the consequence is that the output per unit weight is much greater in an induction motor than in a direct current machine. It is not difficult to construct induction motors of a moderate size, which only weigh 60 or 70 lbs. per horse-power, and this without any sacrifice of efficiency.

Unless considerably overloaded, an induction motor will run quite cool compared with direct-current mctors of the same weight and output. This is due to the large section which can be given to the conductors and to the fact that the winding is so arranged that the copper losses are not localized, whilst the laminated character of the iron facilitates ventilation.

Although polyphase induction motors are termed nonsynchronous, it should be borne in mind that there is always a tendency towards synchronism. The speed of an induction motor is, in fact, almost independent of the load; the variation in speed from no-load to full-load seldom exceeds 5 per cent.

It is usual to make large machines multipolar in order to reduce the peripheral velocity of the rotor. Small machines can safely run at 2000 revolutions or more per minute, and so can be made bi-polar. The angular velocity of the rotor, for a given frequency in the stator currents, varies inversely as the number of poles, so that large machines are necessarily of the multipolar type.

115. Before concluding this chapter, we draw the attention of the reader to a few points which have to be carefully noted in the design of monophase induction motors.

There are two distinct stator windings; one-the running coil -must be such as to provide a counter E.M.F. nearly equal to the applied P.D. and capable of carrying the full-load current continuously; the other the starting-coil—which must likewise provide a counter E.M.F. nearly equal to the applied P.D., but since it is only in circuit for a very short time, it may run at a much higher current density than the running-coil.

Let us examine closely the conditions which the two coils have to satisfy.

When running on load, the power-factor of the motor must be as high as possible. The running-coil, therefore, must be wound so as to have as small an equivalent self-induction as possible; that is to say, the mutual induction between the running coil and the rotor-windings should be as large as possible. The wattless

currents in the stator-windings are solely due to the leakage field, or those magnetic lines which cut the stator coils alone. The running-coil should, therefore, be wound as near to the inside edge of the stator core as possible, and the opening in the slots in which it is wound should be greater than the air-gap between stator and rotor. Further, the winding should be such as to give a fairly high induction density in the iron, otherwise the demagnetizing action of the rotor currents will, at start, have too great an effect, a feeble starting-torque resulting.

The starting-coil, on the contrary, should be wound in closed slots, since in this case a large equivalent self-induction is necessary to give the required difference of phase between the currents in the two stator-windings. Since it is only in circuit for a short time, this coil can have fewer turns than the running-coil, and carry a heavier current.

CHAPTER XV.

Polyphase Transformers-Phase Transformers and Rotary Converters.

POLYPHASE TRANSFORMERS.

116. We have seen that monophase transformers are used for the purpose of transforming alternating electromotive forces from high to low values, or vice versa.

Polyphase transformers may likewise be used for simply transforming the values of the E.M.F.s, or they may be also arranged so as to transform the number of phases of the E.M.F.s.

There is little to be said in the case of polyphase transformers used simply for transforming E.M.F.s, since the same laws relating to ratio of turns hold good here as in monophase transformers; that is, the ratio between the primary and secondary E.M.F.s is approximately the same as that between the number of primary and secondary turns.

We might, of course, in the case of tri-phase currents, employ three single-phase transformers, viz. one in each of the three circuits; but just as it is unnecessary to have three separate circuits with six line-wires for transmission, so is it equally unnecessary to use three separate transformers. All that is necessary is to have three limbs magnetically short-circuited by common yokes, as shown diagrammatically in Fig. 72, in which P1, P2, P3 represent the primary coils, and S1, S2, Sg the secondaries. The phase relationships between the magnetic fluxes in the three cores will be similar to the phase relationships between the three primary

currents.

In the case of di-phase transformers, three cores are again all that is necessary, provided the section of the core wound with the coil which is connected to the common line-wire is √2 times the cross-sections of either of the other two, in order that the induction in all three cores may be the same.

Having due regard to these details, the design of a polyphase transformer differs in no respect from that of a monophase transformer.

117. Phase transformers are used for the purpose of changing alternating current of any number of phases to alternating current of a different number of phases.

The problem of producing such phase transformations was first solved by Professor S. P. Thompson in the following manner :

A ring transformer having a number of coils in closed series was tapped at three equidistant points, and fed thereat by tri-phase currents, with the result that a rotating magnetic field was produced. On further tapping at the extremities of any diameter, single-phase current could be taken from it; or by tapping at four equidistant points, di-phase currents are obtained. In fact, by making a suitable number of tappings, currents of any number of phases could be obtained.

Such a phase transformer is really an auto-transformer of special type, the correct phase relationships of the secondary currents being dependent upon the production of a magnetic field rotating with uniform angular velocity. It cannot, therefore, be used for transforming monophase currents to polyphase.

In the case in which the primary current is tri-phase, the ratio between the primary and secondary E.M.F.s is approximately

proportional to the ratio of one-third of the whole number of turns to the number of turns between the tappings corresponding to one phase of the secondary.

A method of transforming from di-phase to tri-phase currents was subsequently given by Mr. C. F. Scott, of the Westinghouse Company, by the use of two specially wound transformers, as depicted in Fig. 73.

The primaries of the transformers are connected to the terminals of a di-phase generator. In the transformers used by Mr. Scott, one secondary CB was made equal to 100 turns, and was tapped at its middle point, giving 50 turns on each side. The other secondary had 87 turns (= 50√3). One end of it was connected with the middle point of the secondary of the first transformer, and the three free terminals then A, B, C, gave tri-phase E.M.F.s.

The methods given above do not include the most important case from a practical point of view, viz. that of transforming monophase currents to di-phase or tri-phase. Although methods of achieving this have been theoretically propounded, they do not seem to have been realized in practice by any stationary transformer.

ROTARY CONVERTERS.

118. A rotary converter is a machine with but one armature winding, which transforms alternating currents of any number of phases into continuous current, or vice versa.

This definition excludes such combinations as alternatingcurrent motors coupled to direct-current generators, which are called Motor Generators.