Discussion on

Linear induction motors for low-speed and standstill application

J. S. Holmes: My company has been taking an active is a feedback-controlled motor and gearbox, as used in a lift.
interest in the linear induction motor for the past six months, Would the authors comment on whether the linear motor is
and we have now designed and tested development models of a possible or economic alternative, assuming a weight to be
both theflatand tubular types. This paper formed the working lifted of 100kg?
basis for our designs. In general application, could the authors comment on
One point of interest that has emerged as a result of trials whether they consider a linear motor in standstill condition
of our equipment on customers' installations is that the choice to be economically and functionally better than where a
of either a flat motor or a tubular motor is dependent on the solenoid and armature is used at present ?
mechanics of the installation. For example, in applications G. M. Ettinger: Mention has been made of the possi-
such as motorising sliding doors or motor traversing on bility of a train-mounted tubular linear induction motor
overhead hoists, the moving part of the linear motor is whose core would run along the track. May I refer to British
supported and guided by slides or bearings in the equipment Patent Specification 514/1862 by H. W. Cook, which shows
itself; hence the flat motor is used. a d.c. solenoid on a carriage. A flexible core of alternate
In applications where the linear motor is used as a thruster magnetic and nonmagnetic sections passes through the
unit, such as for linkage actuation and packaging, the motor solenoid. Cook suggested that a circuit breaker in the carriage
unit is required to be self contained with its own frame and would interrupt the solenoid current in synchronism with the
bearings. In these cases the tubular motor is employed, as it is passage of magnetic core sections, to achieve unidirectional
far easier mechanically to engineer a bearing system for a tractive effort.
circular rod than for a flat plate. P. G. Kendall: Have single-phase motors or actuators
A second point of interest is the practical implementation been made, and, if so, how does their performance compare
of the goodness factor referred to in the paper. An example with the 3-phase devices described in the paper?
we encountered concerned a flat motor where the air-gap G. F. Nix and E. R. Laithwaite {in reply): We agree
length was dictated by mechanical clearance permitted on the with Mr. Holmes's findings on the question of the type of
particular installation; rotor resistivity was selected to give motor to be used in particular applications. It is true that
the required output performance, and the supply frequency Table 1 of the paper only takes account of stator end-turn
was fixed. Using the goodness-factor equation, the pole pitch resistance to the extent that it is contained in the quantity ps.
was calculated to give the maximum force per watt of input. With small pole numbers, the effective increase in ps due to
This suggested two poles on the particular length of stator the 'inactive' parts of the stator winding is probably caused
block used. However, when a 4-pole winding was also tried, no.t so much by the normal end turns as by the fact that a
it proved to be superior in the terms of force output per watt high proportion of the slot currents must be returned outside
of input. This was due to the decrease in end-turn length with the ends of the block, when a conventional surface winding
the 4-pole winding, resulting in a decrease in stator I2R loss. technique is used. In effect, most of the coil sides in the active
Experimentation was also carried out on flat linear motors part of the machine have opposite sides which do no more
using armatures of plain steel, steel clad with aluminium, and than add to the stator I2R loss. This can be avoided by the
steel clad with copper. As an indication of the variation use of Gramm ring winding, and, of course, the tubular
obtained in performance, using a linear stator block designed motor is immune by its nature. We are grateful to Mr. Holmes
to operate with a steel armature plate clad with 0-040 in of and to his company for the development work which they
copper, a thrust of 251bf was obtained; 181bf thrust was have carried out on these machines.
obtained with the armature clad with 0-040in aluminium, and In reply to Mr. Trinder, there are several methods by which
a thrust of 81bf was obtained with the steel alone. The gap the speed of a linear motor may be controlled, and these
separation between the moving armature and the stator teeth correspond to the methods used in rotating machines. If we
was kept constant. I use the term 'armature' for the item restrict the discussion to machines with solid secondary
termed 'rotor' in the paper, as in the linear motor this part members, the choice lies between frequency control, where
does not rotate. the speed of the field is controllable, and voltage control,
Similar experimentation was carried out on the tubular where the secondary is allowed to slip when working against
motor, but in this case a copperclad steel armature rod was its load. A particularly useful form of the latter method
used and different copper thicknesses were tried, whilst consists of winding the stator for 2-phase operation and using
maintaining a constant mechanical separation between the voltage control on one phase only. This technique is standard
stator and armature rod. practice in closed-loop control systems, and design procedures
Plotting the output force at a constant voltage and fre- based on the use of the goodness factor have been published
quency against thickness of copper showed a distinct turning in an earlier paper.* When fed from a 50c/s supply no 50 or
point where maximum force was produced. With very thin 150c/s vibrations would be imparted to the vehicle, but a
copper coatings, the high surface resistivity limited the force lOOc/s thrust might exist, depending on the number of poles
output. As the copper thickness was increased, the resistivity chosen and the method of speed control; for one of the
decreased and an increase in force output was obtained. As effects of the ends of a linear machine is to superimpose a
the armature was a circular rod, the increase in the copper single-phase field component on to the polyphase-generated
thickness not only increased the magnetic air gap between travelling field, and this can produce double-frequency force
stator and steel rod but also decreased the section of the steel components in the well known manner.
rod. Eventually the turning point was reached where the • Whether a linear motor is a possible or economic solution
increase in magnetic-circuit reluctance outpaced the gain in to a particular problem will always depend on the local
conductivity. conditions. It should be remembered that to lift 100kg is
E. Trinder: In order to study the response of the inner likely to demand an input of at least 8kW, half of which
ear to linear accelerations without exciting by vibration and has to be dissipated in each member.
noise other parts of the nervous system, I have been looking We see the function of the linear induction motor as
for a vehicle accelerator the features of which the linear motor complementary to that of the solenoid and moving-iron
would seem to fulfil with the exception of the control of system, rather than as a competitor, for the latter is virtually
velocity. Now that this can be done, would the authors restricted to applications in which the required movement is
explain this control ? Would they also say if any 50 or 150 Hz only a fraction of an inch. Moreover, unless an elaborate
vibration would be imparted to the vehicle? mechanical design is used, the solenoid will produce a force
The obvious solution to accelerations in the vertical plane whose magnitude is a steep function of armature position,
least force being obtained at the beginning of the operation.
Paper 5023 P by Nix, G. F., and LAITHWAITE, E. R. [see 113, (6), * HESMONOHALGH, D. E., and LAITHWAITE, E. R . : 'Method of analysing the
pp. 1044-1056] properties of 2-phase servo-motors and a.c. tachometers', Proc. IEE, 1963, 110,
Read before the Power Division, 14th December 1966 (11), pp. 2039-2054
1084 PROC. IEE, Vol. 114, No. 8, AUGUST 1967
The linear motor can produce a stroke of inlimited length In reply to Mr. Kendall, single-phase actuators are manu-
and will develop constant thrust throughout the stroke. factured commercially, but, since starting thrust is all impor-
Where the required movement is less than 1 in, a solenoid tant, they are effectively designed as polyphase motors and
will certainly require a smaller input and can make use of make use of capacitors or other phase-splitting devices to
direct current, and will therefore generally be preferred. generate other phases from the supply. Their performance in
The interesting patent to which Dr. Ettinger refers' is, of comparison with a polyphase-fed machine is therefore
course, not for an induction machine at all but for a switched entirely dependent on the effectiveness of the phase splitter.
reluctance motor. We do not think tubular motors could be We are at present experimenting with shaded-pole actuators
used for trains because of the difficulty of supporting the rotor. which we believe to be novel.

Discussion on
Determination of load losses and torques in squirrel-cage
induction motors
B. J. Chalmers: Reference 1 in the paper dealt compre- choice of this value 15 years ago seems to be vindicated by
hensively with the analysis of the main features of high- the present paper.
frequency load losses in cage motors, while the present paper The transient speed/torque curve is not a 'depressed' one;
completes the examination of certain points of finer detail. we should call it a 'retarded' one. The torque at any one
In particular, the authors have now assessed the influence of instant is the torque one should have had a few milliseconds
saturation of the main flux paths. This does not produce one before; the current cannot change instantaneously; the only
single effect only, but rather a plurality of effects, the overall depressed portion is the maximum.
result being a slight reduction as far as load losses are con- B. M. Bird: The paper represents a further significant
cerned. It is interesting to compare the authors' conclusion contribution to the calculation and measurement of load
with the results of experiments conducted at UMIST using losses, and it is a valuable addition to the earlier paper on
an instrumented inverted induction motor. These have the subject by Dr. Christofides.1
enabled the detailed variations of rotor-tooth fluxes, bar The back-to-back test for induction motors described in
currents and air-gap surface fluxes to be investigated under Section 4 is an admirable extension to the mechanical
various conditions of voltage and load current. The overall differential concept and one which readily lends itself to
conclusion of this work is the same as that of the authors. laboratory investigations. The close correspondence between
The authors have clearly devoted appreciable effort to these two tests means that many of the questions raised in
measuring the interbar resistance of their alternative rotors, relation to the mechanical differential test are relevant to the
but such a procedure is not applicable in routine design back-to-back test described in the paper. In particular, it is
analysis. For the latter purpose, knowledge is required of the necessary to assume that the load losses vary according to the
probable value of the interbar resistance for a given range of same laws, regardless of whether a machine is motoring or
machines. Work on this aspect has recently been completed, generating, in order to obtain the stray load loss for an
and it is hoped to present the results for publication. individual machine; but the flux conditions in a generator
We have found the high-frequency losses occurring in the differ from those that exist in a motor, and some evidence
stator to be appreciable in the case of motors with unskewed existsA which indicates that this assumption might well be
rotors, whereas the authors appear to have ignored these questionable.
components. The hot stator resistances of the motor and generator are
W. Hill: Some time ago, I stated that the back-to-back presumably taken at each test point, since a precise knowledge
tests required identical machines. I now withdraw that of these values is essential to the accuracy of the loss measure-
unreservedly. The attempts to deal with saturation in a ments. If the machines are brought to rest to perform these
properly scientific manner have borne fruit, and we are now measurements, it becomes difficult to maintain control over
seeing the results of a systematic study of this phenomenon. the temperature of the machine during the back-to-back test.
Saturation should yield to such treatment. Did the authors attempt any form of temperature measure-
The authors have probably established not so much the ment and control during the back-to-back test? Were any
saturation in the tooth, but that in the layer adjacent to the steps taken to relate the actual rotor temperature to that
air gap, which will actually reduce the zigzag leakage of the assumed for the purposes of computing the losses?
machine. Extrapolation to larger machines will also show In Section 3, the authors describe in detail the accelerometer
that another factor comes in, namely current displacement in arrangement used to measure the torque/speed curve. It is
the actual rotor, whereby the speed/torque curve and the clear, however, that this quasidynamic technique will only
harmonic-speed/torque cusps will behave differently under yield the static torque/speed curve if the time for plugging the
saturated conditions. As the purpose of the investigation is to machine is somehow increased far above the time constant
predict the load loss under full-load conditions, it seems that of the rotor. By what means was this time increased and what
more accurate results will be obtained by basing the harmonic- was the approximate time for plugging? The technique used
torque calculations on the unsaturated value, and not on the to obtain the static torque/speed curve is extremely complex:
saturation one can expect under full voltage during starting. did the authors investigate alternative forms of dynamic
In order to determine the torque curves during the running-up torque/speed measurement such as the drag-cup accelero-
conditions one cannot use this arrangement where the period meter principle ?B
of starting is extended for up to 5 or 6s, because all the K. K. Schwarz: 1 believe the paper should be considered
transient effects of the rotor current, which cannot be changed in the wider context of the general subject of stray losses, and
instantaneously, will be lost. a recapitulation of the position may be opportune:
My own attempts at calculations of saturated leakage-flux
conditions were based on a leakage flux density Bo not of (a) Design parameters affecting all types of stray losses have,
1 -lWb/m2 as in this paper (Fig. 14) but of 1 -23Wb/m2. The as pointed out in Reference 3, been known for a very long
time, although many aspects are, and remain, partly
empirical, and investigations are necessary for checking
Paper 5130 P by CHRISTOFIDES, N., and ADKINS, B. [see 113, (12), individual loss calculations and improving the sophistica-
pp. 1995-2005] tion of a number of aspects.
Read before the Power Division Professional Group PI, 1st February
1967 (b) Methods of checking the total losses, including stray
PROC. IEE, Vol. 114, No. 8, AUGUST 1967 1085