inversion of
Arago's experiment by rotating a magnet beneath a
1820 Oersted was the first to note that a compass needle is pivoted disc causing the disc to rotate. This was truly
deflected when an electric current is applied to a wire induced rotation and just a simple step away from the
close to the compass - the fundamental principle of first induction motor, a step which was not taken for
an electric motor. half a century.
1821 Faraday built two devices to produce what he called 1831 Using an induction ring, Faraday made one of his
electromagnetic rotation: that is a continuous circular greatest discoveries - electromagnetic induction: the
motion from the circular magnetic force around a induction of electricity i n a wire by means of the
wire. This was the initial stage of his pioneering work. electromagnetic effect of a current in another wire.
xiv Preface

The induction ring was the first electric transformer. The speed of these motors depends fundamentally
In a second series of experiments in the same year he upon pole number and supply frequency. Rotor
discovered magneto-electric induction: the produc- resistance control for the slip-ring motor was intro-
tion of a steady electric current. To do this, Faraday duced immediately, but this is equivalent to armature
attached two wires through a sliding contact to a resistance control of a D.C. machine and is inher-
copper disc, the first commutator, an approach sug- ently inefficient.
gested to him by Amp+re. By rotating the disc
between the poles of a horseshoe magnet he obtained By 1890 there was a well established D.C. motor, D.C.
a continuous direct current. This was the first gen- central generating stations, three-phase A.C. generation and
erator. Faraday's scientific work laid the foundations a simple three-phase motor with enormous potential but
of all subsequent electro-technology. From his which was inherently a single-speed machine. There was as
experiments came devices which led directly to the yet no way of efficiently controlling the speed of a motor
modem electric motor, generator and transformer. over the full range, from zero to top speed.
1832 Pixii produced the first magneto-electric machine.
1838 Lenz discovered that a D.C. generator could be used
equally well as a motor. Jacobi used a battery-fed V O L T S VS. OttMS.
D.C. motor to propel a boat on the River Neva. SPEED ]{EOULA'rION OF ]~r.ECTRIC ~fOTC,R~,
Interestingly, Jacobi himself pointed out that bat-
teries were inadequate for propulsion- a problem BY ~I. WARD LEO.N'ARD.
which is still being worked on today.
1845 Wheatstone and Cooke patented the use of electro-
magnets instead of permanent magnets for the field
L
system of the dynamo. Over twenty years were to
elapse before the principle of self excitation was to i
be established by Wilde, Wheatstone, Varley and the t C
Siemens brothers.
1870 Gramme introduced a ring armature somewhat more
advanced than that proposed by Pacinotte in 1860,
which led to the multibar commutator and the
modem D.C. machine.
1873 Gramme demonstrated, at the Vienna Exhibition, the
use of one machine as a generator supplying power Figure P.2 1l Oth meeting of the American Institute of
Electrical Engineers, New York, 18 November
over a distance of 1 km to drive a similar machine as
1896
a motor. This simple experiment did a great deal to
establish the credibility of the D.C. motor.
1879 Bailey developed a motor in which he replaced the
rotating magnet of Babbage and Herschel by a 1896 The work of Ward Leonard clearly marks the birth of
rotating magnetic field, produced by switching of efficient, wide-range, electrical variable-speed
direct current at appropriately staggered intervals to drives. The system he proposed was of course based
four pole pieces. With its rotation induced by a upon the inherently variable-speed D.C. machine
rotating magnetic field it was thus the first commu- (which had hitherto been controlled by variable
tatorless induction motor. armature resistors). His work was not universally
1885 Ferraris produced a motor in which a rotating mag- accepted at the time and attracted much criticism,
netic field was established by passing single-phase understandably, as it required three machines of
alternating current through windings in space quad- similar rating to do the job of one. Today, however,
rature. This was the first alternating-current com- all D.C. drives are based upon his control philoso-
mutatorless induction motor, a single-phase machine phy, with only the implementation changing from
which Dobrowolsky later acknowledged as the multimotor schemes through the era of grid-con-
inspiration for his polyphase machine. trolled mercury-arc rectifiers to thyristors and, more
1886 Tesla developed the first polyphase induction motor. re-cently, in demanding dynamic applications, to bi-
He deliberately generated four-phase polyphase polar transistors, field-effect transistors (FETs),
currents and supplied them to a machine which had a insulated-gate bipolar transistors (IGBTs)...
four-phase stator. He used several types of rotor, 1904 Kramer made the first significant move with respect
including one with a soft-iron salient-pole construc- to frequency changing in 1904 by introducing a
tion - a reluctance m o t o r - and one with two D.C. link between the slip rings and the A.C. supply.
short-circuited windings in space quadrature - the This involved the use of two A.C. ~ D.C. motor
polyphase induction motor. sets. The D.C. link was later to become a familiar
1889 Dobrowolsky, working independently from Tesla, sight in many A.C. drive technologies. Subsequent
introduced the three-phase squirrel-cage induction advances in A.C. motor speed control were based
motor. upon purely electrical means of frequency and vol-
1890 Dobrowolsky introduced a three-phase induction tage conversion. Progress has followed the advances
motor with a polyphase slip-ring rotor into which in the field of semiconductors (power and signal/
resistors could be connected for starting and control. control).
Preface xv

industries such as textiles where a single (bulk)


E inverter was used to feed large numbers of induction
motors (or reluctance motors, despite their low
8
power factor, where synchronisation was required).
1963 Gain-bandwidth relationships of power converters
investigated.
1970 The 1970s saw a new and very significant revolution
2ti-
hit the variable-speed drives market - packaging. Up
until this time the static variable-speed drive design
process had essentially concentrated on perfor-
A A$~rnehronnlotor, l~,i~enerrofulbG~ mance/functionality. Both A.C. and D.C. drives of
B Otoichs~,rom motor. H B'aliptstromerrefd n~.
C F.in,;lkerumform~r, even low rating were broadly speaking custom built/
Ahb. !1. Abb. 12.
hand crafted. This approach resulted in bulky, high-
cost drives the very uniqueness of which often
compromised reliability and meant service support
Neue Methoden zur Regelung yon Asynchron- was difficult. The drives industry was not fulfilling
motoren und ihre Anwendung ffir verschie-
its potential.
dene Zwecke. ~)
1970s A.C. motor drives had made great advances in terms of
Vou Ch. Krilmer.
performance but still lacked the dynamic performance
to really challenge the D.C. drive in demanding pro-
cess applications. Since the early 1970s considerable
Figure P.3 Elektrotechnische Zeitschrift, volume 31, 30 interest was being generated in the field-oriented
July 1908 control of A.C. machines. This technique pioneered
by Blaschke and further developed by Leonhard
1911 Schrage introduced a system based upon an induc- opened up the opportunity for A.C. drives not only to
tion motor with a commutator on the rotor. This match the performance of a D.C. drive but to improve
machine proved to be very popular, requiring no upon it. The processing requirements were such that
auxiliary machines, and was very reliable. It found in its early days commercial exploitation was restric-
large markets particularly in the textile industry and ted to large drives such as mill motor drives, boiler-
some other niche applications, and is still sold today feed pump drives. Siemens was very much in the fore-
although in rapidly reducing numbers. front of commercialising field orientation and was
1923 The introduction of the ignitron made controlled also rationalising the numerous alternative drive topo-
rectification possible. The thyratron and grid-con- logies which had proliferated and, although stimu-
trolled mercury rectifiers made life easier in 1928. lating to the academic, were confusing to drive users:
This made possible the direct control of voltage
D.C. drives
applied to the armature of a D.C. machine so as to
single converter
apply the philosophy of Ward Leonard control
double converter
without additional machines.
1930 The ideas of inversion (D.C. to variable frequency/ - circulating current free
voltage A.C. which is the basis for the present day circulating current
-

inverter) had been established, the use of forced A.C. drives


commutation by means of switched capacitors was voltage (phase) control
introduced. voltage-fed inverters
1931 Direct A.C. to A.C. conversion by means of cyclo- quasi square V/f
-

converters was introduced for the railway service. - quasi square V/f with D.C. link chopper
1932 Nyquist stability criterion developed. - pulse-width modulated (PWM)
1938 Bode stability criterion developed. current-source inverters
1950 The introduction of silicon into power switches - induction motor
replacing the bulky and relatively inefficient mer- - synchronous machine
cury-arc rectifiers (MAR). By 1960 thyristors
static Kramer drive
(SCRs) had become available and the key enabling
cycloconverter
technology for drives had arrived. D.C. drives and
cycloconverters quickly embraced the new silicon 1972 Siemens launched the SIMOPAC integrated motor
technology, at first using techniques with origins in with ratings up to 70 kW. This was a D.C. motor with
the MAR forerunners. The faster switching perfor- integrated converter including line reactors!
mance of the new silicon, however, opened many 1973 A new approach to drives in terms of packaging.
new doors notably in the field of forced commu- Utilising 19-inch rack principles, a cubicle mounting
tation-the way was clear for commercial variable- standard well used in the process industry, compact,
frequency inverters. high-specification ranges of D.C. drives in modular
1957 Back to back reversing D.C. drive introduced. form became available off the shelf. Companies
1960s Power semiconductor voltage and current ratings such as AEG, Thorn Automation, Mawdsley's and
grow and performance characteristics improve. Control Techniques pioneered this work. A new era
Inverters became commercially viable, notably in of drive design had started.
xvi Preface

Figure P.4 D.C. drive module (Control Techniques)

1979 Further advances in packaging design were made


possible by the introduction of isolated thyristor
packages.
1983 In 1983 plastic mouldings made their first significant
impact in drives. Bipolar transistor technology also
arrived, which eliminated bulky auxiliary commu-
tation circuits.
1985 Takahashi and Noguchi published a paper on direct
torque control (DTC) in the IEEE. (This date is Figure P.6 Digital D.C. drive with microprocessor and ASIC
included not because of its technical significance (Control Techniques)
rather as a point of interest as DTC has received
much attention recently.) 1986 Great advances were being made at this time in the
field of microprocessors making possible cost-
effective digital drives at low powers. Further drives
were introduced containing application-specific
integrated circuits (ASIC), which up to that time had
only been used in exceptionally large volume/
domestic applications. Further, new plastic materials
were introduced which gave structural strength,
weight, size, assembly and cost advantage.
1988 IGBT technology was introduced to the drives
market. IGBTs heralded the era of relatively quiet
variable-speed drives (and introduced a few pro-
blems, some of which have led to substantial aca-
demic activity and a very few of which have required
more pragmatic treatment).
1989 The first implementation of the field-orientation or
flux-vector drive was introduced to the high-volume,
lower power market. It found immediate application
in machine tool spindle drives and has grown rapidly
in application (and rating) since. It should be said
that the name vector has been prostituted by some in
the drives industry with voltage vector and other
such names/techniques, causing confusion and frus-
tration to customers.
1990 The trend to smaller drive products which were also
simpler to design was given a significant boost by
Mitsubishi which introduced intelligent power
modules, integrating into the semiconductor package
necessary gate drive and protection functions.
1992 A new packaging trend e m e r g e d - the book-
Figure P.5 Plastic mouldings introduced into drives form shape which had previously been applied to
(Control Techniques) servo drives was now being applied to the broader
Preface xvii

i! i i !i !~i!~i~'i!!i!i!i~i !!ii ~i ~i~!i ~'

iii~?'.....

i~i~ ~i~!~!~ ~?~iii!i~!~i

~ !iI. . . . . ! i,~I ~ . . . . ~ ~.... ~

ii i i ~ ~i i ~ !!~i,~i~' ~iill iii~~:i~ii!~ i!~!!: ~ i~ ~ I~~I~ ~ i~~I~I~I

Figure P.7 I G B T p o w e r stage in an A.C. drive (Control


Techniques) F i g u r e P.9 B o o k f o r m shape of drive (Control Techniques)

industrial A.C. drives market. The trend continues


today but there is not a consensus that this is the most
suitable shape for all market segments.
1993 Another innovation in packaging - at the low-power
end of the spectrum when a DIN rail mounting
0.4 kW inverter package, similar to that used widely
in equipment such as contactors and control relays,
was launched. The first drive with a built-in supply
side filter fully compliant with, the then impending,
EU regulations on conducted EMC was introduced.
1996 The first truly universal drive was launched which
met the diverse requirements of a general purpose
open-loop vector drive, a closed-loop flux-vector
drive, a servo drive and a sinusoidal supply converter
with the selection purely by parameter selection.
1998 The integrated D.C. motor launched in 1972 was not
a great commercial success - much has been learned
since those days. In 1998 integrated A.C. motor
drives were introduced onto the market. These pro-
ducts are, for the most part, open-loop inverter-
driven induction motors and were initially targeted
on replacing mechanical variable-speed drives.
2000 A radical servo drive was introduced with the posi-
tion and speed loop embedded in the encoder hous-
ing on the motor itself. This brought with it the
advantage of processing the position information
close to the source thereby avoiding problems
of noise etc., and allowed dramatic improvements
i~,~ii i~,~,+~i>~,~
ii~iiii~ii~i~!ii '~' ~ in control resolution, stiffness of the drive and
reduced the number of wires between the drive and
F i g u r e P.8 V e c t o r d r i v e ( C o n t r o l Techniques) the motor.
xviii Preface

Figure P. 12 Integrated A.C. motor (courtesy Leroy Somer)

A review of the time lines presented above illustrates that


Figure P. 10 DIN rail mounting drive with built-in EMC development within the drives industry continues at an ever
filter (Control Techniques) increasing pace. Fundamental changes in the product, from a
customer perspective, are still emerging, accessing ever
more applications driven by automation and quality.

This book covers the present state of development, or rather


commercial exploitation, of industrial A.C. and D.C. vari-
able-speed drives and associated systems. It is intended
primarily for the use of professional engineers who specify
or design systems which incorporate variable-speed drives.
The theory of both the driven motor and the drive is
explained in practical terms, with reference to fundamental
theory being made only where necessary. Information on
how to apply drive systems is included, as are examples of
what is available within commercially offered drives and
indications of what can be achieved using them. Emphasis
is placed on industrial drives in the range 0.37 kW to 1 MW.
The practical emphasis of the book has led to two unfor-
tunate but I fear unavoidable consequences. First, some of
the theory behind the technology contained in the book has
had to be omitted or abridged in the interests of simplicity
and volume. Second, in such a practical book it has proved
difficult to avoid reference to proprietary equipment. In
such circumstances a tendency towards referencing the
products of Control Techniques is inevitable. It should be
clear to readers that these products are described for illu-
mination and explanation of the technology. The
lEE, publisher of this book, does not endorse these products
or their use in any way.

This edition of the Control Techniques Drives and Con-


trols Handbook has been created with contributions from
engineers both within Control Techniques itself as well as
Figure P. 11 Universal A.C. drive modules (Control Tech- sister companies within the family that is Emerson. I would
niques) in particular like to thank Dr Pete Barrass, Ray Brister,
Preface xix

Figure P. 13 Speed loop motor (Control Techniques)

Dr Mike Cade, Vikas Desai, Dr Colin Hargis, Jim Lynch,


John Orrells, Bleddyn Powell, Alex Rothwell, Michael
Turner and Peter Worland.

Prof Bill Drury


Contents

Preface xiii-xix Relationships between torque and back e.m.f.


constant 19
Stationary torque characteristics 19
I Industrial motors
Construction of brushless servomotors 20
1 D.C. motors 1 Stator structure 20
General 1 Rotor structure 20
Fundamental equations and performance 2 5 Reluctance motors 21
Wound-field motors 2 6 A.C. commutator motors 22
Permanent-magnet motors 3 7 Mechanical and environmental 22
Operating principles 4 Mounting of the motor 22
Commutation 4 General 22
Rotation 5 IEC 60034-7 standard enclosures 22
Compensation 5 NEMA standard enclosures ~ 22
D.C. PM commutation 6 Degree of protection 24
Construction of the D.C. motor 6 General 24
D.C. motor frame 6 IEC 60034-5 24
D.C. motor armature 6 US practice 25
Brush gear 7 Cooling 26
Mountings 7 General 26
D.C. PM design 7 Air filters 27
Rotor inertia 8 Duty cycles 27
Permanent-magnet materials 8 Continuous duty - S 1 28
2 A.C. induction motors 8 Short-time duty- $2 28
General 8 Intermittent d u t y - $3 28
Fundamental equations and performance 9 Intermittent duty with starting- $4 29
Electrical characteristics of induction motors 9 Intermittent duty with starting and electric
Torque characteristics 10 braking- $5 29
Voltage-frequency relationship 11 Continuous operation periodic d u t y - $6 29
Increased voltage 12 Continuous operation periodic duty with
Reduced frequency 12 electric braking- $7 29
Slip-ring induction motor 13 Continuous operation periodic duty with
Speed-changing motors 13 related load speed changes- $8 30
A.C. induction motor construction 14 Duty with nonperiodic load and speed
3 A.C. synchronous motors 15 variations- $9 30
General 15 Duty with discrete constant loads - S 10 30
A.C. synchronous motor construction 17 Terminal markings and direction of rotation 30
4 Brushless servomotors 17 General 30
General 17 IEC 60034-8 30
Principles of operation of brushless servomotors 18 NEMA 32
Introduction 18 Ambient conditions 33
Torque constant 18 Introduction 33
vi Contents

Temperature 33 MOS-controlled thyristor 66


Altitude 34 MOS tum-off thyristor 66
Power supply system 34 Silicon carbide 67
Noise and vibration 34 Power device packaging 67
General 34 Pressure contact packages 67
Vibration 34 Large wire-bonded packages for power modules 67
Noise 35 Small wire-bonded packages for discrete devices 69
Motors for special applications 37 Applications 70
Geared motors 37 2 Drive converter circuits 72
Brake motors 37 A.C. to D.C. power conversion 72
Torque motors 37 General 72
Motors for hazardous locations 37 Converters for connection to a single-phase
General 37 supply 73
CENELEC 37 Converters for connection to a three-phase
North American standards 39 supply 74
Testing authorities 39 Voltage tipple characteristics 76
8 Effects of semiconductor power converters 40 Practical effects 76
General 40 D.C. motor drive systems 76
Drive converter effects upon D.C. machines 40 D.C. to D.C. power conversion 79
Drive converter effects upon A.C. machines 40 General 79
Introduction 40 Step-down D.C.-D.C. converters 79
Machine rating - thermal effects 40 Step-up D.C.-D.C. converters 81
Machine insulation 40 A.C. to A.C. power converters with intermediate
Beating currents 46 D.C. link 81
Overspeed 48 General 81
Motors for hazardous locations 48 Voltage source inverters 81
Current source inverters 83
2 Power electronics 51 Direct A.C. to A.C. power converters 85
General 85
1 Power semiconductor devices 51 Soft starter/voltage regulator 85
General 51 Cycloconverter 86
Diode rectifier 51 Static Scherbius drive 86
Thyristor 52 Matrix converter 87
Thyristor gating requirements 54
Power losses and current ratings 54
Surge current ratings
3 Speed and position feedback devices 89
55
High-frequency current operation 56 1 D.C. tachometer generator 90
Gate turn-off thyristor 56 General 90
Switching characteristics and gate drive 57 Output voltage ripple 90
Snubber design 57 Temperature effects 91
Voltage and current ratings 57 Linearity and load effects 91
Bipolar Transistor 57 Stability of the output 91
Voltage ratings 58 Maximum terminal voltage 91
Current ratings 58 Maximum operating speed 91
Switching characteristic and base drive 59 Mechanical construction 92
Safe operating areas 60 2 A.C. tachometer generator 92
Short-circuit performance 61 3 Resolver 92
MOSFET 61 Design principles 93
Voltage and current ratings 62 Synchros 93
Switching performance 62 Torque synchros 93
Safe operating area 62 Control synchro 94
Parasitic diode 62 Resolver 94
Insulated-gate bipolar transistor 63 General 94
Operation 63 Computing resolvers 94
Voltage and current ratings 63 Phase shifting 95
Switching behaviour and gate drive 63 Brushless resolvers 95
Safe operating area 64 Multipole resolvers 95
Short-circuit performance 64 A.C. rotary pickoffs 95
Series and parallel operation 64 Resolver-to-digital conversion 96
Integrated-gate commutated thyristor 65 4 Encoder 97
Voltage and current ratings 65 Incremental encoder 97
Switching behaviour and gate drive 65 Absolute encoder 98
Other power devices and materials 66 Sin/cos encoder 99
Contents vii

5 Selection of a feedback device for a Polyphase switched-reluctance machines 127


drive system 99 Losses in the switched-reluctance motor 127
6 Mechanical considerations 100 Excitation frequency 128
7 Glossary of terms 101 Power electronics for the switched-
reluctance motor 128
Power supply and front-end bridge 128
4 Drive control 103
Power switching stage 128
1 General 103 Single-switch-per-phase circuits 128
The ideal control system 103 Multiple-phase operation 129
Open-loop control 103 Single-switch circuit using bifilar winding 129
Closed-loop control 104 Two-switch asymmetrical bridge 130
Criteria for assessing the performance of a Advantages of the switched-reluctance system 131
closed-loop control system 104 Rotor construction 131
2 A.C. motor drive control 105 Stator construction 131
General-purpose open-loop A.C. drive 106 Electronics and system-level benefits 131
Space-vector modulator and inverter 106 Disadvantages of the switched-reluctance system 132
Reference-frame translation 107 Torque tipple 133
Reference-frame generation 107 Acoustic noise 133
Current limit 108 3 Stepper-motor drives 134
Performance and applications 108 Stepping-motor principles 134
Permanent-magnet servodrive 109 The permanent-magnet motor 135
Reference-frame generation 109 The VR motor 135
Current control 109 The hybrid motor 135
Speed control 110 Stepping-motor drive circuits (logic modes) 136
Performance and applications 110 Unipolar switching 137
Closed-loop induction motor drive 110 Bipolar switching 137
Flux calculator and reference-frame generation 111 High-speed stepping- L/R drives 139
Flux control 111 Chopper drives 139
Performance and applications 111 Bilevel drives 139
Operation without position feedback 111 Application notes 139
Four-quadrant operation 112 Effect of inertia 139
Reference-frame generation 112 Resonance 139
Performance and applications 112 Special products 140
Direct torque control 113 Stepper/encoders 140
3 D.C. motor drive control 114 Space-rated steppers 140
Flux controller 114 Fuel-control actuators 140
Torque controller 115
Performance and applications 116
6 Practical drives 141
4 Analysis of and set up of a speed controller 116
Ideal speed controller 116 1 General 141
Calculating the required gains 117 Digital input 143
Nonideal effects in a real speed controller 119 Programmability 143
Typical specification 143
5 Switched-reluctance and stepper-motor Digital output 143
drives 121 Programmability 143
Typical specification 143
1 General 121 Analogue input 143
2 Switched-reluctance motors and controllers 122 Programmability 143
Basic principle of the switched-reluctance motor 122 Typical specification 144
Operation as a motor 123 Analogue output 144
Operation as a brake or generator 123 Programmability 144
To summarise so far 123 Typical specification 144
Relationship between torque polarity and 2 D.C. drives 145
motoring/generating 124 The technology 145
Control of the machine in practice 124 Drive selection 145
Low-speed operation 124 Technical specifications and ratings 148
What happens as speed is increased? 124 Drive set up and commissioning 148
Medium-speed operation 124 Autotune 148
How is performance maintained as speed PC-based commissioning tools- MentorSoft 148
increases? 125 Performance 148
High-speed operation 125 Speed and current-loop response 148
Summary of typical/practical control 126 Typical applications 149
Control of speed and position 126 Low-power analogue D.C. drives 149
viii Contents

The 4Q2 D.C. drive 149 Summary 191


Cheetah-Puma-Lynx 149 2 Network basics 191
3 A.C. drives 151 Physical layer 192
Features common to all A.C. drives 151 Network cables and connectors 192
Power terminal layout 151 Interface circuits 192
Control terminal layout 151 Data encoding 192
Wiring precautions to prevent electromagnetic Network topology 193
compatibility (EMC) issues 153 Data-link layer 194
Open-loop inverters 153 Framing 194
Specifications and ratings 153 Data model 194
Features and options 153 Media access control 194
Methods of control 156 Error handling 195
Performance of the Commander SE open-loop Conclusions 195
drive 156 Application layer 195
Typical applications 157 Device profile 195
The universal A.C. drive 157 3 Simple fieldbus systems 196
The concept of a universal drive 157 Modbus 196
Unidrive option modules 158 Control Techniques' protocol 196
Open-loop operation 159 4 Fieldbus systems 197
Closed-loop operation 159 Requirements for drive applications 197
Servo operation 160 Physical layer 197
Regeneration mode 162 Error detection 197
High-performance servodrives 163 Dynamic performance 197
Performance 166 General message services 197
Summary of practical advantages of SLM Centralised v e r s u s distributed intelligence 197
technology 168 Profibus DP 197
Applications 169 Interbus-S 198
4 Soft-start A.C. motor control 169 CAN 198
Conventional starting 169 DeviceNet 198
Direct-on-line starting (DOL) 169 CANopen 198
Star-delta starting 169 CTNet 198
Auto-transformer starting 170
Disadvantages of conventional starting 170 9 Supply harmonics due to drives 199
Electronic soft start 171
1 Overview 199
Typical applications 171
2 Regulations 200
5 Application boards and software 172
Regulations for installations 200
Applications module 172
Regulations and standards for equipment 200
Software commissioning tools 173
3 Harmonic generation within variable-speed drives 201
Communications modes 173
A.C. drives 201
Drive set-up wizard 173
D.C. drives 203
Commissioning screen 174
Effect of loading 203
Monitoring screen 174
4 The effects of harmonics 203
Parameter list 174
5 Calculation of harmonics 204
Individual drives- D.C. 204
7 Position and motion-control systems 175
Individual drives - A.C. 205
1 General 175 Systems 205
2 Basics of motion control 177 Isolated generators 206
3 Typical motion functions 180 6 Remedial techniques 206
Position lock- electronic gearbox 180 Connect the equipment to a point with a high
Direct positional lock 180 fault level (low impedance) 207
Ramped nonrigid lock 180 Use three-phase drives where possible 207
Ramped rigid lock 180 Use additional inductance 207
Simple single-axis positioning 182 Additional A.C. supply-line inductance 207
CAM functions 182 Additional D.C. inductance 207
Multiaxis positioning 183 Use a higher pulse number (12 pulse or higher) 207
4 Programmability 187 Use a drive with an active input stage 211
5 Summary 188 Use a harmonic filter 211

8 Communications systems 189 10 Electromagnetic compatibility (EMC) 213


1 Introduction 189 1 Introduction 213
Drive set up 190 General 213
Drive control 191 Principles of EMC 214
Contents ix

EMC regulations 214 Limits to cable length 237


2 Regulations and standards 214 Example 240
Regulations 214 Output chokes for long motor cable applications 240
Standards 215 General 240
3 EMC behaviour of variable-speed drives 215 Principles 240
Immunity 215 Calculations 240
Low-frequency emission 216 Example 241
High-frequency emission 216 Position of chokes in multiple motor
4 Installation rules 217 configurations 241
EMC risk assessment 217 Typical recommended cable size 242
Basic rules 217 7 Power supply considerations 243
Segregation 217 High or low line voltage 243
Control of return paths, minimising Supply frequency variations 243
loop areas 217 Supply impedance/fault level 243
Earthing 218 Low supply impedance 243
Simple precautions and fixes 219 High supply impedance 243
Full precautions 219 Multiple drive installations 244
5 Theoretical background 220 8 Thermal design of enclosures 244
Emission modes 220 General 244
Principles of input filters 221 Calculating the size of a sealed enclosure 244
Screened motor cables 222 Example 244
Ferrite ring suppressors 222 Calculating the air flow in a ventilated enclosure 245
Filter earth leakage current 222 Example 245
Filter magnetic saturation 222 9 Installation and maintenance of standard motors
6 Additional guidance on cable screening for and electronic equipment 246
sensitive circuits 223 Motors 246
Cable screening action 223 General 246
Cable screen connections 223 Storage 246
Recommended cable arrangements 225 Installation 246
Maintenance guide 247
Brush gear maintenance 247
11 Systems design 227
Electronic equipment 248
1 General 228 General 248
2 Design matrix 228 Siting of equipment 248
3 Dynamic/resistive braking 228 Ventilator systems and filters 249
General 228 Condensation and humidity 249
D.C. motor braking 228 Fuses 249
Example calculation of a brake resistor of a 10 Common D.C. bus configuration of A.C. drives 250
D.C. motor 231 General 250
A.C. regeneration and braking 231 A simple bulk uncontrolled rectifier 250
Example calculation of a brake resistor Using the mains supply converter in one drive to
of a PWM A.C. induction motor drive system 232 supply all drives from its D.C. link 252
4 Fusing 233 Effectively hard paralleling of all drive input
General 233 rectifiers 252
Protection of mains-drive and drive-motor cabling 233 A bulk four-quadrant controlled rectifier feeding
Protection of drive components 233 the D.C. bus 252
5 Motor overtemperature protection 234 A bulk four-quadrant PWM converter feeding
General 234 the D.C. bus 253
Overtemperature protection of a converter-fed Note on EMC filters for common D.C.
motor 235 bus systems 254
What can be used 235 11 Mechanical vibration, critical speed and
What cannot be used 235 torsional dynamics 255
6 A.C. drive motor cabling 236 General 255
General 236 Example 255
Closed loop- induction motor 236 Causes of shaft vibrations independent of
Cable resistance 236 variable-speed drives 255
Cable-charging currents 236 Subsynchronous vibrations 256
Closed l o o p - P M servomotor 237 Synchronous vibrations 256
Cable resistance 237 Super-synchronous vibrations 256
Cable-charging currents 237 Critical speeds 256
Open-loop current control-induction motor 237 Applications where torque ripple excites a
Cable resistance 237 resonance in the mechanical system 256
Cable-charging currents 237 High-performance closed-loop applications 257
x Contents

Limits to dynamic performance 257 General 284


System control-loop instability 257 Centrifugal pumps 285
Measures for reducing vibration 257 Centrifugal fans and compressors 288
3 Application principles/examples 288
Cranes and hoists 289
12 Applications 259
General 289
1 Typical load characteristics and ratings 259 Planning an installation 289
Metals industries 260 Slewing control 289
Plastics 260 Crane refurbishment for a Norwegian steel wire
Rubber 260 rope maker 290
Chemical 261 Elevators and lifts 291
Materials handling 261 Lift system description 291
Lift, hoist and crane 261 Speed profile generation 292
Concrete pipe manufacture 261 Load weighing devices 292
Fans and blowers 262 Metals and metal forming 292
Pumps 262 Winding, crimping and precise cutting 292
Paper and tissue 262 Roll feed line 295
Printing 262 Wire and cable manufacture 296
Packaging 263 Four-quadrant D.C. drives for a bar mill 296
Engineering industries 263 Wire-drawing machine 296
Wire and cable 263 Paper manufacturing 297
Hydraulics 263 General 297
Electric motors and alternators 264 Sectional drives 297
Textiles 264 Loads and load sharing 298
Foods, biscuit and confection 264 Control and instrumentation 299
2 Techniques common to many applications 264 Winder drives 300
Special D.C. loads 264 Brake generator power and energy 301
Traction motor field control 264 Unwind brake generator control 302
Battery charging 265 Coating machines 302
Electrolytic processes 265 Paper-slitting machine 303
Electric heating and temperature control 265 Paper board machine 303
Digital slaving 266 Building materials 304
General 266 Brick-handling line 304
Drive slaving techniques 266 Roofing-tile manufacturing plant 305
Principle of digital speed/position following 266 Textiles 307
The digital speed/position controller 267 Fabric-dyeing machine 307
Load sharing 268 Quilting machine 308
General 268 Plastics extrusion 308
D.C. thyristor converter-fed system 269 General 308
A.C. inverter-fed systems 271 Basic extruder components 310
High-frequency inverters 272 Overall extruder performance 310
General 272 Energy considerations 310
Frequency control of A.C. induction motors 272 Motors and controls 311
High-frequency purpose-designed motors 273 Food 312
High-frequency inverters 274 Control of hammermills in animal feed
High-frequency applications 274 production 312
Centre winders 275 HVAC 313
General 275 Air conditioning for driver and vehicle
Speed or torque control 276 licensing agency 313
Taper tension 276 Air-handling units at Oxford Brookes
Constant torque and field weakening 276 University students' union 313
Power requirements for centre- Steel 314
driven winders 277 Main mill drives 314
Inertia compensation 277 Auxiliary drives 314
Loss compensation 278 Chemical 315
Flux compensation 278 Enamel painting of fluorescent tubes 315
Drive selection-limiting parameters 278 Marine applications 316
Sectional drive systems 278 Cable laying 316
General 278 Pipe laying 318
Theory of operation 281 Control of lock gates and sluices 319
Using an IEC61131-3 programming tool to Printing 320
configure a sectional drive line 284 Real-time registration and shaftless web
Energy saving 284 tensioning control 320
Contents xi

Offset priming presses 321 Method 346


Stage scenery-film and theatre 322 Calculate speeds and gearing ratio 346
James Bond film stunts 322 Load, force and torque 346
Controlling acoustics 323 Power ratings for the motor and drive 347
Exhibition focal point- the Control Inclined conveyor 348
Techniques' orchestra 323 Hoist 348
Rock concert 325 Data 348
Millennium Dome aerial ballet 326 Velocity ratio (VR) 348
Speed and acceleration of the hook 348
Appendix A Standards for drives 329 Lifting force and torque to accelerate from
rest to full speed 348
A1 IEC (intemational) standards 330 Lifting force and torque to maintain full speed 349
Planned future IEC61800 standards 332 Required motor power rating 349
A2 CENELEC (EC) standards 332 Drive module power rating 349
A3 British standards 334 Screw-feed loads 349
A4 IEEE (USA) standards 335
A5 UL (Underwriters' Laboratories, USA) standards 335
Appendix C Tables 351
A6 Other standards 335
Electricity Association, UK 335 C1 Mechanical conversion tables 351
EIA/TIA (previously RS) 335 Length 351
ANSI 335 Area 352
Volume 352
Appendix B Symbols and formulae 336 Mass 353
Energy 353
B 1 SI units and symbols 336 Inertia 353
SI base units 337 Torque 354
Decimal multiples and submultiples 337 Force 354
Derived units 337 Power 354
Geometrical units 337 C2 General conversion tables 354
Time-related units 337 Length 354
Mechanical units 337 Area 354
B2 Electrical formulae 338 Volume 355
Electrical quantities 338 Mass 355
A.C. three-phase (assuming balanced Force and weight 355
symmetrical waveform) 338 Pressure and stress 355
A.C. single phase 338 Velocity (linear) 355
Three-phase induction motors 338 Velocity (angular) 355
Loads (phase values) 338 Torque 355
Impedance 338 Energy 355
A.C. vector and impedance diagrams 338 Power 355
E.m.f., energy transfer 339 Moment of inertia 356
Mean and r.m.s, values, waveform 340 Temperature 356
Principles 340 Flow 356
Mean D.C. value 340 Torque 356
R.m.s. value 341 Force 356
Other waveforms 341 Moment of inertia 356
Form factor 341 Linear acceleration 356
B3 Mechanical formulae 342 C3 Power/torque/speed nomogram 357
Laws of motion 342
Linear motion 342
Appendix D World industrial electricity
Rotational or angular motion 343
Relationship between linear and supplies (< 1 kV) 358
angular motion 343
The effect of gearing 344 Appendix E Bibliography and further
Friction and losses 344 reading 363
Fluid flow 345
B4 Worked examples of typical mechanical loads 346 Index 365
Conveyor 346
Data 346
CHAPTER 1

Industrial M o t o r s

i!~i~i!i~i 1 D.C. M OTO RS

i iiili 2 A.C. INDUCTION MOTORS 8

iiiii 3 A.C. SYNCHRONOUS MOTORS 15

iiiiiiiiii 4 BRUSHLESS SERVOMOTORS 17

iiiii 5 RELUCTANCE MOTORS 21

6 A.C. COMMUTATOR MOTORS 22

7 MECHANICAL AND ENVIRONMENTAL 22

..!iii
...... 8 EFFECTSOF SEMICONDUCTOR POWER CONVERTERS 40

1 D.C. MOTORS

GENERAL electrical power into mechanical power, and v i c e v e r s a in its


generator form. Inherently straightforward operating char-
History will recognise the vital role played by D.C. motors in acteristics, flexible performance and high efficiency
the development of industrial power transmission systems. encouraged the widespread use of D.C. motors in many
The D.C. machine was the first practical device to convert types of industrial drive application.
2 D.C. MOTORS:General

The later developments of the lower-cost A.C. cage motor It


J

and, more recently, of electronic variable-frequency control


have displaced the D.C. motor to some extent, particularly in
the lower kW range. Nevertheless, the advantages associated
with the inherently stable and relatively simple to control
D.C. machine are indisputable. In its most straightforward
form, speed is approximately proportional to armature vol-
l
v~
l
E~
tage, torque to armature current and there is a one-to-one
relationship between starting torque and starting current.
Modem D.C. motors under thyristor control and with
sophisticated protection continue to provide very sound Figure 1.1 Shunt-wound D.C. motor
1
industrial variable-speed drive performance. High-perfor-
mance test rigs and the higher-kW ratings of drives for the
printing and paper industries, for high-speed passenger lifts,
field flux by:
and drives subject to high transient loading in the metal and
plastics industries; all are likely to continue employing the Ea = klndp (1.1)
D.C. motor with thyristor control for some considerable time
particularly in refurbishment programmes where a D.C. where n = speed of rotation, 0 = field flux and kl--motor
motor exists. The task facing the A.C. drive to completely constant.
oust its D.C. competitor is formidable. Also, the applied, or terminal armature voltage Va is given
D.C. machine tool and servo drives based mainly on by:
chopper technology continue to offer high performance Va = Ea -[- Ia "Ra (1.2)
at low price and at ratings up to approximately 5 kW (con-
tinuous) but here too A.C. technology is making significant where Va - applied armature voltage, Ia - armature current
inroads. and Ra = armature resistance.
The introduction and development of electronic variable- Multiplying each side of equation 1.2 by Ia gives:
speed drives continues to stimulate intensive development of
Vala : Eala + 12Ra (1.3)
motors, both D.C. and A.C. The performance capabilities of
both are being extended as a result, and the D.C. motor is
likely to find assured specialised applications for the fore- total power supplied = power output + armature losses
seeable future.
Interaction of the field flux and armature flux produces an
The majority of standard D.C. motors, both wound field and armature torque. Thus:
permanent magnet, are now designed specifically to take
torque M = k2Ifla (1.4)
advantage of rectified A.C. power supplies. Square, fully
laminated frame construction allows minimal shaft centre where k2=constant, If=main field current and Ia=
height for a given power rating, and affords reduced magnetic armature current.
losses, which in turn greatly improves commutating ability.
This confirms the straightforward and linear characteristic of
Over the last few years the use of permanent-magnet motors, the D.C. motor and consideration of these simple equations
usually in the fractional to 3 kW range, has become com- will show its controllability and inherent stability.
monplace in general-purpose drive applications. In this
design permanent magnets bonded into the motor frame The speed characteristic of a motor is generally represented
replace the conventional wound field. The magnets have by curves of speed against input current or torque and its
a curved face to offer a constant air gap to the conventional shape can be derived from equations 1.1 and 1.2:
armature.
klnq5 = Va - (IaRa)
If the flux is held constant, which is achieved by simply
FUNDAMENTAL EQUATIONS AND holding the field current constant in a properly compensated
PERFORMANCE motor, then:

n -- k3[Va - (IaRa)]
Wound-field Motors
The circuits of shunt-wound and series D.C. motors are
The circuit of a shunt-wound D.C. motor, Figure 1.1, shows
shown in Figures 1.1 and 1.2.
the armature, armature resistance (Ra) and field winding. The
armature supply voltage Va is supplied typically from a With the shunt motor the field flux 0 is only slightly affected
controlled thyristor system and the field supply voltage Vf by armature current, and the value of IaRa at full load rarely
from a separate bridge rectifier. exceeds 5 per cent of Va, giving a speed curve shown in
Figure 1.3, where speed remains sensibly constant over a
As the armature rotates an e.m.f. Ea is induced in the
wide range of load torque.
armature circuit and is called the back e.m.f, since it opposes
the applied voltage Va and the flow of current produced by The series motor curve, Figure 1.4, shows initial flux
Va. This back e.m.f. Ea is related to armature speed and main increase in proportion to current, falling away owing to
Chapter 1.1 3

<

RaI < Ra <


( Va (
v~

Figure 1.2 Schematic of series D.C. motor

Figure 1.5 Compound D.C. motor

13

(3.

-0

(3.
1}

10 20 30 40 50 60 70 80 90 100
torque (%) , | , , , , , , ,

10 20 30 40 50 60 70 80 90 100
Figure 1.3 Torque~speed characteristic of a shunt-wound
torque (%)
D. C motor
Figure 1.6 Torque~speed characteristic of a compound
D.C. motor

Under semiconductor converter control, with speed feed-


back from a tachogenerator, the shape of the speed/load
curve is largely determined within the controller. It has
become standard to use a plain shunt D.C. motor on the basis
of reduced cost, even though the speed/load curve on open-
, , , , l . , . ,
loop control is often slightly rising.
10 20 30 40 50 60 70 80 90 100
torque (%) Permanent-magnet Motors
Figure 1.4 Torque~speed characteristic of a series D.C. Compared with standard types of wound-field D.C. motor,
motor
the conventional D.C. permanent-magnet (D.C. PM) servo-
motor is often designed to exhibit extremely good low-speed
magnetic saturation. In addition, the armature circuit torque tipple characteristics. However, the high-speed
includes the resistance of the field winding and the speed characteristics of a D.C. PM motor are not ideal for all
becomes roughly inversely proportional to the current. If the applications. Because of the mechanical and electrical con-
load falls to a low value the speed increases dramatically, straints set by the motor commutator, increasing the motor
which may be hazardous: the series motor should not speed with a maintained load characteristic soon reveals the
normally be used where there is a possibility of load loss, but commutation limit of the motor. In practice, each motor is
because it produces high values of torque at low speed and designed to work within a safe area of commutation where
its characteristic is falling speed with load increase, it is the available motor torque reduces as motor speed increases,
useful in applications such as traction and hoisting or some in the manner of the shunt-wound D.C. motor, Figure 1.3.
mixing duties where initial stiction is dominant.
The torque/speed curve for a conventional D.C. PM motor is
The compound-wound D.C. machine combines both shown in Figure 1.7. The area of continuous operation
shunt and series characteristics; its circuit is shown in may be defined as the area where the motor can operate on
Figure 1.5. a 24-hour, 100 per cent duty basis with an acceptable
temperature rise.
The exact shape of the torque/speed characteristic is
determined by the resistance values of the shunt and The intermittent duty part of the motor characteristic defines
series fields. The slightly drooping characteristic, Figure the area of operation available on an intermittent basis for
1.6, has the advantage in many applications of reducing the acceleration and deceleration periods. This aspect of rating
mechanical effects of shock loading. is discussed in the section on Duty Cycles (pp. 27-8). The
D.C. MOTORS: F u n d a m e n t a l Equations and Performance

commutation
'~" limit
intermittent zone

{3.



continuous

b ----..~I- . . . . ~ a, I.
torque

Figure 1.7 Characteristic of D.C. PM motor
a peak torque
b stall torque at 40C

commutation limit is most obviously exceeded if brush


sparking becomes excessive. Severe overloading in this

range will cause a complete ring of heavy sparking to run
around the commutator circumference. This phenomenon is
known as brush fire or flashover and must be avoided since it
damages both the commutator and brush gear, reducing the
life expectancy of the motor considerably. Fortunately, the
electronic controller supplying the motor may be easily
specified to prevent overloading.
Good commutation - the ability to give good armature
current reversal without undue sparking at the brushes - is of
particular importance. It should be noted that although the
presence of sparking does give a very good indication of
poor commutation, optimisation of commutation is a very
skilled activity. Indifferent commutating performance places rotation
D.C. motors at some disadvantage when compared with
A.C. cage induction motors, particularly in respect of main-
tenance requirements and costs. The increasing emphasis Figure 1.8 Interaction of stator and rotor fields producing
being placed upon reduced production down time and torque
improved drive system dependability, combined with a a stator field
reduced number of hours available for regular maintenance, b rotor field
are factors tending to favour A.C. c resultant field distortion

OPERATING PRINCIPLES motors act on the same principle except that the field
strength is not variable.
With separately-excited shunt-wound D.C. motors, a steady
D.C. voltage is applied to the field winding and the resulting
current produces a magnetic flux in the motor yoke (main Commutation
frame), the main field poles, the armature and the air gaps,
In order that the motive torque shall be smooth and con-
Figure 1.8a. Field strength is determined by the field-winding
tinuous, the current distribution in the armature coils must
current. The armature winding occupies the rotor peripheral
remain constant. This requires that the current in any given
slots and opposite coils are connected to each other and to a
coil must reverse as that coil passes from the influence of
commutator segment. The brushes provide the means of
one pole to the next in rotation. The process of current
connecting current flow to the rotating armature. This current
reversal is known as commutation and must take place
flow generates an armature flux, Figure 1.8b, with a magnetic
during the time that the coil is short circuited by a brush
axis mechanically fixed at 90 to that of the field flux.
spanning the gap between adjacent commutator segments.
The interaction of these two magnetic fluxes generates a Failure to achieve armature current reversal during this
flux distortion, Figure 1.8c, so that the armature rotates, short-circuit time will cause sparking at the trailing edge of
endeavouring continuously to correct this distortion and the brush as the short circuit is broken. Commutator action is
thus providing motive power. Permanent-magnet D.C. shown in detail in Figures 1.9 and 1.10.
Chapter 1.1 5

//2
rotation [iiiiiiiiii!iiiiiiii!iiiiiiiiiiiiiiiii!iiiiiiii!iii!iil] Itor i ......................................... i .............................

!',, i
i
:
X~ !
' , actual i

r- : ~ :

~ 0
0

i ', i
ideal ',, i
i ', i
"i
b I c rtol ii \i
i, . . .

"
. . . . . . . . .
brush
. .

arcing
. . . .

I/2
contact/~:~ "........... ;eriodoi ........... ~" 'i l
with short circuit
i
period of i
segment c _ current reversal i

I/2 //2 Figure 1.10 Commutator current reversal

segment c makes contact with the brush (Figure 1.9). As the


area in contact with segment c increases, the current flowing
from the brush via segment c to coil A decreases. If the
current through segments b and c is proportional to
brush contact areas then the current through coil A decreases
linearly to zero (direction left to right) and then builds up
linearly in the opposite direction, Figure 1.10. Under these
conditions commutation would be linear and sparkless,
I/2 assisted by the changing ohmic resistance of the carbon-
brush/copper segment contact area. Because of the armature
Figure 1.9 Reversal o f current flow in one armature coil coil inductance, a reactive current is induced in the coil
undergoing commutation, delaying the current reversal in
the short-circuited coil, as shown in Figure 1.10.
There is a physical limit to the speed at which current can be
With heavier motor loading there is an increase of armature
commutated. This commutation limit is often defined as the
current and armature flux, distorting further the flux gener-
product of the motor shaft power and the rotational speed.
ated by the field. This distortion, which is producing the
The limit is widely accepted to be approximately 3 x 106
rotational torque, acts to degrade commutation and to pro-
(kW. m i n - 1). Where greater ratings are required more than
duce brush sparking. The degree of flux distortion from
one armature can be placed on a single shaft, or several
the neutral flux axis is proportional to armature current. The
motors arranged in series (provided the shafts are rated to
effect is termed armature reaction. Referring back to
transmit the total torque).
Figure 1.8, the brushes are shown on the neutral axis, i.e. at
90 to the flux path and midway between the field poles,
Rotation assuming no flux path distortion. Figure 1.8c shows that
distortion of the flux owing to armature reaction causes the
Figure 1.9a shows that coil A is about to be short circuited neutral axis to shift. If the brush axis were moved to position
and is carrying half the armature current (/2) from left to X-Y, Figure 1.8c, the brushes would again lie on the neutral
right, where the current I flows to the brush. Figure 1.9b
axis, that is at 90 to the direction of the main flux path.
shows coil A in the middle of its short-circuit period, where
armature current from the brush does not pass through short-
circuited coil A. In Figure 1.9c, coil A is shown immediately Compensation
following the short circuit, carrying 1/2 current in the right to
To improve commutation, interpoles are introduced to shift
left direction; the current in coil A has been reversed, or
the main flux so that it lies on the geometric neutral axis.
commutated.
These interpoles are interposed between the main field poles
Any failure of the current to fall through zero and rise to the and their coils carry the full armature current, generating a
same value in the reverse direction during this very brief flux proportional to current (torque load) and reducing the
period of commutation results in sparking at the brush as the effect of armature reaction.
inductive circuit of the armature coil is switched.
It follows that interpoles are most effective in their correc-
The current changes are shown in graphical form in Figure tive action when the field flux is at its full level. The com-
1.10. The current in coil A is 1/2 from left to right until monly used technique of field weakening progressively
D.C. MOTORS:Operating Principles

reduces this self-corrective effect; hence the quality of incorporated into the design, the lower the resulting torque
commutation can be affected when a motor operates under ripple will be. However, in practice, this is limited by phy-
the weak field conditions sometimes required in system sical practicalities, cost and available space. Good commu-
application. tation in a D.C. PM motor is therefore a compromise.
An additional method of reducing this troublesome armature
reaction, is to neutralise the armature flux directly under the CONSTRUCTION OF THE D.C. MOTOR
main pole faces by the use of compensating or pole face
windings. These windings are placed in slots cut into the The principal components of the D.C. motor comprise the
main pole air gap faces and carry the full armature current, rotating shaft-mounted wound armature and connected
being connected in series with the armature circuit, and thus commutator, the field system surrounding the armature
giving compensation with current loading. Compensating across the motor air gap, the supporting and enclosing end
windings on the pole faces present problems of both frames with shaft bearings and brush gear to connect the
mechanical security and electrical insulation, a satisfactory commutator to external terminals. These components and
solution often involving significant additional cost. The use others are to be seen in the sectioned assembly of a typical
of a compensated machine is restricted therefore to duties machine, Figure 1.11.
requiring a very wide range of field operation, such as high-
speed machine tool applications and others involving rapidly D.C. Motor Frame
varying loads and speeds, in which the commutating per-
formance of the compensated motor can be extremely good The main flame of a typical modern four-pole D.C. motor is
when compared with that of the more normal non- a square, laminated iron flame with a curved inner face.
compensated machine. Main pole lamination and coil shunt field assembly are
mounted on the inside face at 90 to each other, with inter-
D.C. PM Commutation pole assemblies interposed at 45 between.

The commutator also plays a very important role in the Square-frame motors offer reduced centre height com-
functioning of a D.C. PM motor. To reduce torque ripple at pared with round-flame motors by virtue of the fact that
the motor shaft it is desirable to keep the magnetic interac- the main field poles mounted on the four flat internal
tion between the rotor and stator as evenly distributed as faces are of pancake design and that the four interpoles,
possible. In this case, however, interpoles and winding mounted on diagonal axes, fit into the internal corners of
compensation are not available options. the square.

The prime purpose of the commutator is to switch the current D.C. Motor Armature
to each set of windings in turn on the armature, and if
the optimum relationship between the stator flux and The armature consists of a cylindrical laminated iron core,
the armature flux is to be maintained for the best current shaft mounted, with a series of slots to accommodate the
conversion to torque, higher rates of commutation are armature winding coils, insulated from each other and the
essential to ensure minimum torque ripple in the motor. This core. The coil ends are connected to insulated segments of
means that the higher the number of commutation segments the commutator.

Figure 1.11 High-performance D.C. motor (courtesy Leroy-somer)


Chapter 1.1 7

The commutator, a cylindrical copper segmented construc- Brush Gear


tion, is mounted at one end of the armature core, with
wedge-section segment bars clamped between commutator The number of brushes is based upon the full-load current of
end tings, insulated from each other and the metal of the the armature circuit and the working velocity of the com-
armature. These segments are machined on the outer axial mutator track and the working conditions generally govern
face to provide a smooth concentric surface on which the the grade of carbon used. Brush pressure is usually fixed and
brushes can bear with a tensioned contact. An armature unvarying over the permitted wear length, and is controlled
assembly is shown in Figure 1.12. by tensator constant-tension springs. It is very important to
use the correct grade of brush!
The armature winding slots are skewed by about one slot
pitch to reduce torque ripple (cogging effect) at low speed Mountings
and slot noise generation at high speed. Such a cogging
effect superimposed on the otherwise steady rotation of the The typical D.C. motor shown in Figure 1.11 is a drip-proof
motor shaft would be unacceptable in many applications. (IP23) enclosed foot-mounted D.C. motor with side-moun-
For example, a machine tool axis drive with such a char- ted terminal box and with a feedback tachogenerator coupled
acteristic could well leave undesirable surface patterning on to the nondrive end. Access to the brush gear is through the
the workpiece. four inspection covers around the commutator end shield,
although in practice only three can be used. Any of these
The precise method of skewing presents the motor manu- three inspection apertures can be used for mounting a force
facturer with two problems. The first is in determining the ventilation fan, to allow wide (typically 100: 1) speed range
amount of skew for a particular design of motor. The second operation at constant load torque.
is that it is much more difficult to wind the armature on an
Basic adaptability of the square-frame machine also allows
automatic winding machine when the laminations are
the terminal box to be mounted on any of the three available
skewed in this fashion. Computer-aided design plays a large
faces of the machine, and for the machine itself to be
part in solving the first of these difficulties, and most
mounted by end or foot flange; and since both end shields
motor manufacturers have developed standard programmes
have the mounting feet attached, they can readily be rotated
to speed magnetic calculations. Modem wire insertion
through 90 to allow side foot mounting.
machines make relatively easy work of shooting windings
into the skewed armature slots.
D.C. PM Design
Armature winding coil ends are banded with glass-fibre and
Industrial D.C. permanent-magnet (D.C. PM) motors tend to
epoxy bonding for high speed operation (> 6000 m i n - ~ for
offer a smaller frame size for a given kW rating than their
many machine tool applications) to prevent armature coil
wound-field counterparts, although frame lengths can be
deflection under centrifugal stress. Two balancing discs, one
greater.
behind each beating, permit dynamic balancing to R or the
more stringent S standard over the working speed range of A D.C. PM servomotor is conventional in most aspects of
the machine. construction, but usually incorporates technical refinements

Figure 1.12 D.C. motor armature assembly


8 D.C. MOTORS:Construction of the D.C. Motor

to give performance characteristics of a higher order than determined it is possible to partially demagnetise the motor
those of the general purpose D.C. PM motor. Areas where through overcurrent in the armature or with temperatures in
the motor designer achieves the required sophistication excess of the Curie point of the material, with the result of
include rotor inertia and choice of magnetic materials. reduced torque and usually a higher-than-rated motor speed
on light load, for a given input voltage.
Rotor Inertia Alternative types of magnetic material overcome these
problems with very high demagnetisation levels and oper-
One of the key elements of a D.C. PM motor is the rotor or ating temperature ranges. Such a material is samarium
armature design. It is a normal requirement of a servo system cobalt, which is grouped with the rare-earth category of
to be able to accelerate and decelerate the motor (and load) magnetic materials. There is much development activity
very rapidly. To facilitate rapid changes in speed, and in in this area and new, improved, materials are appearing
order to reduce the input power required to achieve these continually.
changes, the inertia of the rotor is kept to a minimum.
Currently, magnets made from rare-earth materials are
Since inertia increases by the square of the diameter, D.C. expensive and their use can double the cost of a motor, but
PM servomotor rotors tend to be long in relation to their for many applications the choice of such magnetic materials
diameter. Care should be taken not to apply very-low-inertia is unarguably justified; the resulting magnetic flux is much
servomotors to high-inertia loads as the inertia mismatch can denser per unit volume of magnet, and therefore the rotor
lead to very troublesome resonances. inertia can be significantly reduced for a given rated output
A prime reason for choosing a servo system as opposed to a torque. This allows an improved acceleration/deceleration
conventional variable-speed D.C. system is the ability to performance per unit armature current.
operate the motor at or very near to zero speed, while still A magnetic material with the cost advantages of an ordinary
offering full-rated torque. ferrite magnet and the performance advantages of the rare-
earth types has been sought. This has now been achieved
Permanent-magnet Materials with the advent of synthetic magnetic materials, principally
neodymium iron boron, which is fully acceptable to
Choice of magnetic material for the D.C. PM motor is a numerous industries, including the automotive industry, and
major deciding factor in the operational characteristic. provides extremely good performance characteristics with-
Conventional D.C. PM motors use a low-cost ferrite mag- out the large price premium of the rare-earth materials.
netic material which, although more than adequate for
The magnet itself is formed from powders drawn from an
the majority of applications, does have some inherent dis-
ever growing range of magnetic materials and is moulded
advantages.
into the required size, shape and profile using a high-tem-
Mechanically, ferrite magnets are fragile, and so are usually perature specialised sintering process.
bonded to the motor body, which is very secure. However,
Delivered to the motor manufacturer in an unmagnetised
because ferrite magnets are also brittle, they are susceptible
state, the magnet is first bonded into the motor frame using
to mechanical shock and can be fractured unless the motor is
a two-part adhesive with a suitably high melting point.
handled with great care.
The assembly is then magnetised by inserting a close-
One of the more important disadvantages of low-cost mag- fitting metal conductor through the motor frame and intro-
nets, however, is the ease with which they can become ducing a high-current shock pulse through the assembly
demagnetised. Ferrite magnets have a relatively low by discharging a capacitor bank through the conductor.
demagnetising level. Under normal circumstances, and with The current pulse polarises the elements of the sintered
a correctly-determined servo system, demagnetisation of the magnets and saturates them to one end of their magnetisation
motor is not possible. However, if the system is not correctly characteristic.

2 A.C. INDUCTION MOTORS

GENERAL of all electrical power into mechanical energy. This type


of motor nonetheless exhibits some quite unattractive per-
The A.C. squirrel-cage induction motor is the basic, universal formance characteristics in spite of intensive develop-
work horse of industry, converting some 70 to 80 per cent ment, notably instability and a nonlinear load-current
Chapter 1.2 9

characteristic. It is invariably designed for fixed-speed connections for the winding coils are not shown, but R and
operation, larger ratings having such features as deep rotor R1 represent the start and finish of the red phase winding,
bars to limit direct-on-line (DOL) starting currents. Elec- and similarly for the Y and B phase conductors. The R, Y
tronic variable-speed converter technology is able to provide and B phase windings are displaced 120 in space relative to
the necessary variable-voltage/current, variable-frequency one another.
supply which the three-phase A.C. machine requires for
efficient, dynamic and stable variable-speed control. Assuming that when stator current is positive it is flowing
inwards in conductors R, Y and B, and therefore outwards in
Modern electronic control technology is able not only to R1, Y1 and B1, that the current in phase R in Figure 1.13a is
render the A.C. induction motor satisfactory for many at its maximum positive value and that in phase windings Y
modern drive applications but also to greatly extend its and B the currents at the same instant are negative and
application and enable advantage to be taken of its low each equal to half maximum value, then these currents
capital and maintenance costs. produce the magnetic fluxes represented by the arrowed
More striking still, microelectronic developments have lines in Figure 1.13a and the flux axis is horizontal.
made possible the highly dynamic operation of induction
motors by the application of flux-vector control. The prac- Thirty degrees later in the supply cycle, the currents in
tical effect is that it is now possible to drive an A.C. phases R and B are 0.866 (x/~/2) of their maximum, and zero
induction motor in such a way as to obtain a dynamic in the Y phase. The pattern of the flux due to this current is
performance in all respects better than could be obtained shown in Figure 1.13b. It will be noted that the axis of this
with a phase-controlled D.C. drive combination. The various field is now in line with coil Y-Y1 and therefore has turned
forms of A.C. induction motor control are fully described in clockwise through 30 from that of Figure 1.13a.
Chapter 4.
After a further 30 in the supply cycle the current in
phase winding B has reached maximum negative value, and
the currents in R and Y are both positive, at half their
FUNDAMENTAL EQUATIONS AND maximum value. These currents produce the magnetic flux
PERFORMANCE shown in Figure 1.13c, the flux axis being displaced clock-
wise by a further 30 compared with that of Figure 1.13b.
Electrical Characteristics of
Induction Motors Thus, for every time interval corresponding to 30 in the
supply cycle, the axis of the flux in a two-pole A.C. stator
Consider the stator winding of a simple three-phase two-pole rotates 30 in space. With a two-pole stator (one pair of
A.C. cage induction motor, each phase winding having only poles) the flux rotates through one revolution in space in one
one slot per pole per phase as shown in Figure 1.13. End cycle of the power supply. The magnetic flux is said to rotate

+ ,.30~,30~,

- a b c

a S I~

b c

Figure 1.13 Three-phase rotating field


10 A.C. INDUCTION MOTORS: Fundamental Equations and Performance

at synchronous speed, and the rotational speed of the flux is: windage, the rotor could not continue to rotate at synchro-
nous speed. The rotor speed must therefore fall below
f/p revolutions in one second the synchronous speed, and as it does so rotor e.m.f, and
current, and therefore torque, will increase until the speed
matches that required by the losses and by any load on the
where f = supply frequency in Hz, and p = number of pole
motor shaft.
pairs. Note: a two-pole motor has one pole pair!
The difference in rotor speed relative to that of the rotating
It is more usual to express speed in revolutions per minute:
stator flux is known as the slip. In Figure 1.15, for a given
60f/p revolutions per minute (min -1) torque OA, rotor speed is AC and the slip is AD, and:
The e.m.f, generated in a rotor conductor by transformer AD = A B - AC = CB
action is at a maximum in the region of maximum flux
density. With the flux rotation assumed to be clockwise, the It is usual to express slip as a percentage of the synchronous
direction of the e.m.f.s generated in stationary rotor con- speed, i.e. AD/AB x 100. Slip is closely proportional to
ductors (determined by Fleming's fight-hand rule) are torque from zero to full load.
indicated by conventional crosses and dots in Figure 1.14a.
Slip at full load varies from about 7 per cent for small motors
The e.m.f, generated in the single rotor conductor shown in to about 2 per cent for large, and is a good indication of the
Figure 1.14b produces a current, the effect of which is to efficiency of a machine; the lower the slip the higher the
reinforce the flux density on the left-hand side of the con- efficiency. The fact that the A.C. squirrel-cage induction
ductor and to weaken it on the right-hand side. In con- motor is neither a true constant-speed machine nor inher-
sequence, a force is exerted on the rotor tending to move it in ently capable of providing variable-speed operation con-
the direction of the flux rotation. The higher the speed of the stitutes a major limitation.
rotor, the lower the speed of the rotating stator flux field
relative to the rotor winding, and therefore the smaller the Analysis shows that the rotor flux ampere turns travel at the
e.m.f, generated in the rotor winding. same speed as those of the resultant of the stator flux ampere
turns and that both are stationary relative to each other. The
If the speed of the rotor became the same as that of three-phase induction motor can be regarded as a transfor-
the rotating field, i.e. synchronous speed, the rotor con- mer having an air gap between the magnetic circuits of the
ductors would be stationary in relation to the rotating flux. primary (stator) and secondary (rotor) windings. This air
This would produce no e.m.f, and no rotor current and gap, although designed to be as small as practicable, has an
therefore no torque on the rotor. Because of friction and important effect on the characteristics of A.C. squirrel-cage
motors.
The air gap requires larger magnetisation and gives larger
magnetic leakage than a transformer of the same kVA
rating. However, stator and rotor ampere turns have to bal-
ance, as well as supplying magnetisation and no-load
losses, just as in transformer design. So an increase in per-
centage slip due to load is accompanied by an increase in
rotor current (secondary current) and therefore by a corre-
sponding increase in stator current (primary current) above
the no-load current. This no-load current for a typical
squirrel-cage motor lies between 25 and 40 per cent of the
full-load current and is largely due to the magnetic excitation
current required by the air gap. Thus the cage induction
machine over its working load range is self regulating for
input power, this regulation being controlled by the per-
N centage slip with load.
The most popular squirrel-cage induction motor in sizes up
to about 5 kW is of four-pole design. Its synchronous speed
b with a 50 Hz supply is therefore:

60f/p = 60.50/2 = 1500 min -1

Slip accounts for about 5 per cent, and a typical nameplate


speed is 1425 min -1.

Torque Characteristics
Figure 1.14 Torque on the rotor A disadvantage of the squirrel-cage machine is its fixed rotor
a currents induced in rotor bars producing characteristic. The starting torque is directly related to the
field rotor circuit impedance, as is the percentage slip when
b distortion of field running at load and speed.
C h a p t e r 1.2 11

, synchronous speed B

rotor speed
C' i

i
i

i i
i

full-load torque .~'

Figure 1.16 Typical rotor bar profiles

D rotor slip
i
600
o A
torque
ent a
500
Figure 1.15 Speed curve of an induction motor

~400 . current b \
C

o 300
Ideally, a relatively high rotor impedance is required for
good starting performance (torque against current) and low
rotor impedance for low full-load speed slip (and high 200
efficiency).
100
This problem can be overcome to a useful extent, for direct-
on-line (DOL) application, by designing the rotor bars with
special cross sections, Figure 1.16, rather than round or ' ~o ' 166
square, so that eddy-current losses, which these special speed, %
sections cause, increase the impedance at starting when the
rotor flux (slip) frequency is high, and reduce it at normal Figure 1.17 Typical torque~speed and current~speed curves
running speed when the flux frequency is low. a standard m o t o r
b h i g h - t o r q u e m o t o r (6 per cent slip)
Altematively, for special high-starting-torque motors,
two or even three concentric sets of rotor bars are used.
Relatively costly in construction but capable of a substantial
improvement in starting performance, this form of design
produces an increase in full-load slip. Since machine losses the voltage impressed upon the winding by the supply. This
are closely proportional to working speed slip, increased is because the resistance of the winding results in only a
losses may require such a high starting torque machine to small voltage drop, even at full-load current, and therefore in
be derated. the steady state the supply voltage must be balanced by the
e.m.f, induced by the rotating field. This e.m.f, depends on
The curves in Figure 1.17 show indicative squirrel-cage the product of three factors:
motor characteristics. In the general case, the higher the
starting torque the greater the full-load slip. This is one of the
1 The total flux per pole (which is usually determined
important parameters of squirrel-cage design as it is tied to
by the machine designer).
working efficiency and therefore working losses.
2 The total number of turns per phase of the stator winding.
3 The rate of field rotation or frequency.

VOLTAGE-FREQUENCY RELATIONSHIP
Exactly the same factors are valid for transformer design,
except that the field is pulsating instead of rotating.
Nearly all commercially-available industrial induction
motors are wound for direct connection and starting on the In a transformer it can be shown that the e.m.f, induced in a
supply voltage and frequency which prevail in the country winding is given by:
where they will be used. It is a relatively simple matter for
the motor manufacturer to select the number of tums per coil V-K~f
and the wire size to match any voltage within a wide range.
where V is the e.m.f, induced in a coil, K is a constant, ~; is
If it is desired to convert a constant-speed motor operating
the flux through the coil andfis the frequency of the supply.
direct-on-line to a variable-speed drive using an inverter it is
necessary to consider the effect of frequency on flux and This can be rearranged to give:
torque.
(v/f)/x
An induction motor on a normal supply operates with a
rotating field set up by three-phase currents in the stator For constant flux, it can be seen that the ratio V/f must be
winding. The magnitude of the field is controlled broadly by held constant.
12 A.C. INDUCTION MOTORS" Voltage-Frequency Relationship

For inverter operation the speed of field rotation for which


maximum voltage is appropriate is known as the base
speed.
100 %

Increased Voltage
If, again in the steady state, the voltage applied to the stator C~

terminals is increased without a corresponding increase in 0


>

the frequency, only the flux can vary to regain the balance 0

between applied voltage and e.m.f. If the flux is forced to 0


E
increase by applying excessive voltages, the iron core of the
machine is driven progressively into saturation. This not
only increases iron losses due to hysteresis and eddy cur-
rents, but can lead to a very marked increase in stator cur-
rent, with corresponding resistive losses. Since most
machines are designed to work with the minimum of
material, their magnetic circuits are normally very close to ideal
saturation and excessive stator voltage is a condition which
/
100 %
must be carefully avoided.

ical
Reduced Frequency x"
w-
i,_
The consequence of reducing the supply frequency can 0
0
readily be deduced from the relationship described above. E
For the same flux the induced e.m.f, in the stator winding will
be proportional to frequency, hence the voltage supplied to
the machine windings must be correspondingly reduced in
order to avoid heavy saturation of the core. This is valid for
changes in frequency over a wide range. The voltage-
frequency relationship should therefore be linear if a con-
stant flux is to be maintained within the machine, as the
designer intended. If flux is constant so is the motor torque 100 %
for a given stator current, hence the drive has a constant
torque characteristic.
Although constant voltage-frequency (V/f) control is an a;
base speed
important underlying principle, it is appropriate to point m
0
>
out departures from it which are essential if a wide speed J ,
.e'"
range is to be covered. First, operation above base speed 0
E
is easily achieved by increasing the output frequency of the
. ~t ' " " " boost
inverter above the normal mains frequency; two or three ," (exaggerated)
t
times base speed is easily obtained. The output voltage of
an inverter cannot usually be made higher than its input frequency and speed
voltage and therefore the V/f characteristic is typically as Figure 1.18 Voltage-frequency characteristics
shown in Figure 1.18a. Since V is constant above base
a linear VIf below base speed
speed, the flux will fall as the frequency is increased after the
b typical motor flux with linear V/f (showing
output voltage limit is reached. The machine flux falls fall in flux at low frequency as well as
(Figure 1.18b) in direct proportion to the actual V/f ratio. above base speed)
Although this greatly reduces the core losses, the ability c modified V/f characteristic with low-
of the machine to produce torque is impaired and less frequency boost (to compensate for stator
mechanical load is needed to draw full-load current from the resistance effects in steady state)
inverter. The drive is said to have a constant power char-
acteristic above base speed. Many applications not requiring
full torque at high speeds can make use of this extended
speed range.
the stator resistance effect. Indeed, as output frequency
The second operating condition where departure from a approaches zero, the optimum voltage becomes the voltage
constant V/fis beneficial is at very low speeds, whereby the equal to the stator IR drop. Compensation for stator resis-
voltage drop arising from the stator resistance becomes tance is normally referred to as voltage boost and almost
significantly large. This voltage drop is at the expense of all inverters offer some form of adjustment so that the
flux, as shown in Figure 1.18b. To maintain a truly constant degree of voltage boost can be matched to the actual
flux within the machine the terminal voltage must be winding resistance. It is normal for the boost to be gradually
increased above the constant V/f value to compensate for tapered to zero as the frequency progresses towards base
Chapter 1.2 13

speed. Figure 1.18c shows a typical scheme for tapered With the correct value of (usually) resistance inserted in the
boost. It is important to appreciate that the level of rotor circuit, a near-unity relationship between torque and
voltage boost should increase if a high starting torque is supply current at starting can be achieved, i.e. 100 per cent
required, since the IR drop will be greater by virtue of the full-load torque, with 100 per cent full-load current, 200 per
increased stator current. In this case automatic load-depen- cent FLT with 200 per cent FLC etc., (i.e. comparable with
dent boost control is useful in obtaining the desired low- the starting capability of the D.C. machine). Not only the
speed characteristics. Such a strategy is referred to as high starting efficiency but also its smooth controlled accel-
constant V/f (or V/Hz) control and is a feature of most eration historically gave the slip-ring motor great popularity
commercially available A.C. drives, although more advanced for lift, hoist and crane applications. It has had similar
open-loop strategies are now becoming available (refer to popularity with fan engineers, to provide a limited range of
Chapter 4). air volume control, either 2:1 or 3 : 1 reduction, at constant
load, by the use of speed regulating variable resistance in the
So far the techniques described have been based on
rotor circuit. Although a fan possesses a square-law torque-
achieving constant flux within the air gap of the
speed characteristic, so that motor currents fall considerably
machine or, if that is not possible, then the maximum
with speed, losses in the rotor regulator at lower motor
flux. Constant flux is the ideal condition if the largest cap-
speeds are still relatively high, severely limiting the useful
ability of torque is required because the load cannot
speed range.
be predicted with certainty, or if the most rapid possible
acceleration time is desired. A large number of inverters Rotor slip-ring systems, used with this type of motor, offer a
are used, however, for variable air volume applications similar service life to that of the D.C. motor commutator
where control of airflow is obtained by variable-speed fans. system.
The torque required by a fan follows a square-law char-
Efficient variable-speed control of slip-ring motors can be
acteristic, Figure 1.19, with respect to speed, and reducing
achieved by converters using the slip energy recovery
the speed of a fan by 50 per cent will reduce its torque
principle first proposed by Kramer in 1904. Such schemes
requirement to only 25 per cent of its rated torque. As the
are based upon converting the slip frequency on the rotor
load is entirely predictable there is no need for full torque
to supply frequency. (These schemes are described in
capability and hence flux to be maintained, and higher motor
Chapter 2.2.)
efficiency can be obtained by operating at a reduced flux
level. A further benefit is that acoustic noise, a major con- It is also possible to retrofit variable-frequency inverters to
cern in air-conditioning equipment, is significantly reduced. existing slip-ring motors. This can be done simply by
It is therefore common for inverters to have an alternative shorting out the slip-ring terminations (ideally on the rotor
square-law V/f characteristic or, ideally, a self-optimising thereby eliminating the brushes) and treating the motor as a
economy feature so that rapid acceleration to meet a new cage machine.
speed demand is followed by settling to a highly efficient
Variable voltage control of slip-ring motors has been used
operating point.
extensively, notably in crane and lift applications, although
it is now largely being replaced by flux-vector drives and
will therefore not be considered further.
SLIP-RING INDUCTION MOTOR

The wound rotor or slip-ring A.C. machine, although intro-


ducing the negative aspect of brushes, does address some SPEED-CHANGING MOTORS
of the disadvantages of the cage induction motor but with
the handicap of cost compared to the equivalent rated D.C. The cage rotor already described in outline is constructed
shunt-wound machine. with copper rotor bars brazed to shorting end tings, or is
aluminium die cast in a single operation. The construction
is simple, cheap and robust. It has the further advantage
that with various stator winding pole combinations,
whereby the sequence of current reversals in the rotor bars
is altered, the rotor end tings provide a free path for the
current to flow, adapting to the differing number of stator
poles. This allows stator windings of more than one pole
100 %
combination to be wound on the same stator, and such
multispeed A.C. squirrel-cage motors can take one of two
0 ID forms.
~o The simplest is to combine two quite separate stator wind-
s
v ings in the one machine, for example four and eight poles,
"10 0
~E
providing a 2:1 speed range with a constant torque rela-
0
J base speed "'~' tionship between the two speeds. This would suit many
applications in materials handling and in lift and crane
drives, but would be wasteful of torque and therefore of cost
frequency and speed
for fan and pump drives, where the load torque requirement
Figure 1.19 Typical square-law characteristic of a fan or for low-speed operation falls as the square of the speed
pump load reduction.
14 A.C. INDUCTION MOTORS: Speed-Changing Motors
,,

Alternatively, and often more commercially attractive, there A die-cast aluminium rotor cage induction motor construc-
is the design of a single stator winding which allows dif- tion is generally used up to about 50 kW and allows best
ferent pole combinations by switching external connections. economics in manufacture. Rotor losses with aluminium
This is termed consequence pole switching and is restricted bars are bigger than those with copper and copper tends to be
to 2:1 speed range combinations, i.e. 2/4 pole or 4/8 pole, substituted at 50 kW or thereabouts.
few other 2:1 combinations being practicable. Since this
The use of aluminium in the frame construction of smaller
uses a single winding only, a constant torque relationship
machines up to 37 kW rating is becoming standard practice
between the two speeds is available.
on account of the better thermal conductivity when com-
More than two speeds can be made available from A.C. pared with cast iron. It offers benefits in the areas of resis-
squirrel-cage motors by combining separately-wound and tance to corrosion, ease of machining, reduced weight and
consequence pole switching in the same machine. Pole generally improved appearance. The better heat conduction
amplitude modulated (PAM) speed change, as this method is of aluminium and the much improved force-ventilated
known, is also available for machines of larger kW ratings cooling of the standard IP54 enclosed machine of current
and is somewhat similar in principle to consequence pole design show significant benefits, although such cooling still
switching. falls short of the effectiveness of screen-protected IP22
enclosures, which exchange the cooling air around the
stator and rotor windings directly with the surrounding
A.C. INDUCTION MOTOR CONSTRUCTION
ambient air.
The A.C. cage induction motor stator coils are wound on a Above 37 kW, aluminium construction appears to lose its
laminated iron core formed into a pack usually by seam advantages, particularly in terms of the mechanical strength
welding. Semiclosed slots accommodate the windings, required for the higher torque loading involved. Cast iron
which are normally wound as a concentric three-phase, and steel construction becomes the standard as does
typically four-pole, arrangement suitably insulated per coil detailed attention to effective cooling of the IP54 enclosed
set, per phase and from earth. machine. At these higher kW ratings the inferior efficiency
of the A.C. cage machine compared with the equivalent rating
A laminated cylindrical rotor core is mounted centrally in
of D.C. motor and brushless synchronous A.C. motor gives
the stator with the smallest practical radial air gap and with
the larger cage machine additional kW losses to dissipate.
semiclosed or closed rotor slots around the periphery to
accommodate the rotor winding. This winding comprises However, a major advantage of the A.C. cage machine is
bare or lightly-insulated aluminium or copper bars, occu- that it is readily available in enclosures to match the
pying the slots and connected to shorting end rings at each requirements of difficult and hazardous environments and
end face of the rotor pack. A typical die-cast aluminium can be specified to accept much higher levels of external
rotor construction is shown (with stator) in Figure 1.20, for vibration and shock loading than is possible for a compar-
comparison with a copper bar cage rotor, Figure 1.21. For able D.C. machine.
lower ratings, the rotor laminated core and cage construction
is replaced by a cylinder of mild steel, which works quite Standard cage induction machines are available to comply
well but with greatly reduced output power relative to size. with a wide variety of international standards.

i~i!i: ~!~: i!i: !~!: .... ? ::~ ~ ::%::: ~:i:i~....


: : . . . . .... ~:~~ ii~!i~i:i~

i ::~i: !i(~
iii!~:iiii:i)/i!i:i:~iiii!i~iii~!i~:iii!~!

ili i i!i il i ii! i ! !iil i !~i " : .... !i


ii:~i
~!:iiiil:i:i~i.... !
~iI:il~i I!~:'~Ii:iii~i!ii~i iiii~i~i
i:i:i
iii!

Figure 1.20 Typical squirrel-cage induction motor die-cast rotor with its laminated stator
(courtesy Leroy-somer)
Chapter 1.3 15

Figure 1.21 Typical squirrel-cage induction motor rotor with inserted copper rotor bars
(courtesy Leroy-Somer)

3 A.C. SYNCHRONOUS MOTORS

GENERAL permanent magnets into the rotor system, Figure 1.23, or by


borrowing A.C. alternator technology and making the rotor
Considering a rotor of the solid pattern it is possible to of wound multipole design, either with a separate D.C.
visualise a series of flats machined around the rotor block supply for excitation or of brushless alternator design with
periphery, the number of flats corresponding to the number shaft-mounted exciter.
of poles with which the three-phase stator winding is
A.C. synchronous motor application engineering needs
wound, and as shown in Figure 1.22 for a four-pole
particular care to ensure that the machine will accelerate its
(1500 m i n - 1 50 Hz) design.
load to synchronous speed and be capable of holding syn-
A motor fitted with such a rotor would run at synchronous chronous speed under transient load conditions. Such a
speed with no load on the shaft and, for a typical low kW motor can prove commercially attractive, particularly where
design, would pull out of, or lose, synchronism at a torque precise speed holding or speed matching between drives is
loading in excess of about 10 per cent of that of the essential and where, by borrowing alternator technology,
equivalent frame size asynchronous design. Having pulled advantage can be taken of volume manufacture and the
out of the synchronous condition the percentage slip would availability of an IP23 protection enclosure, which allows
be high, the motor essentially reduced to a crawl if the load is improved cooling.
not removed, causing significant losses and excessive tem-
Small salient-pole synchronous motors possess limitations
perature rise.
for general and for variable-speed drive use. For this reason,
The performance of such a simple synchronous cage machine further comment refers to the larger salient-pole and to
would be poor but could be greatly improved by designing cylindrical-rotor synchronous motors, which are likely to
16 A.C. SYNCHRONOUSMOTORS: General

frequency related to the flux speed difference, the mean


torque being zero. Such conditions can occur due to load
transients and during starting.
Wound-rotor synchronous motors have a stator design
similar to that of a standard A.C. cage induction motor with a
three-phase balanced winding of either two or four-pole
configuration. It is the rotor design, which can take alter-
native forms, that affects the performance of a given size of
synchronous machine.
Cylindrical-rotor synchronous motors have not only a uni-
form air gap but also a rotor flux distributed sinusoidally in
space. This combines with the three-phase stator flux, which
in the air gap produces a sinusoidal pattern of mutual flux,
Figure 1.22 Synchronous motor rotor with flats machined resulting in a sinusoidal torque pattern.
on a solid core
Such a rotor has a suitably-distributed three-phase winding
excited by D.C., usually one rotor phase carrying the full
D.C. current value, the other two phases each carrying half
current, in a series/parallel connection. An advantage of the
permanent magnets synchronous motor is that it does not require reactive current
to magnetise the air gap so there is little restriction on the
radial air gap when compared with the induction cage
machine, where air gap is of great importance.
A common alternative to the cylindrical rotor is the salient-
pole machine. Generally, a three-phase four-pole winding is
employed, except on some small machines where a two-pole
winding is more convenient. Salient-pole rotors provide a
large space for the windings but generally give a flux wave
pattern rich in lower-order harmonics, although these can be
somewhat reduced by correct design of the pole tips. The
general symmetry per pole, with differing reluctances
between stator and rotor, results in a tendency for stator and
Figure 1.23 Synchronous motor rotor with permanent rotor poles to align. This reluctance torque allows the rotor
magnets to accelerate and run synchronously without rotor excitation
at small to moderate loads.
Normal excitation supplements the torque produced by
reluctance and allows the machine to hold synchronism with
have wide application in A.C. variable-speed drives above
transient load changes. To improve starting, rotor bars are
about 40kW. Permanent-magnet motors are considered
often built into the salient-pole faces with similar design
here, but only their more common application in the spe-
considerations as to the profiles of the slots and bars as in
cialised field of servomotors. Switched reluctance motors
a standard cage machine. With the salient-pole machine
are considered entirely separately in Chapter 5.2.
running at synchronous speed, for reasons already con-
A.C. asynchronous induction motors produce shaft torque, sidered, little current will circulate in the rotor bars. That
which is proportional to percentage slip, implying that with which does tends to reduce harmonic currents.
zero slip the machine produces zero torque. Synchronous
torque can be produced at speed: In the same way as the design and arrangement in modern
A.C. generators provide rotor excitation without the use of
either slip rings or commutators, such self-excited designs
=f/p
have now become general for A.C. synchronous motors in
where ns =synchronous rotational speed in min-1, f = view of the poor maintenance record of slip-ring systems.
frequency in Hz and p = number of field pole pairs. Self excitation is additionally attractive, since rotor D.C.
excitation involves relatively little power, perhaps only up to
This is achieved by a field winding, for convenience of about 2 per cent of the machine rating, derived from a
design generally wound on the rotor and D.C. excited so that common shaft-mounted A.C. exciter with rectifiers. The
it produces a rotor flux which is stationary relative to the excitation is capable of being regulated externally to the
rotor. Torque is produced when the rotating three-phase field machine. This is necessary where advantage is to be taken
and the rotor field are stationary relative to each other, hence of the power factor control capability of this type of
there must be physical rotation of the rotor at speed ns in machine, and where inverter-based variable-speed control is
order that its field travels in step with the stator field axis. At required.
any other speed a rotor pole flux would approach alternately
a stator north-pole flux, then a south-pole flux, changing the The so-called V curves for a typical synchronous motor
resulting torque from a positive to a negative value at a are shown in Figure 1.24. The operating power factor is
Chapter 1.4 17

increases the excitation field voltage proportionately, to


\ ,,"" full load
compensate for the effects of flux neutralisation.
t-
eD
The A.C. synchronous motor appears to have some attractive
L_
~ / o full load features for inverter variable-speed drive applications, par-
0
C~ ticularly at ratings of 40 kW and above. Not least is overall
-

cost when compared with an A.C. cage motor plus inverter,


c-
.m no load
or D.C. shunt-wound motor and converter alternatives. In
c"
applications requiring a synchronous speed relationship
E between multiple drives or precise speed control of single
large drives the A.C. synchronous motor plus inverter con-
trol system appears attractive: freedom from brush gear
lag .~ power"
-,-factor ,~ lead
! maintenance, good working efficiency and power factor are
I

field excitation current the main considerations.

Figure 1.24 V curves of a typical synchronous motor


A.C. SYNCHRONOUS MOTOR
CONSTRUCTION
determined by the relationship between load current and
excitation current. The dotted line indicates the minimum
With the wound-rotor A.C. synchronous motor the stator
stator current at the various loads, that is with excitation
remains generally the same as for the A.C. cage motor, with
adjusted to give unity power factor. For steady-state operation,
the rotor slots now accommodating a fully insulated wire or
adjustment of the excitation current will change the working
bar-wound rotor winding of the same pole number as the
power factor. With variable loading, and where variable-speed
associated stator with connections brought out to external
inverter control is involved and operating power factor is
terminals via slip rings and brushes.
important, conventional power-factor correction techniques
are employed. Motor excitation level is set to give optimum Synchronous A.C. motors above the low kW rating sizes are
power factor for the working conditions. similar in construction to A.C. alternators. Again, the stator
winding is similar to that of the A.C. cage motor, the rotor
In common with D.C. motors, increased line current partially
carrying the field excitation winding, typically of brushless
neutralises the excitation field m.m.f. Since shaft torque is
design, wound either upon a cylindrical rotor (similar to the
approximately related to the product of line and field current
wound rotor arrangement mentioned above) or upon salient
fluxes, increased line current can give a reducing torque
poles.
increase. A typical inverter for variable-speed control auto-
matically regulates the main stator voltage to be in proportion It is interesting to note that slip-ring induction motors can
to motor frequency. It is possible to arrange an excitation also be made to operate synchronously by supplying the
control loop, which monitors the main stator voltage and rotor with D.C. current through the slip rings.

4 BRUSHLESS SERVOMOTORS

GENERAL The synchronous motor stays in synchronism with the


supply, although there is a limit to the maximum torque
In Chapter 1.2 we saw that a three-phase stator winding of an which can be developed before the rotor is forced out of
induction motor produces a sinusoidal rotating magnetic synchronism. Pull-out torque will be typically between one
field in the air gap. The speed of rotation of the magnetic and a half and four times the continuously-rated torque. The
field is directly proportional to the supply frequency. In torque speed curve is therefore simply a vertical line, which
Chapter 1.3 we saw that the same is true for a synchronous indicates that if we try to force the machine above the syn-
motor. In this case the rotor has a D.C. excited winding or chronous speed it will become a generator.
permanent magnets designed to lock on or synchronise with
the rotating magnetic field. The synchronous machine with The industrial application of brushless servomotors has
permanent magnets on the rotor is the heart of the modem grown significantly for several reasons: reduction of price
brushless servomotor. of power conversion products; establishment of advanced
18 BRUSHLESS
SERVOMOTORS"General

motoring

o"
i-
o ns speed

generating
b S W-

Figure 1.25 Steady-state torque speed curve for a


synchronous motor supplied at constant
frequency

control of PWM inverters; development of new, more


powerful and easier to use permanent-magnet materials; C
e
development of highly accurate position controllers; the ! i

manufacture of all these components in a very compact


form. They are, in principle, easy to control because the
torque is generated in proportion to the current. In addition,
they have high efficiency and high dynamic responses can
be achieved.
Brushless servomotors are often called brushless D.C. ser- |

vomotors because their structure is different from that of A B


ordinary D.C. servomotors. Brushless servomotors switch
current by means of transistor switching within the asso- Figure 1.26 Principle of a rotating field
ciated drive/amplifier, instead of a commutator as used in
D.C. servomotors. In order to confuse, brushless servomo-
tors are also called A.C. servomotors because brushless at point A, and phases V and W are both negative. Therefore,
servomotors of synchronous type with a permanent magnet the direction of the current of each coil is as shown in
rotor detect the position of the rotational magnetic field to Figure 1.26a and the composite vector of the magnetic flux
control the three-phase current of the armature. It is now induced by the current is generated in the direction from N
widely recognised that brushless A.C. refers to a motor with towards S. If there is a rotor field intersecting the magnetic
a sinusoidal stator winding distribution which is designed for flux at right angles at that time, torque is generated to rotate
use on a sinusoidal or PWM inverter supply voltage. the rotor clockwise owing to the repulsive and attractive
Brushless D.C. refers to a motor with a trapezoidal stator forces between the magnets. At point B, magnetic flux is
winding distribution, which is designed for use on a square- generated 60 degrees further clockwise.
wave or block commutation inverter supply voltage.
It follows from the above that a continuously rotating field
can be obtained by making three-phase currents flow in the
PRINCIPLES OF OPERATION OF stator coil. If the sine wave phase and rotational position can
BRUSHLESS SERVOMOTORS be made to be always at right angles, it becomes possible to
make a highly efficient motor producing smooth torque
Introduction without using brushes.

The brushless servomotor lacks the commutator of the D.C.


motor, and has a device (the drive sometimes referred to as Torque Constant
the amplifier) for making the current flow according to the
In the armature of the motor of Figure 1.27, the current
rotor position. In the D.C. motor, torque variation is reduced
distribution is as illustrated. If the current flowing in the
by increasing the number of commutator segments. In the
conductors to the right of the symmetrical axis OO t is in
brushless motor, torque variation is reduced by making
the direction of (away from the reader), then current in the
the coil three phase and, in the steady state, by controlling
conductors to the left flows in the opposite direction of
the current of each phase into a sine wave.
(towards the reader). Assume that all the conductors in the
Figures 1.26a and b are cross-sectional views of a three- fight-hand half are under the north pole and all the con-
phase synchronous motor, with U +, U - , V +, V - , W + and ductors in the left-hand half are under the south pole of the
W - indicating the beginning or the end of the coil of each permanent magnet and the magnetic flux density has an
phase. When a motor is energised by three-phase alternating average value of B to simplify the discussion. Then the
currents as shown in Figure 1.26c, only phase U is positive torque RBIL should work on every conductor and the whole
Chapter 1.4 19

!
O force (back e.m.f.), and the direction of this force is opposite
to the terminal voltage applied. This value is directly pro,
portional to the rotational speed 9t (in radians per second)
and is given by the following equation:
E=Ke~
The proportional constant Ke in this equation is the back
e.m.f, constant.
S I~ I !~1 N
Note: the back e.m.f, constant is usually expressed in units
of V/kr.p.m. (where the voltage is the r.m.s, voltage and
kmin-1 is in thousands of revolutions per minute).
The back e.m.f, constant Ke can be expressed in terms of
other parameters. If the rotor is revolving at a speed of f~
radians per second, the speed of the conductor v is:
R v = f~R
Therefore, the back electromotive force e generated in a
conductor is:
Figure 1.27 Field flux and current distribution in a rotor
e = ~ RBL
If the total number of conductors is Z, then the number of
torque T around the axis will be: conductors in series connection is Z/2 and the total back
e.m.f. E at the motor terminals is:
T = ZRBLI
E = ~ RBLZ/2
= (ZRBLIa) / 2
By substitution we can express E in terms of the flux q) as:
where Z is the total number of conductors, R is the radius of
E = (q,Z/ZTr)f~
the rotor, B is the flux density linking the stator windings, L
is the inductance of the winding and Ia is the current from the We can therefore obtain for the back e.m.f, constant Ke:
motor terminal, which is equal to 2L
Ke = (Z/ZTr)ff
In this model, the magnetic flux is equal to:
It should be noted that Kt and Ke are only equal when a self-
-- 7rRLB consistent unit system is used. The international system of
units (SI) is one such system. For example, if Kt is equal to
Therefore, by substitution, we get:
0.05NmA -1, then Ke is equal to 0.05 Vsrad -1. As stated
T = (Z/Tr)~Ia/2 earlier, although it is normal to express the torque constant in
terms of NmA-1 it is more usual to express the back e.m.f.
Now let us consider this equation. The number of conductors
constant in terms of V/kmin-1 (volts per thousand revolu-
Z never changes in a finished motor. Because the magnetic
tions per minute).
flux ~I, is determined by the motor dimensions and state
of magnetisation, (Z/Tr)~b is a constant. Therefore, we can Stationary Torque Characteristics
conclude that the torque T is proportional to the armature
current Ia. A motor which uses permanent magnets to supply the field
flux is represented by the simple equivalent circuit of
We can therefore define the torque constant Kt as"
Figure 1.28. This is a series circuit of the armature resist-
Kt- (Z/27r)~ ance, Ra, and back e.m.f., E.

Therefore, we obtain If we ignore the voltage drop across the transistors, the
equation for the voltage is"
T = Ktla
V -- gala + ge[~
The torque constant is usually expressed in units of NmA-
(where the current is the r.m.s, current). The armature current Ia is:

It should be noted that the above torque equation is identical Ia-- (V-ge~)/Ra
to that of a D.C. motor with constant field (equation 1.4).
Therefore, from above, the torque T is:

Relationships between Torque and T - Kda


Back E.M.F. Constant -- ( K t / R a ) ( V - Ke[~)

As each conductor passes through the north and south Figure 1.29 shows the relationship between T (torque) and f~
magnetic poles, the electromotive force changes succes- (rotational speed) at two different voltages. The torque
sively. The total electromotive force on each coil merges to decreases linearly as the speed increases. The slope of this
the motor terminals. This voltage is the back electromotive function is a constant KtKe/Ra and is independent of the
20 BRUSHLESSSERVOMOTORS:Principles of Operation of Brushless Servomotors

terminal voltage and the speed. Such characteristics make the stator of a D.C. motor. In other words, the magnetic
the speed or position control of a D.C. motor relatively easy. field for generating torque is stationary in D.C. motors, but
rotates in brushless servomotors; conversely, the armature
The starting torque and the no-load speed (assuming no
revolves in D.C. motors and is stationary as a stator in
beating friction and windage loss) are given by:
brushless servomotors.

r, = K, z/1o Servomotors require rapid acceleration and deceleration,


and the maximum torque therefore has to be several times
~o = V/Ke
larger than the rated torque. As brushless servomotors,
unlike D.C. motors, do not have a commutation limit, they
CONSTRUCTION OF BRUSHLESS can be operated up to the boundary of high-speed rotation
SERVOMOTORS without decreasing the maximum torque. Further, in a
brushless servomotor, the primary area of heat dissipation
Figure 1.30 shows the structure of a typical brushless ser- occurs not on the rotating part but on the armature in the
vomotor. The windings are in the static armature, which stator, since the permanent magnets are mounted on the axis
is part of the stator, and therefore the rotor of a brushless of rotation. The heat dissipated in the stator diffuses into the
servomotor can be considered to be the equivalent of air through the flame. It is therefore relatively easy to cool
brushless servomotors. Moreover, brushless servomotors
provide more precise overload protection, because the
R~ temperature of the hottest part can be detected directly.

0 |i [ Stator Structure
A typical stator consists of an armature core and armature

TE= KeD windings. The armature core is made of laminating punched


silicon steel sheet of 0.35 to 0.5 mm thickness (laminated
core). In many cases, the armature core is slotted and skewed
to reduce torque tipple, which results in speed tipple. The
0
armature windings are similar to that in an A.C. motor and
Figure 1.28 Simplified equivalent circuit are usually of the distributed three-phase type. The windings
are usually designed according to the drive specification,
which requires either a sinusoidal or trapezoidal back e.m.f.
waveform. The factors governing the design are stator slot
number, pole shape, windings coil pitch, rotor pole number
slope= Kt Ke and magnet shape.
e~

IID vl Rotor Structure


E
0
B
The structure of a typical rotor is characterised by perma-
nent magnets fixed on the axis of rotation. The shapes of
the permanent magnets vary according to design and
can be classified into two principal types: cylindrical and
salient pole.
Fixing of the permanent magnets to the rotor is critical in
speed
the design of brushless servomotors. Various techniques
Figure 1.29 Torque speed characteristic have been applied to the adhesion of magnets in order to

Figure 1.30 Structure of a brushless servomotor


C h a p t e r 1.5 21

prevent the destruction of motors from centrifugal force the same time they need to be stable against any thermal
in high-speed rotation or that caused by repetitive rapid change.
acceleration and deceleration. Common methods used to
prevent separation of permanent magnet from the rotor Recent development in rare-earth magnets has also con-
surface are: binding the outer surface of the permanent tributed greatly to improve brushless servomotor perfor-
magnet with glass-fibre tape or yarn; using a thin stainless- mance. Rare-earth magnets have nearly the same residual
steel cylinder as a sleeve to cover the outer surface of magnetic flux density as that of an Alnico magnet and two to
the permanent magnet. Adhesives are used in combination three times higher coercive force than that of a ferrite
with either of these methods and are chosen with a linear magnet. These features greatly help in making modem
expansion coefficient which is comparatively close to that permanent magnet brushless servomotors of light weight and
of the permanent magnet and that of the axis of rotation. At high performance.

5 RELUCTANCE MOTORS

The reluctance motor is arguably the simplest synchronous synchronously in much the same way as for a permanent
motor of all, the rotor consisting of a set of iron laminations magnet rotor.
shaped so that it tends to align itself with the field produced
Reluctance motors may be used on both fixed-frequency
by the stator.
(mains) supplies and inverter supplies. These motors tend to
The stator winding is identical to that of a three-phase be one frame size larger than a similarly rated induction motor
induction motor. However, the rotor is different in that it and have low power factor (perhaps as low as 0.4) and poor
contains saliency (a preferred path for the flux); this is the pull in performance. As a result of these limitations their
feature which tends to align the rotor with the rotating industrial use has not been widespread except for some spe-
magnetic field, making it a synchronous machine. The cial applications such as textile machines where large num-
practical need to start the motor means that a form of bers of reluctance motors may be connected to a single bulk
starting cage also needs to be incorporated into the rotor inverter and maintain synchronism. Even in this application,
design, and the motor is started as an induction motor, as the cost of inverters has reduced, bulk inverters are infre-
the reluctance torque then pulling in the rotor to run quently used and the reluctance motor is now rarely seen.

saturable bridge

Q-axis teeth
fhJX
barri~
....

Q-axis

filled
with
aluminium
structurally reinforced
for mechanical strength

Figure 1.31 Rotor punching for a four-pole reluctance motor


22 A.C. COMMUTATORMOTORS

6 A.C. C O M M U T A T O R MOTORS

Single-phase A.C. commutator motors of small kW ratings Again, although many ingenious designs have been produced
are manufactured in large quantities, particularly for in the search for improved variable-speed performance, the
domestic applications and power tools. Development gen- most successful for ratings between 5 kW and 150 kW has
erally has been in the direction of improved commutation been the Schrage or rotor-fed machine, followed by the
and reduced cost. The series A.C. machine is basically stator-fed or induction regulator motor, which has been
similar to the series D.C. motor and remains the more manufactured in ratings in excess of 2000 kW.
important. Although not capable of commutation to the
Both designs combine commutator frequency changing and
standards required for industrial drives, the type can provide
variable-speed motor in the same frame, generating the
approximate speed control simply by crude voltage regula-
required slip frequency voltage in a primary winding, for
tion often facilitated by a single semiconductor switch.
injection into a secondary winding.
An alternative design of some interest is the single-phase The kW rating of the Schrage motor is restricted by the slip
A.C. commutator repulsion motor in which the armature rings, carrying as they do the total power of the motor.
brushes are shorted together, with the A.C. supply taken to
the field windings only. Limited additional speed control can The stator-fed or induction regulator motor is a some-
be achieved by angular brush shift - an inconvenient what similar arrangement which converts the mains fre-
method, particularly in machines of low rating and small quency injected voltage to slip frequency voltage for the
size. Starting torque is high for the repulsion motor and its rotor coils. A step-down variable ratio transformer is gen-
design survives nowadays in the form of the repulsion-start, erally connected between the line supply and the commu-
induction-run single-phase machine, where a centrifugal tator connection to give voltage/phase control. The parallel
device shorts out the commutator segments when the motor transformer connection gives shunt characteristics and
is up to speed. the series transformer connection, series characteristics.
The stator power feed of this design allows, as has been
More important is the A.C. three-phase commutator motor, mentioned, ratings over 2000 kW to be designed success-
which over past years has provided outstanding service fully although commutators now appear to be an increasing
where variable speed is required. disadvantage.

7 MECHANICAL AND ENVIRONMENTAL

MOUNTING OF THE MOTOR IEC 60034-7 standard enclosures


General Electrical machines have been categorised within this stan-
dard by the prefix IM (international mounting), a letter and
Internationally agreed coding applies to a range of standard one or two subsequent digits. It is unusual for the prefix IM
mountings for electric motors, D.C. and A.C. which covers to be used and it is more usual to see only a letter followed
all the commonly required commercial arrangements. by one or two digits e.g. B 3 - foot mounting.
Incorrect mounting of a motor can cause premature failure The most usual types of construction for small and medium-
and loss of production. All motor manufacturers will provide sized motors are shown in Table 1.1.
advice on the suitability of a particular build for a specific
application.
NEMA standard enclosures
Within IEC 60034-7 (EN 60034-7) there are examples of all
practical methods of mounting motors. NEMA publishes Motor mounting and location of the terminal box location is
alternative standards within NEMA standards publication designated in accordance with the arrangements shown in
no. MG 1 - Motors and generators. Table 1.2.
Chapter 1.7 23

Table 1.1 Usual types of construction for small and medium-sized motors

foot mounted
two bearing plates

B3 B6 B7

I iI 51.
/. . I/I/. a , l

B8
11 ! V6
V5
flange mounted
M...
(with through holes | |

on the beating plate) I I ,I i a

two beating plates Ii


I
1
I
.
!
[ I
.
!
I
{ Ii "1-"
l

B5 V1 V3
attachment to front
end of casting (as B5,
,-4-, ]
V1 or V3 with beating
plate removed) only one o 1
bearing plate J
B9 V8 V9
two beating plates flange
with holes on the
machine casing

B10 VIO V14


two beating plates flange
with threaded holes on the .4-,
beating plate I I 1

[il
l ; J

B14
T t
"-t"
I !

V18 V19
two beating plates feet plus
flange with through holes
on bearing plate

t.....,--....,.2.a

B3/B5
two beating plates feet plus
flange with threaded holes
on bearing plate

B3/B14
24 MECHANICAL AND ENVIRONMENTAL: M o u n t i n g o f t h e m o t o r

Table 1.2 NEMA standard enclosures

floor mountings

U.II=
F-1 F-2
wall mountings
~('~

I_Wd L.W-I
(ll
~Ut~
W-1 W-2 W-3

MU~
k 1

W-4 W-5 W-6


( '"~

k J

W-7 W-8
ceiling mountings
/ m N m

C-1 C-2

DEGREE OF PROTECTION IEC 60034-5

General It has already been stated that the designation used as defined
in IEC 60034-5 consists of the letters IP followed by two
All types of electric motor are classified in accordance with numerals signifying conformance with specific conditions.
the provisions which specify a standard coding to indicate
the degree of protection, afforded by any design, against IP 4 4
mechanical contact and against various degrees of ambient
contamination. characteristic letters
The designation used as defined in IEC 60034-5, (EN 1st characteristic numeral
60034-5) consists of the letters IP (international protection) 2nd characteristic numeral
followed by two numerals signifying conformance with
specific conditions. Additional information may be included Example of designation
by a supplementary letter following the second numeral.

Interestingly, this system is contained within NEMA MG1 The first characteristic numeral indicates the degree of
but is not widely adopted by the industry in the United protection provided by the enclosure with respect to persons
States. and also to the parts of the machine inside the enclosure.
Chapter 1.7 25

Table 1.3 First characteristic numeral

First characteristic numeral Brief description Definition

nonprotected machine no special protection


machine protected against solid no accidental or inadvertent contact with or approach to live and
objects greater than 50 mm diameter moving parts inside the enclosure by a large surface of the
human body, such as a hand (but no protection against
deliberate access); no ingress of solid objects exceeding 50 mm
in diameter
machine protected against solid no contact by fingers or similar objects not exceeding 80 mm in
objects greater than 12 mm in diameter length with or approaching live or moving parts inside the
enclosure; no ingress of solid objects exceeding 12 mm
in diameter
machine protected against solid objects no contact with or approaching live or moving parts inside the
greater than 2.5 mm in diameter enclosure by tools or wires exceeding 2.5 mm in diameter;
no ingress of solid objects exceeding 2.5 mm in diameter
machine protected against solid objects no contact with or approaching live or moving parts inside the
greater than 1 mm in diameter enclosure by wires or strips of thickness greater than 1 mm
in diameter
dust-protected machine no contact with or approaching live or moving parts within the
machine; ingress of dust is not totally prevented but dust does
not enter in sufficient quantity to interfere with the satisfactory
operation of the machine
dust-tight machine no contact with or approach to live or moving parts inside the
enclosure; no ingress of dust

Table 1.4 Second characteristic numeral

Second characteristic numeral Brief description Definition

nonprotected machine no special protection


machine protected against dripping water (vertically falling drops) shall have
dripping water no harmful effects
machine protected against dripping vertically dripping water shall have no harmful
water when tilted up to 15 from the vertical effect when the machine is tilted at any angle up to
15 from the vertical, from its normal position
machine protected against spraying water water falling as a spray at an angle up to 60 from
the vertical shall have no harmful effect
machine protected against splashing water water splashing against the machine from any
direction shall have no harmful effect
machine protected against water jets water projected by a nozzle against the machine
from any direction shall have no harmful effect
machine protected against heavy seas water from heavy seas or water projected in
powerful jets shall not enter the machine in
harmful quantities
machine protected against the effects of
immersion in water to depths of between
0.15 m and 1 m
protected against the effects of prolonged
immersion at depth

The second characteristic numeral indicates the degree of The most frequently used degrees o f protection for electrical
protection provided by the enclosure with respect to harmful machines are as given in Table 1.5.
effects due to ingress o f water.
Brushless servomotors are normally IP65 - this single fea-
ture is often the key reason for a user to select brushless
For open internally air-cooled machines suitable for use servomotors for specific applications such as in the food
under specific weather conditions and provided with addi- industry where w a s h d o w n is a c o m m o n requirement.
tional protective features or processes, the letter W m a y
be used. This additional letter is placed immediately after US Practice
the IP e.g. IPW 54. Similarly, the letter R is used to indicate
duct-ventilated machines (in such cases the air discharge It is c o m m o n practice for manufacturers o f electrical
must be located outside the r o o m where the motor is machines in the United States to adopt less formal desig-
installed). nations, as given in Table 1.6.
26 MECHANICALAND ENVIRONMENTAL:Degree of Protection

Table 1.5 Frequently used degrees of protection for electrical machines

First numeral Second numeral


0 1 2 3 4 5 6 7 8

IPO0 IP02
IPl l IP12 IP13 IP17 IP18
IP21 IP22 IP23

IP44
IP54 IP55 IP56
IP65

Table 1.6 Protection categories in the US

open drip proof (ODP) a machine in which the ventilating openings are so constructed that successful operation
is not interfered with when drops of liquid or solid particles strike or enter the
enclosure at any angle from 0 to 15 downward from the vertical; these are
motors with ventilating openings, which permit passage of external cooling air
over and around the windings
totally enclosed fan cooled (TEFC) the TEFC-type enclosure prevents free air exchange but still breathes air; a fan is
attached to the shaft, which pushes air over the flame during operation to help the
cooling process
totally enclosed air over (TEAO) the TEAO enclosure does not utilise a fan for cooling, but is used in situations where
air is being blown over the motor shell for cooling such as in a fan application; in such
cases the external air characteristics must be specified
totally enclosed nonventilated (TENV) a TENV-type enclosure does not have a fan
washdown duty (W) an enclosure designed for use in the food processing industry and other applications
that are routinely exposed to washdown, chemicals, humidity and other severe
environments

COOLING constant-torque load requiring constant motor current, as


discussed under A.C. motor characteristics, the motor has
General approximately constant losses over its speed range.

Very much related to the enclosure of the machine, but not The motor will require significant derating of its output, or
synonymous with it, is the method of cooling. All rotating cooling with constant-velocity cooling air provided by a
electrical machines designed for economy of materials and fan driven at constant speed independently of the motor
dimensions require an effective form of cooling to ensure shaft speed.
that internal losses are dissipated within the limits of the
D.C. motors and synchronous motors are generally of
maximum temperature rise for the class of winding insula- IP23 enclosure and therefore suitable for constant-velocity
tion employed, and so that bearing and surface temperature
forced ventilation cooling, either from a frame-mounted
rise figures are kept within safe limits. fan as already mentioned or, where the working environ-
The different historical periods of development between mental conditions are so difficult as to require an IP54
the D.C. and the A.C. motor resulted in the two having enclosure, from a remote fan mounted in a clean-air position,
different enclosures and cooling arrangements, the D.C. with ventilating ductwork between motor and cooling
motor as standard being drip proof (IP23) and force venti- fan unit.
lated, the A.C. cage machine being totally enclosed (IP44),
The D.C. IP23 machine is restricted to use in a clean-air
fan cooled (TEFC), with a shaft-mounted fan at the nondrive
environment or to one in which an air filter on the vent fan
end running within a cowl to duct the cooling air over
inlet gives sufficient protection to the winding and com-
a finned motor body. The result is that the standard D.C.
mutator. Where the working environment is so difficult as to
motor is readily force ventilated internally through its
require a totally enclosed machine, an attractive form of
winding space, Figure 1.32, to allow a wide speed range,
cooling is single-pipe motor ventilation. This employs a
whereas the standard A.C. cage motor is not.
remote fan, drawing air from a clean source and delivering it
It should be noted that for variable speed A.C. applications, through a pressurised duct system to an adapter on the
the shaft-driven internal or external fans of the standard commutator-end end shield (IC17 of IEC 60034-6/EN
IP23 or IP54 A.C. cage machine can produce a cooling 60034-6). Alternatively, double-pipe ventilation supplies
problem, as their cooling performance varies inversely as the cooling air from a remote fan as described, and in addition a
square of the shaft speed. At half speed they provide only discharge duct takes the used and warmed air away from the
25 per cent of the full-speed cooling effect. On a typical motor (IC37).
Chapter 1.7 27

: ~i :i~: : i

Figure 1.32 Typical arrangement of a forced ventilating fan on a D.C. motor


(courtesy Leroy-Somer)

Ducted ventilation possesses the advantage that injurious available both for the difficult and for the explosion-hazard
gases and contaminants are unlikely to invade the motor environment in the flameproof enclosure form. Flameproof
winding space from the working environment. Additionally, D.C. motors are of restricted availability, particularly above
the use of double-pipe ventilation and the introduction of 20kW, and on account of small demand are relatively
(say) a ten-minute purging period before main motor starting expensive.
may sometimes allow the IP55 enclosed D.C. motor to be
The more common forms of motor cooling are specified by
used in an atmosphere where there is an explosive gas risk,
IEC 60034-6 (EN 60034-6).
for example in the printing industry, where the volatiles from
some inks constitute a hazard.
Air Filters
Although ductwork costs have to be taken into account, such
single or double-pipe ventilated D.C. motors can prove an The fitting of an air filter to a forced ventilation fan can
attractive alternative to totally enclosed D.C. motors, with provide useful protection against internal motor con-
either closed-air circuit-air cooling (CACA) or closed-air tamination. However, heavy contamination of the filtering
circuit-water cooling (CACW), IC0161, IC3666 and IC3166. element can reduce cooling airflow markedly. Although
thermal devices protect against this circumstance, closer,
For the screen-protected IP23 D.C. machine, it is relatively
more direct protection is possible by using an airflow, or
easy to arrange the machine with a frame-mounted fan,
air-proving switch. This is generally arranged to monitor
blowing constant velocity air through the winding space. For
the air pressure driving the air through the motor winding
the much less popular IP54 totally enclosed fan-cooled
space. As filter contamination gradually builds up, pressure
machine, the fan being motor-shaft mounted, the mechanical
falls and the air switch will indicate an alarm condition
arrangement is less satisfactory, requiring a separate cooling
when the preset limit is detected. The drive system can be
fan motor to be mounted externally to the motor but inside
arranged either to shut down immediately or after a preset
the cooling fan cowl.
time interval to allow the driven machine to be cleared of its
This in turn can pose a problem when the application product or to give a warning to the machine or process
requires a tachogenerator feedback signal, since the tacho- operator of imminent shutdown.
generator is normally mounted coaxially at the rear of the
By providing accurate and reliable protection, particularly
motor. With close-coupled bearingless tachogenerators of
for A.C. cage motors (which on account of their nonlinear
short axial depth, however, it is possible to arrange satis-
electrical characteristics require special care in protection),
factory mounting which is compatible with the forced
the facilities inherent in modem electronic motor controls
cooling arrangement.
and associated sensors have enhanced the standard of
CACW and CACA machines are available as both A.C. and reliability of modem industrial drive systems.
D.C. designs. Although a relatively expensive solution, this
is worth investigating where difficult ambient conditions
DUTY CYCLES
preclude lower cost alternative enclosures.
The differing patterns of development mentioned give The capacity of an electrical machine is very often tem-
the A.C. cage motor the enormous advantage of being perature dependent, and therefore the duty cycle of the
28 MECHANICAL AND ENVIRONMENTAL: Duty Cycles

where PN and n N are the rated power and speed, respectively,


of the motor.
load
Short-time D u t y - $2
losses "////~//////////////////J///////////~ With a short-time duty cycle the on-load period is too short
for the motor to reach a steady thermal condition and the
subsequent off-load period is long enough for the tempera-
Omax

temperature
# ture of the motor to drop to that of the cooling medium (even
with the motor at rest). Starting and braking are not taken
into consideration. With $2 duty the load torque may be
greater than the rated torque, but only for an appropriately
short period.
Figure 1.33 Load, lossesand temperature for duty type S1 When specifying short-time duty $2 it is also necessary to
state the on period e.g. $2-30 min. The standard specifica-
tions recognise the following on periods: 10min, 30min,
60 min and 90 min.
application may significantly affect the rating IEC 60034-1
defines duty cycles as follows.
Intermittent D u t y - $3
Continuous D u t y - $1 An intermittent duty rating refers to a sequence of identical
duty cycles. Each cycle consists of an on-load and off-load
Continuous duty rating S 1 relates to a duty where the on- period with the motor coming to rest during the latter The
load period is long enough for the motor to attain a steady on-load period during one cycle is too short for the motor to
thermal condition. With rated load, this refers to its max- reach a steady thermal condition, and the off-load period is
imum permitted temperature. Starting and braking are not likewise too short for the motor to cool to the temperature of
taken into consideration on the assumption that the duration the cooling medium.
of these events is too short to have any effect on the tem-
perature rise of the motor. Short time overloads are accep- Starting and braking are not taken into account on the
table. An off-load dwell period is of no significance if it is assumption that the times taken up by these events are too
followed by a load run. The load torque must not exceed short in comparison with the on-load period, and therefore
rated torque. do not appreciably affect the heating of the motor.

The load torque during one cycle may be greater than the
Example o f motor selection for continuous duty S1
rated torque of the motor.
The application calls for a power P1 over a speed range of
When stating the motor power for this form of duty, it is also
nal ~_ rta ~_ na2 (min- 1). The motor is rated according to the
necessary to state the cyclic duration factor:
maximum load torque and the maximum speed na2.

The selected motor must comply with the following cyclic duration factor = (on time/cycle time) 100%
requirements:
The standards specify that the duration of one cycle must be
rated power: PN > P1 x (na2/nal) kW
--1
shorter than l Omin. Cases where the duty cycle is longer
rated speed: nN > na2 mm
than 10 min must be brought to the attention of the motor
rated torque: PN/nN >_P1/nal kW/min- 1
manufacturer.

load
l! load

losses losses

Omax ~max
temperature
temperature

Figure 1.34 Load, lossesand temperature for duty type $2 Figure1.35 Load, lossesand temperature for duty type $3
Chapter 1.7 29

load
load

losses losses

6)max

/ TL
(gmax

temperature \ temperature

Figure 1.36 Load, lossesand temperature for duty type $4 Figure 1.38 Load, lossesand temperature for duty type $6

With short cycle times starting and braking must be If the motor had a ks3 of 1.4 then we could choose a motor
taken into account and the motor temperature rise must be with an equivalent continuous duty rating of 56.52/
checked (see $5). 1 . 4 - 41 kW at 1200 min-1.

The standards specify certain preferred cyclic duration fac-


tors: 15, 25, 40 and 60 per cent. Intermittent Duty with Starting- S4
The motor rating for intermittent duty $3 may be increased $4 is similar to $3 but taking starting into account.
over that applied to S 1 duty by a factor ks3. This factor may
be of the order of magnitude 1.4. Motor manufacturers can
advise on this. Intermittent Duty with Starting and
Example of motor selection for continuous duty 53 Electric Braking- S5
torque M1- 0.5 kNm for a period tl -- 10 s $5 is similar to $3 but taking starting and electric braking
torque M 2 - 0.4 kNm for a period t2--30 s into account.
torque M 3 - 0.6 kNm for a period t3 = 5 s
cycle time T - 120s; motor speed = 1200min -1
Continuous Operation Periodic Duty- S6
cyclic duration factor = [(10 + 30 + 5) / 120] x 100 %
$6 is similar to $3 except that the duty cycle is such that the
= 37.5% motor has not returned to the temperature of the cooling
medium by the end of the off period. The cycle should be
select cyclic duration time of 40 per cent:
repeated until the temperature at the same point on the duty
r.m.s, motor torque - v/{[(0.52 x 10) + (0.42 x 30) cycle has a gradient of less than 2C per hour.

+ (0.62 5)]/[10 + 30 + 5]}


= 0.45 kNm Continuous Operation Periodic Duty with
Electric Braking- S7
motor rating PN = (27r/60) x 0.45 x 1200
= 56.52 kW @ 40 % cyclic duration time $7 is as $6 but taking electric braking into account.

load ~//~ ~~/~


load

losses
losses

~gmax
~gmax
temperature
temperature

Figure 1.37 Load, lossesand temperature for duty type 55 Figure 1.39 Load, lossesand temperature for duty type $7
30 MECHANICAL AND ENVIRONMENTAL: Duty Cycles

Duty with Nonperiodic Load and


load ~/////~ Speed Variations- $9
$9 duty cycles should be discussed with the motor
manufacturer as the effects of this duty cycle are heavily
losses dependent upon specific motor design philosophies.
I II II II I
~) rnax
temperature
Duty with Discrete Constant Loads - Sl0
As with the $9 duty cycle, $10 duty cycles should be
speed
t__ 4111 discussed with the motor manufacturer.

TERMINAL MARKINGS AND DIRECTION


Figure 1.40 Load, lossesand temperature for duty type $8 OF ROTATION

General
speed / Terminal markings and directions of rotation are set out in
the international standard IEC 60034-8 (EN 60034-8).
load NEMA also defines the terminal markings in NEMA stan-
dards publication no. MG1.
For clarity the two standards will be discussed separately.
losses

Omax IEC 60034-8


temperature ~..A ~ .... GENERAL

Figure 1.41 Load, lossesand temperature for duty type $9 IEC 60034-8 describes the terminal windings and direction
of rotation of rotating machines. A number of broad con-
ventions are followed:
windings are distinguished by a CAPITAL letter (e.g.
.p
PN U, V, W)
end points or intermediate points of a winding are dis-
tinguished by adding a numeral to the winding (e.g.
load
U 1, U2, U3)

0
I- "1" I 1 t DIRECTION OF ROTATION
Tc
Pv
IEC 60034-8 defines the direction of rotation as an
observer facing the shaft of the motor (viewing from the

losses

o
I 1 t
drive end).
In A.C. polyphase machines (without a commutator) the
direction of rotation will be clockwise when the alphabetical
sequence of the terminal letters of a phase group corresponds
0 01 $A02 Tc
with the time sequence of the supply/terminal voltages.

temperature
oN '1 = =

AO4
For D.C. machines see below.

TERMINAL MARKINGS FOR A.C. MACHINES


o 1 ff
rc The terminal markings for three-phase stator windings of
synchronous and asynchronous machines are marked as
Figure 1.42 Load, lossesand temperature for duty type SI0
shown in Figures 1.43 to 1.46.

TERMINAL MARKINGSFOR D.C. MACHINES


Continuous Operation Periodic Duty with
Related Load Speed Changes- $8 Individual windings

$8 is as $7 but with defined and cyclic load speed The terminal markings for D.C. commutator machines are as
changes. shown in Figures 1.47 to 1.51.
Chapter 1.7 31

U1 Vl Wl E I ~ OE2
Figure 1.51 Separately excited field winding with two
terminals

Compensated motor with compensating and


commutating windings for clockwise rotation
U2 V2 W2
Figure 1.43 Single winding with six terminals
E1

U V W
I E2
A1 ~ A 2
C1 C2
l
AlE1
I
or AE
I
C2E2
or CE
or A or C
or A1 or C2
Figure 1.44 Delta connection with three terminals
Figure 1.52 Compensated motor

U V W N
( ()
Compound motor with commutating windings
for clockwise rotation

1
w w

Figure 1.45 Star connection with four terminals

i A1 ( ~ A 2 B1 B2
F10 ~ OF2
Figure 1.46 D.C. excitation winding of a synchronous
AlE1 D2E2
ou/or AE ou/or DE
ou/or A ou/or D
ou/or A1 ou/or D2
A10 O OA2 Figure 1.53 Compound motor

Figure 1.47 Armature winding with two terminals


Shunt-wound D.C. m o t o r - connections for
clockwise rotation
B10 / r ' Y ' ~ ' - " ~ O B2

Figure 1.48 Commutating winding with two terminals

C10 ~ C2
A1
O A2

Figure 1.49 Compensating winding with two terminals E1 E2

Figure 1.54 Shunt-wound motor

D10 ~ O D2
Note: the direction of rotation will be clockwise, regardless
Figure 1.50 Series excitation winding with two terminals of voltage polarity, if connections are made as in Figure 1.54.
32 MECHANICAL AND ENVIRONMENTAL: T e r m i n a l M a r k i n g s and D i r e c t i o n o f R o t a t i o n

Series D.C. m o t o r - connections for Table 1.7 General guidance for terminal markings
clockwise rotation
armature A1, A2, A3, A4 etc.
brake B1, B2, B3, B4 etc.
D2
alternating-current rotor M1, M2, M3, M4 etc.
windings (slip rings)
capacitor J1, J2, J3, J4 etc.
control signal lead C
attached to commutating
winding
dynamic braking resistor BR1, BR2, BR3, BR4 etc.
D
field (series) S1, $2, $3, $4 etc.
field (shunt) F1, F2, F3, F4 etc.
line L1, L2, L3, L4 etc.
A2 magnetising winding El, E2, E3, E4 etc.
d (for initial and maintenance note: El,E3 and other
magnetisation and odd-numbered
Figure 1.55 Series motor demagnetisation of terminals should be attached to
Note: the direction of rotation will be clockwise, regardless permanent magnet fields) the positive terminal of the
of voltage polarity, if connections are made as in Figure 1.55. magnetising supply for
magnetisation
resistance (armature R1, R2, R3, R4 etc.
Series D.C. generator- connections for and miscellaneous)
clockwise rotation resistance (shunt field V1, V2, V3, V4 etc.
adjusting)
D1 shunt braking resistor DR1, DR2, DR3, DR4 etc.
space (anticondensation) H1, H2, H3, H4 etc.
heaters
stator T1, T2, T3, T4 etc.
starting switch K
thermal protector P1, P2, P3, P4 etc.
(e.g. thermistor)
D equalising lead = (equals sign)
neutral connection terminal letter with numeral 0

O A2
J

Figure 1.56 Series generator


Note: the direction of rotation will be clockwise, regardless of F1 shunt
voltage polarity, if connections are made as in Figure 1.56. field
O'
~D
e--
i
i
i.
comp comm
NEMA field field

GENERAL

NEMA MG1 provides for general guidance for terminal A1 C A2 F2


Figure 1.57 Shunt-wound motor anticlockwise rotation
markings for motors, generators and their auxiliary devices
facing nondrive end
as given in Table 1.7.

D.C. MOTOR DIRECTION OF ROTATION

The standard direction of rotation of the shaft for D.C.


motors is anticlockwise (counter clockwise) as viewed from
the end opposite to the motor shaft. F'Ip shunt
field
The direction of rotation depends upon the relative polarities t"
|
:)

of the field and armature, therefore if the polarities are both _


|
|

reversed then the direction of rotation will be unchanged. comm


Reversal can be obtained by transposing the two armature field
leads or the two field leads. The standard direction of rota-
tion for generators will of course be the opposite of the case -I-

for motors. A2 A1 F2

Connections for the three main types of D.C. motor are Figure 1.58 Shunt-wound motor clockwise rotation facing
shown in Figures 1.57 to 1.62. nondrive end
Chapter 1.7 33

The first of these categories includes temperature, altitude


comp comm series
and the effects of the weather (electric storms, for example).
field field field
The second includes effects of the electrical supply system
including system faults, voltage surges, voltage dips and
power switching effects.
Again, there are different standards. Countries using a 60 Hz
A1 C A2 $1 $2 supply system tend to follow the ANSI or NEMA stan-
dards and the remaining countries tend to follow the IEC
Figure 1.59 Series-wound motor anticlockwise rotation
facing nondrive end standards.

Temperature
comm comp series
field field field One of the key factors which determines the rating of
an electric motor is the temperature rise at full load of
its active materials. The permissible temperature rises for
various classes of insulation material are specified in the
+
standards. Both NEMA and IEC use 40C as the base
A2 C A" $1 $2 ambient temperature.
Figure 1.60 Series-wound motor clockwise rotation facing The life of a motor is equal to the life of its winding insu-
nondrive end lation (disregarding the wear of bearings, brushes, slip rings
F 1.[ . shunt or commutator which can be replaced at relatively low cost).
o '' field Any service conditions influencing temperature rise and thus
"" comp comm series the condition of the insulation must be given particular
"~|
|
field field field attention.
|
|
r . . . . . . . . . . . m
If a motor is installed in an ambient temperature above its
|

rated value, the permissible temperature rise will need to be


A1 C A2 $1 $2 F2
reduced to keep the absolute value of the maximum tem-
Figure 1.61 Compound-wound motor anticlockwise perature at its design level. This is a key consideration for all
rotation facing nondrive end users and high ambient temperatures must be discussed with
the supplier in order to ensure high availability and long life.

F 1_[ . shunt The temperature rise in the motor results from the losses
0 I i I caused by the conversion of energy (electrical to mechan-
( D i l l
I-- ical) which can be expressed in the following equation:
L
1.1.1 comm comp series
i
i field field field
i
i
i
Ploss - Pelec -- Pshaft
i

A2 c A1 Sl $2 F2
In practice, it is not the losses of a motor but its efficiency 07)
which is quoted; this is calculated as follows:
Figure 1.62 Compound-wound motor clockwise rotation
facing nondrive end Z] -- (Pshaft 100)/Pelec

The resulting energy losses are stored in the motor and the
A . C . MOTOR DIRECTION OF ROTATION greater part is dissipated to the surrounding atmosphere by
ventilation, the condition depending upon the heat storage
NEMA MG1 states quite clearly that terminal markings of
capacity of the motor and the temperature rise. With con-
polyphase induction machines are not related to the direction
stant load, the steady-state condition is reached when the
of rotation.
amount of heat produced by the losses in the machine is
For synchronous machines, numerals 1, 2, 3 etc. indicate equal to the heat dissipated. In continuous duty, this state of
the order in which the terminals reach the maximum equilibrium is typically reached for industrial motors after
positive values (phase sequence) with clockwise shaft about three to five hours. The resulting temperature rise of
rotation. the winding and other parts of the motor is the difference
between the temperature of the particular motor part and the
Again, it is best to consult with the manufacturer.
coolant temperature. It may be determined from the increase
in resistance of the winding:
AMBIENT CONDITIONS
0 -- [(Rw - RK)/RI,:] x (235 + tcold -- tcoolant)
Introduction where O is the temperature rise of the winding (C), t c o l d is
the temperature of the cold winding (C), tcoolant is the
There are two main categories of ambient condition:
temperature of the coolant (C), Rw is the motor winding
1 those due to geographic conditions resistance at operating temperature (~) and Rx is the motor
2 those which are man made winding resistance when cold (f~).
34 MECHANICAL
AND ENVIRONMENTAL:Ambient Conditions

As a rule of thumb the life of typical winding insulation 1.0


decreases by about 50 per cent for each 10C.
It should be noted that the flame temperature is neither a 0
-~ 0.9
criterion for the quality of the motor nor for the temperature
0'}
rise of the winding. An extremely cold motor may have c'-

higher losses and higher winding temperatures than an ~ o.8


extremely warm motor.

0.7 I I I I ~-
Altitude 0 1 2 3 4 5
voltage unbalance, %
Owing to the fact that air density reduces with increasing
altitude, it is necessary to allow for the resulting reduction in Figure 1.63 Derating factor due to unbalanced supply
cooling capacity of the air when motors are operated at
altitudes in excess of their rating.
It is normal for motors to be rated for a maximum altitude of
1000 m.
SUPPLY VOLTAGE UNBALANCE
Some manufacturers rate their machines for combinations
of ambient temperature and altitude without derating, for Voltage unbalance in per cent may be defined as:
example 40C/1000 m or 30C/2000 m or 20C/3000 m.
percentage voltage unbalance
It should be remembered that, although the outdoor
maximum voltage deviation from average
temperature at higher altitudes is usually low, the motors = 100x
average voltage
will probably be installed indoors at higher ambient
temperatures.
For example, with voltages of 400 V, 408 V and 392 V, the
The IEC recommendation is to reduce the permissible average is 400 V, the maximum deviation from the average
temperature rise by 1 per cent per 100 m above 1000 m. is 8 V and the percentage unbalance is [100 x (8/400)] =
2 per cent.
The operating altitude is important and should be specified
when purchasing a motor.
Voltage unbalance can produce serious overheating in A.C.
motors due to the high negative sequence currents which
Power Supply System flow with a relatively small out of balance voltage compo-
nent. This negative sequence voltage produces in the air gap
GENERAL of the motor a flux rotating against the rotation of the rotor,
tending to produce high currents. It is important, on all
The windings of any electrical machine must be designed to supplies where voltage unbalance may be a problem, to
operate on the supply to which it is to be connected. provide protective devices to trip the motor if the sustained
Further, it is necessary to coordinate the protection of the unbalance exceeds 3 per cent.
motor and its cables within the system.
A derating curve is given by NEMA in MG1 which
The constraint frequently imposed on any user of a supply should be applied to motors operated on an unbalanced
system is the maximum current or kVA which can be supply. The standard also recommends that motors
drawn during starting. This constraint may lead the user to should not be operated on supplies with a voltage unbalance
consider altematives to direct-on-line starting (DOL) such in excess of 5 per cent. The derating curve is given in
as, in order of increasing flexibility, star-delta starting, an Figure 1.63.
electronic soft start or a variable-frequency inverter. When
considering the supply constraints at start, it is important
to take into account the supply impedance to ensure that NOISE AND VIBRATION
there is sufficient voltage at the machine terminals to
provide sufficient torque capability to overcome the load General
torque.
Noise and vibration are both unwanted cyclic oscillations.
SUPPLY VOLTAGE VARIATION Vibration can be considered as structure-borne noise as
opposed to airborne noise.
It is laid down in IEC 60034-1 that motors must be capable
of delivering their rated output at supply voltages between
95 per cent and 105 per cent of the rated value. Vibration
NEMA MG1 describes conditions of supply variation Magnetic, mechanical and airflow inaccuracies due to con-
between 90 per cent and 110 per cent of the rated value. It struction lead to sinusoidal and pseudosinusoidal vibrations
states that operation under such conditions for extended over a wide range of frequencies. Other sources of vibration
periods of time may accelerate the deterioration of the can also affect motor operation, such as poor mounting,
insulation system of the motor. incorrect drive coupling, end-shield misalignment etc.
Chapter 1.7 35

Consider the vibrations emitted at the operating frequency, If the vibratory displacement is measured against frequency,
corresponding to an unbalanced load the amplitude of which the measured value decreases with the frequency. High-
swamps all other frequencies and on which the dynamic frequency vibrations are not taken into account.
balancing of the mass in rotation has a decisive effect.
If the vibratory acceleration is measured against frequency,
In accordance with ISO 8821, rotating machines can be the measured value increases with frequency. Low-frequency
balanced with or without a key or a half key on the shaft vibrations (unbalanced loads) cannot be measured.
extension. ISO 8821 requires the balancing method to be
marked on the shaft extension as follows: The maximum value of r.m.s, speed of vibration is the
variable chosen by the standards and is generally classified
H half-key balancing as in Table 1.8 for medium-sized machines.
F full-key balancing
N no-key balancing
The testing of the vibration levels is undertaken with either
Noise
the motor suspended, Figure 1.64, or mounted on flexible It is inevitable that even an economically designed and effi-
mountings, Figure 1.65. cient electric motor will produce audible noise, due for ex-
The vibration speed can be chosen as the variable to be ample to magnetic torsions and distortions, bearings or
measured (in mms-1). This is the speed at which the airflow. The latter type of noise is most predominant in two-
machine moves either side of its static position. pole and four-pole machines with shaft-mounted fans and
most D.C. machines where a separately-mounted fan is most
As the vibratory movements are complex and nonharmonic, common.
it is the r.m.s, value of the speed of vibration that is used to
express the vibration level. Procedures for testing for motor noise are clearly laid out
in the standards; although there are detailed differences
Other variables, which could be measured, are the vibratory
between them, the principles are the same.
displacement amplitude (in microns) or vibratory accelera-
tion (in m s-Z). Referring to Figure 1.66, a series of background sound-
pressure readings is taken at the prescribed points. The
motor will then be run on no load and at full speed. A.C.
motors will be supplied at rated voltage and frequency.
Synchronous machines will be run at unity power factor.

I Certain correction factors may be applied where the test


reading is close to the background reading.
In most industrial applications the motor is not the pre-
dominant source of noise and the overall situation must be
taken into account when making a noise analysis.

Many steps can be taken to reduce the noise of electrical


machines including:

the use of oil-lubricated sleeve bearings which are much


quieter than most other bearing types
careful choice of bearing lubricant which can affect the
noise of the machine
Figure 1.64 System for suspended machines (measuring careful design of the machine air circuits to minimise
points as indicated) ventilation noise

Although the D.C. motor is generally less troublesome than

I the induction motor, there are applications where special


action is needed. The following measures can be made to
I r i reduce noise:

i use a reduced magnetic loading (lower flux densities)


I~ ] ~'~ increase the number of armature slots
f ~
skew the armature slots (or, less commonly, the pole
shoes)
use continuously graded main pole gaps or flare the gaps
at the edges of the main pole
increase the air gap
brace the commutating poles against the main poles
use semiclosed or closed slots for the compensating
winding
Figure 1.65 System for machines with flexible mountings select the pitch of the compensating winding slots to give
(measuring points as indicated) minimum variation in air-gap permeance
36 MECHANICAL
AND ENVIRONMENTAL:Noise and Vibration

Table 1.8 Maximum r.m.s, speed o f vibration

Class Speed n (min- ]) Frame size H (mm)


8 0 < H < 132 132 < H < 225 225 < H < 315

N (normal) 600 < N < 3600 1.76 2.83 4.45


R (reduced) 600 < N < 1800 0.70 1.13 1.76
1800 < N <m 3600 1.13 1.76 2.83
S (special) 600 < N < 1800 0.44 0.70 1.13
1800 < N < 3600 0.70 1.13 1.76

a i

i i lm i

h : i :

~-- - -O- . . . . . . O- . . . . . . . ----0 ........ 0"

i
i
i

i
i
lm
i

O
b
i

....h jL
i

I
|
| i
i

I
i i
| i

b
O

:I
|
|
, lm |
i
|
I
|
...... - ~ .... 4 ........

Figure 1.66 Location of measuring points for horizontal machines


a vertical plane
b horizontal plane

use a twelve-pulse rather than a six-pulse D.C. drive or


fit a choke/reactor in series with the machine to reduce
the current ripple
mechanical
vibration
After the designer has taken whatever steps to minimise the
noise generation at source, it may still be higher than is
absorbed
energy(heat)
acceptable. To achieve the specification it may now be
necessary to apply external silencing. This may be in the
form of inlet or outlet air duct silencers or even the fitting of
a complete enclosure.
incident e n ~ transmitted energy
(airbornen o ~ (airbornenoise)
The use of acoustic partitioning requires a good know-
ledge of at least the octave band sound pressure levels
present in order that good silencing can be achieved. Air- reflected ~
borne noise striking an acoustic partition will, like other energy
forms of energy, be dissipated in various ways as shown in
Figure 1.67. Figure 1.67 Energy flow in an acoustically excited partition
Chapter 1.7 37

MOTORS FOR SPECIAL APPLICATIONS 100

Geared Motors 80

o~ 60
Standard industrial motors are often unsuitable as direct
drives for low-speed applications. Moreover, the use of O
L"
N 40
motors with low and medium ratings is uneconomic at low
speeds. Geared motors are available for such applications. 20
These units consist of a high-speed motor and a gear reducer
assembled to form an integral unit. 0 I
0 20 40 60 80 1O0
The hardened teeth of the gear wheel resist high stressing speed, %
and ensure long life of the assembly.
Figure 1.68 Torque~speed characteristic of a torque motor
Geared motors are widely used on single machines such
as tower cranes, lifts/elevators, construction machinery, in
agriculture etc., as well as in industrial plants.
environment, where explosive gas-air mixtures may occur in
dangerous concentrations.
Brake Motors
The decision as to whether an outdoor area or an enclosed
Mechanical brakes are often used in conjunction with motors location should be considered subject to an explosion hazard
instead of, or as well as, electrical braking circuits. These as defined by the relevant regulations and specifications
units consist of a motor and a brake assembled to form an rests entirely with the user or, in case of doubt, the compe-
integral unit. tent inspecting authority.
It is important to note that the brake may be rated to brake It should be realised that the degree of hazard is variable and
the motor and its load or may be rated to provide a holding this has led to the concept of area classification and the
duty only. A holding only brake will be quickly destroyed if development of design techniques to ensure that electrical
it is used to brake a load from speed. It is common practice equipment will operate safely in the specified hazardous area
particularly on brushless servomotors that the brake be rated zones.
for holding duty only.
Harmonised standards exist in Europe under the guidance of
In order to rate a brake motor correctly the following CENELEC. In North America, emphasis is also placed on
information about the application is necessary: test and accreditation, under the guidance of Underwriters'
Laboratories (UL) and CSA; however the nomenclature is
type of load and the type of duty of the unit
different from European practice.
frequency of braking cycles per hour
total inertia (motor, brake, gearbox (if fitted) and load)
CENELEC
referred to motor speed
load torque as a function of speed, referred to the motor The seven commonly recognised methods of protection, as
shaft published by CENELEC are as follows:
whether the load torque has a braking or accelerating
effect EN 50 014 general
braking time and braking torque required EN50015 (Ex)o
EN50016 (Ex)p
Brake motors can be designed to be fail safe - if the power is EN50017 (Ex)q
lost than the brake will automatically be applied. EN50018 (Ex)d
EN50019 (Ex)e
EN50014 (Ex)i
Torque Motors
Although comprehensive guidance on the selection of
Torque motors have been developed from the basic designs explosion-protected equipment is contained within the
of three-phase squirrel-cage induction motors. They are standards, the information to be considered falls into four
not designed for a definite output but for a maximum categories.
torque, which they are capable of delivering at standstill
(i) The type of protection of the apparatus in relation to
and/or at low speed (when supplied from a fixed-frequency
supply). They have a torque speed curve of the form shown the zonal classification of the hazardous area.
in Figure 1.68. The degree of protection required is dependent upon
the presence of ignitable contaminations of inflatable
gas or vapour in relation to the length of time that the
MOTORS FOR HAZARDOUS LOCATIONS explosive atmosphere may exist, and this is defined as
in Table 1.9.
General
(ii) The temperature classification of the apparatus in
Manufacturing processes in many sectors of industry can be relation to the ignition temperature of the gases and
described as hazardous by the nature of the operating vapours involved.
38 MECHANICAL
AND ENVIRONMENTAL:Motors for Hazardous Locations

Table 1.9 Protection in hazardous locations Table 1.12 Equipment suitable for different zones
Zone 0 a zone in which an explosive gas-air mixture Zone Type of protection
is continuously present, or present for long periods
Zone 1 a zone in which an explosive gas-air mixture is 0 (Ex)ia
likely to occur in normal operation (Ex)s - specifically certified
Zone 2 a zone in which an explosive gas-air mixture is not 1 any of the above plus
likely to occur in normal operation, (Ex)d; (Ex)ib; (Ex)p; (Ex)e; (Ex)s
and if it occurs will only exist for a short time 2 any of the above plus
(Ex)N or (Ex)n; (Ex)o; (Ex)q

Table 1.10 Temperature classification These recognised types of protection are as follows:

Maximum surface (Ex)o - oil immersed


temperature (C) All or part of the apparatus is immersed in oil to prevent
T1 450 ignition.
T2 300
(Ex)p - pressurised
T3 200
T4 135 Since it is not practical to manufacture explosion-proof
T5 100 motors in large sizes, it is common practice to employ
T6 85 pressurised motors for zone 1 applications. These motors
tend to be manufactured to normal industrial standards
except that special attention is paid to the sealing of all
Table I. 11 Ignition temperatures removable covers and to shaft seals. The motors must be
,,

totally enclosed. Cooling must be by air-to-air or air-to-water


Example of Ignition Suitable equipment regarding heat exchangers.
compound temperature temperature classification
(of) Before the motor is energised it must be purged with at least
five times its own volume of clean air to remove any
Acetone 535 T1 T2 T3 T4 T5 T6
flammable gases which may be present.
Butane 365 T2 T3 T4 T5 T6
Hydrogen 270 T3 T4 T5 T6
sulphide (Ex)q - p o w d e r filled
Diethyl ether 170 T4 T5 T6
Carbon 100 This form of protection is very unusual in rotating
T5 T6
disulphide machines.

( E x ) d - explosion p r o o f
Electrical apparatus must be selected to ensure that the All parts of the motor where igniting arcs or sparks
maximum surface temperature is below the ignition may be produced are housed in a flameproof enclosure.
temperature of the specified gas. EN50 014 gives temp- The sealing faces, cable entries, shaft glands etc. are made
erature classifications for equipment as in Table 1.10. with comparatively large gap length and limited gap clear-
ances to prevent ignition of the surrounding explosive
These classifications can be related to the ignition
atmosphere. During operation explosive mixtures penet-
temperatures given in Table 1.11.
rate only seldom into the enclosure. Should an internal
(iii) The apparatus subgroup (if applicable) in relation to the explosion occur, it is prevented from spreading to the
relevant properties of the gases and vapours involved. external atmosphere.
Explosion-protected electrical apparatus is divided into
two main groups: (Ex)e - increased safety
group I mining applications This type of construction is used for motors without com-
group II all nonmining applications mutators or slip rings, which do not produce sparks during
normal operation.
For some types of protection, notably flameproof
enclosures, it is necessary to subdivide group II This type of motor may be used in zone 1, with some qua-
according to the properties of the gases, vapours or lifications, and in zone 2 areas. It is required that all surface
liquids, since apparatus certified and tested for, say, a temperatures are kept within the ignition temperature of the
pentane-air mixture will not be safe in a more easily specified gas under all conditions of operation or fault.
ignitable hydrogen-air mixture. This has led to appa-
To avoid the danger of ignition in the event of a fault, sui-
ratus subgrouping i.e. IIA, liB and IIC.
table protective devices such as circuit breakers with mat-
(iv) The suitability of the apparatus for the proposed ched thermal characteristics should be used to protect the
environment. motor against overheating.
Explosion-protected apparatus appropriate to a parti- The worst abnormal condition that can occur without per-
cular zone is readily identified by reference, Table 1.12. manently damaging the motor is a stalled condition. Most
Chapter 1.7 39

motor designs are rotor critical, the rotor temperature Testing authorities
increasing more rapidly than the stator under stalled condi-
tions. The surface temperature of the rotor conductors is the The main US and EEC testing authorities are shown in
critical and limiting factor in this type of motor. Tables 1.15 and 1.16.

The te characteristic of (Ex)e motors is important and must Table 1.13 US and European temperature classification
be quoted on the nameplate. It is defined as the time taken
for a winding, when carrying the worst case current, to be European North American Maximum surface
heated up from the temperature under rated operating classification classification temperature (C)
(EN50 014) (NEC NFPA 70)
conditions to the limiting temperature, te must never be less
than 5 s. T1 T1 450
T2 T2 300
T2A 280
(Ex)i - intrinsically safe
T2B 260
The concept of intrinsic safety is based upon restricting the T2C 230
electrical energy within the apparatus and its associated T2D 215
wiring to prevent the occurrence in normal operation of T3 T3 200
incendive arcs, sparks or hot surfaces. It is necessary to ensure T3A 180
T3B 165
that high voltages cannot be induced into the intrinsically safe
T3C 160
circuit. Shunt diodes are usually employed as barriers
T4 T4 135
between the intrinsically safe and the hazardous areas. T4A 120
This method is used in signalling, measuring and control T5 T5 100
circuits but is not practical for motors. T6 T6 85

(Ex)s- special protection


Table 1.14 US hazardous area classification
This protection concept allows for certification of equipment
Class I gas or group C ethyl-ether, ethylene,
which does not comply with the specific requirements of the
vapour cycle propane
established forms of protection. gasoline, hexane, naphtha,
group D
benzene, butane, propane,
(Ex) N a n d (Ex)n alcohol, lacquer vapours,
and natural gas
In the designation of type-N apparatus, the upper case N is Class II hazardous group F carbon black, coal or
used in the UK, but the lower case n has been proposed for dusts coke dust
the European standard having a similar concept. group G flour, starch or grain dust
Nonsparking motors, which are suitable for use in zone 2 Class III easily fibres easily ignitable but
ignitable not able to be suspended in
areas, are supplied in the UK to BS5000:Part 16. No part of
fibres air to produce ignitable
the motor may exceed 200C during normal operation but it
mixtures, such as rayon,
may do so during starting. nylon, cotton, saw dust
and wood chips

North American Standards


The principles applied in North America are broadly similar Table 1.15 European Economic Community approved
to those in Europe. The key differences are as follows. testing authorities

Temperature classification Country Name Location

Although the basic temperature classifications of the Belgium 1NIEX Paturages


Denmark DEMKO Herlev
european standards are retained, interpolation has occurred
France Cerchar Verneuil
between some T classifications giving greater resolution.
LCIE Paris
Table 1.13 shows the result (with cross reference to the Germany BVS Dortmund-Derne
European classification). PTB Braunschweig
Italy CESI Milan
The apparatus subgroup UK BASEEFA Buxton
In the US, the system of area classification and gas group-
ing is again different to European practice. Here the hazar-
dous area is divided into flammable gases or vapours Table 1.16 Main North American testing authorities
and combustible dusts. The key classifications are given in
Country Name Location
Table 1.14.
Canada CSA Toronto
Class I, group D is approximately the equivalent of the
USA Factory Mutual Norwood
European Group IIA with temperature classification to suit Underwriters' Laboratory Northbrook
the specific explosion hazard.
40 EFFECTSOF SEMICONDUCTORPOWER CONVERTERS:General

8 EFFECTS OF SEMICONDUCTOR POWER CONVERTERS

GENERAL Machine Rating-Thermal Effects


The control of electric motors by means of power electronic Operation of A.C. machines on a nonsinusoidal supply
converters has a number of significant effects. These are inevitably results in additional losses in the machine. These
primarily due to the introduction of harmonic components losses fall into three main categories:
into the voltage and/or current waveforms applied to the (i) Stator copper loss - this is proportional to the square
motor. In the case of A.C. machines which are normally of the r.m.s, current. Additional losses due to skin
considered to be of fixed speed there are additional impli- effect must also be considered.
cations which need to be considered including mechanical (ii) Rotor copper loss - the rotor resistance is different for
speed limits and the possible presence of critical speeds each harmonic current present in the rotor. This is
within the operating speed range. due to skin effect and is particularly pronounced in
deep bar rotors. Since the rotor resistance is a function
of frequency, the rotor copper loss must be calculated
DRIVE CONVERTER EFFECTS UPON independently for each harmonic. The increase in
D.C. MACHINES rotor copper loss caused by harmonic currents is very
often a significant component of the total losses, par-
The effects due to deviation from a smooth D.C. supply are, ticularly with PWM inverters which have significant
in general, well understood by drive and motor manufac- higher harmonics for which slip and rotor resistance
turers. The impact of ripple in the D.C. current clearly are high.
increases the r.m.s, current, which leads to increased losses (iii) Iron loss - this is increased by the harmonic compo-
and hence reduced torque capacity. The harmonics asso- nents in the supply voltage. The increase in iron loss
ciated with the current ripple lead to the now universal owing to the main fluxes is usually negligible, but
practice of using laminated magnetic circuits, which are there is a significant increase in losses due to end
designed to minimise eddy currents. With chopper con- winding leakage and slew leakage fluxes at the har-
verters, which are used in servo amplifiers and traction monic frequencies.
drives, frequencies in excess of 2 kHz can be impressed on
the motor. Special care is needed to select a motor with The total increase in losses does not directly relate to a
sufficiently thin laminations. derating factor for standard machines since the harmonic
losses are not evenly distributed through the machine. The
The ripple content of the D.C. currents significantly affects nonfundamental/harmonic losses mostly occur in the rotor
commutation within a D.C. machine. The provision of a and have the effect of raising the rotor temperature. Whether
smoothing choke can be extremely important in this respect, or not the machine was designed to be stator critical (stator
and recommendation should be made by the motor manu- temperature defining the thermal limit) or rotor critical
facturer depending upon the supply converter used. temperature clearly has a significant impact on the need for,
Apart from the thermal and commutation impacts, the ripple or magnitude of, any derating.
current also results in pulsating torque, which can cause Many fixed-speed motors have shaft-mounted cooling fans.
resonance in the drive train. Laminating the armature not Operation below the rated speed of the motor therefore
only improves the thermal characteristic of the motor but results in reduced cooling. Operation above the rated speed
also its dynamic behaviour by decreasing the motor time results in increased cooling. This needs to be taken into
constant. account by the motor manufacturer when specifying a motor
for variable-speed duty.

DRIVE CONVERTER EFFECTS UPON


A.C. MACHINES Machine Insulation
CURRENT-SOURCE INVERTERS
Introduction
Current-source inverters feeding induction motors have
It is often stated that standard off-the-shelf A.C. motors
motor terminal voltages characterised as a sine wave with
can be used without problem on modem PWM inverters.
the superposition of voltage spikes caused by the rise or fall
Although such claims may be largely justified switch-
of the machine current at each commutation. The rate of rise
ing converters do have an impact and certain limitations
and fall of these voltage spikes is relatively slow and only
do exist.
the peak magnitude of the voltage is of practical importance
NEMA MG1-1987, Part 17A gives guidance on operation in considering the impact on machine insulation. The
of constant-speed squirrel-cage induction motors for use on supply voltage never exceeds twice the crest voltage of
a sinusoidal bus with harmonic content and general purpose the sinusoidal waveform, and is consequently below
motors used with variable-voltage or variable-frequency almost all recognised insulation test levels for standard
controls or both. machines.
Chapter 1.8 41

Current-source inverters feeding synchronous machines For supply voltages less than 500 V A. C.
are even gentler on insulation systems, as the sinusoidal Check that the motor has the capability to operate with a
terminal voltage is reduced during commutation producing PWM drive. Most reputable motor manufacturers have
the same effect as notching on the supply associated with assessed their products for drive applications and can give
supply converters. an assurance of compatibility.
Alternatively, Figure 1.69 shows the peak voltage/rise-time
VOLTAGE-SOURCEINVERTERS
withstand profile, which is required for reliable operation.
PWM inverter drives are used with standard induction The motor supplier should be asked to confirm this cap-
motors in very large numbers throughout the world, and their ability. Figure 1.69 also shows the capability of a typical
advantages are well known in terms of improved energy good quality motor, which comfortably exceeds the
efficiency and flexibility of control. Occasionally drive users requirement. However, note that conformity of the motor
are advised to take special precautions over the motor with IEC 60034-17 alone is not sufficient.
terminal voltage because of an effect sometimes referred to For supply voltages in the range 500 V-690 V A. C.
as spikes or dv/dt which could possibly damage the motor Select an inverter-rated motor. An enhanced insula-
insulation. This section explains the effect and prescribes the tion system is required. The permitted voltage/rise-time
steps which should be taken to ensure that the motor insu- curve should equal or exceed that shown in Figure
lation system gives a long reliable life when used with a 1.70. Figure 1.70 also shows the capability of a typical
PWM drive. inverter-rated motor for use up to 690 V, which comfort-
ably exceeds the requirement. Note, however, that con-
Summary
formity of the motor with NEMA MG31 alone is not
The main effects of PWM drive waveforms on motor insu- sufficient.
lation are as follows: (b) Alternative a p p r o a c h - use additional preventative
Motor winding insulation experiences higher voltages methods
when used with a PWM inverter drive than when driven It may not be possible to follow the above recommen-
directly from the A.C. mains supply. This effect is dations, for example because the drive is to be retro-
caused by the fast-rising PWM voltage pulses which fitted with an existing motor or data is not available
result in a transiently uneven voltage distribution across for the motor concerned.
the winding, as well as short duration voltage overshoots
because of reflection effects in the motor cable. It is a In this case, additional preventative measures are
system effect that is caused by the behaviour of the drive, recommended. The most cost-effective measures are
cable and motor together. usually drive output line chokes, for lower power sys-
For supply voltages up to 500V A.C., the voltage tems, and motor cable termination networks, for higher
imposed by a correctly designed inverter is well powers. More details are given later.
within the capability of a standard motor of reputable (c) Factors affecting motor selection:
manufacture.
For supply voltages over 500V A.C., an improved star windings are preferable to delta windings
winding insulation system is generally required to ensure windings with single conductors are preferable to
that the intended working life of the motor is achieved. those containing parallel paths
When the motor used is of uncertain quality or cap- motor loading and duty should be carefully assessed
ability, additional circuit components can be added to to ensure that the motor does not over heat -
protect it. the insulation system is degraded by excessive
temperature.

Guidance for avoiding problems and Special cases


explanation of the phenomena involved 1 High braking duty
Where the drive spends a large part of its operating time
1 The voltage at the drive terminals is limited within tight in braking mode, the effect is similar to increasing the
bounds by the drive circuit. The motor cable increases supply voltage by up to 20 per cent and the relevant
the peak motor voltage. In applications with short motor precautions must be taken for the higher voltage.
cables (i.e. 10 m or less) no special considerations of any
2 Active front end (regenerative/sinusoidal/unity power
kind are required.
factor input drives)
2 Output inductors (chokes) or output filters are sometimes For drives with active front ends (regenerative and/or
used with drives for reasons such as long-cable driving unity power factor) the effective supply voltage is
capability or radio frequency suppression. In such cases increased by up to 15 per cent and the relevant precau-
no further precautions are required because these devices tions must be taken for the higher voltage.
also reduce the peak motor voltage and increase its rise
time. Special control schemes
Some drive designs using flux vector control with fast
3 In all other cases the following guidance should be
acting flux orientation can generate continuous double
followed:
pulses where the output voltage changes by twice the
(a) Preferred approach - select a suitable motor D.C. link voltage in a single step. This can result in four
42 EFFECTS OF SEMICONDUCTOR POWER CONVERTERS: Drive C o n v e r t e r Effects u p o n A.C. M a c h i n e s

2.6
2.4
2.2
2.0
> 1.8
~11"~ l/ typical standard motor for up to 500 V

'

,01---
0.8
0.6
0.4
02
"0 i t-
O 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 2.2
voltage pulse rise time, ps (IEC definition)

Figure 1.69 Peak voltage~rise-time profile requirements for supplies up to 500 V A.C.

2.6 I
2.4
2.2
2.0 minimum requirement up to 690 V

> 1.8

; 1.2-I II
,otll
0.8
0.6
0.4
0.2
0 l l l l

0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 2.2
voltage pulse rise time, Its (IEC definition)

Figure 1.70 Peak voltage~rise-time profile requirements for supplies up to 690 V A.C.

times the D.C. link voltage appearing at the motor The use of an inductor with 2 per cent impedance at the
terminals, causing increased stress and possibly pre- maximum output frequency is sufficient to lengthen the
mature motor failure. The stress is so extreme that a rise time to a point where it is no longer a consideration -
combination of inverter-rated motor and additional 5 ps is easily attainable. The natural high-frequency
measures such as line chokes may be required to prevent loss in a standard iron-cored inductor gives sufficient
motor damage. The drive supplier should be consulted damping, and this is more cost effective than using a
for detailed guidance in this case. Please note that Control low-loss inductor with separate damping components.
Techniques drives do not use this form of control.
Commercially available dv/dt and sinusoidal filters
should not normally be considered purely for motor
Additional preventive measures
protection, since their cost is excessive. They may,
The two most cost-effective techniques are:
however, be specified for other reasons such as EMC or
Output inductors (chokes) and output filters motor acoustic noise.
These are all connected at the drive in series with its
Choice of inductor
output. They all work by forming a low-pass filter in
The inductance should be chosen so that the impedance
conjunction with the motor cable impedance, thus
does not exceed 3 per cent pu at maximum frequency,
reducing the rate of rise of the drive output voltage.
otherwise the voltage drop will cause significant loss of
Some overshoot still occurs, which is controlled by
torque at high speed.
damping or clamping. This results in some power loss,
which must be allowed for in sizing the inductors or Conventional iron-cored inductors are suitable. Allow-
selecting the filter. The loss is roughly proportional to the ance should be made for additional core loss because
motor cable length and the drive switching frequency. of the presence of high frequencies. Special low-loss
Chapter 1.8 43

.. .... ,. .... , .... ,. .... ! .... ... ~ ~ '. ,.


high-frequency inductors should not be used because
severe resonance problems can occur.
Individual phase inductors or three-phase inductors are
equally effective.
Other benefits
Reduced loading effect on the drive from the cable
........... .,771"11i1i..71 ii,
capacitance
Reduced radiofrequency emission from the motor cable
(EMC) ll
i ........ .......... .... i ........ i ...... . .......
Disadvantages .......... : ......... l ......... : ......... i ........ ~ ........ ": ......... "t ........ ": ......... i...

Voltage drop oov looiJs i


. . . . i . . . . : . . . . . . . . . ; . . . . . . . . . . . . . . . . . . ; . . . . . . . . .
Power loss
Cost is modest at low current ratings but increases Figure 1.71 Motor terminal voltage with inductor
rapidly with increasing rating
For power levels above 100 kW the inductance at 3 per
cent pu may be insufficient for the purpose ' ' ' ' i ' ' " ' I . . . . i ' ' ' ' I . . . . | ' ' ' ' i ' ' ' ' I ' ' ' ' i . . . . ! . . . .

+
A.

2 Motor cable terminating unit


. . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . ~I . . . . . . . . . . . . . . . . . . . . . . . . . .

With increasing drive rating the above methods become . . . . T


t
+
" " "

. . . . . . .

increasingly expensive since they have to pass the entire . . . . . . . . . . . . . . . . . . . . . . . . . ,-I . . . . . . . . . . . . . . . . . . . . . . . . . . .

: : : : + : : :
drive output current. For powers exceeding about 70 kW :I: . . . .

it may be more cost effective to use a terminating unit. . . . . . . . . . . . . . . . MI max ~ . ~. . . . . . .. . . .. . . .. . . . . . . . . . . . . . . .

This is a resistor-capacitor network, which is connected .


. .
.
. .
. 800 v
.
. . .l-
.~.
.
.
. . . .

at the motor terminals in parallel with the power con- -~++-~ + + + ~ - ~ - ~ - + ~ ~ '-O-~H-~-H ++-,-+i + + + ~
nections and presents an impedance approximately 1

matching that of the cable during the pulse edges. This ..... i .... : .... i .... i .... ~- ....
,L
i .... : .... i .... : ......
. . . . ! " i i :
suppresses the reflection. It does not change the rise time
but it virtually eliminates the overshoot. It has the lUl
advantage of not carrying the drive output current, but
power loss tends to be greater than that for an inductor i .... i .... i .... i ......
and mounting at the motor terminals may be incon- / ~ , , , i . . . . i . . . . i . . . . i . . . . "t . . . . i . . . . ' . . . . " . . . . " . . . .

venient and require a special sealed construction to


match that of the motor. Figure 1.72 Motor terminal voltage with terminating unit
(note different time scale)
Figures 1.71 and 1.72 show typical waveforms produced
by these methods.
Precautions Table 1.17 Relative costs of alternative techniques
The unit must have an enclosure rating (e.g. IP Motor Drive Motor Output dv/dt Sinusoidal Terminator
number or NEMA category) suitable for the motor rating inductor filter filter
application.
(400 V supply)
Other benefits 2.2kW 350% 100% 74% 443% 334% 170%
None 75kW 220% 100% 14% 99% 146% 9%
250kW 120% 100% 5% 65% not 3%
Disadvantages practical
May affect the control of some kinds of flux vector or
other closed-loop controllers
Power loss
Technical explanation of the phenomena
Additional cost and inconvenience of motor terminal
mounting Review of PWM principles
The output voltage of the drive is a series of pulses
3 Output filters
with magnitude either + VD.C. or --VD.C., where VD.C. is
More advanced output devices are available, in the form
the drive D.C. link voltage, with pulse-width modula-
of dv/dt filters and sinusoidal filters. They have similar
tion (PWM). Because the motor load has inductance, the
benefits for motor terminal voltage, but since they are
current flowing and the magnetic flux in the motor comprise
relatively expensive they are unlikely to be cost effective
mainly the underlying low frequency of the pulse-width
unless they are also needed for other reasons.
modulation with a small ripple component at the switching
frequency. Figure 1.73 illustrates in simplified form a part of
Table 1.17 shows some relative costs of typical examples of
the output voltage waveform, with the associated motor
these alternative techniques. From this it may be concluded
magnetic flux.
that output inductors are the most economic measure for
systems rated up to about 70 kW, beyond which terminators VD.C. is typically about 1.35 times the r.m.s, supply
become more attractive. voltage, for example 540 V with a 400 V supply.
44 EFFECTSOFSEMICONDUCTOR
POWERCONVERTERS:
Drive Converter Effects upon A.C. Machines

overshoot lasts for about twice the time of flight in the


+VD.c.
voltage cable. If the rise time of the pulse is longer than twice
the time of flight in the cable, then the overshoot is
i1~~[~ I ~ time cancelled before it reaches 100 per cent.

UUI U
For a single pulse of magnitude VD.C., regardless of the
motor cable length, the overshoot can never exceed 100
--VD.c. per cent of VD.C.. However, the duration of the overshoot
L,__ does increase with increasing cable length.
For an ideal lossless cable, the rise time of the pulse is
Figure 1.73 PWM inverter output voltage and current maintained along the cable so that the rate of change of
voltage at the motor terminals (dv/dt) approaches twice
that at the drive. However, in practice, the cable exhibits
Table 1.18 Typical frequencies and times high-frequency loss, which causes an increase in the rise
time. This also means that the rise time at the motor
Frequency (Hz) Period/time
terminals is fixed mainly by the high-frequency behaviour
Power output 50 20 ms of the cable, so that contrary to statements sometimes
Switching 3000 333 #s made it is not the case that the introduction of new faster-
Pulse rise time - 100 ns switching power devices increases the stress on the motor.
Note that bipolar pulses, which have pulse edge magni-
Typical frequencies and times are given in Table 1.18. tudes of 2 VD.C., are also increased by 100 per cent so that
the total voltage during the reflection is then 4 VD.C.. These
Note the timescales. The rise time is five orders of magni-
can be generated by some kinds of drive with special
tude shorter than the output period.
vector control schemes.
Drive designers generally aim to use the highest practical
Figure 1.74 shows some typical measured voltage wave-
switching frequency, since this has a variety of benefits
forms, which illustrate the effect in practice. Even with
including reducing the audible noise from the motor. This
4 m of cable some overshoot is apparent. With 42 m the
means they are constantly seeking to use faster power
overshoot is virtually 100 per cent.
switching devices, which give lower switching losses
through shorter rise times. Winding voltage
All of the pulse edges in Figure 1.73 have amplitude equal to The voltage overshoot has little effect on the main motor
the D.C. link voltage. Standard PWM controllers only gen- insulation systems between phases and from phase to
erate these unipolar pulses. Some special control schemes earth, which are designed to withstand large overvoltage
without PWM modulators can generate bipolar pulses, pulses. Typical dielectric strengths for motors of reputable
which change from + VD.C. to -- VD.C. in one transition. origin are about 10 kV. However, some small low-cost
motors may have had economies made in the interphase
Motor voltage
insulation, which can lead to premature insulation failure.
The PWM pulse rise times are so short that the time for the
pulse to travel down the motor cable can easily exceed Because of its short rise time the pulse also affects the
the rise time. For example, the velocity of the pulse is insulation between turns, and especially between coil
typically 1.7 x 10s m s - 1, SOin 100 ns it has travelled only ends. The voltage pulse travels around the motor winding
17m. When this happens, analysis needs to rely on as it does along the motor cable. Figure 1.75 illustrates
transmission-line theory. Full details are beyond the scope how this results in a large part of the pulse appearing
of this guide, but the essential mechanism is as follows: across the ends of a coil during the time between it
At each pulse edge the drive has to charge the entering one end and leaving the other.
inductance and capacitance of the cable, so a pulse of In practice even in the largest low-voltage motors the
energy is delivered into the cable. The pulse travels at voltage between electrically adjacent turns is insignificant,
a velocity which is characteristic of the cable and is but between the ends of the coil it may briefly reach a
typically 1.7 x 108 ms-1.
substantial part of the pulse magnitude. In this simplified
When the edge reaches the motor terminals, a illustration the entire pulse voltage appears across the coil.
reflection occurs because the motor surge impedance In practice magnetic coupling between turns reduces this.
is higher than that of the cable (this is true for most Figure 1.76 gives a summary of the results ofmeasurements
low-voltage motors although the impedance does fall made with a range of rise times on a variety of motors.
as the motor rating increases). The voltage tends
With a sinusoidal supply voltage the coil ends only
towards double the step magnitude; i.e. there is an
experience a fraction of the phase voltage, as determined
overshoot approaching 100 per cent.
by the number of series coils. With a drive, therefore,
The reflection returns to the drive where it is again there is a considerable increase in the voltage stress
reflected, but in the negative sense because of the low between the coil ends. The effect of this depends on the
impedance of the drive. motor construction.
When this second reflection returns to the motor Large motors using form winding are constructed so that
terminals, it cancels the overvoltage. Therefore the the coil ends are not in contact. The interturn insulation
C h a p t e r 1.8 45

order of 1400 V, as could be generated by a drive with a


500 V A.C. supply. In the USA where supplies of this
level are common, many motor manufacturers routinely
use an inverter grade wire with further enhanced insula-
.................. !....M2;max...........................................................
6154 V tion withstanding at least 1600 V.
There is a possibility of a low-energy electrical discharge
'i'i'i'i'i'i'i'i'i
................................................... , , , , . i . . | . i . . . . , l , ,
effect called partial discharge, which can occur in voids
between wires. This is because of the electric field

"....
"....
'.'."
....
"......
'".'..'
.'.." concentration in such voids where the permittivity of the
gas or air is lower than that of the insulation material. At
every pulse edge a small discharge of energy occurs,
II
i :
which may gradually degrade the insulation system. If the
effect is excessive, the motor fails prematurely with an
20d V 400fis

interturn fault. Resin impregnation suppresses this effect,
. . . . . . . . . . . . . . . . . . . : . . . . . . . . . . . . . . . . . . . . . . . . . . .

as well as contributing to the physical stability of the


winding under high mechanical stress or vibration.
. . . . , . . . . ! . . . . ! . . . . ! . . . . ~ . . . . i . . . . I . . . . I . . . . I . . . .

For supply voltages higher than 500 V further measures


i ........ i......... i......... i ......... i .................
t i ......... i ......... i ......... i ........ are required to prevent partial discharge. Inverter rated
i i i M~ maxl 1 i i i i motors use inverter-grade winding wire, which is resistant
........ ; ......... !A .... ~ 4 v i ........ + ......... i ......... i ......... i ......... i .........
to partial discharge, as well as multiple impregnation
regimes to minimise voids, and enhanced interphase
insulation.
: : : : : : : : : : : : : : : : : : : : : : : : ...... ; .... ; .... ; .... i .... '

Motor standards
.................. i ..............................................................................

. . . . . . . . . . . J
The international standard IEC60034-17 gives a pro-
m file for the withstand capability of a minimum standard
motor in the form of a graph of peak terminal voltage
against voltage rise time. This replaces the older IEC34-
200 V 400ns

17 standard, which gave a rather arbitrary 500 V/~ts limit
without a clear rationale. The new standard is based on
. .
research into the behaviour of motors constructed with the

i
:
:
i
" '
i minimum acceptable level of insulation within the IEC
i
........ i ...... i ..... ~ ...... " ................... .: ......... " ......... ~ ......... " ......... motors standard family. There is a great deal of published
:: :: i i. i: :i
:
technical information on this subject. The best description
......... . ......... ~ M2m.~ ~ . . . . . . . . . . . . . . . . . . . i ......... ~ ......... i ......... ~ .........
! 13t~:d) kV i i ! : !
is contained in a paper written by workers at Dresden
University who carried out a major research exercise on
the subject: 'Failure mechanism of the interturn insulation
of low-voltage electric machines fed by pulse-controlled

m
". I inverters', M. Kaufhold et al., IEEE Electrical insulation
magazine, vol. 12, no. 5, 1996.

Tests show that standard PWM drives with cable lengths


of 20 m or more produce voltages outside the IEC60034-
!........i.........iiiiiiiiiiiiiiiiiiiiiiiiiiiiiiii'i'i'i'[iiiiiiiiiiiiiiiiiiiiiiiiiiiiii'i'i'il 17 profile. However, most motor manufacturers produce
! 06 i2.00.; i i i [ as standard motors the capability of which substantially
exceeds the requirements of IEC60034-17. Figures 1.69
Figure 1.74 Motor terminal voltage waveforms for vary- and 1.70 give the actual requirements for supply voltages
ing cable lengths (note scale changes) up to and exceeding 500 V, respectively. Standard motors
a cable length = 0.5 m are widely available to meet the requirements of Figure
b cable length = 4.0 m 1.69. Usually, a special inverter-rated motor is needed to
c cable length = 42.0 m meet the requirements of Figure 1.70. Such motors carry a
price premium of between 3 and 10 per cent depending on
then does not experience the high-voltage pulse. Smaller the rating.
random-wound motors may however have coil end wires Figure 1.77 gives some measured voltages for a typical
in contact, so then special attention is required to the system, showing that they exceed the IEC60034-17 limits
quality of the interturn insulation. but they do not exceed the capability profile of a typical
standard motor from a well known manufacturer.
Motor interturn insulation design
Modern motors of good quality manufacture use advanced This graph illustrates clearly the effect of lengthening the
winding wire, which has a multilayer insulation system motor cable. The rise time increases steadily with
and is easily capable of withstanding peak voltages of the increasing length, and the overshoot falls off after a peak
46 EFFECTSOF SEMICONDUCTORPOWERCONVERTERS:Drive Converter Effects upon A.C. Machines

motor termin~l=J ,
V VJ V I V F
,,.h
t=0 t=0!2T t=0'03T
i

)/ :
turns

;/
Voltage between coil ends: i/--~ coils
(7=propagation time around coil) ~ ' L

i i

',7",,

Figure 1.75 Propagation of pulse through motor windings

1.0 from one end of the rotor shaft to the other. If the bearing
breakover voltage is exceeded this will result in a current
flowing through both bearings. In some large machines it is
0.8
common practice to fit an insulated bearing, usually on the
nondrive end, to stop such currents flowing.
~ 0.6 This mains frequency issue is well understood and with
d. 0.4 modem motors such problems are rare.

0.2 Supply asymmetry


0 s eady s a e .~
An ideal power supply is balanced and symmetrical.
o ~ ~ Further, the neutral is at zero potential with respect to the
rise time, gs
system earth. With all modem PWM inverter supplies,
Figure 1.76 First coil voltage distribution against incident although it can be assumed that the supply feeding the motor
voltage rise time is indeed balanced and symmetrical in peak and r.m.s.
amplitudes, it is impossible to achieve perfect balance
at about 50 m. The voltage stress on the motor therefore between the phases instantaneously, when pulses of different
falls above quite moderate cable lengths. widths are produced. The resulting neutral voltage is not
NEMA publishes similar limits in the USA in MG1 part zero with respect to earth, and its presence equates to that of
31, shown in Figure 1.78. The measurements suggest that a common-mode voltage source. This is sometimes referred
these limits are insufficient for drives operating much to as a zero sequence voltage. It is proportional in magnitude
above 500 V. However, inverter-rated motors are readily to the D.C. link voltage (itself proportional to the supply
available with much improved capability, as shown. voltage), and a frequency equal to the switching frequency
of the inverter.
Bearing Currents This common-mode voltage will lead to the flow of currents
through stray impedances between the inverter phase con-
In theory the sum of the three stator currents in an A.C.
nections and earth. This includes motor cables.
motor is zero and there is no further path of current flow
outside the motor. In practice, however, there are conditions Considerable research into this complex subject has shown
which will result in currents flowing in or rather through the that the common-mode currents can be usefully considered
bearings of A.C. motors even when fed with a sinusoidal 50 in three distinct frequency ranges:
or 60 Hz supply.
a) Supply frequency (typically up to 100 Hz)
ROOT CAUSESOF BEARINGCURRENTS This type of current flow is usually related to motor
cable asymmetry and not power supply characteristics.
Magnetic asymmetry

It is well understood that an asymmetric flux distribution b) The switching frequency of the inverter (typically 1 to
within an electrical machine can result in an induced voltage 20 kHz)
Chapter 1.8 47

2.6
2,4 -

2,2 -

2.0 -

1,8 -

> 1.6-
I _/'- typical standard motor for up to 500 V
1.4
m
o 1.2 m.,0 m .....
," .[20 ,rrff"..:. . . . . . . . . . . . . . "0" lEG 34-17 (new version, up to 690 V)
1.0 100 m
i ...............
7
0.8-'"
0.2_0.4_0.
- 6 5Om
.......... i.................................................................................................... up(lversion,to
d 690 V)

,T" I I I I I I I I I I "~

0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 2.2
voltage pulse rise time, its (IEC definition)

Figure 1.77 IEC limits, manufacturers" limits and measurements


Test results at voltages with SWA cable lengths as indicated: 415 V A. C. supply; II 480 V A. C. supply

2.6 typical inverter-rated motor for 500V-690 V


2.4 s
s
s
s
2.2 $
$
$
2.0 AA A
A
1.8
NEMA (MG1 part 31) inverter-fed motor for up to 690 V
> 1.6
typical standard motor for up to 500 V
~ 1.4 mJ m
I
o> 1.2
m30 m50.m . . . . . . . ....................
.,""'1'
i "c"a'4: i T iX ; ; "u'p" "gg"o" i "
1.0 100 m .............(EC 34-17
II -

0.8 ................................ (Oldup


toVersion'6V)
90
5m
0.6
0.4
0.2
~- I I I I I I I I I I I ~"

0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 2.2
voltage pulse rise time, #s (IEC definition)

Figure 1.78 Limits and measurements for motors rated over 500 V, and NEMA MG1
Test results at voltages with SWA cable lengths as indicated: II 480 V A. C. supply; A 690 V A. C. supply

As described above, this is the fundamental frequency of Rather than consider the phenomena in further detail, it is
the common-mode voltage. Owing to the relatively low more helpful to consider what can be done to protect against
frequency, most of the currents which flow at this fre- the risk of bearing currents.
quency return to the inverter without passing through
motor bearings due to their high impedance. GOOD PRACTICESTO REDUCE
THE RISK OF BEARING CURRENTS
c) Common-mode resonant frequencies stimulated by
inverter switching (typically 50 kHz to 5 MHz) The first thing to remember is that although a great deal
This is considered to be the most critical and is has been discussed and published on the subject of
responsible for nearly all nonmechanically induced bearing currents associated with inverter-fed motors, it is in
bearing problems in inverter-fed motors. practice a rare event where the particular combination of
Frequency range can be limited by limiting the switch- motor construction, installation and inverter has caused a
ing time of dv/dt of the PWM pulses. problem.
48 EFFECTSOF SEMICONDUCTORPOWERCONVERTERS:Drive Converter Effects upon A.C. Machines

That said, any motor may be subject to beating currents Table 1.19 Maximum motor speeds and balancing for L-S
MV (2/4/6 pole) motors
if its shaft is connected to machinery at a different
ground potential than that of the motor frame. In order Motor type Maximum speed Balancing
to eliminate motor flame voltage it is necessary that a (min-1)
grounding strategy is adopted to keep all system components
80 15 000 S
grounded at the same potential. This needs to be achieved
90 12 000 S
for all frequencies, not just the 50/60Hz which many
100 10 000 S
grounding practices were based upon. This means avoid- 112 10 000 S
ing high-inductance paths - keeping cable runs as short as 132 7500 S
possible. 160 6000 R
160 LU 5600 R
Define a low-impedance path for the common-mode cur- 180 5600 R
rents to flow back to the inverter. As the common-mode 200 4500 R
current flows through the three motor conductors (cable), the 225 ST/MT/MR 4100 R
best return path would be through a shield around that cable. 225M/MK 4100 R
This could be in the form of a screen. Obviously, it is 250 4100 R
280 SP 3600 R
necessary to connect the screen at both the motor and the
315 3000 R
inverter, although this is in conflict with conventional
practice on screening. Such measures are well defined by
most reputable manufacturers in their EMC guidance. A
conduit would act in the same way, but it is important to The following issues should be considered by the motor
ensure that the conduit is capable of providing a reliably manufacturer when sanctioning use in the overspeed range:
continuous high-frequency path. Conduit is designed to
mechanical stress at the rotor bore and assurance that the
provide mechanical protection and may not be electrically
shaft to core fit is secure
continuous. Further, care needs to be taken to ensure that the
beating life, which is a function of the speed for anti-
cable lies within conduit for the entire distance between
friction beatings; each beating has overspeed and tem-
motor and inverter.
perature limits which need to be reviewed
In all cases take great care with the terminations of the beating lubrication, which is also a function of speed and
screen. All terminations must be of low resistance and low operating temperature; grease may not adhere properly
inductance or the benefit may be greatly attenuated. and oil may churn or froth
vibration, which is a function of the square of the angular
An obvious action is to use symmetrical motor cables. Take velocity; care must be given to ensure no operation near
care to ensure that the ground cores in the cable are sym- system critical or natural frequencies
metrically arranged to avoid asymmetrical induced currents airborne noise can be dramatically increased at higher
in the motor cable. speeds
winding stress caused by vibration of the windings at
SUMMARY high frequency may require additional winding bracing
and treatment
Bearing currents in inverter-fed motors is a complex proper attachment of balancing weights affixed to the
subject area. It has received wide publicity and is a prac- rotor or fan assembly; at high speeds shear stress levels
tical problem but only in very limited situations. The may be exceeded
number of motor bearing failures due to beating currents is methods of shaft coupling should be reviewed; this
very small compared with mostly mechanical reasons for would apply to any other auxiliary devices attached to
failure. the motor shaft, notably including speed and position
transducers
The higher the supply voltage the greater the potential risk.
the speed rating and energy absorption capability of
Good system grounding and cabling practice are critical in brakes
defining the risk of beating current flow. maintaining acceptable internal stress levels and fits of
cooling fans
When a beating failure has occurred and beating currents are
decreased motor efficiency caused by increased losses
suspected, detailed analysis at an experienced tribology
motor torque capability at increased speed
laboratory is necessary to identify the cause.
As an example, Table 1.19 gives the maximum speeds which
can be tolerated by Leroy-Somer MV motors in horizontal
Overspeed and vertical operation, directly coupled to the load and with
Most standard industrial motors may be capable of operating no radial or axial loading.
at speeds above their 50/60 Hz rating; however it is impor-
tant to have the manufacturer's assurance on the suitability
of any motor for operation above base speed. MOTORS FOR HAZARDOUS LOCATIONS
The bearings and type of balancing of the standard rotor The effects of converter supplies described above are, of
dictate a maximum mechanical speed, which cannot be course, applicable to motors designed for use in hazardous
exceeded without endangering the motor or its expected life. locations. Traditionally, it has been the responsibility of the
Chapter 1.8 49

user to ensure that the motor does not overheat as a result of power supply or both, should not be used in division 1
misuse. This has been achieved through the use of devices hazardous (classified) locations unless:
such as current-sensing protection relays. With converter
supplies, the situation is somewhat more complex and it is The motor is identified on the nameplate as acceptable
necessary for the motor manufacturer to assume the for variable-speed operation when used in division 1
responsibility for rating the motor correctly for variable- hazardous (classified) locations.
frequency inverter supplies. Thermistor sensors are mounted
at critical points in the motor, and used to monitor motor The actual operating speed range is not outside the
temperature during operation. Trip relays are used to remove permissible operating speed range marked on the motor
the supply from the motor if any one thermistor reaches the nameplate.
tripping temperature.
The actual power supply is consistent with the type of
In North America, NEMA MG 1-1987 17A.04.10 states that power supply identified in information, which is sup-
motors operated from variable-frequency or variable-voltage plied by the motor manufacturer.
CHAPTER 2

Power Electronics

@ 1 POWER SEMICONDUCTOR DEVICES 51

2 DRIVE CONVERTER CIRCUITS 72

1 POWER SEMICONDUCTOR DEVICES

GENERAL many of them are, in general, complex with integration of


many protection features such as overcurrent. For this
All A.C. and D.C. drives use power semiconductor devices reason, details of these circuits have been, for the most part,
to convert and control electrical power. The devices operate limited to a description of the requirements to gate the
in the switching mode (either on or off) which causes the devices.
losses to be reduced and conversion efficiency to be
improved compared to operation in linear mode.
DIODE RECTIFIER
The practically important power semiconductor devices in
relation to motor drives can be considered as follows:
The PN junction diode, Figure 2.1, is the simplest of all
diode rectifier semiconductor devices. It may be considered as an electro-
thyristor (includes phase control, fast and asymmetric nic switch the conduction state of which depends on the
types) polarity of an externally applied voltage. When a sufficiently
gate turn-off thyristor (GTO) high positive voltage is applied to the anode with respect to
bipolar junction transistor the cathode, current will flow in a forward direction, the
MOSFET device acting as a closed switch. The forward voltage drop
insulated-gate bipolar transistor (IGBT) across the device is typically one to two volts. Conversely,
integrated-gate commutated thyristor (IGCT) when a negative voltage is applied, current flow is prevented
other devices and the diode is able to block voltages up to a certain level,
This section reviews the important characteristics of these VRRM, which is the maximum reverse voltage that can be
devices. The electronic gate drive circuits for operating applied repetitively if breakdown of the PN junction is to be
52 POWER SEMICONDUCTOR DEVICES: D i o d e Rectifier

b forward current
anode forward volt

cathode
i, ~] applied to turn
rrm- " d i o d e off
b
C I

forward volt drop


diode on diode off

Figure 2.2 Diode switch o f f and reverse recovery


,i{--- forward
conduction
optimised for speed they tend to have higher forward voltage
leakage drops which restricts their current rating for a given chip
VRRM
current
/ size. Fast-recovery diodes find their main use in free-wheel
functions (where they must quickly commutate current
I
J I v
from and to primary switching devices) and high-frequency
v
t =IV
rectification.

~_ reve rse THYRISTOR


breakdown
The thyristor, or silicon-controlled rectifier (SCR), is a four-
reverse layer PNPN device shown in Figure 2.3. In the off state the
blocking
device can be considered as three diodes in series, so that
current conduction is prevented in either direction. Figure
2.3c shows that the reverse characteristic (cathode positive
Figure 2.1 Diode with respect to anode) is similar to that of the diode; how-
a symbol ever, the forward characteristic exhibits no current flow
b PN junction other than leakage current until the central control junction,
c VII curve J2, breaks over. Anode current/14 is then able to flow, limited
solely by the external load and supply capacity.
The forward breakover voltage is equal in magnitude to the
prevented. The voltage-current characteristic of the diode is reverse voltage, because in the blocking state J1 supports
shown in Figure 2.1 c, illustrating the two modes of operation. almost all the voltage, junction J3 breaking over at about
10 V. Once breakover in the forward direction occurs, the
Unfortunately, the diode does not behave as a perfect switch
thyristor behaves rather like a diode which has two junctions
when it is forced from a conducting state to a blocking state. A
(J1 and J3) because the gate P region is neutralised by for-
reverse current, with peak value Irrm, flOWS during the reverse
ward current flow. The overall forward voltage drop is
recovery time t,~, see Figure 2.2. During this time, stored
therefore between 1.5 and 2 V.
charges, responsible for forward current flow, are removed
from the PN junction. The total charge recovered, Qrr, along With the thyristor forward biased it is normally tumed on by
with lrrm and t,~, forms part of the diode specification; these injecting a positive pulse of current, Io, into the gate, causing
are important parameters in many applications as they deter- J2 to break down (assuming that the device has been forward
mine the energy loss in the diode each time it switches off. blocking). Once the anode current has exceeded the latching
current, the gate pulse can be removed. Typical waveforms
The reverse recovery time may be reduced by careful design
of gate current, anode current and anode cathode voltage
of the doping profile of the PN junction and by measures
during turn on are shown in Figure 2.4. Typical turn-on time
such as doping with particular elements or irradiating the
for a thyristor is several microseconds, depending on anode
junction with an electron beam. These features are designed
current.
to reduce the number of charge carriers in the diode and also
to reduce their lifetime so that I , ~ and trr are both reduced. For the thyristor to remain in the conducting state, the anode
A side effect of this is that the forward voltage drop current must reach the latching current level, It, and not fall
increases so there is a trade off between it and speed. below the holding current, 1/4, IL being greater than 1/4.
Generally, a diode used for low-frequency rectification of The thyristor is normally turned off by forcing the anode
A.C. to D.C. power has a long &r, and high Qrr as it has been current to zero by applying a reverse voltage for a minimum
optimised for minimum forward voltage drop. Diodes of this period of time before it can regain its forward-blocking state,
type are available in ratings as high as 8000 V and 6000 A. as shown in Figure 2.4. During the first stage of tum off a
Diodes with fast characteristics, that is short trr and low Q~r, reverse current flows because of stored charges for a time,
are referred to as fast-recovery diodes. As these have been trr, while junctions J1 and J3 recover. This process is similar
Chapter 2.1 53

b anode A
&

J1

l 1
Vforward Vreverse
gate G J2
J3

forward on-state volt drop q{-- values of V when thyristor gated

holding current
reverse leakage latching current
current ~L - -

_ )~) ) J
J
r
forward leakage forward breakdown
current
reverse
blocking

reverse breakdown

Figure 2.3
a thyristor symbol
b thyristor structure
c thyristor characteristic

- o._!~ I ,

stffte J
r e c ~ t q / reapplied

1
C ~
_r "v''"
' Iv. '
turn ~ VRRM
off I on j. on 4 turn off J. off

Figure 2.4 Thyristor switching waveforms


a gate current
b anode current
c anode voltage
54 POWER SEMICONDUCTORDEVICES:Thyristor

to a diode turning off. The control junction J2 needs an minimum levels, dependent on junction temperature,
additional time to recover, called the recombination time, tqr, which lie between the upper and lower resistance limits
and only then can a forward voltage be reapplied at a maxi- shown in Figure 2.5a. It is also necessary to ensure that
mum specified rate. The total turn-off time, tq is an important the peak gate power (Vc Ic) is not exceeded. Figure 2.5b
parameter for thyristors in fast switching applications. shows a typical gating characteristic, illustrating the
boundary conditions.
Thyristor Gating Requirements A simplified example of a pulse-transformer-based firing
The gate cathode characteristic of a thyristor resembles circuit is shown in Figure 2.6. Resistor R1 limits the gate
that of a poor PN junction and will vary between pro- current while R2 provides a low impedance across the gate
duction batches for a given type. To be certain of turning to attenuate any gate voltage when the thyristor is in the off
on the thyristor, the gate current and voltage must attain state. To achieve short turn-on times, the gate current is
required to rise at a minimum of 1 A/Its. A succession of gate
pulses, Figure 2.6b, supplied by the gate drive circuit, causes
a gate power boundary characteristicsof firing to occur when external conditions are suitable for
individual conduction.
VG~ .,. ~ \ thyristors
/ ~ ~,\~,"~ \ of same type Power Losses and Current Ratings
In normal operation the thyristor dissipates power in the

-2!5
/ J form of heat resulting mainly from:
forward conduction loss, which is a function of on-state
voltage and forward current
+25C switching loss, which is energy dissipated during turn on
and turn off
I / / / J I minimumlimit to ensure blocking leakage loss, which is a function of the off-
~ firingat given Tj state forward or reverse leakage current and blocking
voltage

IG The heat generated must be removed by a cooling system in

bl,Ov
order that the maximum junction temperature of the device
is not exceeded (usually 125C for a thyristor). The fol-
lowing equation describes the relationship between junction
temperature Tj and power dissipation PD for any semi-
VG conductor:
Tj = PD Rthjc + Tcase
where Rth/c is the thermal resistance, junction-to-case, and
Tcase is the case temperature of the device.
Clearly, the more efficient the cooling system, the grea-
ter the power that can be dissipated for a given case
temperature, which leads to greater current capability.

\
R1

24V t
0.1
10-3 10-2 10-1 100 101 102
A
IG ,,~

Gate characteristic and peak gate power dissipation


b 1A
Parameter A B C D
Pulse duration tp ms 10 1 0.5 0.1
Maximum allowable peak gate power W 40 80 100 150 o

Figure 2.5 Thyristor gate curves Figure 2.6 Thyristor gate circuit
a range of characteristics and limits a simplified firing circuit
b example characteristic b 20 kHz gate current pulse train
C h a p t e r 2.1 55

There is a finite limit, however, expressed by the r.m.s. Surge Current Ratings
current rating. This ensures that excessive heating of
internal joints and bonding wires is prevented. Thyristors It is possible for the junction temperature to be exceeded for
are therefore given an average current rating based on short periods of time under fault or overload conditions. The
practical case temperatures for a defined waveform, and an thyristor becomes predominantly resistive and it can be
r.m.s, rating, both of which must not be exceeded. Manu- shown that the temperature rise is proportional to:
facturers also provide graphs of average power versus IZ x tp
average current, Figure 2.7a, and allowable case tem-
where tp is the pulse duration and I is the r.m.s, value of the
perature versus average current for various waveforms,
pulse current calculated over time tp.
Figure 2.7b. From the graph in Figure 2.7a it can be seen
that average current decreases with duty cycle, the lines The equation assumes that all the heat generated is not
terminating at the point when the maximum r.m.s, current dissipated but stored in the mass of silicon. An IZt rating can
is reached. The graph in Figure 2.7b shows that the be specified and is a useful parameter in determining the
maximum junction temperature is 130C and all the curves size of fuse for overload protection. The thyristor is also
converge to this point as the current reduces. This type of given a nonrepetitive surge rating, half sine wave in shape,
information is essential for the selection of thyristor size for 10 ms and is typically equal to about ten times the r.m.s.
and for the design of the heat sink. current rating.

400

D.C.
180sin
120 J-I.
i 300 60 J-l.
30 Yl.

200

E
<
~ lOO
\

0 25 50 75 100 125 150

temperature, C (Tc)

b 400
RthKAK/W
l 0.1
0.2
0.3
0.4
0.6
~ 0.8
1.0
d.c.
in
6oO;
120o.1"1.

0 100 200 300 A 0 25 50 75 100 125 150


v

current, A(ITAvMIIFAvM) temperature, C (TA)

Figure 2.7 Thyristor on state current and loss


a average on-state p o w e r dissipation for rectangular current w a v e f o r m
b m a x i m u m allowable case temperature Tcasefor rectangular current w a v e f o r m
56 POWERSEMICONDUCTORDEVICES:Thyristor

The conventional thyristor, which is turned off by applica- forced commutation. High switching frequencies coupled with
tion of reverse voltage, has a structure which can be altered high rate of rise of on-state current cause the turn-on losses
to have characteristics to suit specific applications. Three to reach very significant levels: in some situations the device
types of thyristor can be distinguished: does not fully turn on owing to the limited current spread-
ing over the chip, demanding stringent current derating.
(i) Thyristor for A.C. line commutation (phase control Resonant load circuits, such as those used in induction
thyristor) - thyristors used in A.C. applications are heating, overcome the problems of excessive switching loss
turned off, or commutated, naturally by the existence of by switching on at the point where the load passes through
the A.C. supply, which changes polarity in alternate half either zero current or zero voltage. This technique allows
cycles. The thyristor is designed to have a low on-state fast thyristors to be used at up to 50 kHz. Furthermore, the
voltage, thereby maximising current rating at the load resonant circuit forms part of the forced commutation
expense of relatively long turn-off times (typically 100 circuit and in some circuit topologies additional auxiliary
to 200 Its). This does not matter because the thyristor thyristors are not required for commutation as this function
switching frequency is low. is performed by the main thyristors.
Equal forward and reverse voltage up to 12 000 V are
possible for large phase-control thyristors. For applica- GATE TURN-OFF THYRISTOR
tions onA.C, supplies up to 500 V A.C., it is usual to specify
1400 V types, to allow for an overload factor of two. The gate turn-off thyristor (GTO), like the conventional
It is common practice to use RC networks and varistors thyristor, can be latched into conduction by a short positive
across the thyristor to give additional protection. gate signal but, unlike the thyristor, the GTO can revert to
the forward blocking state by the application of a negative
(ii) Fast thyristors - these devices are generally used in gate signal. The GTO can therefore replace the fast thyristor
D.C. circuits such as choppers or inverters, although and its associated commutation circuits in D.C. switching
their use is now less frequent as more modern devices applications. The circuit symbol and the more complex GTO
such as IGBTs have replaced them in many applica- structure are shown in Figure 2.8. It is still a four-layer
tions. Within a D.C. circuit there is no natural reversal device with similar voltage blocking capabilities to those of
of the supply for thyristor commutation, therefore it the thyristor. The gate region is highly interdigitated with the
must be derived by external circuits. The process oftum cathode, producing a patterned structure which is designed
off under these conditions is called forced commutation to give high current turn-off capability by preventing con-
in contrast to line commutation. Typical commutating duction continuing in the cathode islands between the gate
circuits are expensive because they consist of inductors, contacts. This also ensures even current density across the
capacitors and auxiliary thyristors; however their size die during turn off. Although the physical operation of
can be reduced if tq, the total turn-off time, is kept to a the GTO is very complex, it is helpful to refer to the two-
minimum. The design of the thyristor is therefore transistor model of the GTO, Figure 2.8c, to understand how
optimised for low tq (typical values, 15 to 30 Its) but turn off is achieved. The devices may be considered as two
unfortunately this has the undesirable effect of increas- interconnected transistors which have regenerative action:
ing the on-state voltage drop, which consequently the collector current of one feeds the base current for the
lowers the current rating. other transistor. It can be shown that (neglecting leakage
(iii) Asymmetric thyristor - in many fast-switching appli- currents) the anode current IA is given by:
cations the reverse blocking capability of the thyristor
I G OLNpN
is not required because an antiparallel diode must be IA=
1 -- (OgupN -Jr- Olpup)
connected across the device for reactive current con-
duction. Manufacturers have exploited this relaxation
where Ia is the gate current, OlNp N and OLpNP are the common
by offering the asymmetric thyristor or ASCR, which
base gains, where a =/3/(1 +/3) and/3 is transistor current
has even lower tq times than the fast thyristor but at the
gain, Ic/I8. The current gains are dependent on the collector
expense of very limited reverse blocking. Turn-off current and increase as the current increases from zero.
times as low as 8 Its are possible while still retaining an
Conduction of the thyristor is initiated by a gate current
acceptable 15 V reverse blocking.
pulse which raises the loop gain (O~NPN--[-OZpNp)to unity so
Another technique for enhancing the performance of that from the above equation IA is infinite.
both ASCRs and fast thyristors is to use an interdigitated In practice, the anode current is limited by the load. Tum-off
gate structure which considerably increases the device
action is produced by extracting sufficient current from the
di/dt rating at turn on. This technique effectively gate to cause the loop gain to fall to a point where regen-
enlarges the turn-on area of silicon available at the start
erative action ceases. The turn-off gain ~ o f f is the ratio of
of gate firing, thus preventing excessive current density
anode current being controlled to negative gate current
near the gate which could lead to device failure.
required to produce turn off, and is an important parameter.
Typical values of ~off lie between three and five. To reduce
High-Frequency Current Operation the loop gain, and hence increase the tum-off gain, the gate
is often connected to the cathode with a low resistance
Although the ASCR and fast thyristor have been designed for within the package, emitter shorts, which has the side effect
high-speed operation, the maximum frequency that can be of making the gate less sensitive for tum on. Unfortunately,
switched in practice is limited to approximately 1-2 kHz with there is a limit to the maximum anode current that can be
Chapter 2.1 57

anode A current flowing at the time. The anode current then falls

I rapidly, in several microseconds, as the device recovers its


blocking capability. After the bulk of anode current has been
commutated, a small tail current flows due to trapped
// charges in the base N region, adding to the turn-off loss.
gate After turn off it is normal to reverse bias the gate cathode
with a 15 V supply to prevent retriggering of the device.
cathode C During switch on and switch off considerable power is dis-
interdigitated gate sipated as heat, limiting practical GTO switching fre-
quencies to around 2 kHz.
st ru ~//u/,l~re
/lilll,sct 0 A
Snubber Design
(base region) The GTO must have a snubber circuit connected across it to
limit the rate of rise of voltage at turn off. An example is
~lV p shown in Figure 2.10. Unlike the thyristor, forward voltage
is reapplied immediately after turn off (compare Figure 2.4c)
which means that the dv/dt limit usually has to be lower,
, T J necessitating a larger and more lossy snubber. It is also
important that the voltage spike, marked Vs in Figure 2.9a,
which is due to snubber circuit inductance, is minimised
A A through good layout. In fact, manufacturers stipulate that the
I maximum controllable anode current cannot be guaranteed
unless a certain size of snubber capacitor is used and Vs does
not exceed a specific level.

C c Voltage and Current Ratings


I~T 2 ~PNP = Like the thyristor, the GTO can be designed to have equal
IT
forward and reverse blocking capabilities, or limited
PNP(~ ~ | O~NPN=~E
1
reverse blocking with the advantage of improved turn-off
times. The latter type is sometimes referred to as an
c2 ~
,4
~IE1
` T1NPN
anode-short GTO and is designed to be used with an
antiparallel diode. The forward on-state voltage drop is of
~c the order of 2 V, giving current ratings similar to those for
Figure 2.8 The gate turn-off thyristor the fast thyristor.
a circuit symbol
b structure BIPOLAR TRANSISTOR
c two-transistor model of the GTO
The most common type of power bipolar transistor, some-
times referred to as the bipolar junction transistor or BJT, is a
switched off; generally this is about four times the average
three-layer NPN device as shown in Figure 2.11. PNP types
current.
are available but they tend to have inferior voltage and
current ratings.
Switching Characteristics and Gate Drive To conduct a collector current Ic, the transistor must be
supplied with a continuous base current 18, depending upon
Example anode-current, anode-voltage and gate-current
the voltage level between collector and emitter. This rela-
waveforms are shown in Figure 2.9. Turn on is initiated by a
tionship is shown in Figure 2.12.
relatively high amplitude gate pulse and takes about 3 to
5 las. Despite the highly interdigitated gate structure there is The ratio Ic/18 is called the current gain hFE and may be less
a practical limit to the rate of rise of anode current: therefore than l0 for a 1000 V transistor. The gain can be greatly
external inductors are sometimes used to limit di/dt and turn- improved if the base current is obtained from another tran-
on losses. After turn on, the gate current may fall to a lower sistor using the Darlington connection shown in Figure 2.13.
level, the 'back-porch' current, sufficient to minimise the The three transistor stages are integrated on the same silicon
on-state voltage drop. If a continuous gate drive were not chip giving an overall gain of several hundred.
provided, the volt drop would tend to increase at low anode
A typical curve of gain versus Ic for a 50 A 1000 V three-
currents owing to the fall in loop gain.
stage (Mitsubishi QM50DY-2H) Darlington transistor is
To turn the GTO off, the forward gate current is removed shown in Figure 2.14. This illustrates the dependence of gain
and a negative voltage source ( - 1 2 to - 1 5 V) is applied. on collector current. In practice, for motor drive applications
The resultant negative gate current increases rapidly until the power transistor is always operated as a switch. When
large enough to stop GTO regenerative action. This may take closed sufficient base current is provided to ensure that the
20 rts (the storage time ts) depending on the amount of anode transistor operates in the saturated or quasisaturated mode.
58 POWER SEMICONDUCTOR DEVICES: B i p o l a r Transistor

anode current

anode voltage

IGM IG
diGIdt~__~
...J=.~-,,,.~~ gatecurrent

diGQ/dt

IGQ

Figure 2.9 GTO gate waveforms

a collectorC

~
base B
iode
Rs

,[ emitter E
C

T
Figure 2.10 Typical GTO snubber lvc
Ic

Typically, VcE is 2 to 3 V for a Darlington operating in the


saturated region. When open a small negative bias is applied
to the base to ensure minimum leakage current.

Voltage Ratings
The transistor will break down if a sufficiently high collector Figure 2.11 Bipolar transistor
emitter voltage is applied, causing permanent damage. To
a structure
prevent this, various voltage ratings are defined: the two b circuit symbol with current directions
most important are:
1 VcEv- the maximum voltage between collector and
emitter with the base reverse biased (usually - 2 V). Current Ratings
This is the highest rating and is used to classify the
The continuous or D.C. current rating is the principal rating
device. VcEv ratings up to 1400 V are available.
used for classification. Although quoted at 25C case tem-
2 VcEo- the maximum collector-emitter voltage with the perature it is usually possible to achieve this rating at 100C
base open circuited. This rating tends to be the lowest without exceeding the maximum junction temperature, tjmax,
rating, being 50 to 90 per cent of VcEv. It is more assuming adequate cooling. The transistor can also be pulsed
important than VcEv and indicates how rugged the with twice the D.C. rating or ICM for 1 ms at 50 per cent duty
switching capability of the transistor is. cycle provided that peak collector power and tj'ma x a r e not
Chapter 2.1 59

continuous collector power dissipation

e-
/
~ s SI
x_
0
I!
0
s S
0

I -at fixed values of Ie


saturation
voltage %%% I

I
~ Ii
i

f
leakage current
J' I

~ IB=O

reverse c o l l e c t o r emitter v o l t a g e VCE lforward breakdown


breakdown voltage

,,-lOV __~ ~,l~ linear region


saturated region

Figure 2.12 Characteristic VCE curve

I I I I
VcE=5.0 v
T 103 s
s
7 I s

~ 5
s
s
f
- ~,"~:,
.i
~ 4 Vce- 2.8 V
E
~ 3 f
\\\\
E 2

O
\X',
d 102
a 7 \v,
i I
I I I I
i
. ---5=25oc , I

.... ~. =125C ,
Figure 2.13 Three-stage Darlington arrangement I I I I
100 2 3 45 7101 2 3 45 7102
collector current I o A
exceeded. Ratings up to 1000 A with VcEv equal to 1200 V
Figure 2.14 D.C. current gain curve
are available in a single module.

Switching C h a r a c t e r i s t i c a n d Base D r i v e
reverse current IB2 the level of which is controlled by the
Typical operating waveforms are shown in Figure 2.15 for
base drive circuit flows out of the base for a period called the
the power transistor switching an inductive load clamped by
storage time during which stored charges in the base region
a free-wheel diode - a situation encountered in chopper
are removed. Once this is complete, the collector voltage
circuits and voltage source PWM inverters. When the tran-
rises and then the collector current falls in 0.5 to 2.0 Hs.
sistor turns on (ton typically 0.5 to 2 ps) the collector current
Typical storage and fall times are 10 ps and 2 ItS, respec-
will see a peak current higher than the load current due to the
tively. The storage time may be reduced by operating the
reverse recovery current of the free-wheel diode. The peak
transistor in a quasisaturated state. This is achieved by
current must be less than IcM. Forward base current I81 is
controlling the base current so that only just enough current
then maintained to keep the transistor at or near saturation
is supplied to hold VcE at a value just above the saturated
level.
value. This may be achieved using a relatively simple diode
Turn off is initiated by quickly removing the forward base clamp circuit known as a Baker clamp. As the collector
current and applying a reverse voltage to the base emitter. A current falls any stray inductance present in the power circuit
60 POWER SEMICONDUCTORDEVICES: Bipolar Transistor

will cause the collector emitter voltage to overshoot as Safe Operating Areas
shown in Figure 2.15, making snubber circuits necessary.
During turn on and turn off, the instantaneous collector
The total power loss in the transistor will comprise switch- current and voltage are both very high for short periods of
on, switch-off and conduction losses. Switching losses are time, hence the switching losses. This is permissible pro-
almost double at 125C junction temperatures compared vided that the locus of Ic and Vce lies within specified safe
with those at 25C because of the temperature dependence of operating curves defined by the manufacturer. When the
the switching times, but the conduction loss remains nearly transistor is forward biased, for example during turn on, the
constant. Typical calculations for power and switching forward-biased safe operating area (FBSOA) curve is
losses are shown in Table 2.1. Generally, switching fre- applicable. For very short pulses it has a current boundary
quency is limited to between 2 and 5 kHz, switching loss and limited by lcu and a voltage boundary limited by the VcEo
storage time being the determining factors. rating as shown in Figure. 2.16a. For longer pulses the

1
I
L

.L
-C a
I
I +~,~ T Vcc

o peak current due to


reverse recovery of
J freewheel diode
ve~h;ot

t~
ts - - ~

%~ ~--
IB

~,. %2
to.

turn-on A turn-off
loss
loss " ~ ' ~ A conductionloss

Figure 2.15 Bipolar transistor half-bridge switching an inductive load


a test circuit
b collector current and voltage waveforms
c base drive
d power loss over one switching cycle
Chapter 2.1 61

Table 2.1 Calculation of switching and conduction losses a


forbidden area
for a bipolar transistor /c
Energy dissipated during turn on Wo, ~ 0.5 x t~c x Vcc x Ic ICM
Energy dissipated during turn off Woj:f ~ 0.5 x tic x Vcc x Ic (2 xlc)
Energy dissipated during Wcond ~ tcond X VCEsat X I c
conduction period forward bias
Switching loss (watts) esw :fsw X (Won "JI-moff) E)3
(-
safe operating area
Conduction loss (watts) PcoM = f~w x Wco,,a repetitive
Total average power loss Ptotal = P~w + PcoM
= Ic Tj ~< ~2SC
0
(-
c- tp defined
O
maximum power is limited by the junction temperature 0 (typically <1 ~s)
thermal limit and a complex phenomenon called secondary
breakdown. Secondary breakdown can be broadly described
as excessive localised heating in small areas of the
chip where current density is high due to uneven current VCEO(sus)VCE
distribution.
At turn off the transistor is reverse biased, therefore the
tcM
reverse-biased safe operating area (RBSOA) curve is I
applicable, Figure 2.16b. This curve is defined by the
maximum voltage and current ratings and secondary E:
reverse bias
breakdown effects due to uneven current distribution during safe operating area
(/) repetitive
turn off.
o tc
:3

:3
t-
.m
.i.-,
Short-Circuit P e r f o r m a n c e t-
O
Ie2 defined
o
v

If the transistor load is accidentally short circuited the col- Tj ~< 125 C
lector current will invariably exceed its allowable peak
rating pulling the transistor out of saturation. The device can
survive nonrepetitive surges of this nature provided that:
VCEV VCE
a surge is detected and interrupted within a specified VcEr(,u~)
(guide only)
time
VCE and I c do not exceed specified levels 160 _Tj'- 1:)5~_ I I
140 I I I
Most manufacturers guarantee that the device will survive a < IB2/-~ A-
limited number of short circuits under defined test condi- G 120 /

tions. A typical transistor may be rated for a peak current of E 100 b


II
four to six times rated current, with VCE of half-rated voltage
o 80 ,'~" A
and a short-circuit duration of 30 ~ts. This type of data is
o 60 /82=-3 A \%,
essential for the design of practical protection schemes
required by chopper and inverter applications. o 40 "~

20
MOSFET
0 200 400 600 800 1000
collector-emitter voltage VCE,V
The metal oxide semiconductor field-effect transistor
(MOSFET) was developed into a useful fast-switching Figure 2.16 Safe operating areas for bipolar transistors
power device in the 1980s. There are both N-channel and a f o r w a r d - b i a s safe o p e r a t i n g area (FBSOA)
P-channel devices available, but for power applications b reverse-bias safe o p e r a t i n g area (RBSOA)
N-channel devices predominate due to their lower losses. c typical RBSOA
The device symbol for an N-channel MOSFET is shown in
Figure 2.17.
current will flow until the gate threshold voltage, typically 2
When a voltage, VGs, is applied between the gate and source
to 3 V, is reached. With Vcs above the threshold voltage the
an electric field is set up within the device. The field modu-
characteristic has two distinct regions: a constant resistance
lates the resistance between the drain and source, permitting
region with channel resistance Ros:o,), and a constant current
a current to flow in the drain in response to the applied
region where the transconductance of the device is almost
drain circuit voltage. The transfer characteristic is shown in
constant. RDs:o,) is a key parameter and will determine the
Figure 2.18.
forward voltage drop (Ros:o,) I9) and ultimately the cur-
With zero VGS applied, a positive voltage will be blocked at rent rating of the device. Operation within the constant
the drain until the breakdown limit is reached. Any reverse current region is normally avoided (to minimise conduction
voltage applied will be clamped by the presence of a para- losses) by setting Vcs high enough for the load current, a
sitic diode in the device structure. If VGs is raised no drain value of 10 V is usually sufficient. With VGs above threshold
62 POWER SEMICONDUCTOR DEVICES: MOSFET

drain In conventional high-voltage MOSFETs (>200 V) the lower


doping density and thicker die required for higher voltages
,1 has resulted in devices which have an on resistance pro-
'r - - -
I
/ I
I
2.5
portional to V{)ss, where VDSS is the voltage rating. This is
. - 1, - internal why there are relatively few devices rated above 600 V, and
t I
_ -,~ parasitic

gate diode
devices rated above 1000 V are very rare. A process tech-
0 I nology, marketed as CoolMOS by Infineon, uses a new

~
three-dimensional doping profile which allows the doping
density to be increased and chip thickness reduced compared
sourcet - - - J ID to conventional MOSFETs so that the on resistance only
increases linearly with voltage rating. This allows a smaller
p o s i t i v e voltage at g a t e
produces current in drain
l Vos
die to be used for a given rating although the more com-
plicated chip fabrication increases costs.
For low-voltage MOSFETs (< 50 V) the channel resistance
is significant, as opposed to high-voltage devices where the
body or drift region dominates. To reduce the channel
+ T resistance another process technology has been used: trench-
gate structure. A trench-gate device, as its name suggests,
has the gate etched down into the chip, rather than being a
Figure 2.17 MOSFET
planar feature near the surface of the chip. This reduces the
a circuit symbol channel length and so resistance, and also allows the cell size
b electrical circuit to be reduced thus giving more cells per unit area and hence
reduced resistance per unit area.

Switching Performance
VGS=20 V (maximum)
The MOSFET has two very important differences to the
constant current
BJT; first it is a voltage-controlled device rather than being a
current-controlled device, and second it is a majority carrier
device rather than a minority carrier device. Being voltage
controlled with an oxide insulated gate, very little power is
VGS=10 V
needed to control the device, current only being required to
q)
charge and discharge the gate capacitance. A majority car-
VGS= 6 V tier device only conducts due to intrinsic charge carriers.
This enables the device to switch off very quickly because
VGS=3 V no time is required for the removal and recombination of
minority charge carriers, as there is in a bipolar transistor or
thyristor. Fast switching allows the MOSFET to be used in
drain-source voltage, VDS vlsr0oss applications with switching frequencies of 100 kHz or more.
parasitic diode volt drop The down side of only having intrinsic charge carriers
available for conduction is that the conductivity of the sili-
con is lower and so a larger chip is required for a given
current rating.
Figure 2.18 Power MOSFET characteristic
Safe Operating Area
voltage the channel is able to conduct current in the reverse
direction as well as the normal forward direction. This fea- The MOSFET does not exhibit the phenomenon of sec-
ture can be exploited by using MOSFETs as very-low- ondary breakdown, which means that the safe operating area
voltage drop diodes. curve for all operating modes is square. It extends to VDSS
along the voltage axis and up to four times ID in the current
Voltage and Current Ratings axis. The only limitation to switching at these levels is that
the maximum junction temperature must not be exceeded.
Current ratings are usually given for case temperatures of
25C; at the more practical temperature of 100C the current Parasitic Diode
rating is reduced because of an increase in RDs(on). The
variation in RDS(onjwith temperature depends on the voltage The parasitic or body-drain diode that exists within the
rating of the device; lower voltage devices have a lower MOSFET structure has a slow switching characteristic
temperature dependence, typically a factor of 1.2 from 25C compared with the MOSFET channel itself. The switching
to 125C; for high-voltage devices the ratio may be two or frequency of circuits which make use of the diode (for
more. The positive temperature coefficient of on resistance example ultrasonic inverters) may be limited solely by the
has the benefit that devices may be connected in parallel and diode and not the MOSFET. This is mainly due to very
will share current equally. A chip is in fact made up of significant diode switching losses, which are a function of
thousands of MOSFET cells connected in parallel. reverse recovery charge, and operating frequency. In some
Chapter 2.1 63

a collector

low on-state 5
voltage diode , gate O

emitter
/ ~ - vdei or dy efast
n
. . . .

I C )
I

~k
b

)lvc
-- _~"qL __ parasitic VBE Ic
I diode
q I
'
I

.i

I Vos IpNp

Figure 2.19 MOSFETwith antiparallel diode GO

applications the body diode is actually bypassed by a fast


diode as shown in Figure 2.19; the series-connected diode VCE-" V DS+ (]MOS X RMOD) + VBE
Ei
prevents current from flowing in the body diode.
Figure 2.20 IGBT
a circuit symbol
b steady-state equivalent circuit
INSULATED-GATE BIPOLAR TRANSISTOR

The insulated-gate bipolar transistor (IGBT) combines


the best features of a MOSFET and a BJT to give a collector doping density than do NPT types. NPT devices
voltage-controlled device with low state losses. The circuit generally have faster temperature independent switching and
symbol and terminal designations are shown in Figure 2.20a. a positive temperature coefficient of VcE(saO, and PT devices
have a strong temperature dependence of switching loss and
a negative temperature coefficient of VCE(saO.
Operation
Unlike the MOSFET, the IGBT does not have a parasitic
The on resistance of a high-voltage MOSFET is mostly due diode inherent within its structure, so that a suitable anti-
to the drain drift region as explained in the previous section. parallel fast diode may be selected to match the speed of the
In an IGBT the resistivity of this region is substantially IGBT.
reduced by the injection of charge carriers from an addi-
tional semiconductor layer in the device. The process is
called conductivity modulation, and results in on-state losses Voltage and Current Ratings
comparable to those of the BJT. An IGBT may be modelled
The principal voltage rating of the IGBT is the collector-
as a low-gain PNP transistor and MOSFET connected as
emitter breakdown voltage VcEs specified with zero gate
shown in Figure 2.20b, the additional semiconductor layer
emitter voltage. Devices are available with ratings from
being the emitter of the PNP transistor, which is also the
250 V up to 6500 V. Although devices are available with
collector of the IGBT.
terminal current ratings up to 2400 A, these devices use
The on-state voltage VCE(saO of the IGBT has three many IGBT chips connected in parallel. Single chips typi-
components: cally have a maximum rating of 300 A.

VCE(sat) -- VBE + (Iz) X RDeaFr) + (Iv x RcI-I)


Switching Behaviour and Gate Drive
where Vee is the forward voltage drop of the PNP transistor,
Like the MOSFET, the IGBT is voltage controlled and the
RDRIFT is the drift region resistance and is much smaller than
gate presents a capacitive load to the drive circuit. Turn on
in an equivalent MOSFET due to the conductivity modula-
takes place when the capacitance has been charged to above
tion, and Rci-i is the MOSFET channel resistance. Typical
the gate-emitter threshold voltage, which is usually 4 to 5 V.
values for VCE(saO a r e 2 to 3 V at rated current and 25C for a
Typically a 1200 V 100 A IGBT will turn on in less than
1200 V rated device.
200 ns. During the conduction time, the gate emitter voltage
Trench-gate IGBTs have been developed recently which is held at between 13 and 17 V so that VCE(saO is kept as low
with an optimised geometry reduce both RDRIFr and Rci4 as possible to minimise conduction losses.
giving a lower VCE(saOfor a given current density.
Turn off is initiated by discharging the gate emitter capaci-
There are two basic types of IGBT: punch through (PT) and tance. The MOSFET structure turns off first, allowing Ic to
nonpunch through (NPT). The difference relates to the fall rapidly to an intermediate level. A slower fall of current
doping and thickness of the drift region and the collector then follows as the PNP structure turns off. Unfortunately,
layer, PT types having a thinner drift region and higher the gate drive circuit can only control the MOSFET turn off
64 POWER SEMICONDUCTORDEVICES: Insulated-Gate Bipolar Transistor

and has no influence on PNP behaviour. Turn-off delay time a short period of very high dissipation without damage. A
and current fall time are much shorter than for equivalent typical device will be rated for a short-circuit duration of
bipolar transistors owing to the low gain of the PNP structure 10 ~ts with test conditions of a current of ten times ID, with a
and processing steps taken to reduce the carrier lifetime. collector-emitter voltage of half rated voltage, and 125C.
Both these measures have the effect of increasing Vce(saOso During this fault condition the junction temperature may
there is a trade off between low switching loss and low exceed 300C. The control and gate drive circuit must detect
conduction loss. To give optimum performance in low and the overcurrent condition and switch off the IGBT within the
high-frequency applications IGBTs are available in different rated time to avoid damage. The short-circuit current may be
families, optimised for either low conduction or low controlled by adjusting the gate voltage. For the Eupec DN2
switching loss. series devices a Vce between 10 and 17 V will produce
short-circuit currents of between two and ten times ID. To
Both turn-on and turn-off times can be adjusted by selection
reduce the inductive voltage overshoot when turning off the
of gate resistor value; this controls the rate at which the input
large short-circuit current, the gate drive circuit may slow
capacitance is charged or discharged. This is a very impor-
down the turn off under these conditions by limiting the gate
tant feature as it allows turn on to be set at a rate which suits
discharge current.
the reverse recovery characteristic of the free-wheeling
diode, and allows the rate of turn off to be reduced if
required to limit inductive voltage overshoot. Series and Parallel Operation
During the off state the IGBT gate emitter is normally held at Although devices are available up to 6500V, the more
a minimum o f - 5 V to ensure that the device cannot be common value is 3300 V. For some applications this is not
spuriously tumed on. sufficient and so devices must be connected in series. This is
easier to achieve with IGBTs than with many other power
devices due to the following combination of characteristics:
Safe Operating Area (SOA) voltage control, fast switching, square SOA. Figure 2.22
The FBSOA and RBSOA curves for an IGBT chip are both shows how two IGBTs may be connected in series. The
square, and bounded by the rated breakdown voltage, and the resistors maintain steady-state voltage balance by compen-
pulsed collector current, IDM, usually twice the rated D.C. sating for differences in device leakage currents. Very fast
current ID. The square SOA makes the IGBT a very robust transients are balanced by the capacitors, which can be much
device and allows operation without snubber circuits, thus smaller than for a similar circuit using BJTs or thyristors.
reducing system losses and size. Although the chip RBSOA During switching, differences in delay from one device to
is square the device RBSOA may have some reduction in another will tend to lead to unbalanced voltages. The active
VcE at high currents. This is due to wiring inductance within clamp circuit, formed by the zener string and IGBT, limits
the package, which during turn off increases the voltage at VcE to a value a little above the zener breakdown voltage.
the chip above that measured at the terminals. Modern The voltage-controlled gate means that little current is
IGBTs do not suffer from secondary breakdown unlike needed and fast switching ensures a fast response. The IGBT
BJTs. The RBSOA for an Eupec BSM100GD120DLC is is able to survive this operation because of the square SOA.
shown in Figure 2.21. Devices may be connected in parallel to make up power
switches with ratings of many kA. When using devices with
Short-Circuit Performance a positive temperature coefficient little or no derating is
required. When using devices with a negative temperature
Most IGBTs produced for high-power applications are short- coefficient it is usual to select devices with the same Vce(saO
circuit rated. This means that the device is able to withstand and apply a derating factor.

240

200

160
IC, module
<120 [.... IC, chip

80

40

0 r
o 6c o 1 oo 12oo 14oo
VcE(V)

Figure 2.21 Reverse-bias safe operating area for Eupec BSM100GD120DLC


C h a p t e r 2.1 65

2000

ance resistor
VDM <~ VDRM
1600
A.C. balance rj = o - ~ s o c
capacitor

1200

<
v
O

800

400

Figure 2.22 Series connection IGBT with active clamp


0 1()00 2000 3()00 4000
vo (v)
anode Figure 2.24 IGCTSOA

This voltage is greater than the maximum voltage which can


oate t I'/1 I cathode
be sustained during switch off. The on-state average and
r.m.s, current ratings, ITAVE and IrnMs, determine the maxi-
Figure 2.23 IGCTsymbol mum load current. In the on state the device can withstand
high surge currents, but the device cannot switch off these
currents. The maximum current, which can be turned off
INTEGRATED-GATE COMMUTATED under defined conditions, is IrGQM. Devices are available
THYRISTOR with ratings up to 4000 A. The maximum controllable cur-
rent is approximately twice the on-state current.
The integrated-gate commutated thyristor (IGCT) is a
development of the GTO and has essentially the same device Switching Behaviour and Gate Drive
structure. By operating with a turn-off gain of 1, rather than
3 to 5 for a GTO, faster turn off is achieved and the Turn on is initiated by application of a positive gate current
requirements for snubbers much reduced. In order to have typically up to 100 amps. Lower gate currents may be used
the correct conditions for turn off the gate circuit inductance but this increases the turn-on time and limits di/dt capability.
must be very low. To achieve this the gate circuit is inte- Once conduction is initiated the gate current may be reduced
grated into the device - hence the name integrated-gate to the back-porch current of a few amps. As with a GTO this
commutated thyristor. IGCTs are available with or without is necessary due to the low device loop gain at low anode
an antiparallel diode. currents.

As with a GTO, the IGCT is a four-layer device with a highly At turn on the anode current di/dt must be limited to prevent
interdigitated gate structure. The device can be triggered into hot spots as the conduction spreads out from the gate, and
the on state by applying a positive gate current allowing also in applications with a free-wheel diode to limit the
current to flow between the anode and cathode. Once swit- reverse recovery di/dt. However, once turn on is initiated
ched on the current is determined only by the external cir- through the gate, the anode current di/dt cannot be controlled
cuit. The on-state voltage is typically 3 V for a 4500 V rated via the gate, and so an inductor and associated snubber must
device. Turn off may be achieved by reverse biasing the be added.
main power circuit, or more normally by extracting current In order to turn the device off sufficient gate current must be
from the gate. Sufficient current is extracted from the gate so extracted from the device for the regenerative action to stop.
that no current flows across the gate to the cathode junction, To allow operation with little or no snubber and minimise
and the device behaves as a PNP transistor. This gives fast switch-off losses a turn-off gain of 1 is used, (i.e. all anode
turn offwith a safe operating area (SOA) similar to that for a current is diverted out of the gate), and this must be
bipolar transistor. accomplished fast enough to prevent current redistribution in
the device which would lead to hot spots and a reduction in
Voltage and Current Ratings device turn-off rating. In a standard GTO package the gate
inductance is of the order of 50nil. To divert an anode
The maximum repetitive voltage, which can be blocked, is current of 2000 A out ofthe gate in 1 ~ts would require 100 V.
represented by VDR~, with devices available up to 5500 V. This would lead to very high gate drive power losses, and
66 POWER SEMICONDUCTORDEVICES: Integrated-Gate C o m m u t a t e d Thyristor

also exceeds the gate reverse breakdown voltage of anode


approximately 20 V. By integrating the gate drive unit into
the IGCT assembly a gate circuit inductance of 5 nH is gate %
achieved. Now only 10 V is required to achieve the required
gate current di/dt, gate drive power is reduced and the
breakdown voltage is not exceeded.
To switch the large gate current required an array of low-
voltage MOSFETs is used coupled to a bank of low-impe-
dance electrolytic capacitors. Gate drive power consumption
can be significant, approximately 80 W for a 700 A device
switching at 500 Hz.
As with a GTO, an IGCT has a minimum on and off time. cathode
The minimum times of 10 gs are determined by a combi-
anode
nation of time required for current density across the chip to
stabilise and for the gate drive circuit to prepare for the next
switching event. A typical maximum average switching

,K
frequency is 500 Hz, although this may be increased to 2 kHz
with a suitable gate unit and some reduction in current gate &
rating.
In the event of a short-circuit fault the device must be on FET off FET

switched off before the anode current rises above the maxi-
mum controllable current or control will be lost and the
device destroyed.

OTHER POWER DEVICES AND MATERIALS

MOS-Controlled Thyristor cathode

Figure 2.25 MCT symbol and circuit


The MOS-controlled thyristor (MCT) was developed to
exploit the low conduction loss of a thyristor with the low
gate drive power and fully controlled behaviour of a
MOSFET. The symbol and simplified equivalent circuit of
an MCT is shown in Figure 2.25. anode
Operation is most easily understood from the equivalent
circuit. In the off state the turn-off MOSFET is held on, this
keeps the PNP device off and the MCT can block positive
anode cathode voltage. To turn on, the turn-off FET is
switched off and the turn-on FET switched on. This provides
base current to the NPN transistor and regenerative action
then takes place as with a standard thyristor. Once in the on
state the MCT has a similar surge capability to a thyristor,
and low voltage drop of 1 to 2 V. To turn off the MCT the
on gate(~
"'
cathode
off gate

turn-off FET is again turned on. This short circuits the PNP
transistor base and so regenerative action stops and the anode
device turns off in a manner very similar to that for a GTO.
Although the MCT offers low on-state loss and low gate
drive power, there are important limitations which have
prevented widespread use of the device; the gate bias must
be maintained at all times to ensure that the device remains
off, and the switching safe operating area is limited to half
rated voltage at rated current.
Devices have been made with maximum ratings of
approximately 1500 V, 100 A.
on gateC) I
MOS Turn-off Thyristor off gate ( ~
1
The MOS tum-off thyristor (MTO) is very similar in many off FET
respects to the MCT and IGCT. The circuit symbol and
cathode
equivalent circuit are shown in Figure 2.26. In contrast to
other power devices the MTO is a four-terminal device; in Figure 2.26 MTO symbol and equivalent circuit
Chapter 2.1 67

addition to the main power terminals there are two gate Pressure Contact Packages
terminals, one for turn on and another for turn off.
Pressure contact packages are in general only used for very
As can be seen from the equivalent circuit, the MTO com-
high-power applications, and devices rated above 4.5 kV or
prises a GTO, which is responsible for turn on and con-
2.5 kA are in general only available in these packages. They
duction, and a MOSFET, which is only used during turn off.
are well suited to devices with a large single chip such as
To turn on, a gate current of several tens of amps is injected
thyristors, although they are also now being applied to
into the turn-on gate. Once conduction is initiated regen-
multichip devices such as IGBTs.
erative action starts and the anode current is limited only by
the external circuit. As with GTOs and IGCTs, a back-porch
current of several amps is required to ensure that the device CONSTRUCTION
stays on with minimum voltage drop. To turn off, the back-
porch current is switched off and a positive voltage applied As shown in Figure 2.27e, a pressure contact package
to the turn-off gate terminal. This switches on the MOSFET consists of two large copper pole pieces between which the
which diverts current out of the gate forcing regenerative chip is sandwiched. Externally applied pressure ensures
action to stop and the GTO to turn off. As with an IGCT, contact between the chip and the contact plates. For the
sufficient gate current is removed to give a turn-off gain of gate connection an internal pressure contact is provided. A
near unity, which gives much faster turn off than for a GTO ceramic body provides the rest of the enclosure and may be
operating at a turn-off gain of 3 to 5. As the turn-off ribbed to increase the creepage distance over the surface.
MOSFET is integrated into the MTO, the inductance is low After assembly the joints are welded, the package evac-
enough to ensure that the gate current can be removed fast uated and sealed. In use the device is clamped between two
enough to prevent current redistribution during turn off. plates, usually the heatsinks, and a known pressure applied.
As the pole pieces are in direct contact with the chip the
Devices have been made with ratings up to 4500 V, 500 A heatsinks and mounting plates are all live. Great care must
with plans for ratings up to those of conventional GTOs. be taken to ensure the correct even contact pressure; this
makes the mechanics of pressure contact devices quite
Silicon Carbide complex.

The vast majority of semiconductor devices use silicon (Si)


FEATURES
as their base material. Although the process technology is
well developed, Si has some limitations for power devices, Pressure contact devices have several features which
most notably a maximum junction temperature of 125C to make them well suited to very high-power, high-reliability
175C. By using silicon carbide (SIC) as the base material, applications:
devices may be operated with junction temperatures of
250C to 350C. Furthermore, SiC has a breakdown field Double-sided cooling gives significantly reduced ther-
strength five times that of Si and a thermal conductivity mal impedance compared to single-sided cooling. This
three times that of Si. Overall SiC offers lower loss devices, allows a device to operate at a high loss per unit area.
faster switching and much higher operating temperatures. The two main causes of power device wear out are the
Despite the significant technical gains in performance pos- failure of wire bonds and soldered contacts, due to
sible, SiC is much more difficult to process than Si and the thermally induced mechanical stress. As a pressure
fabrication technology is relatively young. This makes contact device has no wire bonds or soldered joints it
devices rated at more than a few amps not commercially has a very good thermal cycling capability which is
viable, although this is expected to change over the next five especially important in applications such as railway
to ten years. traction.
In the event of a failure the device will go short circuit.
This allows redundancy to be built into high-voltage
POWER DEVICE PACKAGING applications which have several devices connected in
series.
Power devices are available in a very wide range of The package has a high rupturing IZt so that with correct
packages, a selection of which is shown in Figure 2.27. fusing it is possible to prevent rupture of the device in
The very smallest packages are the wafer-scale packages, the event of failure. It is very important in high-voltage
which only have a layer of passivation on the surface of applications (>3 kV) to limit damage to other
the chip. These must be soldered to a printed circuit board equipment.
(PCB) or other substrate and switch ratings are limited to a
few hundred VA. Packaged devices, normally housing a
single chip, for through-hole or surface mounting, are
available in industry standard sizes such as D2-Pak, TO- Large Wire-Bonded Packages for
220, TO-247 in ratings up to 10 kVA. For larger devices Power Modules
with ratings up to 1 MVA isolated base modules with
moulded plastic cases are used. There is a wide variety of Large wire-bonded packages are used for power modules
packages containing one or more devices but with little housing single devices rated over 50 A or multiple devices
standardisation in sizes. For the highest power devices, up rated over 10 A. Maximum ratings are typically 4.5 kV and
to 10MVA, pressure contact packages are used. These 2.5 kA. They are very widely used and are much cheaper
only house a single device. than the equivalent device in a pressure pack.
68 POWER SEMICONDUCTOR DEVICES: P o w e r D e v i c e Packaging

a 1 to lOmm
glass passivation

silicon chip

solder bumps

10 to 40 mm

bond wire epoxy silicon chip

leads
copper slug

/
30 to 150 mm

lead.,

~lastic case

)3 DBC substrate
Cu

40 to 200 mm

leads I"1 bond wire soft-gel fill silicon chip

plastic case

DBC substrate
~ copper base plate

40 to 120 mm
v

spring contact copper pole piece silicon chip

gate lead

ceramic body

copper pole piece

Figure 2.27 Package cross-sections types


a w a f e r scale
b TO-220
c direct b o n d e d c o p p e r (DBC)
d c o p p e r base
e pressure
Chapter 2.1 69

CONSTRUCTION variety of locations and features may be incorporated to


increase the creepage and clearance distances between
There are two basic types of package, either with or without terminals.
a copper base plate, as shown in Figure 2.27c and d. The
copper base plate gives better transient thermal impedance, The DBC substrate allows several chips to be placed within
aids heat spreading across the heatsink and makes the device the package and connected in a variety of ways to form
less prone to damage due to incorrect mounting. either high-current single switches or complete power cir-
cuits, as in the case of a PIM or CIB module. Where addi-
The chip or chips are first soldered to the direct bonded tional control or monitoring circuits are required, these may
copper (DBC) substrate. This consists of two layers of be easily incorporated in the DBC or a separate PCB within
copper, between which there is an Aluminium Oxide the housing.
(A1203) or Aluminium Nitride (A1N) insulator. The DBC
isolates the chips and the power circuit from the base plate of Devices with solder terminals are only designed to be sol-
the device and so the heatsink may be earthed. The top layer dered to a PCB. Screw-terminal packages are more flexible
of copper is etched to form an interconnect pattern similar to as they may be screwed to a PCB for low or medium-current
a printed circuit board. Wire bonding is then used to connect applications or to busbars for higher currents.
the substrate, chips and package terminals. A plastic package Compared to pressure packs, wire-bonded packages have
supports the power terminals and provides mechanical pro- some disadvantages for very high power applications: surge
tection. To provide electrical insulation and environmental current rating may be wire-bond limited, in the event of a
protection the chips and bond wires are covered in an failure the device will go open circuit as the wire bonds blow
insulating gel. off, and in the event of a major device failure the package
In use, the package is fixed to a heatsink using screws may rupture.
with electrical contacts and either soldered or screw
terminals.
Small Wire-Bonded Packages for
PACKAGE TYPES Discrete Devices
Single power devices with screw power terminals are Small wire-bonded packages for discrete devices are made
available in ratings from approximately 50 A to 2500 A. in very large volumes and have the lowest production
For currents above 800 to 1000 A several parallel contacts costs. In general, they only contain a single chip although
are used. There is a wide variety of package sizes, from some packages are available with two chips, for example
20 mm x 92 mm to 140 mm x 190 mm, some sizes being an IGBT with antiparallel diode. The ratings of these
adopted as a de facto standard. devices are limited by the available chip area and the
current rating of the leads. Surface mount (SM) packages
Screw terminal packages are also available with multiple are commonly available up to 30A, with through-hole
power devices. These may range from a half bridge (two devices up to 70 A.
power switches) to a three-phase inverter. Ratings are
available up to 450 A, 1200 V as a three-phase inverter.
CONSTRUCTION
For packages which contain multiple devices at current
ratings up to 150 A, all solder terminals tend to be the most The chip is first soldered to a copper slug before wire
cost effective. Two industry standard packages are the bonding to the lead frame. An epoxy resin is then moulded
Econo2 and Econo3. These are available from a large over the chip to provide mechanical support for the leads,
number of manufacturers and may house a three-phase electrical isolation and environmental protection.
inverter, three-phase rectifier, or rectifier, inverter and brake
chopper. The latter type are often known as power integra-
tion modules (PIM) or converter inverter brake (CIB) PACKAGE TYPES
modules.
There is a wide variety of packages, many of which are
Packages containing multiple devices are also available industry standard. Packages, which are designed for screw
with screw terminals for the power connections or pressure mounting, such as TO-220 or TO-247, can also be clip
contacts. mounted. Using a clip saves parts and labour and gives
In addition to the main power devices the control, moni- more even contact pressure to the heatsink. By having a
toring and protection circuits may also be mounted within package without a mounting hole, the popularity of clip
the module. These parts are commonly known as intelligent mounting is being exploited to fit larger chips within a
or integrated power modules (IPMs). The additional circuits given footprint.
are normally assembled onto a small PCB, which is wire In many power applications it is desirable to have isolation
bonded to the main power devices. for the heatsink. Traditionally this has been achieved by
putting an insulating washer between the device and heat-
FEATURES sink. By moulding the epoxy resin around the whole copper
slug and chip a fully insulated package can be produced. As
Large wire-bonded packages with moulded plastic hous- the epoxy is a poor thermal conductor, device ratings are
ings are the most flexible of power device packages. The reduced. To overcome this limitation, devices are also
moulded case allows many terminals to be placed in a available which have a small DBC substrate rather than a
70 POWERSEMICONDUCTORDEVICES:Power Device Packaging

copper slug. This provides electrical isolation and much APPLICATIONS


better thermal conductivity than epoxy resin.
Cost and performance requirements are paramount in
FEATURES deciding the most suitable power semiconductor device for a
particular application. The following tables summarise
Packages are soldered to a printed circuit board, which in the
device characteristics, packages and applications for the
case of surface-mount packages provides the cooling path in
most popular power devices.
addition to electrical connection.

Table 2.2 Comparison of power semiconductor devices

Property SCR GTO BJT MOSFET IGBT IGCT

Self-commutation no yes yes yes yes yes


ability
Maximum r.m.s. 5000 A 2000 A 1000 A 300 A 2400 A 1700 A
current rating
Maximum voltage 12 000 V 6000 V 1600 V 1500 V 6500 V 5500 V
rating
Maximum switched VA 30 MVA 30 MVA 1 MVA 30 kVA 4 MVA 12 MVA
rating
Surge current ability excellent excellent limited limited limited excellent
(15 x Imvts) (15 x Im4s) (2 x IRMS) (4 X InMS) (2 X IRMS) (15 X InMS)
Operating current 140 A/cm 2 30 A/cm2 40 A / c m 2 75 A / c m 2 140 A / c m 2 30 A / c m 2
density at rated at 2 kV at 4.5 kV at 1000 V at 200 V at 1200 V at 4.5 kV
device voltage
15 A/cm 2 35 A/cm 2
at 800 V at 3.3 kV
Maximum junction 125C 125C 150C 150C 150C 115C
temperature C
On-state losses low medium medium high medium low
Switching losses very high very high high very low low medium
Turn-on ability medium medium good very good very good medium
(di/dt limit) (di/dt limit) (di/dt limit)
Turn-off ability none via gate poor-slow medium-long very good very good good
and lossy storage time
Turn-off safe na poor medium excellent excellent medium
operating area (50 %) (75 %) (100 %) (100 %) (70 %)
(percentage of
rated voltage at
rated r.m.s, current)
Load short-circuit none poor medium medium excellent poor
turn-off ability (2 x IR.vts) (4 X IRMS) (4 x It~ts) (10 x IR~ts) (2 x IRMS)
Snubbers usually yes yes yes no no no
required
Minimum on or 10-100 Its 10-50 Its <liLts < 100 ns < 1 Its lOgs
off time
Maximum 250 Hz 500 Hz 5000 Hz 100 000 Hz 10 000 Hz 500 Hz
switching
frequency
Switching time no no yes yes yes no
controllable
from drive circuit
Drive circuit power low high high low low high
Drive circuit complexity low high medium low low high
Series and parallel device selection very difficult series difficult; fairly simple fairly simple fairly simple
operation and passive series or parallel requires series and series and in series,
components parallel device selection parallel parallel, selection more difficult
required may be needed in parallel
for parallel
Chapter 2.1 71

Table 2.3 Typical power semiconductor device application areas as a function of system voltage and equipment VA rating

Application Supply voltage and equipment VA RATING

up to 240 V A.C., 400 V D.C. from 240 V A.C., 400 V D.C. above 690 V A.C., 1200 V D.C.
up to 690 V A.C., 1200 V D.C.
up to 1 kVA from 1 kVA up to 1 MVA above 1 MVA

a.c. motor drives


voltage source inverter MOSFET, IGBT IGBT, BJT IGBT, IGCT, GTO
current source inverter SCR, BJT, GTO SCR, GTO, IGCT
cycloconverter SCR
soft starters SCR SCR SCR
d.c. motor drives
force commutated MOSFET, IGBT, BJT IGBT, BJT IGBT, IGCT, GTO
line commutated SCR SCR SCR

Table 2.4 Availability of power device types and ratings in a variety of packages

Property Discrete Power module Pressure pack

Available devices SCR, GTO, SCR, BJT, SCR, GTO,


BJT, MOSFET, IGBT MOSFET, IGBT BJT, IGBT, IGCT
Maximum voltage rating 2000 V 6500 V 12 000 V
Maximum current rating 100A 2500 A 6000 A
Electrical failure mode open circuit open circuit short circuit
Power circuit connections solder solder, screw, or pressure contact
pressure contact
Control circuit connections solder solder, screw, or flying leads
pressure contact
Mounting method solder, screw or clip screw pressure plate
Cooling method convection to air, single-sided conduction double-sided conduction
conduction to PCB to heatsink to heatsink
or heatsink
Isolation from heatsink only with selected yes no
packages
Package rupture current low medium high

Table 2.5 Power semiconductor symbols

Meaning Specific to: Meaning Specific to:

aNPN large signal current gain GTO IGM maximum peak positive GTO
of a common-base NPN gate current
transistor; ratio of collector Iao maximum negative gate GTO
current to emitter current current during turn-off interval
aPNP as NPN, but for a GTO 1tt holding current
common-base PNP transistor IL latching current
/3off turn-off gain; ratio of anode GTO IMOS drain current of MOSFET in IGBT
current being controlled IGBT equivalent circuit
to negative gate current IpNP collector current of parasitic IGBT
required to produce turn off PNP transistor in IGBT
0 conduction angle equivalent circuit
hfe static forward current transfer Lr reverse recovery current
ratio of a common emitter IT direct on-state current
transistor; ratio of d.c. ITAVF_. the average on-state current IGCT
output current to d.c. input rating
current ITRMS the r.m.s, on-state current rating IGCT
f switching frequency IFGQ nonrepetitive peak controllable
IA anode current on-state current
18 continuous base current ITGQM the maximum current which IGCT
IB1 forward base current can be turned off under defined
182 reverse base current conditions
Ic continuous collector current BJT, IGBT PD power dissipation
ICM peak collector current Qrr reverse recovery charge
ID peak value of pulsed drain current RCH MOSFET channel resistance IGBT
Ia gate current RDRIFT drift region resistance IGBT
72 POWERSEMICONDUCTORDEVICES:Applications

Table 2.5 (Contd.)

Meaning Specific to: Meaning Specific to:

RDS(on) drain source on resistance VCE(sat) collector-emitter saturation


RMOD resistance of epitaxy region IGBT voltage
in IGBT equivalent circuit VcEo(sus) collector-emitter sustaining
R thjc thermal resistance, junction voltage, with open base, Ic
to case specified
case case temperature VcEs collector-emitter breakdown IGBT
junction temperature voltage, specified with zero
bmax maximum allowable junction gate emitter voltage
temperature VcEp peak collector-emitter voltage BJT, IGBT
tcond conduction time during short-circuit current
tl fall time fall time
tic turn-off crossover time BJT VcEv collector-emitter voltage with
tgq turn-off time GTO reverse-biased base-emitter
t~t turn-on time GTO junction
toe turn-off time VD direct off-state voltage SCR
ton turn-on time VDS drain-source voltage MOSFET
tp pulse duration Va gate voltage
tq total turn-off time SCR VaE gate-emitter voltage
tq~ recombination time Vas gate-source voltage
trc turn-on crossover time BJT VR reverse voltage
trr reverse recovery time V~ repetitive peak reverse voltage
ts storage time Vs voltage spike between anode GTO
VBE base-emitter voltage and cathode during fall time
V(SR)CES collector-emitter breakdown IGBT Wco,,d energy dissipation during the
voltage, gate-emitter short conduction time
circuited Woff energy dissipation during the
V(BR)DSS drain-source breakdown MOSFET turn-off time
voltage, gate-source Won energy dissipation during the
short circuited turn-on time
Vcc collector-emitter supply voltage
VCE collector-emitter voltage

2 DRIVE CONVERTER CIRCUITS

In order to concentrate on the salient characteristics of the overall A.C. power factor defined as the ratio of mean
many alternative power conversion circuits used in power power (W) to volt-amperes (VA) [note: the power factor
electronic drives today, it is necessary to make a number of thus defined equals the product of the displacement
simplifying assumptions. A somewhat idealised theory will factor (the fundamental power factor or cos4~ i.e. the
therefore be presented. Practical aspects, such as switching phase shift of the fundamental current with respect to
delays, will only be discussed where they have practical the A.C. supply) and the distortion factor (the ratio of
significance. The diversity of machines which can be used the r.m.s, of the fundamental current to the r.m.s, of the
with alternative power converters will be limited to those total current)]
seen as being of greatest practical importance. maximum attainable A.C. power factor that can be
achieved using capacitors only to counter the funda-
A.C. TO D.C. POWER CONVERSION mental VAR consumption of the converter
A.C. supply current harmonics for a constant D.C.
General current Id
D.C. voltage as a function of A.C. voltage
Prior to considering the relative merits of alternative con-
D.C. power for a constant D.C. current Ia
verters it is necessary to establish meaningful performance
voltage tipple-although form factor or peak-peak
parameters. Useful to the user and system designer are:
values are often used as a measure of voltage ripple, it is
r.m.s, value of the A.C. current for a constant D.C. more useful in practice to consider a factor M, which is a
current Id measure of the volt-second integral of the voltage pulses;
Chapter 2.2 73

the peak-peak current ripple A i d is then given by: In the half-controlled circuit, two of the thyristors shown in
Figure 2.30 are replaced by diodes. A number of diode
A i d -- MVdo combinations are possible: (Ap + Bp), (An + Bn), (Ap + An)
or (Bp + Bn). Figure 2.32 shows typical waveforms for the
where Vao = the maximum attainable D.C. voltage,
case when Ap and Bp are diodes. Note the absence of any
f = the A.C. frequency and L = D.C. circuit inductance
negative voltage, and the reduction in A.C. current periods.
pulse number p is the number of pulses of D.C. voltage
Only energy flow from the A.C. to D.C. side is possible.
during one complete A.C. cycle (1 cycle = 20ms for
50 Hz supply, 16.67 ms for a 60 Hz supply) SINE WAVE INPUTCONVERTERS

Converters for Connection to a With growing demands to reduce harmonic effects on the
supply, and improve the supply power factor, circuits are
Single-Phase Supply
now coming to the market, notably in switch-mode power
UNCONTROLLEDCONVERTERS supplies (SMPS), which draw near sinusoidal currents from
the A.C. supply at unity power factor. A typical circuit is
Only two power circuits need be considered in this category. shown in Figure 2.33.
The half-wave configuration of Figure 2.28 is not particu- It is expected that, as greater regulation is introduced in
larly useful for power applications, but is included as an respect to allowable harmonic content of drive systems, this
introduction to semiconductor behaviour in bridge circuits. type of circuit will become more common. However, in such
situations it is often more cost effective to have a single
The single device is available for conduction from 0 to 180
(the positive half cycle). A freewheeling diode may be added
across the load to conduct the load current during the
negative half cycle, and prevent it being reduced to zero. For

1
long time-constant loads, the load current can be considered
to be continuous: derived from the supply during the positive
half cycle and carried by the diode during the negative half
cycle.
When this circuit is used, a capacitor often replaces the
t
VD.C.
freewheeling diode, maintaining the output during the idle
half cycle. A notable application of this particular circuit is
for high-frequency SMPS secondaries.
The detailed characteristics of this circuit will not be con-
sidered further.
Figm'e 2.29 shows full-wave rectification from an A.C.
l
supply in its most popular form. The four diodes conduct in
diagonal pairs during every alternate half cycle of A.C. line
__1 I
i7'
I
I
I
0 ~ 2~
voltage.
Figure 2.29 Single-phase full-wave uncontrolled bridge
CONTROLLEDCONVERTERS

Again, only two power circuits are of practical importance. P


O
Figure 2.30 shows the power circuit for the fully controlled t~
c ~ Vo.c.
bridge together with associated A.C./D.C. relationships.
Bp
Figure 2.31 shows how the D.C. voltage can be varied by
adjusting the firing delay angle c~. Since negative D.C.
voltages are possible, energy flow from the D.C. side to the
.)
A.C. side is possible.

v
> o

( ) ) T
Va
conducting thyristors
top row bottom row
A.C./D.C. relationships
VD.C

Ap An 0 0
BR Vab + [d
' l Bp An
Bn 0
Vba --]d
0
l o
Figure 2.28 Single-phase half-wave uncontrolled bridge Figure 2.30 Single-phase fully controlled bridge
74 DRIVE CONVERTER CIRCUITS: A.C. t o D.C. P o w e r Conversion

Ap Bp

Bn An

~l--ll
t:z
conduction periods power
Va~ factor
controller

m I/ab Vba

Figure 2.33 Single-phase sine wave input converter


o 21c..

"', ..
i ! v

D.C. side voltage waveforms



i\I /\
AL
IF

a--i I
J I i
A.C. supply current waveforms A
W
~c
Figure 2.31 Single-phase fully controlled bridge, D.C.
voltage control

Ap Bp ~ Ap
lw

A n
gn I
I
Figure 2.34 Three-phase full-wave uncontrolled bridge

conduction periods More complex, and expensive, converter arrangements exist


which monitor the supply voltage and control currents in
such a way as to minimise supply voltage distortion. Such
i i systems are rare.
n o" i . o " " r ~ vba i

Converters for Connection to a


Three-Phase Supply
VDC i"
UNCONTROLLED CONVERTERS
. , . .
! ..!
A variety of uncontrolled converters is available, however
I "'-. .... -'" I only one is of practical importance in regard to drive sys-
I tems, namely the full-wave bridge converter. This arrange-
D.C. side voltage waveformi
, ment is as shown in Figure 2.34.
I i In contrast to the single-phase bridge, where altemate pairs
i

of devices conduct, switching in the three-phase bridge
i I alternates between the upper and lower row of devices. This
means that there are six conduction periods per A.C. cycle,
A.C. supply current waveform
each device conducting for a period of 27r/3 (120 elect.).
Figure 2.32 Single-phase half-controlled bridge, D.C.
voltage control CONTROLLED CONVERTERS

Two power circuits are of practical importance:


supply converter with a common D.C. bus feeding multiple
The fully controlled circuit is by far the most important
drives.
practical bridge arrangement. Figure 2.35 shows the power
Commercial three-phase sine wave input converters tend to circuit together with associated A.C./D.C. relationships.
take the form of an IGBT bridge connected to the supply Figure 2.36 shows how the D.C. voltage can be varied by
operated as a current controller to draw sinusoidal currents. adjusting the firing delay angle a. The pulse number, p, of
A practical implementation of such a scheme is shown in this bridge equals 6. Energy flow can be from A.C. to D.C.
Chapter 6.3. or D.C. to A.C.
Chapter 2.2 75

Table 2.6 Salient characteristics of the main single-phase A.C to D.C converters
Bridge Fully controlled Half controlled

Firing angle a a

Vdo ~2N2 Vs ~2N2 Vs


71" 71-

Pdo ~2N2Vs/d ~2N2Vs/d

Vd/Vdo cos a 0.5 (1 4- cosa)

Is/Id 1 N { (Tr - c~)/Tv}

Overall power ~2 N
c o2s a N{Z/Tr} (1 + cos a )
factor 7r N (Tr - a )

Maximum corrected cos a 1 + cos a


power factor N(Tr2/8 - sin 2 a) N{Tr/2. (Tr- a) - sin 2 a }

Input power/Pdo cosa 0.5 (1 + cosa)

Input VARs/Pdo sina 0.5 sina

Supply current 0 for n even 0 for n even


nth harmonic/Id
0.9/n for n odd 0.64/n- N(1 + cos na)
for n odd

Phase of supply na na/2


current harmonics

Va$ P

/a I 2 pi
T
v ~ ~ v~= o.75 V~o
n I. n2~Cwn
N Vt,a Vca Vc~ Va~ Vac Vbc Vba Vca Vo~

conducting thyristors A.C./D.C. relationships ia]l I I I I I

top bottom Vd ia ib ic
row row

Ap An 0 0 0 0
Bn Vab + id -- i d 0
Cn Vac + id 0 -- id
Bp An Vba -- id + id 0 Vd ..'" Vd=0.75 Vdo
Bn 0 0 0 0
Cn Vbc 0 + id -- i d
vt,. Vc. Vcb Vab Vac V.c vb. vo. Vcb
Cp An V~a -- id 0 + id
Bn Vcb 0 -- id 4- i d ial I I I I I F
Cn 0 0 0 0
Figure 2.36 Three-phase fully controlled bridge, D.C
Figure 2.35 Three-phase fully controlled bridge voltage control
76 DRIVE CONVERTERCIRCUITS: A.C. to D.C. Power Conversion

In the half-controlled circuit either the top three devices of current ripple on A.C. supply harmonics is of great practical
Figure 2.35 (Ap, Bp and Cp) or the bottom three devices (An, industrial importance mainly in relation to three-phase
Bn and Cn) are replaced by diodes. The pulse number of this bridges (ignoring the single-phase traction requirement).
bridge equals 3. Only energy flow from A.C. to D.C. is Practical experience has led to the adoption by many of the
possible. The voltage tipple is much greater than in the case following values:
of the fully controlled bridge, but the A.C. current drawn is
lower at reduced D.C. voltage. Figures 2.37 and 2.38 show 15 = 0.2511 (ideal = 0.211)
the half-controlled bridge. 17 - 0.1311 (ideal = 0.1411)
111 -- 0.0911 (ideal - 0.1111)
Voltage Ripple Characteristics 113 -- 0.0711 (ideal - 0.0811)

The voltage ripple characteristics for the most significant In general, the amplitudes of higher harmonics are rarely of
bridge configurations are shown in Figure 2.39. It should be significance, in regard to supply distortion. Under conditions
noted that these characteristics are for idealised conditions of of very high D.C. current ripple, the fifth harmonic can
smooth D.C. current and zero supply impedance. Increases assume a considerably higher value than that quoted above.
in supply impedance generally tend to result in somewhat A practical example would be an application with a very
lower D.C. voltage tipple levels. capacitive D.C. load (e.g. a voltage source inverter): in such
a case where no smoothing choke is used 15 could be as high
as 0.511.
Practical Effects
D.C. Motor Drive Systems
The characteristics presented above have, for the most part,
been based upon idealised conditions of negligible A.C. In principle, little has changed since 1896 when Harry Ward
inductance and constant D.C. current. Although these Leonard presented his historic paper 'Volts versus ohms -
assumptions provide a convenient means for comparison the speed regulation of electric motors'. In practice, however,
they are not often valid in practice. It is not practicable many advances have been made from auxiliary machines
to consider all such effects here. The effect of D.C. link through mercury-arc rectifiers to thyristors.

~dc
0.25

o
A
w
I
I
A
i v~ > _ ~ A - - - - - z~ z~ z~ n

-1.0 0.0 1.0


mean D.C. voltage/Vdo

Figure 2.39 Ideal voltage ripple characteristics


1 phase fully-controlled bridge
1 phase half-controlled bridge x
Ani Bn.L Cn,,L 3 phase fully-controlled bridge A
v i
3 phase half-controlled bridge
Figure 2.37 Three-phase half controlled bridge

ap
,,~, Vl

i o I I I
I On I I I On I
- - ~ - an

I', I\ , ", /x. ,, ,, ,,-, /\ /"\ .,,,,. ,,"\

I/- ,'~_ ,, /'\ ,~, _., .>", ,~ ,<, /~

I,, ,I, I I I I
J I I
Figure 2.38 Three-phase half controlled bridge, D.C. voltage control
Chapter 2.2 77

Table 2. 7 Salient characteristics of the three-phase A.C. to D. C converters described in the


text

Bridge Fully controlled Half controlled


Firing angle a a
a > 7#3 a < 7r/3
Vao
-3N2
-G -3N2
-G
7r 7r

Pdo
3N2 3N2
--G& --G~

gd/Vdo COS OL 0.5 (1 + c o s a )

ls/Ia N(3/2) N {(Tr- a)/Tr} N(3/2)

Overall power factor (3/70 cos a N3. (1 + cos a) (3/270 - (1 + cos a)


N(71 - ol)

Maximum corrected
COS O~ 1 + cosa 1 + cosa
power factor
N(Tr2/9 - sin2 a) N{27r(Tr - a) - sin2 a} N{47r2/9 - sin2a}

Input power/Pao cos a 0.5 (1 +cosc 0

Input VARs/Pao sin a 0.5 sin a

Supply current 0 for n = 3, 6, 9 . . . 0 for n = 3 , 6, 9 . . .


nth harmonic//d 0 for n even N(3/mr) .N(1 - cos nc0 for n even
N6/mr for n odd N(3/mr).N(1 + cos na) for n odd

Phase of supply na ncd2


current harmonics

\ I
I \I
X
I\
/ \

Figure 2.40 Single converter D.C. drive


I 1 Figure 2.41 Single converter reversing~regenerative D.C
drive

In applications where motor resistance varies with tem-


perature, or on sites with poorly regulated supplies which
The D.C. motor is still a versatile machine for variable-speed
result in unacceptable variations in field current, a controlled
drive systems and is often the preferred choice when con-
power converter is employed with current control. Such field
siderations such as freedom from maintenance or operation
controllers are further discussed later as applied to field
under adverse conditions are not paramount.
weakening control.
In Chapter 1.1 it has been shown that complete control o f a
D.C. machine can be achieved b y controlling the armature SINGLE-CONVERTERDRIVES
voltage, Va, and the field current, I f Two power converters
Figure 2.40 shows a single-converter D.C. drive. In its most
are employed for this purpose in most variable-speed drives
basic form the motor will drive the load in one direction only
which employ the separately excited D.C. machine. (In
without braking or reverse running. This is said to be a
referring to the number o f converters in a drive, it is c o m m o n
single-quadrant drive, only operating in one quadrant o f the
to ignore the field converter - this nomenclature will be
torque-speed characteristic.
adopted below). It is relatively c o m m o n in simple drives for
the field converter to be a single-phase uncontrolled bridge Such drives have wide application from simple machine tools
thereby applying fixed field voltage. to fans, pumps, extruders, agitators, printing machines etc.
78 DRIVE CONVERTERCIRCUITS:A.C. to D.C. Power Conversion

If the drive is required to operate in both the forward and +1500rev/min can still be achieved in approximately
reverse directions and/or provide regenerative braking, a 200 ms.
single fully controlled converter can still be used. However,
some means of reversing either the field or armature con- This circulating-current-free dual converter is by far the
nections, as shown in Figure 2.41, must be added. most common industrial four-quadrant D.C. drive and is
used in many demanding applications - paper, plastics and
Reversal of armature current can involve bulky (high- textile machines where rapid control of tension is required
current) reversing switches but, due to the low inductance are good examples.
of the armature circuit, can be completed in typically
0.2 seconds. Field current reversal takes longer, typically of
FIELD CONTROL
the order of 1 second; however lower cost reversing switches
may be used. The field reversal time can be reduced by using The output power of a motor is the product of torque and
higher voltage field converters to force the current. Forcing speed. If torque reduces in proportion to increases in speed,
voltages up to four per unit are used but care must be taken the motor is said to have a constant power characteristic.
not to over stress the machine. Obviously this increased
voltage cannot be applied continuously and requires either a In applications where material is coiled or uncoiled at con-
switched A.C. supply or a controlled field converter. stant tension, the torque required to produce that tension
varies in proportion to coil diameter, and rotational speed
Armature and field reversal techniques are used where required to maintain a constant peripheral speed (and
torque reversals are infrequent such as in hoists, presses, therefore line speed) is inversely proportional to diameter. A
lathes and centrifuges. motor having a constant power characteristic is well suited
to this type of application, the advantage being that a smaller
DUAL-CONVERTER DRIVES
motor can be used than would otherwise be the case.
When a four-quadrant drive is required to change the Machine tool drives also make use of constant power
direction of torque rapidly, the delays associated with operation, since loads are small at high speeds and heavy
reversing switches described above may be unacceptable. A work is done at low speed.
dual converter comprising two fully controlled power con- As explained in Chapter 1.1, the torque produced by a D.C.
verters connected in inverse parallel can be used as shown in motor is proportional to the product of armature current and
Figure 2.42. Bridge 1 conducts when the armature current IA field flux. By weakening the field as speed increases, a
is required to be positive, bridge 2 when it is required to be constant power characteristic can be achieved.
negative.
In practice, there are two major techniques for field weak-
There are two common forms of dual converter. In the ening, both of which rely on a field controller which itself is
first, both bridges are controlled simultaneously to give the a simple thyristor converter operating in a current control
same mean output voltage. However, the instantaneous mode.
voltages from the rectifying and inverting bridges cannot
be identical, and reactors Lp are included to limit the In the first method, suitable for coiler and uncoiler applica-
current circulating between them. The principal advantage tions, the field current reference is arranged to be inversely
of this system is that when the motor torque, and hence proportional to coil diameter (measured directly, or calcu-
current, is required to change direction (polarity), there lated from the ratio of line speed to motor speed). Since flux
need be no delay between the conduction of one bridge is not strictly proportional to field current, this method does
and the other. This is the dual-converter bridge with cir- not give a true constant power characteristic unless com-
culating current. pensation for the nonlinear part of the motor field curve
is applied.
In the circulating-current-free dual converter, only one
bridge at a time is allowed to conduct. The cost and losses The second method is to use the field controller with an outer
associated with the Lp reactors can then be eliminated, and voltage loop having a fixed reference, and to use motor
economies can also be made in the drive control circuits. armature voltage as the feedback signal. At low speeds, the
However, the penalty is a short time delay, as the current voltage loop saturates, providing maximum field current,
passes through zero, while the thyristors in one bridge safely since armature voltage is below the set value. As speed
turn off before those in the second are fired. This delay increases, the armature voltage rises to the point where
introduces a torque-free period of typically 10ms. Speed it matches the preset reference in the field controller.
reversal for a 3 kW drive of this type from - 1 5 0 0 to Above that speed an error signal is produced by the voltage
loop, which causes the field controller to weaken the motor
field current and thereby restore armature voltage to the
v set-point level. The resulting characteristics are shown in
I Figure 2.43.
Lp
t Lp
As regenerative braking depends on the return of power from
the motor to the mains, it cannot work if the mains supply
fails due to a blown fuse or a power cut. Dynamic braking of
four-quadrant drives is often encountered as a fail-safe
2 means of stopping the motor and its load, and as the only
means of (reverse) braking of single-ended drives. This
Figure 2.42 Single-phase dual-converter D.C. drive involves switching in a resistor across the D.C. motor.
Chapter 2.2 79

I approximately 100 Hz, which is too low for many servo-


constant field I field weakening
drive applications.
Thyristor-controlled A.C.-D.C. converters have an
O
inherently poor input power factor at low output vol-
L tages. (Near unity power factor can be achieved using an
uncontrolled rectifier feeding a D.C.-D.C. converter.)
E
Electronic short-circuit protection is not economically
possible with thyristor converters. Protection is normally
speed accomplished by fuses.
D.C.-D.C. converters are however more complex and
somewhat less efficient than A.C.-D.C. converters. They
find application mainly in D.C. servodrives, rail traction
drives and small fractional kW drives employing permanent-
0
constant constant power magnet motors.
torque Since step-down converters are of greatest practical impor-
"'I
0 tance emphasis shall be placed on their consideration.
0
0 For the purpose of illustration bipolar transistors will be
E
considered, however MOSFET, IGBT and GTOs are
widely used.
base speed speed

Figure 2.43 Constant power operation using a field Step Down D.C.-D.C. Converters
controller
SINGLE-QUADRANTD.C.-D.C. CONVERTER
Since the kinetic energy of the motor and its load is con- The most basic D.C. to D.C. converter is shown in Figure
verted into heat by the braking resistor, it is important to rate 2.44. The output voltage is changed by pulse-width mod-
it correctly for the duty it is expected to perform, taking ulation (PWM) - that is, by varying the time for which the
account of load inertia and the number of stops per hour. transistor T is turned on. The voltage applied to the motor is
therefore in the form of a square wave of varying periodicity.
D.C. TO D.C. POWER CONVERSION Because the motor is inductive the current waveform is
smoothed, the flywheel diode D carrying the current while
General the transistor is turned off.
The basic formulae relating the variables in this circuit are as
D.C.-D.C. power converters (often referred to as choppers)
follows:
provide the means to change one D.C. voltage to another. It
is more usual for the conversion to be to a lower voltage, Va = VD.C. " t ' f

although step-up converters are available and have sig- A Ia = VD.C./(4" La "f)
nificant potential for the future. wherefis the frequency of transistor on pulse (Hz), A/a is the
D.C.-D.C. power converters are fed from a D.C. supply maximum deviation of armature current, La is the motor
usually comprising an uncontrolled A.C. to D.C. converter inductance, t is the on pulse duration (s) and VD.C. is the
or alternatively a battery supply; the controlled D.C. output source D.C. voltage.
can then be used to control a D.C. machine as in the case of The circuit is only capable of supplying unidirectional cur-
the controlled A.C. to D.C. converters. rent and voltage to the motor and is therefore not capable of
D.C. drives employing controlled A.C. to D.C. converters four-quadrant operation, that is reversing or regenerating.
have several important limitations, which are overcome by Industrial applications for this circuit are normally limited to
the D.C.-D.C. converter: drives below 5 kW and simple variable-speed applications.
The inability of a thyristor to interrupt current means that In traction applications, however, drives of this fundamental
an alternating supply is necessary to commutate the type are designed at ratings of many hundreds of kW.
converter- this precludes operation from a D.C. supply.
This is a common requirement on battery vehicles and Two-QUADRANT D.C.-D.C. CONVERTER
D.C.-fed rail traction.
The D.C. ripple frequency is determined by the A.C. and In order to achieve full four-quadrant operation a converter
is, for a 50 Hz supply frequency, 100 Hz for single-phase must be capable of supplying reversible voltage and current
and 300 Hz for three-phase fully controlled bridges. This to the motor. A circuit that is capable of two-quadrant
means that additional smoothing components are often operation - that is motoring and braking in one direction
required when using high-speed machines, permanent- only - is shown in Figure 2.45. This converter is able to
magnet motors or other special motors with low arma- reverse the current flow to the motor but unable to reverse
ture inductance. the motor terminal voltage and hence the speed. During
As a result of the delay inherent in thyristor switching motoring, the converter operates as the basic chopper with
(3.3 ms in a 50 Hz three-phase converter) the current T1 and D2 carrying the current. During braking (or regen-
control loop band width of the converter is limited to eration) T1 is inoperative and T2 controls the current.
8U[)IEJ q asJaAaJ 9 somos se qons S0A!.Ip "~)'O oouetmojaod
8u!Jo),ouJ pJEMJO~t q q~!q u.t posn iop!Ax oae Sa0U0AUO0 ~ueapenb-anoj osoq
~!n:)J!:) e
ouonbo.tj oldd.u 3 o ~u!lqno p ~ pu~ O~EIIOA oldd.u
Ja~a4uo9 0"0-'9"0 .~ue.~penb-.mo_-I 9t,,'rc a.m6t...-I
u! uo.uonpoa ~ s! mo33o ~ou oq/, "C[ pue IZ v!,t so~inoa!o
~uoano l o O q ~ l j oq~ leql os 'uo ,i pu~ JJO potum s! kL
"] '~ugonpuoo uooq OAeq lz,L pue IZ :tmje 'oidtuexo aod "pouod
j j o oql ~upnp opo!p e pue aols!sue:n e v.~,t pomlnoa!o s!
J ]uo~no iooqA~ij oq~ onb.mtloo] s.tq] qlt.A~ -uogelnpom po~po
-olqno p se UA~OmI onb!uttoo ~ e ~q 0tUO0a0AO 0q U~0 tu0Iq
-"e ! ~ m
/ -oad s!qz "0SaOA~o:toJoaoql s! oldd.u ]uo.uno oq] pue 'a0U0A
-uo0 oldtu!s oql j o ~eq~ O9.IAVISt offmiOA oldd V mdmo oq~
A o j o oPmiidtue oql ~,eq~,s.t .IOLIOAHO0sIq~,oi0 o~elueApes!p ou 0
".,(IU0A0 ;~.lOttl S09.IA0p
:toaxod oq~,u.t sosso I oq~, m,nq.m,s!p o~, :top:to m smm,s~s ie!o:totu
-tuoo ~uetu u! pmdope oae soouonbos ~u.tqm,iaxs p00ueApe
.1 OaOlAI "pou.tm,qo s.t uo.n,e:todo meapenb-anoj iin_J oq~, oouoq
pue p0S/0AOI St :luoMn9 oql '~u.tlonpuoo pue EL s:to~,s.ts
\ -[II~l1 qlI.l~ "~LII.7~I?,I;~I,I;~;glS.I ;gA.Up~)ql pue p;gSJ;~A0/SI. poads
oql oouoq '.uogooa!p otuus oql u! luo~no ql!ax mq ;DS.I;~A;9.I
ol 0~Ul[OA mdmo oq~, osneo ii!m ~,uo3 aod 0~; aXoloq oioo ~ n p
j o uop, onpoa ~u V "op~o Lmp ~,uoo aod 0g ~,e oaoz ol poonpoa
0 s[ 0~lZlIOA 0ql .,~[luonbosuo9 pul~ p0SlOAOI S.I iddns :tmotu
( oq~, honpuoo lzO pue E(I uoqA,~ "po.uodjjo oq~,~u!:mp ~,onpuoo
lzO pue E(I sopo!p pue 'povod uo oql ~u!:mp uo p0qm!A~S
oae 17 pue I s:tms!su~a mdmo OA.II.Isod '~U.UOlOtU ~u.un(I
uoge:touo~oa pue
0 ~UISa0AO:tJO "O'! ;uo~na pue 0~ellOA ~lq!SI~AOI ~Uilddns jo
olqedea aOU0AUO0 ,ue:tpenb-anoj a!seq e saxoqs 91z'Z oan~!:I
fiU!)leJ q pJeAAJO~t 9
fiU!JO:~OUJ pJEAAJO~tq
8:I.LS:IANO~) 9 c ] - 3 a J.NVaavnb-~no-I
3,!n)J !) e
.~a.,~aAuo:) 0"0-'.9"0 ~ue.~penb-oNtl ~P'E a.m6t.d
~[ddns plog ao o:mmtme oq] 0SIOAO.I0~l qg].IA,XS.IOAO0~Ul~q9
JO sueotu ~q mo po!~eo s[ ~UtSa0A0:~ oaoqax suopeoildde
uo!;oea] ui posn sotugotuos s[ ~! ~;io![dtu!s m! jo osneooq
'aOAOaXOH "suogeoildde ieuo!;vo:qp!un u! posn ~Iuo ~iietmou
s! ~! l~S~OAOapoods ao~otu j o olq~d~o ~ou s! ~!no~!o s!q~ oouIS
% oql oouoq '.Xlddns oql olu! ~Ioeq I C[ via s! ~uo~no oql :~oj qled
Xluo oql 'jjo stum E i uoq/~ ""7 ooumonpu! aolotu Kq pmitu!i
'XIOAII~ffOU dn spi!n q luo:uno aolotu 'spoFmd uo sl! ffupno
uo ~Cl la~aAuo:) D'G-D'G .~uelpenb-al6Ut.S ~t;,'E aJn6t._,-I
1
uo ~CI
%
~ m
~A
/'--..... /
/
uo ~I~
q aslnd uo
- o
O
v~
e7 .. ClX
o e7 J.
k
@
UO!SJOAUOZ) JaAAOd ":)'(30~. "D'(] :S.LIn981~) 8:II83ANOZ) 3A18(] 0g
Chapter 2.2 81

Step-up D.C.-D.C. Converters considered, and consequently are very flexible in their
application. Major inherent features include:
As with step-down converters, many alternative configura-
tions exist for step-up converters. Although a full description Multimotor loads can be a p p l i e d - this can be very
is not necessary, the principle is of value. economical in applications such as roller table drives,
Figure 2.47 shows a much simplified arrangement of a step- spinning machines etc.
up converter. When T is turned on, current builds up in Inverter operation is not dependent upon the
inductor L. When T is turned off, the energy stored in L is machine - indeed, various machines (induction, syn-
transferred to capacitor C via D. When the capacitor voltage, chronous or even reluctance) can be used provided that
which is the same as the motor armature voltage, reaches the the current drawn is within the current rating of the
desired level, T is turned on once more. C cannot discharge inverter. Care should be taken where a low power-factor
via T as diode D is reverse biased. motor is used (e.g. reluctance) to ensure that the inverter
can provide the required VARs.
In this way a stabilised voltage typically twice the input D.C. Inherent open-circuit protection, very useful in applica-
voltage can be obtained. This circuit is particularly useful tions where the cables between the inverter and motor
when operating on low-voltage supplies and can lead to very are in some way insecure (e.g. fed via slip-tings, subject
cost-effective converter designs. to damage etc.).
Facility to ride through mains dips can easily be pro-
A.C. TO A.C. POWER CONVERTERS WITH vided by buffering the D.C. voltage link with capaci-
tance or, where necessary, a battery.
INTERMEDIATE D.C. LINK
Motoring operation only in both directions is possible
without the addition of resistive dumps for braking
General
energy or expensive regenerative converters to feed
This category of A.C. drive, commonly termed a variable energy back to the supply.
frequency inverter, is by far the most important in respect of
the majority of industrial applications. It is considered here
as a complete converter; however the input stages have SIX-STEP SQUARE-WAVE INVERTER
been considered earlier in isolation, and their individual
A typical D.C. link square-wave voltage-fed inverter drive is
characteristics so described are, of course, applicable.
shown in Figure 2.48. The three-phase A.C. supply is con-
Alternative input stages to some of the drives are applicable.
verted to D.C. in the phase-controlled rectifier stage. The
Also, some converters may be used with a variety of rectified D.C. power is then filtered and fed to the inverter.
machine types. Only practically important combinations are
described. Note that the D.C. link reactor is small compared to those
used in current-source designs. Indeed, in drives up to about
The concept of these inverter drives is well understood - 4 kW it is not practically necessary. Some manufacturers
rectification of fixed frequency, smoothing and then omit the reactor in designs to 400 kW and above; however
inverting to give variable frequency/variable voltage to feed this has a significant effect upon supply harmonics and
an A.C. machine. Within this broad concept two major unduly stresses the rectifier and filter capacitor.
categories of drive exist:
The inverter switching elements shown as transistors T1 to
voltage source in which the converter impresses a vol- T6 are gated at 60 intervals in the sequence in which they
tage on the machine, and the machine impedances are numbered in the diagram, and each transistor conducts
determine the current for 180 . The feedback diodes D1 to D6 are connected in
current source in which the converter impresses a current inverse parallel with the transistors, and permit the retum of
on the machine, and the machine impedances determine energy from reactive or regenerative loads through the
the voltage inverter to the D.C. link.
For a star-connected motor, synthesis of inverter output
Voltage Source Inverters voltage waveforms is shown in Figure 2.49. The phase-to-
neutral voltage of the inverter has a six-step waveshape, and
GENERALCHARACTERISTICS
the corresponding phase-to-phase voltage has 120 con-
The fixed-frequency mains supply is a voltage source behind duction angle. The output frequency is controlled by the rate
an impedance. Voltage source inverters can be similarly at which the inverter transistors are triggered into conduction
by the inverter control circuitry. Reversing the firing
L D sequence of transistors in the inverter changes the direction
of rotation of the motor, and no switching of power leads,

i
either on the incoming supply or to the motor itself, is
necessary.
~c
The phase-controlled rectifier regulates the D.C. link voltage
and this, in tum, determines the magnitude of the output volt-
age from the inverter. Thus, the output voltage/frequency
relationship may be controlled to regulate the motor flux in
Figure 2.47 Step-up D.C.-D.C. converter the desired manner.
82 DRIVE CONVERTERCIRCUITS: A.C. t o A.C. P o w e r Converters w i t h Intermediate D.C. Link

D.C. link inverter

T3 I I T5

0
,

I ,

rectifier

Figure 2.48 Square-wave voltage-fed inverter

Very high-speed motor operation is possible by increasing


the output frequency. Faster switching devices such as MOS
transistors and insulated-gate bipolar transistors (IGBT) can
be used to achieve this performance.

I, It is known that the square-wave inverter gives objectionable


torque pulsations at low-frequency operation, below
VBc approximately 5 Hz. This pulsating torque is due to the
interactions of low-order harmonics with the fundamental
voltage, causing a stepping or cogging motion to the rotor
running at low speed. Hence, the pulsating torque limits the
low-frequency operation of the square-wave inverter.
Appropriate feedback control techniques or flux weakening
can attenuate the low-speed pulsating torque problems.
VCA
The existence of a phase-controlled rectifier to control the
voltage of the inverter as illustrated in Figure 2.48 is
an inherent weakness of this circuit. The phase-controlled
rectifier will present a low power factor to the A.C. supply,
at low speeds, and the D.C. link filter capacitor is large and
phase-to-phase voltages reduces the response time of the system to voltage and hence
speed changes. If the drive system is one for which regen-
VAN
c~
erative braking operation is a requirement, the rectifier has
I\ ,,,,,,, to be of inverse-parallel type. The input power factor and
response time of the drive can be improved by replacing the
//
'/ ]\\\ I// f II phase-controlled rectifier with a diode rectifier feeding a
r%. I D.C. chopper which regulates the input voltage to the
.I
\ I ~ inverter. For recovering regenerative energy of the load, a
phase-to-neutral voltage
two-quadrant chopper will be necessary. The alternative
Figure 2.49 Square-wave voltage-fed inverter, output supply converter arrangement of a diode bridge plus chopper
voltage and current also provides a fixed voltage link, which is more econom-
ically buffered, if mains dip ride through is required.
The advantages of the square-wave inverter are high effi-
The voltage-fed square-wave drive is usually used in low-
ciency (98 per cent), suitability to standard motors, potential
power industrial applications where the speed range is lim-
good reliability and high-speed capability. However, it suf-
ited to ten to one and dynamic performance is not important.
fers from low-speed torque pulsations and possible low-
Recently, this type of drive has largely been superseded by
speed instability.
PWM-type voltage-fed inverters. Nevertheless, the voltage-
In a square-wave inverter, each harmonic voltage amplitude fed square-wave inverter can be easily adapted to multi-
is inversely proportional to the harmonic order and hence motor drives where the speed of a number of induction
there are no pronounced high-order harmonics. These are motors can be closely tracked. It is also used in some high-
filtered out by the motor leakage inductances. frequency ( > 1 kHz) and some high-power applications.
Chapter 2.2 83

PULSE-WIDTH MODULATED (PWM)INVERTER _ carrierwave


A ~ ~ r ] ~ ~ ~ : sinewave
In the PWM inverter drive, the D.C. link voltage is uncon- ~/~ ~ ~ ~ ~ ~ ~ ~ar,,,~k~referencesignal
trolled and derived from a simple diode bridge. The output
voltage can be controlled electronically within the inverter i/A'
V',',V,
,, ,,'V"
, ,",,,,, V",,,,,", ,' ', V '8
by using PWM techniques. In this method, the transistors are , , ,
I
, t
II
~
II I I
,~_upperthyristor
I I I 'I ,, , , on
switched on and off many times within a half cycle to gen-
erate a variable-voltage output which is normally low in
+sV"l Fll II II II- -!
harmonic content. 0,v.I IIU II U_
A PWM waveform is illustrated in Figure 2.50. per phase output voltage, _~ ~_ lower thyristor on
(with respect to mid D.C. link point)
A large number of PWM techniques exist each having dif-

VIH
ferent performance notably in respect to the stability and
audible noise of the driven motor.
Using the PWM technique, low-speed torque pulsations are
virtually eliminated since negligible low-order harmonics
are present. Hence, this is an ideal solution where a drive
system is to be used across a wide speed range.
Since voltage and frequency are both controlled with the
PWM, quick response to changes in demand voltage and
PWM line voltage
U LIUHH
frequency can be achieved. Furthermore, with a diode rec-
tifier as the input circuit a high power factor, approaching
unity, is offered to the incoming A.C. supply over the entire
speed and load range.
PWM inverter drive efficiency typically approaches 98 per
nr F U1
cent but this figure is heavily affected by the choice of
PWM phase voltage
switching frequency- the higher the switching frequency
the higher the losses in the drive. In practice, the maximum Figure 2.50 Sinusoidal PWM line and phase voltages
fundamental output frequency is usually restricted to 100 Hz
in the case of gate turn-off thyristors (GTO) or about 1 kHz
for a transistor-based system. The upper frequency limit may
be improved by making a transition to a less sophisticated CONVERTER-FED SYNCHRONOUS MACHINE
PWM waveform with a lower switching frequency and
ultimately to a square wave if the application demands. Once rotating, a synchronous machine generates A.C. vol-
However, with the introduction of faster switching power tages which can be used for the natural commutation of a
semiconductors, these restrictions to switching frequency converter connected to its terminals. Indeed, the connected
and minimum pulse width have been eased. synchronous machine behaves as the mains in respect of the
A.C. to D.C. converters described earlier.
In general, a motor with a large leakage reactance is desir-
able to limit the flow of harmonic currents and thereby
Figure 2.51 shows the basic components of the drive system.
minimise losses.
A low-impedance or stiff D.C. current source is required and
is obtained from a controlled rectifier and a series reactor.
MULTILEVEL CONVERTERS
With a stiff current source, the output current wave is not
At high powers and/or high voltages it is not possible to greatly affected by the magnitude of the load.
implement PWM strategies with high switching frequencies.
The synchronous machine can be approximately represented
The waveform of such systems can be improved by pro-
by a counter e.m.f, in series with an equivalent leakage
viding intermediate D.C. voltage levels. Commercial sys-
inductance. The D.C. current is switched through the
tems of this type exist, but their application is quite rare.
inverter thyristors so as to establish three-phase, six-stepped
symmetrical line current waves. Each thyristor conducts for
Current Source Inverters 120 and at any instant one upper thyristor and one lower
thyristor remain in conduction.
GENERAL CHARACTERISTICS
It is necessary to maintain an approximately constant
Whereas each voltage-fed inverter can be used with most
angular relationship between the rotor and stator e.m.f.s and
forms of A.C. machine, a different design of current-source
hence automatically maintain the correct inverter frequency.
inverter is usually adopted for synchronous and induction
This is an important point. The inverter does not impose a
motors. Current-source drives are usually, but not always,
frequency upon the machine, rather the machine itself
single-motor systems, and since current is controlled, have
determines the frequency. The motor cannot therefore pole
simple short-circuit protection.
slip. The drive is accelerated by increasing the current fed to
In contrast to voltage-source inverters full four-quadrant the motor, which then accelerates and thereby increases the
operation is inherently possible. frequency.
84 DRIVE CONVERTERCIRCUITS:A.C. to A.C. Power Converters with Intermediate D.C. Link

TH5 TH2
,/ I I

v[
Ld Id

TH5' TH1

,C'
TI-
TH2'
|
supply rectifier inverter synchronous
machine

Figure 2.51 Converter-fed synchronous machine

As in the D.C. drives, the A.C. supply power factor is poor at upon the converter-fed synchronous machine drive having
low speeds. Full four-quadrant operation is possible without additional components to provide VAR compensation.
additional components.
Figure 2.52 shows a basic power circuit. The diagram
Special procedures are necessary for starting these drives somewhat belies the potential complexity of the VAR
because at standstill the machine voltage is not available to compensator. In its simplest form, this could comprise
commutate the current. In essence this is usually achieved by capacitors plus appropriate switches. Control of such a
momentarily switching off the D.C. link current every sixth system is somewhat involved and it is often better to use a
of a cycle. This allows the thyristors in the inverter to turn cycloconverter or even an auxiliary synchronous machine to
off so that the next pair can be fired. Above approximately 5 provide the commutation and motor VARs.
per cent of rated speed the machine generates sufficient This system is only appropriate for high-power drives,
voltage for natural commutation and control is undertaken in generally above 4MW, where an induction motor is
a similar manner to that of a D.C. drive. preferred.
Applications for this type of drive fall into two main cate-
gories. First, starting converters for large synchronous FORCED-COMMUTATED INDUCTION MOTOR DRIVE
machines, the converter being rated only for a fraction of the
machine rating. Second, as large high-power (and sometimes A forced-commutated induction motor drive is the most
high-speed) variable-speed drives for a variety of applica- widely used current source inverter at power levels in the
tions. Power ratings typically from 1.5 to 30 MW at speeds range 50-3500 kW at voltages up to normally 690 V (high-
up to 8000 r.p.m, are available. Also of import is the fact that voltage versions 3.3 kV/6.6 kV have been developed, how-
high-voltage drives are offered with supply voltages up to ever they have not proved to be economically attractive).
typically 5 kV, but systems up to 25 kV are in service where
Figure 2.53 shows the inverter and motor of the drive. The
the high-voltage converter technology is similar to that used
for HVDC power converters. D.C. link current Id, taken from a stiff current source, is
sequentially switched at the required frequency into the
CONVERTER-FED INDUCTION MOTOR DRIVE stator windings of the induction motor. The motor voltage
waveform is approximately sinusoidal apart from the
Unlike the synchronous machine, the induction motor is superposition of voltage spikes caused by the rise and fall
unable to provide the VARs or terminal voltage to com- of machine current at each commutation. Further distor-
mutate a converter connected to its terminals. Commercial tion is caused by the effects of slot tipple and D.C. current
schemes are available, however, which are closely based tipple.

II. . . . . . . . . . . . . . . . . . . . ~VAR
I r , compensator
r" . . . . . . . . . . . .

I I Ld Id
1
supply I 'I II 2 .~ { _7

M
1

induction
inverter motor
rectifier

Figure 2.52 Converter-fed induction motor


Chapter 2.2 85

Power can flow in only one direction via the diode bridge,
O which means that motoring torque can be developed only at
subsynchronous speeds. For reverse running it is necessary
TH 1 FH3 TH5 to reverse the phase sequence of the stator supply.
37 x7 x7
(; II ,, || This drive can be very economic when designed for opera-
II I|
,) tion over a limited speed range below synchronous s p e e d -
this is the useful operating region for fans, pumps etc. The
~7D1 \ 703 ~7D5
uo A ,L
B J.
C';
M) converter bridges required for such limited speed range
operation need only be rated at a fraction of that of the
machine which it is controlling. It is necessary in such
7D6 7D2 designs to provide a starter, usually a resistance to run the
~ZD4 I I motor up to the lowest controllable speed. This means that
'; II " should there be a fault with the converter equipment, the
II
system can be easily designed to run at full speed without the
ZTH4 7TH6 7TH2 controller.
Note that the supply harmonic currents and VARs asso-
IF
ciated with the converter part of the drive may be sub-
Figure 2.53 Forced-commutated induction motor stantially reduced by adopting a limited speed range
solution.
The operating frequency range is typically 5 to 60 Hz, the The static Kramer drive finds application mainly at ratings
upper limit being set by the relatively slow commutation between 1 and 20 MW, with induction motors with four or
process. Special motors with low leakage inductance do more poles (stability problems exist with two-pole motors,
offer advantage with this converter and allow reduced which can only be resolved with care). Speed ranges of 30
capacitance in the inverter and/or higher operating fre- per cent are typical (i.e. 70-100 per cent rated speed). The
quency. Below 5 Hz, torque pulsations can be problematic induction motor stator can be wound for any conventional
but PWM of the current can be used at low frequencies to voltage e.g. 6.6 kV, 11 kV.
ease the problem.
This system is most commonly used for single-motor DIRECT A.C. TO A.C. POWER
applications such as fans, pumps, extruders, compressors
CONVERTERS
etc. where very good dynamic performance is not necessary
and a supply power factor, which decreases with speed, is
General
acceptable.
This final category of power converter converts the fixed-
STATICKRAMERDRIVE frequency, fixed-voltage A.C. supply to a variable fre-
quency and/or variable voltage without an intermediate
The static Kramer drive is shown in Figure 2.54 and com-
D.C. link.
prises a slip-ring (wound-rotor) induction motor together
with an uncontrolled converter, D.C. smoothing reactor and
a fully controlled converter in the rotor circuit. Soft Starter/Voltage Regulator
The diode bridge gives an output voltage Va that is propor-
Figure 2.55 shows a typical soft start comprising inverse
tional to the slip of the motor. Vd is opposed by the D.C.
parallel connected thyristors in each supply line to an
voltage of the fully controlled bridge, a small potential dif-
induction motor. Alternative connections are available but
ference being sufficient to circulate current corresponding to
the principles are similar.
the required load torque. Ideally, neglecting losses, the fully
controlled bridge D.C. voltage sets the speed to which the The converter shown in Figure 2.55 is used to control the
motor will accelerate. Control is therefore very similar to voltage applied to the motor and in this way soften the
that for a D.C. drive. effects of switching an induction motor direct-on-line

Ls

v~

Figure 2.54 Static Kramer drive


86 DRIVE CONVERTERCIRCUITS:Direct A.C. to A.C. Power Converters

-__k

r~

AL
\
v

Figure 2.55 Typical soft start

Figure 2.56 Cycloconverter

(DOL). Although the converter will control the cur-


rent drawn from the supply, its most usual application is in
controlling torque to provide smooth jolt-free acceleration.
Because the stator frequency is unchanged, a reduced run-
ning voltage, and hence flux, equates to a large slip which
J Ap J Bp J Gp
results in additional rotor losses - care must therefore be
i,o
taken in its application.
A

In a number of specialised cases, purpose-designed high-


resistance rotors (or slip-ring motors with external rotor res- JOAn J
,~Bn
istors) are used to form a variable-speed drive - the rationale ~ ) On output
for such a system is based more on history than technology.
More recently, such converters have been employed as
combined soft starters/power-factor controllers/energy-
saving devices. The case for significant energy saving by
J
oAm J
08m J
/0 Grn
this form of converter is often hard to prove.
A crude form of frequency control is possible by modulating
(varying cyclically) the thyristor firing angles at the required
output frequency. Commercial systems are available but
they are of limited value as supply current and motor torque A B C
are of poor quality. input

Cycloconverter Figure 2.57 Matrix converter

A typical scheme for a cycloconverter drive is shown in


Figure 2.56. Each motor phase is supplied, in effect, from a
mills, minewinders etc. They are also used to feed multi-
dual A.C. to D.C. converter which was described earlier. It is
motor loads such as roller tables.
usual to employ circulating current-free converters. To avoid
line to line short circuits, isolating transformers are used on Owing to the modulation of the converter firing angles, the
the supply side. By modulating the firing angles of the dual- harmonic content of the A.C. supply is complex and designs
bridge converters, a controllable three-phase set of voltages for appropriate harmonic filters somewhat involved.
can be produced suitable for feeding polyphase A.C.
The cycloconverter is suitable for feeding both induction and
machines.
synchronous machines. In specialised applications such as
The drive is inherently four quadrant. The maximum output wind generators, cycloconverters have been placed in the
frequency is limited to approximately half the supply fre- rotor circuit of a slip-ring induction motor. Such a system,
quency by considerations related to harmonics in the motor known as a static Scherbius drive, is detailed below.
currents and torque, stability and dimensions of the drive
components. The cycloconverter therefore finds application Static Scherbius Drive
in low-speed drives. The complexity of the drive also means
that only high-power systems ( > 1 MW), or specialised The static Scherbius drive is closely related to the static
applications (e.g. conveyor drives for use in hazardous Kramer drive, with the single-quadrant diode bridge in the
environments), are economic. They are used on large ball rotor circuit replaced by a cycloconverter.
Chapter 2.2 87

The cycloconverter is used as the voltage and frequency new, recent advances in power devices offer the potential to
changer between the rotor and the supply. The cyclo- overcome many of the drawbacks inherent in the circuit
converter is inherently regenerative, and the output can be when the switches comprise inverse parallel thyristors.
controlled up to half the supply frequency in both phase Limitations in the maximum output voltage (86 per cent
sequences. It is thus possible for the system to operate as a input) means that its application in the commercial industrial
full four-quadrant drive. For a given converter rating the market could be problematic. There are prospects for use in
range of speed control is therefore twice that of a static integrated motors and some servo systems where machine
Kramer drive. voltage is not seen as critical.
The relative complexity of the drive limits its application to Commercial systems are available only for very specialised
somewhat specialised high-power applications where a very applications at present. It has yet to be proven to be a
limited speed range only is required and perhaps stringent practical and cost-effective industrial drive.
harmonic current limits have been imposed by the supply
authority.

Matrix Converter
Recently, attention has been refocused on the matrix con-
verter shown in Figure 2.57. Although the basic circuit is not
CHAPTER 3

Speed and Position Feedback Devices

i 1 D.C. TACHOMETER GENERATOR 90

m 2 A.C. TACHOMETER GENERATOR 92

3 RESOLVER 92

4 ENCODER 97

@ s SELECTION OF A FEEDBACK DEVICE FOR A DRIVE SYSTEM 99

B 6 MECHANICAL CONSIDERATIONS 100

/I 7 GLOSSARY OF TERMS 101

The precise control of speed, position or acceleration Before considering the various forms of sensor, it is neces-
requires appropriate measuring systems to be applied to the sary to clarify the, often misunderstood, difference between
controlled variable. Although considerable progress has resolution and accuracy of a feedback device:
been made in the development of sensorless drive systems,
these schemes tend to result in improvements in dynamic Resolution
and shaft performance as opposed to precise positional or
The resolution of a feedback device is most simply described
speed accuracy.
as the number of measuring steps in one revolution of the
This chapter deals with sensors extemal to the drive module motor shaft. For an incremental encoder it is the number of
itself. pulses per revolution.
90 D.C. TACHOMETERGENERATOR: General

Resolution is often described in terms of number of bits. This shaft, the graduation, the optics and the electronic signal
is related to the twos complement; for example, a 12-bit processing.
feedback device has the equivalent of 212= 4096 measuring For analogue feedback devices, the accuracy is influenced
steps per revolution. primarily by the winding distribution, eccentricity of the air
gap, uniformity of the air-gap flux and the elasticity of the
For analogue feedback devices such as resolvers, where a
resolver shaft and its coupling to the motor shaft. When
resolution is quoted it usually refers to an associated
used, the accuracy of electronic signal processing can also
resolver-to-digital converter.
heavily influence performance.
Accuracy Electronics can be used to improve accuracy. This is pos-
sible as many of the causes of error are fixed and correction
The accuracy of a sensor is best described as the position
techniques such as calibrated look-up tables can be used.
deviation within one signal period or measuring step.
To ensure smooth drive performance, a speed feedback
In an encoder the accuracy is influenced primarily by device must give high resolution. At low speeds the position
the eccentricity of the graduation to the bearing, the elasti- deviation of the encoder within one signal period, i.e.
city of the encoder shaft and its coupling to the motor accuracy, affects speed stability.

1 D.C. TACHOMETER GENERATOR

GENERAL OUTPUT VOLTAGE RIPPLE

The D.C. tachogenerator, also known as a tachometer or The ripple voltage is defined as the variation in output vol-
more commonly as a tacho, is an electromagnetic transducer tage caused by the number of armature coils for a given
that converts mechanical rotation into a D.C. output voltage. design. Hence, the basic ripple frequency is dependent upon
This voltage is directly proportional to rotational speed. In the number of coils and is a measure of the number of cycles
addition, it has a polarity sense dependent on the direction of in one revolution.
rotation.
The fundamental ripple can be distorted by intemal random
The basic theory of the D.C. motor described in Chapter 1.1 and cyclic electrical noise, such as commutation due to
applies to the tacho. There are, however, some aspects which varying contact resistance between the brush and commutator
differentiate the tacho from the generic D.C. motor: interface and brush bounce. In addition, misalignment of the
driving medium and the output shaft of the tachometer would
1 Linearity of the output- the tacho must provide a D.C.
cause serious modulation in the output signal.
output voltage proportional to the shaft speed with a
defined linearity.
2 Smooth output - the output voltage must be relatively
smooth/free from ripple, particularly in the frequency
range in which the drive is operating.
3 The output voltage for a given speed should be constant
with changing temperature.
The tacho output is directly proportional to speed. t-~

O
This relationship can be expressed in the form: O
"t~ ---- vg/n
e-
(D
O')
0
where Vg is the output voltage of the tacho, Kg is the transfer c-
O
constant of the tacho and n is the speed of the tacho. Kg is
often quoted in volts per 1000 min-1.
There are two basic forms of D.C. tachogenerator, shunt-
wound and permanent-magnet types. Most modem tachos shaft speed n
are of the permanent-magnet type, which are compact and Figure 3.1 Tachogenerator output voltage versus shaft
robust. speed
Chapter 3.1 91

To overcome the internal distortion, high-quality brushes Linearity of output voltage is also affected by the output
are used, having low contact resistance and sufficient current drain. This current in turn depends on the voltage
brush force to maintain positive contact stability without gradient of the tacho, the mechanical speed and the load
degrading the brush life. Enough cannot be said to ensure that impedance.
the end user correctly aligns the drive and driven shafts
A further impact could be armature reaction and saturation
because it becomes critical if a good signal is required at the
effects particularly in very high-speed applications.
lower speeds.
Linearity should not be confused with the tolerance given
The ripple voltage is usually measured in r.m.s, although
to the output voltage which is normally given as +10
peak to peak values are sometimes quoted.
per cent.
High-frequency brush ripple and its harmonics is rarely
quoted in performance data since a minimum of electronic STABILITY OF THE OUTPUT
smoothing effectively removes these components without
adding a significant time delay to transient events.
The stability of the output voltage is mainly determined by
the state of the commutator surface. To maintain the com-
mutator in good condition the recommended load RL should
TEMPERATURE EFFECTS be typically between 50 and 500 times RA, where RA is the
armature resistance of the tacho. The use of high values of
Tachogenerators are specified in respect of their temperature
RL, within the above limits, will improve linearity and sta-
error. The quality of this parameter is closely related to the bility, particularly under conditions of fluctuating load. The
design principle and materials used in the manufacture of the use of values below or above the limits can lead to excessive
tacho. In order of increasing cost:
brush and commutator wear, or to increasing instability of
Uncompensated generators the output.

This is the lowest cost type, having a temperature error up to The ideal commutator condition is with a carbon track
0.2 per cent per K of magnet temperature. deposited on the commutator surface which has a shiny, dry
appearance, evenly distributed around the commutator cir-
Compensated generators cumference. A thick deposit of oily or powdery carbon
suggests that contamination has occurred and the output
This type has a temperature error in the area of 0.05 per cent
will be unstable. A bright, polished copper surface is not
per K of magnet temperature. Low-cost ceramic magnet
recommended and commutator cleaning sticks, abrasive
materials are often used in combination with thermistor-
powders or commutator lubricants must not be used. If
based compensating networks. This temperature compen-
cleaning becomes essential, the brushes must be removed,
sation system limits the application of this type of tacho to a
the armature may be rotated typically up to 1000min-1
temperature range up to 75C.
and a hard felt pad moistened with a nonoily cleaning
The output impedance is generally higher than that of the fluid held against the commutator surface to remove car-
more expensive types. bon and debris. It is important to allow the commutator
surface and the intersegment insulation to dry thoroughly
Stable generators before reassembly, and to fit new brushes if possible
This type has a temperature error in the area of 0.02 per cent since the old ones may well be impregnated with the
per K of magnet temperature. They are usually based on harmful contaminant. In this case it is desirable to run the
Alnico 5 magnets. tacho at speed with a low value of load resistance to bed in
the brushes.
Ultrastable generators
This type has a temperature error in the area of 0.01 per cent M A X I M U M TERMINAL VOLTAGE
per K of magnet temperature. To achieve this low level of
temperature sensitivity Alnico 8 or Alnico 5 magnetic The maximum terminal voltage is a function of the bar to bar
material in combination with a compensating alloy in the voltage seen at the commutator and a value of 12 volts bar to
magnetic field circuit is used. bar is typical to ensure stability in the output signal. To
exceed this value will result in poor stability and, in the long
term, destruction of the commutator bars and brushes due to
LINEARITY AND LOAD EFFECTS the high self-induced currents in the windings under com-
mutation. The maximum voltage is rarely quoted, rather the
Linearity is defined as the deviation from a straight line maximum speed.
voltage/speed relationship at any given point within the
speed range, as a percentage of the theoretical output at that M A X I M U M OPERATING SPEED
point.
Either voltage or commutation will be the limiting factor on
A limiting factor for linearity is the speed at which
the maximum operating speed. These values are normally
the machine is rotating. A deviation will occur at speeds
given in the manufacturer's data sheets.
beyond the machine's capability, caused by aerodynamic
lift of the brushes etc. and the hysteresis losses in the Care should also be taken on very high-speed applications
armature core. that the mechanical speed limit is not exceeded.
92 D.C. TACHOMETERGENERATOR:Mechanical Construction

brush assembly

bearing

- - t(-.-------.-)

armature assembly field assembly

Figure 3.2 Typical construction of a D.C. tachogenerator

MECHANICAL CONSTRUCTION even are wound onto the same armature as the motor they are
monitoring.
The mechanical construction of tachogenerators varies Figure 3.2 shows a typical arrangement used in many drive
considerably. Many are separate devices mounted on the applications.
outside of a motor or machine, and others are integral, some

2 A.C. TACHOMETER GENERATOR

A.C. tachogenerators/tachometers generate a three-phase linearity error of approximately 1.5 volts due to the forward
A.C. voltage proportional to speed, which is rectified into voltage drop of two diodes. This error is essentially constant
a D.C. voltage via an integral, usually three-phase, diode throughout the speed range.
bridge. The polarity of the D.C. output voltage is not
dependent upon the direction of rotation so can only be Output voltage ripple is typically in the order of 4 per cent.
used on drives having only one direction of rotation. The
These are low-cost units with moderate performance used on
advantage of such generators is that they are almost main-
unidirectional applications.
tenance free, being of brushless design. The rectifier has a

3 RESOLVER

The resolver is a specific member of the synchro family. A sensing devices which can operate together to provide a rota-
synchro is a general term for a family of angular position tional torque, for light loads. Alternatively, they can provide
Chapter 3.3 93

(CX or TX) (CT or TR)


transmitter receiver supplyfor
,, S3 ~ torquechain
= ,R1 ,_,1~s3
input ,"N~ N

,n0uti
angle @
"r Sl
I
' ~
i T S2 "', outputsignal
; ~
output
, angle
"'. for controlchain
"

Figure 3.3 A simple synchro

torquetransmitter
(TX) torquedifferential
transmitter(TDX) torquereceiver
(TR)
$1 $3
(TX) inp
shafta
)
I
r TDXA L O / ~ . ~TgX T ~ ~ l . . ( T R , . s h a f f ,
stator~ R"~..kZ.k~'
~ "~l:pul: angle
TXl
rTtXr ~S2 S2~
~_......I~ ~S2 [rTtRri
R2 R1 inputangle R2 R1
to TDXshaft
Figure 3.4 Torque transmitter~differential transmitter~receiver

a signal with direction and magnitude information enabling another similar three-phase stator, an alternating field will be
a servomotor system to provide rotational torque. A resolver set up in alignment with the first. A rotor within that stator
is a modified form of synchro used for resolving angular field will have a voltage induced in it the amplitude and
position into coordinate data for use in control systems. phase of which will be an indication of the input rotor
position.
DESIGN PRINCIPLES If the output rotor is energised from the same source as the
input rotor, it will develop an alignment torque.
All synchros and resolvers have the same basic construction,
a wound rotor, carried on precision bearings, free to rotate
inside a fixed, wound stator. Operation is based on the
TORQUE SYNCHROS
principle of a rotary transformer, so that when windings are
in electrical alignment a maximum voltage is induced from
Torque synchros are applied when it is required to transmit
the primary to the secondary. When the two windings are light torque without additional servo components. Their use
situated at right angles to each other, there is no coupling and
is normally limited to driving instrument pointers, compass
no voltage is induced in the secondary winding. At any other cards and similar applications. Typical components used in
angle a voltage will be generated which takes the form of a
systems are:
sine wave as the rotor shaft moves through 360 electrical
degrees. torque transmitter (TX)
torque receiver (TR)
SYNCHROS torque differential transmitter (TDX)
These components may be used together in a torque chain as
This simplified description ignores the fact that a single follows.
winding on a rotor and stator would not give angular rotation
information, so the synchro family is based on a three-phase From Figure 3.4 it can be seen that when the two rotors are
transmission system. The voltages induced in the three stator energised from the same source and the stators are connected
windings are displaced with respect to each other by 120 as shown, there will be a torque developed by interaction
electrical degrees. If the three stator coils are connected to of the stator field and the rotor field of the receiver. The
94 RESOLVER: TorqueSynchros
torque will be to bring the rotor of the receiver into align- rotoroutput
ment with that of the transmitter. When aligned, the
torque is zero. The torque/misalignment curve is of sinu- ~ . Ii VrIteasigdeu
(arl'm's')/7~
soidal form between zero and 180 with its maximum
value at 90 .

Typical alignment accuracy for torque synchro chains is


+ 1. Where this is not sufficiently accurate gear reduction
systems may be used, and then the accuracy can be increased
by the gear ratio, less a factor for gear error and backlash.

Typical performance figures for torque synchros: clrotati


ockwionse" minimium
coupl ng "-counter
clockwise
rotation
transmitter accuracy: + 6 . . . 10 arc min
receiver accuracy: + 5/60 arc min Figure 3.5 Effect of null output voltage
receiver/transmitter: + 6/45 arc min
differential transmitter: 6 . . . 10 arc min

R1, $4
CONTROL SYNCHRO
rotorI I stator
The control synchro is similar to the torque synchro but with R3 =$2
high impedance windings, which reduce the system current
loading, magnetic nonlinearity errors and temperature rise.
This often enables smaller components to be used in a given
_.$1
application. The magnetic circuit is modified to provide stator
uniformity of output, rather than torque output. -$3
The components of the control synchro are the control Figure 3.6 The basic configuration of a resolver
transmitter (CX), control transformer (CT) and control dif-
ferential transmitter (CDX), and they may be used in the
same way as torque synchros except that the rotor of the
control transformer provides an electrical output signal as
Furthermore, the windings are displaced by 90 to each other
the input to a servomotor drive amplifier, which provides the
compared to the 120 spacing of the synchro. The basic
error correction torque.
configuration is as shown in Figure 3.6.
When the shafts of CX and CT are aligned, a null voltage The prime purpose of a resolver is to provide Cartesian
will appear at the rotor terminals of the CT. A slight coordinate output signals from a polar coordinate input,
deviation from this alignment will produce a signal in the plus the ability to add mechanical rotation of the resolver
rotor winding with phase relative to the reference voltage shaft.
which will depend on the direction of deviation.
Depending on the control circuit requirements,
The control transformer can therefore be considered a null resolvers may be supplied with one or two windings on the
detector and it is most often used in this way. However, the stator. The input may be fed to the rotor or the stator,
null is never zero, there is always a residual voltage present, although rotor-fed units are more usual. It is often con-
as shown in Figure 3.5. This is due to stray coupling within venient to build in a voltage reduction ratio between the
the laminated stator that results in an in-phase voltage, primary and the secondary windings known as the trans-
a quadrature voltage, both at fundamental frequency, plus formation ratio (TR), to match into the next stage of the
a number of harmonics. The level of this residual voltage servo control electronics.
is typically in the range 3 0 - 1 0 0 m V depending upon the
supply voltage and individual specification, with some 50 to
Computing Resolvers
80 per cent at fundamental frequency.
Computing resolvers are primarily used for calculating
Typical performance figures for control synchros:
trigonometrical functions and they are available with a
transformer accuracy: + 6 . . . 7arc min feedback winding built in, or without. They are normally
transmitter accuracy: + 6 . . . 8 arc min stator fed and the nonfeedback versions are for operating on
differential transmitter: + 6 . . . 10 arc min substantially constant voltage sources. When there is a
likelihood of the source voltage changing, feedback versions
can be supplied. These are units with additional windings in
RESOLVER the stator slots, in which a voltage is generated and com-
pared with the input; the difference is amplified by a high-
General gain amplifier and fed to the main windings. This feature
compensates for variation in source voltage, and also redu-
The resolver is a special version of a synchro and has ces the effect of variation in frequency, winding temperature
two windings on the stator and one or two on the rotor. and load impedance.
Chapter 3.3 95

Typical performance figures for computing resolvers: Multipole Resoivers


Nonfeedback
Multipole resolvers may be used where higher accuracies
phase shift: 2 . . . 2 0
are required and emulate the geared systems previously used
null output: 1 0 . . . 60mV
but without the additional hardware and associated gear
sine deviation: 0 . 1 . . . 0.25 per cent
errors.
Feedback
phase shift: 5 . . . 2 0 Typical performance figures for multipole resolvers:
null output: 1 5 . . . 80mV phase shift: 6 . . . +16
sine deviation: 0 . 0 5 . . . 0.2 per cent null output: 1 ... 15 mV
electrical error: 3 ... 7 arc min
Phase Shifting A.C. Rotary Pickoffs
If a two-phase, or quadrature, supply is applied to a two- A.C. rotary pickoffs are transducers having a single-
phase stator winding the electromagnetic field in the air gap phase input and output. They normally operate on 400 Hz
will rotate and generate in the rotor voltages of equal mag- or higher frequencies, and the output voltage/rotor angle is
nitude but 90 displaced. The phase relationship of the rotor reasonably linear. There are essentially two types, one
to the stator voltages will be determined by the rotor angle. having an angular range up to 25 and the other with an
This is a useful technique for generating a supply of variable angle up to 65 . The narrow-angle version has four stator
phase for a number of test applications. coils connected in a balanced bridge configuration as shown
in Figure 3.8.

Brushless Resolvers The output voltage is the result of the impedance


change within the magnetic circuit caused by the position/
Originally all synchro devices employed slip rings and rotation of the salient pole rotor. The voltage output sensi-
brushes to feed current to the rotor, but brushless resolvers tivity in volts per degree is proportional to the excitation
have been developed in which current is fed to the rotor voltage.
through a circular transformer mounted at the end of the unit.
Generally, this technique is used only for rotor-fed units For the wide-angle rotary pickoff, the primary excita-
having a single rotor winding. Two rotor windings can be tion on the stator is separate from the secondary winding
provided but a second transformer is then necessary and the on the rotor. The rotor is eccentric in the stator bore resulting
construction can become uneconomic. in a variation of flux linkage as the rotor rotates. The sen-
sitivity is proportional to the transformation ratio between
primary and secondary turns and may be made larger than
Typical performance figures for brushless resolvers:
that for the narrow-angle type. A centre tap can be provided
phase shift: - 8 . . . +30 in the secondary winding to give two balanced outputs,
null output: 1 5 . . . 120mV which are in antiphase. A typical output curve is given in
electrical error: 3 . . . 15 arc min Figure 3.10.

resolver stator assembly transformer stator


resolver body / assembly

IL
rotor assembly

resolver rotor transformer rotor


stack assembly assembly

Figure 3.7 Sectioned view o f a brushless resolver


96 RESOLVER:Resolver

output Resolver-to-Digital Conversion


red 1 Of all angle-measuring transducers the resolver is inherently
the most robust and stable in long-term performance and in
relatively hostile operating environments. Where a digital
oreen4 output signal is required the analogue resolver can still be
used in conjunction with resolver-to-digital (RJD) conver-
sion electronics.
R/D converters can be designed to accept input from both
black synchros and resolvers. An internal transformation circuit is
used to convert synchro signals to a resolver/quadrature
Figure 3.8 Narrow-angle A.C rotary pickoff, stator coil signal format.
configuration
A number of techniques have been applied to the R/D
conversion. Common is the use of single or double RC
phase-shift circuits in which the zero crossing times of
stator the two resolver format signals are compared. The differ-
excitation ence is used to gate a clock, the frequency of which can
R1 $1
be scaled to indicate the digitally coded angle.
T T Another is the real-time function generator, in which
two resolver format signals are applied to trigonometric
"', L ~ ~ bidirectional
.]]~i tatin function generators, providing an analogue output volt-
age that is integrated and digitised, then fed back to
balance a bridge from which the value of the angle could
be derived.
An improved version is the ratio-bounded harmonic oscil-
- ~', rotor lator, although it suffers some disadvantages in that it is not
" l i d "', output a real-time measurement so, like the RC converters, it
$2 experiences staleness errors.
Figure 3.9 Wide-angle A.C. rotary pickoff, stator coil With the introduction of monolithic semiconductor devices,
configuration ratiometric tracking converters were widely adopted. This
technique is based upon a servo loop in which the converter
tracks the input continually while there is any change taking
place in the ratio of the sine and cosine signals. Since the
typical output
characteristics system is ratiometric, neither changes of input voltage, nor
~10- voltage drop in the lines between resolver and converter are
of significance.
~ 8
7
For high-accuracy R/D systems the repeatable nature of
5 a resolver's error profile can be compensated for by means
shaft GW 4
degrees a3
80706050403020 1 phase shift
I I I I I I I I I I I I I
(o lead)
2I0 20304050607080 degrees 5
3 shaft CCW d 3.0
4 6 2.5
5 i_
2.0 transformation ratio / ~
6 output volts c
0
7 in-phase ..~
t~ 1.5 ~,..versus
frequency ( l e f t - h a n ~
8 component E
9
~ 1.0
- 10 " 0.5 15
i__

Figure 3.10 Typical output curve of wide-angle A.C. rotary " 0 J,O
30 100 300 ~ 1Ok 30k lOOk
pickoff Hz
phase shift ~"~uency'
versus frequency (right-hand scale) ~ 90

Typical performance figures for A.C. pickoffs:

phase shift: 2 . . . 3 0 80
null output: 2 5 . . . 4 5 m V Figure 3.11 Performance characteristics of resolvers with
voltage sensitivity: 220...750mV/degree variations of temperature and carrier (supply)
output linearity: 0.4... 3 per cent frequency
Chapter 3.4 97

of a look-up table. This technique can be applied to complex range 50 Hz to 20 kHz with a bandwidth of 1 kHz. It is of
system installations or to simple R/D conversions to enhance value therefore to illustrate the dynamic characteristic of a
the performance. typical resolver over a range of frequencies and severe
Although it is often convenient to utilise a reference supply operating temperatures.
voltage and frequency in common with other servo com- From the typical curves in Figure 3.11 it can be seen that the
ponents, operation at higher frequencies can be of benefit for optimum range is from 400 Hz to 5 kHz.
R/D conversion. Many designs of converter operate in the

4 ENCODER

The encoder principle is simple: a photocell is etching techniques. Imperfections in the gratings can be
positioned behind a slotted disc, or more commonly a averaged out increasing the overall measuring accuracy and
transparent disc with lines photographically drawn on it. A reliability.
light source shines through the disc into the cell. The pho-
tocell output is monitored as the disc is rotated with In practical encoders a second signal is produced phase
the encoder shaft. The number of pulses per revolution shifted from the first by 90 . These signals are then digitised
(p.p.r.) can range from tens to several thousand. The choice (squared) and fed into a counter in which counting pulses are
of encoder is important. Physically, the encoder must be derived from the square-wave edges. Simultaneously, the
compatible with the environment and also with the square wave signals are applied to a direction-determining
mechanical system. circuit, which transmits the counting pulses to the positive or
negative port of the bidirectional counter. This method of
Encoders have the advantage that their output is in the form determining linear or angular displacement by counting
of a pulse train the frequency of which is not affected by pulses is commonly referred to as the incremental measuring
temperature or attenuation of long cable runs as is the ana- method.
logue signal of a tachogenerator or resolver. It is therefore
potentially capable of contributing to extremely accurate Typically, the finest grating used is 8 l.tm. Finer gratings can
digital speed control. be manufactured but cannot be scanned in the same manner
as it would be necessary to maintain too high a tolerance on
Encoders can be classified into two primary types: the gap between the two gratings. Scanning signals of a
relatively coarse grating are therefore fed into a circuit
(i) incremental encoders that produce digital signals
which interpolates between signals increasing the apparent
which increase or decrease the measured value in
resolution by a factor of, typically, five.
incremental steps
(ii) absolute encoders that produce a code value, which
The encoder produces a train of pulses the frequency of
represents the absolute position directly
which is proportional to speed. The direction of rotation
can be determined if necessary by pulses from a second
These two types of encoder have somewhat different tech-
photocell, displaced 90 in phase from the first, as shown in
nologies at their heart and may therefore be considered
Figure 3.12.
separately.

INCREMENTAL ENCODER
a I I I I I I
The basic technology of incremental encoders is often
referred to as the Moir6 principle - a photoelectric scan- bl I I ! I I I
ning method to produce periodic signals using two fine clockwise rotation
gratings, which are positioned closely to each other and have
approximately parallel and equally spaced lines. If the I I I I I I
gratings are moved relative to each other periodic fluctua- b I I I I I I I
tions in brightness can be seen. These fluctuations are con-
anticlockwise rotation
verted to electrical signals via photoelectric sensors.
Using this method it is possible to scan very fine graduations Figure 3.12 Pulse trains corresponding to bidirectional
which can be produced very accurately using photographic rotation o f an encoder
98 ENCODER:Incremental Encoder

Incremental encoders are available in the following common be turned at low speed by one revolution or until the marker
forms: pulse is located.

Incremental - readily available from a wide variety of Where it is necessary to know the absolute position at tum
suppliers. Almost any line count available up to 5000 per on and without rotating the motor, it is possible to have
revolution. Special line counts and output options are a battery-backed encoder and control electronics, or to use
easily obtained. an absolute encoder.

Incremental with block commutation signals - again,


readily available but somewhat constrained by lack of ABSOLUTE ENCODER
industry standards. Mounting configuration, signal con-
ditioning and power supplies vary widely. Available in In absolute encoders the value of the actual position is
line counts up to 8000, plus block commutation signals immediately measured when the system is switched on.
(120 elect blocks) for two, four, six, and eight-pole Absolute encoders do not therefore need a counter since the
motors. They are being developed in both hollow-shaft measured value is derived directly from the graduation
and modular versions by a variety of encoder suppliers. pattern. In most cases the output from the encoder is in the
form of a pure binary code or in Gray code.
Incremental with sine wave c o m m u t a t i o n - this type of
encoder generally has sinusoidal quadrature outputs, A further coding exists which is a direct derivative of the
with a 1 volt peak-peak amplitude. Commutation is Gray code, and is called the Gray excess code. This com-
accomplished using a quadrature one cycle per revolu- prises a section from the middle of a Gray code pattern and
tion output. permits the transmission of other than 2 and yet remains a
unit-distance code.
Marker pulses are available on incremental encoders to give
a precise mechanical reference position. This is used when In the example in Figure 3.15 where the last two code
the absolute position of the motor shaft is required. In such values have been omitted from the graduation pattern, to
cases at start up of a machine a procedure often referred to as give 28-position resolution, the code would be described as a
homing is undertaken which typically requires the motor to 28-excess-3 Gray code.

mmmmm ml~mml~mirJI H m,lmm~mml;mmi~mmm~E ~ ~ ~ mhl m[a mu ~ ~ ~ ~ r~4 w l ~ ~ ~ ~ ~ ~ ~ W


m m m m m m m m m m m m m m m m
mm mm mm mm mm mm mm mm
mmmm mmmm mmmm mmmm
mmmmmmmm
Figure 3.13 Pure (natural) binary code

mm mm mm nm mm mm mm mm
mmmm mmmm mmmm mmmm

Figure 3.14 Gray code

28 positions

Figure 3.15 Gray excess code


Chapter 3.5 99

SIN/COS ENCODER where incr is the timer count or incremental count, N is the
line count of the encoder and ff~0is the zero position.
The optical arrangement of a sin/cos encoder is similar to One incremental step is equivalent to a 90 (electrical) phase
that of an incremental encoder. They provide the differential shift of the signals, A and B.
analogue output signals ( i A, 4- B, and 4-R) equivalent to
those of an incremental encoder but of sinusoidal wave The phase ~ of the sinusoidal signals A and B can be used to
shape, and of magnitude typically 1 V peak to peak. interpolate the position between two consecutive line counts
or four incremental steps, which are equivalent to each other.
The sinusodial wave shape allows high-resolution position It can be calculated as
determination. Direction and speed of rotation is detected as
in the incremental encoder. ~5 = 90 + arctan (B/A), if A > 0
The angular position can be determined by knowing the line q~ = 270 + arctan(B/A), if A < 0
count of the encoder and, when between two consecutive
increments or lines, deriving the phase from the analogue which has the advantage that the absolute amplitudes of A
signals A and B. The reference mark R provides absolute and B, which are a common function of the encoder's
position determination, if the angle at which the encoder is rotation speed and supply voltage, do not affect the ratio and
mounted is known. hence the angle. Since the arctan-function is ambiguous, one
has to check the sign of the sinusoidal signals A and B to
The incremental count and hence the incremental posi- identify the correct quadrant.
tion can be determined by a timer that counts up when
A is the leading sequence and counts down when B is It is this ability to interpolate between the line counts asso-
the leading sequence. When digitised, both edges of A ciated with incremental encoders which gives the sin/cos
and B are counted, thus the incremental position t~incr is encoder such an advantage in applications where high
given by resolution is required. Care must be taken in the practical
implementation of sin/cos encoders due to the low signal
~incr-- [360/(4 x N) x incr] + ~o levels which are commonly used.

5 SELECTION OF A FEEDBACK DEVICE FOR A DRIVE SYSTEM

The combination of any motor drive package must consider High-speed o p e r a t i o n - any feedback system will
a number of criteria including: generate data based upon the resolution of the sensor,
multiplied by the interpolation factor of the interface
Absolute versus incremental p o s i t i o n - the need for electronics (interpolation is a method for producing
absolute or incremental position information is generally measurement steps which are smaller than a quarter of
application dependent, and many drive systems can the scanning signal period).
develop or handle either type of interface.
For most resolver-based systems, the feedback element
Tachometers inherently give speed information and do provides one cycle of output per revolution and so is
not give position. Position can only be derived by inte- an absolute position device. Because the position infor-
grating the speed signal. mation is produced by amplitude modulation of a
Resolvers inherently provide absolute feedback. In cases carrier signal, it usually needs to be converted to a
where absolute position information is needed, e.g. for digital format before it can be used by a modem motor
robotics applications, a resolver can therefore be a controller. Resolver-to-digital converters (RDC) inter-
good choice. polate the resolver output signals and provide typically
10, 12, 14, or 16-bit results, depending upon the con-
Rotary encoders can provide either incremental or verter used. When coupled to a 12-bit RDC, a position
absolute outputs and, in fact, most absolute encoders measurement resolution of 360o/212 or 5.27 arc minutes
now provide both. is obtained.
100 SELECTION
OF A FEEDBACK DEVICE FOR A DRIVE SYSTEM

For an encoder, the maximum speed is limited primarily Hall effect switches are in the same range, and are
by the frequency response of the sensor. Encoders gen- typically very difficult to align.
erally have a maximum frequency response capability
in the range 200-400 kHz, which allows a 4096 cycle Resolver and optical encoder commutation can be good
per revolution encoder to turn at 6000min-1 without to 10 arc min with some careful planning, and they are
missing counts. Higher performance is available at very simple to align.
higher cost, with a maximum frequency response cap-
e Accuracy- in general, optical encoders have an advan-
ability of approximately 1 MHz.
tage in this area. Specially fabricated resolvers can
be obtained which are accurate to 4-2 arc min, but in
Low-speed o p e r a t i o n - low-speed operation requires
general resolvers have accuracy ratings of 4-2 to
high resolution and accuracy from the sensor.
4- 20 arc min. The resolver-to-digital converter (RDC)
Many encoder-based drive systems have simply selected adds an uncertainty of 2 to 4- 8 arc min. Resolver errors
the encoder resolution/pulse number based upon the also have both static and dynamic contributions, which
maximum speed of the application and the frequency result from the acceleration error in the RDC tracking
response of the encoder input of the drive. For example, loop, offset voltages that are uncompensated, phase
many drives use a 2000 pulse per revolution encoder shift between the signals and the reference voltage, and
with a 200 kHz maximum frequency input capability of capacitive or inductive crosstalk between the resolver
the drive. This allows the motor to run at 6000 min-1 signals and the reference cabling. Noise in the inter-
without missing counts. connect or on the reference will generate speed-depen-
In a digital system, where encoder pulses are counted dent errors proportional to the phase shift in reference
and inversely proportional to the reference frequency.
over a defined time period, quantisation effects result in
what can best be described as a ripple on the speed Modular incremental rotary encoders usually have an
accuracy rating of two arc minutes or better over all
feedback signal. Whilst this occurs at all speeds, the
operating conditions.
effect (percentage ripple) is higher at low speeds. Whilst
this is rarely a practical problem, in some applications it f High-temperature operation - resolvers are capable of
can be important. In such applications a sine-cosine operating at high temperatures, and in many cases can be
encoder should be used. operated at up to 150C.
d Motor efficiency- in permanent magnet servodrives Standard encoders are typically specified for use up to
where the switching of current to the machine windings 100C... 125C.
is determined by information from the position sensor,
Although the encoder is usually mounted within the
motor efficiency is a function of the accuracy of the
motor housing, many measures are normally taken to
position information.
thermally isolate it from the motor core. As a result, it is
Single cycle sine-wave-type commutation signals gen- usually not necessary to require a feedback device to be
erally have accuracies to about 5 mechanical. operated above 100C.

6 MECHANICAL CONSIDERATIONS

Speed and position feedback devices are obviously subject Moment of inertia of the rotor and the coupling of the
to the acceleration and speed associated with the motor or feedback device form a single vibrating mass system
load that they are measuring. with a natural frequency, which needs to be outside the
Encoders: operation frequency range of the drive. A large number of
angular acceleration 104 rad s -2 different coupling types exist, but in general a diaphragm
type is usually preferred to give the highest natural
max acceleration of 1000 m s -2 for shock and impact valid
frequency.
for 11 ms
vibration resistance 100 m s -2 from 50 to 2000 Hz The natural frequency is given by the equation:
Maximum speed is generally related to bearing life and
forms one of the selection criteria for specific applications. f = (1/27c). N(C/I)
Chapter 3.7 101

where f is the natural frequency in Hz, C is the torsional adequate. Because both encoders and resolvers can be
rigidity of the coupling (Nmrad -1) and I is the moment of purchased with or without bearings, neither can claim an
inertia of the sensor motor (kg m2). advantage in this respect; however, frameless resolvers
may have slightly less sensitivity to axial play than a mod-
It is interesting to note that although many suppliers quote
ular encoder would. With respect to electronics, it is true
current loop torque bandwidths of servo amplifiers of 2 kHz
that the resolver electronics can be mounted remotely from
and above, the natural frequency of the very best coupling is
the motor, in a less extreme environment. However, they are
approximately 2 kHz. It is not possible to take advantage of
much more complex than those of the encoder. Typical
better converter performance.
encoder designs use a small number of very basic compo-
Resolvers: nents. It is possible that for extremely high-impact shock
No one can discount the sturdiness of the resolver. It is a environments a resolver could claim some advantage over an
simple device with a similar make up to the motor, con- encoder. However, if resolutions of less than 1000 counts per
sisting of windings, bearings etc. However, for the majority revolution are desired then an encoder with a metal or mylar
of the environments encountered, an encoder is completely code wheel can compete favourably with resolver designs.

7 GLOSSARY OF TERMS

Absolute measuring system - the measuring value is Encoder - apparatus consisting of a measuring standard and
determined by reading information from a scale, without a scanning unit (transducer, sensor).
counting. The measuring value is immediately available
after switch on. Gray code - unit-distance code system in which only one
code signal changes with the transition from one measuring
Accuracy grade - grade of quality, determined by the step to the next.
maximum permissible measuring deviations within a pre-
determined measuring range (e.g. 1 m). Incremental measuring system - measuring method by
which the measuring value is derived by the summation
Amplitude evaluation - method of evaluating signals gen- (counting) of increments (measuring steps).
erated by dynamic scanning (with cartier frequency): the
amplitude variation of two alternating voltages of the same Integral coupling - innovative angle encoder design with
frequency is used to determine the measuring value. built-in coupling, located preferably on the stator side.

Angle encoder - angle-measuring device, converts the shaft Interferential measuring system - photoelectric measuring
rotation angle into electrical signals (can be incremental or system with a phase grating scale where scanning signals are
absolute). produced via the interference of diffracted beams.

Carrier frequency method - scanning method used mainly Interpolation - method for producing measuring incre-
with magnetic and inductive measuring systems (see ments which are smaller than a fourth of the scanning signal
dynamic scanning). period.

Direction discriminator - part of a bidirectional counter, Measuring system - consists of an encoder and associated
which determines the counting direction. electronics incorporating interpolation, counter, readout
and/or data interface.
Distance-coded reference marks - incremental measuring
method, whereby the absolute position can be determined by Modular angle encoder - angle encoder, consisting essen-
evaluating the systematically varying distances between tially of disk and scanning unit assemblies (rotor and stator)
consecutive reference marks. which are integrated into a machine or a rotary table.

Dynamic scanning - scanning method by which two alter- Moir~ principle - photoelectric scanning method to produce
nating signals of constant amplitude and slightly different periodic signals using two fine gratings, which are closely
frequencies are generated and where the phase between the positioned to each other and have approximately parallel and
two signals represents the measuring value. equally spaced lines.

Eccentricity e r r o r - measuring error of an angle encoder Multiturn rotary encoder - absolute rotary encoder which
caused by an eccentricity in the mounting of the circular determines the angular position of the shaft and the number
graduation. of shaft rotations.
102 GLOSSARY
OF TERMS

Phase evaluation - method of determining position by Resolver - inductive angle-measuring device, producing
detecting the phase between alternating voltages having a two alternating voltages the amplitudes or phases of which
slight variation in frequency. depend on the (shaft) rotation angle.
Phase grating scale - scale with step grating which diffracts Reversal error - measuring error which results from
the transmitted or reflected light into two or more orders. approaching a position from different directions.
Radian - standard unit of angle: the angle at which the arc of
Scanning frequency - response level which limits the
circle has the same length as the radius.
velocity of an incremental measuring system.
Reference m a r k - random graduation pattern which, when
traversed over, produces a signal peak, which may be used to Static scanning - scanning method, which generates peri-
determine an absolute datum within an incremental mea- odic signals during movement. The signal periods and
suring system. fractions thereof correspond to a definite linear or angular
displacement.
Reference pulse - square-wave signal produced when the
scale reference mark is traversed over; normally one mea- Systematic error - reproducible measuring deviation, which
suring step wide; may be used to define an absolute datum
can be compensated for by e.g. computation.
within an incremental measuring system.
Resolution- measuring step, smallest digital unit of the Torsional stiffness - rotational rigidity of a precision cou-
measuring value. pling governing the reversal error of a rotary encoder.
CHAPTER 4

Drive Control

~ 1 GENERAL 103
iiiiiiiiiiiiii~iiiiiii

~'!iiiiiiiiiiiiiiiiiii!ii!i
~'~"~ 2 A.C. MOTOR DRIVE CONTROL 105

iiii 3 D.C. MOTOR DRIVE CONTROL 114

:,ii!iii
iii!iiiiiiiiiii!iiiiiiii!
4 ANALYSIS OF AND SET UP OF A SPEED CONTROLLER 116

1 GENERAL

THE IDEAL CONTROL SYSTEM OPEN-LOOP CONTROL

Many applications exist where something has to be con- Unfortunately, the transfer function of many systems is not
trolled to follow a reference quantity. For example, the speed a constant and so, without any form of feedback from the
of a large motor may be set from a low-power control signal.
This can be done using a variable-speed drive as shown in
mains power supply
Figure 4.1.

Ideally, the relationship between the reference and the


motor speed should be linear and the control system should
respond instantly to changes in the reference. Any control
speed reference
W*
] v=r,=b,e-I motor M
_1. I

speed
system can be represented, as shown in Figure 4.1b, with
W
an input reference signal, in this case a speed reference, a
b
transfer function G and an output, in this case the speed of speed reference ,.] speed
rotation of the motor shaft W. For the system to be W* I v W
ideal, the transfer function G would be a simple constant,
so that the output is proportional to the reference with no
delay. Figure 4.1 Variable-speed drive and motor
104 GENERAL: O p e n - L o o p C o n t r o l

output to correct for the nonideal nature of the transfer CLOSED-LOOP CONTROL
function, the output does not remain proportional to the
control signal. Using an induction motor supplied by a The simple open-loop drive of the previous section is
simple open-loop variable-speed drive as an example, the replaced with a control system as shown in Figure 4.2. The
following are some unwanted effects that can occur in speed of the motor shaft is measured and compared with the
control systems. speed reference to give a speed error. The error is modified
by a transfer function F to give a current reference i* at the
Regulation input to the current control block. Various methods of cur-
The output of a simple open-loop drive can be a fixed fre- rent control for motors are discussed in the next section;
quency, which is proportional to the speed demand signal. however, for now it should be assumed that the motor cur-
Therefore, the frequency applied to the motor remains rent can be controlled to give a torque which is proportional
constant, for a constant speed demand. However, the speed to the current reference. If the speed of the motor varies from
of the motor drops as load is applied because of the char- the reference level a speed error is produced and the torque
acteristic slip of an induction motor, and so the speed does applied to the load is modified to bring the speed back to the
not remain at the demanded level. reference level.
It is necessary to choose a suitable transfer function F to
Instability obtain the required performance from the closed-loop con-
It is possible under certain load conditions and at certain trol system. The function could be a simple gain, therefore
frequencies for the motor speed to oscillate around the the current reference i*= Kp x Werr. This would give some
required speed even though the applied frequency is con- degree of control over the output speed, but the speed error
stant. Another major source of instability in rotating mech- must have a nonzero value if any torque is required to hold
anical systems is low loss elastic couplings and shafts. the motor speed. If the speed error is not zero, then the speed
would not be at the required reference level, and so the speed
Nonlinearity would vary with load. By adding an integral term so that the
current reference i*= Kp x Werr+ Kifwerrdt, it is no longer
There are many possible sources of nonlinearity; if, for
necessary to have any speed error even when torque is
example, the motor is connected to a gearbox the speed at
required to drive the load at the reference speed. The integral
the output of the gearbox could be affected by backlash
term accumulates any speed error over time and builds up a
between the gears.
current reference to provide the necessary torque. A closed-
loop control system with proportional and integral terms is
Variations with temperature
called a PI controller. Although there are many types of
Some aspects of the system transfer function may vary with closed-loop controller, the PI controller is the most com-
temperature, for example the slip of an induction motor monly used because it is simple to implement, relatively
increases as the motor heats up, and so for a given load the easy to set up and well understood by most engineers.
motor speed may reduce from the starting speed when the
motor was cold.
CRITERIA FOR ASSESSING THE
Delay
PERFORMANCE OF A CLOSED-LOOP
With a simple open-loop inverter and induction motor there CONTROL SYSTEM
can be a delay before the motor speed reaches the demanded
level after a change in the demand. In very simple applica- The step response is one method of assessing the ability of
tions, e.g. controlling the speed of a conveyor belt, this type a closed-loop control system to follow a step change in
of delay may not be a problem. In more complex systems, the reference. Some example step responses are shown in
such as machine tools, delays have a significant effect on the Figure 4.3 for a simple second-order system.
quality of the system.
If the output reaches the reference in the shortest possible
These are some of the unwanted effects that can be produced time without any overshoot the response is described as
if an open-loop control system is used. One method that can being critically damped. If overshoot is not acceptable,
be used to improve the quality of the controller is to use a then this represents the best possible response giving the
measure of the output quantity to apply some feedback to minimum delay between the input and the output of the
give closed-loop control. system. If the system damping is increased the response

speed error current reference current torque


Werr i* i Te

speed reference
W~
*~ F current motor load
control

Figure 4.2 Closed-loop control system


Chapter 4.2 105

5
100
IIII
,, 1000
IIII
lOO(
~ ,~nder damped ,8 III "bUl IIII
-5
III \ II
dB-10 N

-15 III \ II
-20 III II I
~: ed -25 III "l
-30 III rad s-1
I
0- - "t3"~ [O]) lO00q
10 IIII
-10 1oo "~
I \
I I I I
-20
-30
\
\
Figure 4.3 Step responses deg-40 \
-50 \

-60
becomes slower and is described as being over damped.
Alternatively, if the system is under damped the response
-70 \
-80
includes some overshoot and may oscillate about the
-90
required reference before settling. These results, which are rad s-1
for a simple second-order system, show that increasing the
Figure 4.4 Bode plot of system gain and phase
damping reduces overshoot and slows down the system
response. As will be demonstrated in Chapter 4.4, real sys-
tems can be more complex and increasing the damping does of the system, which is often defined as the - 3 dB point
not always give this result. for the gain characteristic. In the example this occurs at
670 rad s -1. The corresponding phase delay varies depend-
The step response may be the closed-loop response of the ing on the order of the system. A first-order system has a
system, where a change to the reference level in the mini- delay of 45 at the - 3 dB point, whereas a second-order
mum time is required. Alternatively, the step response may system like the one in the example has a delay of 60 at the
show the change of output to some other stimulus, such as a - 3 dB point. The transport delays associated with digital
load torque transient. In this case the response should be as systems can further increase the delay at the - 3 dB point.
small as possible.
In many cases the bandwidth is quoted as an indication of the
The step response can be used to assess the controller per- dynamic performance of a control system, i.e. the higher the
formance when it is used in isolation. However, if the con- bandwidth the better the system performance. This band-
troller itself is to be included within the closed control loop width is usually quoted as the frequency at the - 3 dB point
of another system, the gain and delay of the original system of the gain characteristic, and particularly with a digital
are important as they affect the performance of the outer implementation may give no indication of the quality of
system. The gain and delay can be measured by producing a the control system at all. If the controller is to be included
bode plot showing the gain and frequency response as shown within another closed-loop control system the phase delay
in Figure 4.4. is important. If the delay is too large it may be necessary
to detune the outer loop to maintain stability. The amount
An ideal controller would have unity gain and zero phase at of overshoot in the step response is also important in
all frequencies; however, in a real system as shown in Figure many applications. Increasing the frequency of the - 3 dB
4.4 the gain reduces and the phase delay increases at higher point of the gain characteristic may result in unacceptable
frequencies. A measure of these effects is the bandwidth overshoot.

2 A.C. MOTOR DRIVE CONTROL

There are many types of variable-speed drive each given of typical control systems for a range of differ-
suited to different applications or for operation with ent types of variable-speed drive operating with A.C.
different types of motor. In this section descriptions are motors.
106 A.C. MOTORDRIVECONTROL:General-Purpose Open-Loop A.C. Drive

GENERAL-PURPOSE OPEN-LOOP vectors (U 1 to U6) , and the remaining two states (Uo and U7)
A.C. DRIVE give zero voltage vectors where all three output phases are at
the same voltage level. By switching rapidly between the
The open-loop A.C. drive is a high-power variable-fre- various states the average output from the inverter can be a
quency voltage source. In its simplest form, the output fre- voltage vector at any angle with the required magnitude. In
quency is defined by the user's reference and a suitable fixed the example given in Figure 4.6 a voltage vector at angle c~
frequency to voltage characteristic used to define the output could be synthesised by producing u~ and U2 each for 50 per
voltage. Although this type of drive is normally designed to cent of the time. Changing the ratio of the time during which
supply one or more induction motors connected in parallel, it u~ and I!2 are active allows the vector to be moved from u 1 to
can also supply other types of A.C. motor or it can be used as u2 with a locus which forms one side of a hexagon con-
a variable-frequency/variable-voltage power supply. The necting the tips oful and u2. The magnitude of the vector can
following description relates to the operation of an open- be controlled to be anywhere inside the hexagon connecting
loop A.C. drive with an induction motor. the tips of the active vectors by introducing periods of zero
voltage with vectors Uo and u7. The active vector periods and
Figure 4.5 shows a block diagram of a typical general-pur- inverter output waveforms for the required vector Us are
pose open-loop drive. Each of the blocks within the control shown in Figure 4.6c.
system is described in the following sections. For this type of
control system feedback is required from the output current A space-vector modulator produces sinusoidal output vol-
of at least two phases of the inverter and the voltage from the tages provided that the vector locus is a circle. The limit for
D.C. side of the inverter. These feedback signals can be sinusoidal operation is a circular locus that just fits inside the
derived from within the drive itself, and so no external hexagon in Figure 4.6b. Operation outside this circular locus
feedback, such as the motor speed or position, is required. is over modulation and, although the fundamental output
The control system allows a simple open-loop vector strat- voltage can be increased, odd harmonics of the fundamental
egy to be implemented. In common with all other A.C. drive output voltage are also produced. At full over modulation,
control strategies described in this chapter the control sys- i.e. when the locus of the voltage vector is the hexagon
tem is based on a reference frame as described later. shown in Figure 4.6b, the average output voltages are tra-
pezoidal waveforms. It is possible to further increase the
Space-vector Modulator and Inverter output voltage by making the angular velocity of the voltage
vector vary as it moves round until the output voltages are
The power circuit is a diode rectifier and voltage source square waves. The fundamental output voltage for these
PWM inverter. The space-vector modulator converts a ranges is given in Table 4.1 as a ratio between the drive
modulation depth, m, and modulation angle, Om, into control supply voltage and the output voltage for an ideal drive,
signals to turn the six devices in the inverter on or off. The which has a simple diode rectifier at its input.
modulation depth defines the output voltage from the
The waveforms in Figure 4.6c show that an upper or lower
inverter as a proportion of the range available without over
power device in each output phase of the inverter is always
modulation, and the modulation angle defines the angle of
on. This may be an IGBT or a diode depending on the
the space-vector representation of the three-phase voltages
direction of the output current from each phase. When the
at the inverter terminals. The modulator is called a space-
control system switches off an IGBT it must allow a short
vector modulator because it applies voltages to the three
time, the safety margin, before turning on the other device in
phase windings of a motor to give a voltage vector in space
the same phase to prevent short pulses of current flowing
with respect to the body of the motor. This voltage vector is
from the positive to the negative rails of the inverter through
conceptual rather than real as voltage does not exist as a
both devices in that phase. During this safety margin, the
space quantity, such as, for example, flux.
dead time, the output voltage of the inverter is undefined as
There are eight possible states for the devices in the inverter, it is dependent on the direction of the output current.
excluding those that result in shoot through, as shown in Although the dead time is short it distorts the inverter output
Figure 4.6. These states give six possible active voltage voltage and has the most significant effect at low modulation

I Ivo M
f*
I fcornp*
voltage
I
I
\
~
I
~
soace
' ' vector
modulator
YI i characteristic Vs I R \
~
~ and
inverter ]
] Y ,._1
vl
motor

I [~ f IT T II . "r Oref(reference
~rframeangle)

s,i01
compensation['-
' I xy "~J.~ 'sQI D Q ' ~ , Isv

Figure 4.5 Open-loop drive control system


Chapter 4.2 107

depths. It is possible to reduce this effect by using dead-time control system used with this type of drive only ensures
compensation to modify the widths of the IGBT control that the reference frame aligns with the flux under steady-
signals depending on the direction of the inverter output state and not during transient conditions. By using this
currents. technique the voltages and currents appear as D.C. quantities
under steady-state operation, even though the motor currents
The modulation depth does not define the absolute output
and voltages are A.C. quantities. The x and y axis currents
voltage of the inverter because the D.C. voltage at the input
relate to the production of flux and torque in the motor,
to the inverter can vary with drive supply voltage or if the
respectively. A reference frame is simply a set of axes as
drive operates in braking mode. Therefore the required
shown in Figure 4.7.
voltage magnitude [v*[ must be converted to modulation
using the following equation: The method used to calculate the position of the reference
frame and to control the motor voltage is covered in the next
m- v/3 x v*l/V~.c. (4.1) section. First, in this section, the translation of the current
feedback into the required reference frame and the transla-
where VD.C. is the D.C. link voltage and Iv*l is scaled to a
tion of the voltage references from the reference frame into
value equivalent to the peak phase voltage.
an angle and voltage for the modulator are described.
Reference-frame Translation The current feedback is taken from two of the three inverter
output phases (is, and isv). These can be converted into two-
It is helpful to apply voltages and measure currents in a phase quadrature currents (isD and isQ) and then into current
reference frame aligned with some quantity in the motor. In components in the x and y axis of the required reference
an open-loop drive the reference frame is usually aligned frame as follows"
with the stator flux. As will be explained later, the type of
isD- isu (4.2)
isQ -- (i,, + 2isv)/v/3 (4.3)
t'~ -- i~D COSOref + isQ sin O,.~f (4.4)

a ..... ~_J++J~V+?+ imotl-lor isy -- isQ cos Oref - isO sin Oref (4.5)

where 0ref is the angle of the required reference frame.


inputrectifier D.C.linkand inverter The voltage reference components (vsx and Vsy) can be
brakingchopper converted to a magnitude and angle by simple rectangular to
polar conversion.
u3(U LVUW L) u2(UuVuWL)

V s l - v/(v~2 + v~2) (4.6)


Ov -- tan -1 (V~y/V~x) (4.7)

zerovi~a~!W~)ct ul (UuVLWL)
Reference-frame Generation
uo(U LV LWLi
It is a convention with open-loop drives to use the syn-
uT(UuVuWu) u5(U Lv LW U) u6(UuVLW U)
chronous frequency related to the required speed of the
I I I motor as the reference, i.e. 50 Hz for 1500 min -1 using a
i i i
i i i
, , ,
four-pole motor. The reference shown in Figure 4.5, f*, is
i i | compensated with a frequency from the slip compensation

| | !

U0 U1 U2 U7 U7 U2 Ul U0

Figure 4.6 Space-vector modulation


a power circuit Vsy ~ R s i s y
b space vectors
c inverter output voltages

Table 4.1 Fundamental output voltages


Range Maximum output/input
isy is

voltage ratio
Space-vector modulation 1.00
without over modulation
Space-vector modulation 1.05
with over modulation stator flux (in steady state)
Square-wave operation 1.10
Figure 4.7 Stator flux reference frame
108 A.C. MOTOR DRIVE CONTROL: General-Purpose Open-Loop A.C. Drive

block. When the reference frame is aligned with the stator the drive when fast acceleration is required or a large tran-
flux the y axis current, i,y, is a measure of motor torque, and sient load is applied to the motor. Therefore, a current-limit
so the slip compensation block can estimate the change of system is required to decrease the frequency reference when
motor speed with load applied due to motor slip (see Chapter a large decelerating load is applied, or increase the fre-
1.2). By adding the compensation value to the frequency quency when a large accelerating load is applied. Figure 4.8
reference, the reference is increased as load is applied to the shows a current-limit system commonly used with open-
motor, f~omp is integrated to give the reference frame angle loop drives.
Orefi and so the reference frame rotates at frequency fcomp. If The frequency reference, f*, is derived from the user fre-
fixed values of V~x and V~y were used the drive would apply
quency reference via a ramp block which limits the rate
voltages to the motor at the frequency f~omp, which should
of change of frequency and hence the rate of change of
hold the motor speed constant even when a load is applied.
motor speed. If the torque-producing current in the motor
The frequency adjustment applied by the slip compensation
exceeds the current limit then the PI controller is enabled
block may not be very accurate as the motor slip varies with
to modify the frequency reference. If the motor is produc-
temperature, but this method gives moderate control of
ing too much accelerating torque because a large load has
speed without any form of motor position or speed feedback.
been applied the frequency reference is reduced towards
The motor stator flux is defined as: zero and unless the load reduces with speed the motor
stalls. If the motor is producing too much accelerating
torque because the ramp rate is too fast the PI controller
~s -- / Vs - Rsisdt (4.8)
reduces the rate of change of frequency reference so that
the current is limited because the acceleration ramp is
where qDs is the stator flux, vs is the voltage vector repre- extended. If the decelerating torque is too large because a
senting the motor terminal voltages and is is a vector large accelerating load has been applied or the decrease in
representing the motor currents. In the steady state the vol- frequency reference is too fast, the PI controller works in the
tage vector applied to produce the stator flux leads the flux opposite way and either accelerates the motor or reduces the
by 90 and is given by: deceleration rate.

V~s - - Vs - Rsis (4.9)


Performance and Applications
In Figure 4.7, V~x - Rsisx and V~y = v~s + Rsisy. If V;x and V~y The performance characteristics of the open-loop drive
are derived in this way then the x axis of the reference frame control system applied to an induction motor can be sum-
is aligned with the stator flux, and the y-axis current can be marised as:
used as a measure of load for slip compensation etc. For
constant stator flux, Iv l must be proportional to frequency Moderate transient performance.
up to the rated frequency of the motor. Above rated fre- Full torque production down to approximately 3 per cent
quency, Iv ,l is held constant so that the flux falls with of rated motor speed.
increasing frequency to give flux-weakening operation. Although a good estimate of stator resistance (R,)
improves torque production at low speeds, the control
This method of control gives good and robust motor control system will work with an inaccurate estimate, albeit with
down to approximately 3 per cent of rated motor frequency. reduced torque. The stator resistance can be measured by
The motor speed can be controlled below this frequency, but
the drive with a simple test.
usually at reduced torque. This method gives significantly Although a good estimate of motor slip improves the
better performance than the fixed-voltage control methods ability of the drive to hold the reference speed, the
described in Chapter 1.2. control system will work with an inaccurate estimate,
albeit with poorer speed holding. The motor slip depends
Current Limit on the rotor time constant of the motor (Tr) and this
cannot be measured easily.
The motor current cannot be allowed to increase with load No position or speed feedback is required from the
without any limit or the inverter protection system will trip motor shaft.

user
frequency ramps
reference

maximum
allowed value
PI
of isy I

I
current
above limit

Figure 4.8 Current-limit system


Chapter 4.2 109

This type of drive is used in many applications where the torque is proportional to i~y. The rotor flux position is
moderate performance is required and where providing easy to calculate from the absolute rotor position as shown
position feedback would be unacceptable because of the below.
environment or cost, or is simply not necessary. The fol-
lowing are some examples of applications where open-loop It is clear that the absolute position within each electrical
drives are used: revolution is required, and so the position feedback device
must give some absolute position information. Traditionally,
fans and pumps resolvers have been used as they can give the absolute
conveyors position of the rotor (see Chapter 3.3), but these are being
centrifuges replaced by incremental encoders which give more precise
feedback with lower cost interface electronics in the drive.
However, an incremental encoder does not give the absolute
PERMANENT-MAGNET SERVODRIVE position of the rotor, and so additional Gray code signals are
required to find the absolute position at power up. For a
The permanent-magnet servodrive is generally used for three-phase six-pole motor a three bit Gray code is repeated
applications requiring high performance where motor shaft three times per mechanical revolution giving the position to
position feedback can be used. Because the rotor is not within 60 during each electrical revolution. Provided that
symmetrical this feedback must give absolute position the angle between the Gray code signals and the magnets is
within each electrical revolution of the motor. Figure 4.9 known the reference frame can be placed within +30 of the
shows the control system of a permanent magnet servodrive. correct position. Once the motor has rotated past a change
The inverter control and reference frame transformation from one Gray code value to another, the position is known
is the same as for the open-loop drive. The other blocks exactly and the incremental signals can be used to track the
within the control system are described below. absolute position until the drive is powered down again. The
incremental encoder with additional Gray code signals is a
Reference-frame Generation relatively cost-effective form of feedback. Other types of
encoder that can provide the absolute position of the rotor
In common with the open-loop drive the reference frame are described in Chapter 3.4.
must be aligned with the motor flux so that the current
in the x and y axes controls the flux and torque, respect-
ively. In the open-loop drive the stator flux is used because Current Control
it is easy to calculate its position without rotor posi-
tion feedback. However, the effects of the current in each Unlike the open-loop drive the permanent-magnet servo-
axis are not completely decoupled because the x-axis drive uses closed-loop current control giving the drive the
current changes with motor torque even if the stator flux ability to change the current and hence the torque produced
is constant. It is preferable to align the reference frame by the motor very rapidly. A simple PI controller is provided
with the rotor flux as this gives complete decoupling so that for each axis of the reference frame. The reference for the x-
the x-axis current controls the rotor flux and the y-axis axis (flux-producing current) is zero because the flux is
current controls the torque produced by the motor. In the provided by the magnets on the rotor. The y-axis reference
control system shown the x-axis current reference is set to defines the torque-producing current and hence the torque
zero, and so all the rotor flux in the x axis is produced by the produced by the motor. Although the currents and voltages
magnets. The torque produced by the motor is given by are D.C. quantities in the steady state, high dynamic per-
Te = K1 ~rxisy, where K1 is a constant, ~rx is the component of formance is still required so that the phase delay of the
rotor flux in the x axis and isy is the component of current in current controllers is as small as possible and the effect on
the y axis. As ~rx is produced by the magnets and is constant, the outer speed controller is minimised.

i*sx=O ;[soace F
vector (3 ~] mo 3r
k/
[modulator
and
Te=i*sy inverter
"1
Oref(reference
~ frame angle) absolute
position
feedback

oo ,sv

Figure 4.9 Permanent-magnet servomotor drive control system


110 A.C. MOTOR DRIVE CONTROL: P e r m a n e n t - M a g n e t Servodrive

The control system described has a zero current refer-


ence in the x axis, and so the rotor flux cannot be
altered by the drive. This limits the maximum speed of
rotor flux operation as the motor voltage increases with speed.
Field weakening is possible but, owing to the large
effective air gap of a permanent-magnet motor, a large
amount of current is required to reduce the flux, making
the drive inefficient in the field-weakening region. It is
also necessary to limit the maximum speed with field
weakening because, although the motor terminal vol-
tages are reduced by reducing the flux while the drive
is enabled, when the drive is disabled the voltages
return to the level that would be produced without field
Figure 4.10 Rotor flux reference frame for a permanent- weakening and may damage the power electronics in
magnet motor the drive.
The drive can be used in simple torque control, without
The description of the current controllers has been kept the speed controller shown in Figure 4.9.
simple and does not include anything about compensation
for cross coupling or back e.m.f. The following are some examples of applications where
permanent-magnet servodrives are used:
Speed Control * machine tools where precise and dynamic performance
is required
The outer speed controller is shown as a PI controller using
pick and place applications where the requirements are
the differential of the position as speed feedback. A more
less precise, but where rapid movements are required
detailed description of this type of controller is given in
more recently, permanent-magnet motors with high
Chapter 4.3.
numbers of poles (e.g. 32 poles) have been used in
low-speed applications such as direct (gearless) drives
Performance and Applications for lifts.
The performance characteristics of the permanent-magnet
servodrive can be summarised as:
CLOSED-LOOP INDUCTION MOTOR DRIVE
Good dynamic performance at all speeds.
Full torque operation down to standstill. The closed-loop induction motor drive, often referred to as a
Permanent-magnet servomotors usually have low iner- closed-loop vector drive, is used in many applications
tia. Combined with a fast sample rate for the speed requiring better performance than that of an open-loop drive
controller and fast torque control, this gives a speed with an induction motor. To obtain the best performance
controller with a very high bandwidth. with this type of drive position feedback is required from the
Position feedback is required that gives the absolute rotor, but unlike the permanent-magnet servodrive only the
electrical position of the motor. change of position and not the absolute position is required.
Permanent-magnet motors exhibit an effect called cog- The control system is similar to that used with a permanent-
ging related to the geometry of the motor, which results magnet servomotor as shown in Figure 4.11. Descriptions
in ripple in the motor torque. This effect can be mini- are only given for blocks that have not already been covered
raised by good motor design, but can still be a problem. in previous sections.

vo.c.

flux
controller space
vector
modulator
and
W* inverter

Omr(reference
frame angle) incremental
3osition
feedback
i~x ,~1 flux
i* "'1 calculator
sy ~1

I
Figure 4.11 Closed-loop induction motor drive control system
Chapter 4.2 111

Flux Calculator and Reference-frame Wide power range of motors available so that this type of
Generation drive can be used for applications requiring less than
1 kW up to more than 1 MW.
In common with the permanent-magnet servodrive the Suitable for field-weakening applications where motors
reference frame is aligned with the rotor flux. Unlike the can be operated up to many times base speed.
rotor flux in a permanent-magnet motor, which remains at a The drive can be used in torque control, without the
fixed position with respect to the rotor and at a constant level speed controller shown in Figure 4.11. The estimate of
defined by the magnets when operating without field the flux position to align the reference frame is important
weakening, the rotor flux in an induction motor moves at slip as this affects the absolute level of torque produced for
frequency with respect to the rotor and the flux is provided a given torque reference. The flux position calculation
by the x-axis current. Therefore, a flux calculator is required is dependent on an estimate of the rotor time constant,
which derives the magnitude and angle of a vector to which varies significantly with rotor temperature.
represent the rotor flux. By convention a current vector is However, it is possible to include a rotor time constant
defined called the rotor-magnetising-current space vector, estimator in the drive control system, so that the drive
imr, which represents the rotor flux. This vector is aligned gives consistent torque control.
with the flux and has a magnitude proportional to the flux,
Closed-loop induction motor drives are used in many
but in units of current. The magnitude and angle of this
applications where good dynamic performance is required
vector are given by the following equations where Or is the and especially where an induction motor drive is required to
rotor position and Tr is the rotor time constant of the motor: give full torque at standstill. The following are some
]imr[-- i~x/(1 + sty) (4.10) examples of applications where this type of drive is used:
cranes and hoists
lifts
Omr -- Or q- f i~y/(Trlimrl)dt (4.11) high-speed spindle applications
material winding
As with the permanent-magnet motor the torque is propor-
tional to the x-axis component of flux and the y-axis com-
ponent of current, and so Te--Kz[imr[isy , where K2 is a Operation without Position Feedback
constant. To achieve a linear relationship between the
The control system described above requires incremental
demanded torque at the output of the speed controller, Te,
rotor position feedback, but it is possible to implement the
and the torque produced by the motor the torque demand
scheme without any physical feedback device. This can be
must be modified using ]]mr[ to give the y-axis current
done by estimating the rotor position from information
demand, isy, as shown in Figure 4.11. The angle Omrgives the
available to the drive through the motor voltages and cur-
angle of the rotor flux and can be used as the reference-frame
rents. One class of methods used to determine the rotor
angle for translation of currents and voltages. As the rotor of
position, referred to as model-based methods, uses a model
an induction motor is simply a symmetrical conductive cage,
of the motor to calculate the rotor speed and hence the
the absolute position is not important, and so only the
incremental rotor position. When a physical position feed-
incremental position is required to define Or. Therefore,
back device is used the drive gives good dynamic perfor-
a simple incremental encoder can be used for position
mance and operates with full torque even at standstill. When
feedback.
a position estimator is used the dynamic performance is
reduced and the minimum speed for full torque operation is
Flux Control similar to that of the open-loop drive. However, a closed-
As the flux is produced by the x-axis component of the stator loop induction moto