252 ALTERNATORS. 



The curve rises very steeply from zero, since the impedance 

 at low speeds is small. The dotted portion could not be 

 obtained experimentally, as it was not possible to drive the 

 machine steadily at such low speeds. The approximate 

 position of this part of the curve was obtained by calculation 

 from the value of the armature impedance previously cal- 

 culated, and the calculated voltage at low speeds derived from 

 the measured values at normal speed. 



Efficiency Tests of Alternators. The direct method of deter- 

 mining the efficiency of an alternator, by measuring the 

 power supplied to drive it and the power given out in 

 the form of electrical horse-power, is not generally adopted 

 on account of the difficulty of measuring accurately the 

 mechanical power supplied. The nearest approach to this 

 method of measurement which is frequently adopted in 

 testing small machines, is to drive the alternator by means of 

 a motor whose efficiency is known at various loads. By 

 measuring the power supplied to the motor, and multiplying 

 this by the motor efficiency, the power transmitted to the 

 alternator is obtained. 



In the case of large alternators the power necessary to 

 drive the machine at full load is costly, and the absorption 

 of the load current becomes a difficult matter. Also the 

 measurement of the output with a sufficient degree of accuracy 

 to make the value obtained for the efficiency reliable is by 

 no means an easy matter. 



It is consequently usual to calculate the efficiency of 

 alternators from measurements of the losses occurring in the 

 various portions of the machine, and to divide the output 

 by the calculated quantity (output -(- losses) at various 

 loads to arrive at a curve of efficiency. 



Losses in an Alternator. The chief losses in an alternator 

 may be summarised under four headings : (a) Armature 

 resistance loss, (6) iron losses (hysteresis and eddy currents), 

 (c) excitation losses, (d) friction losses (air and bearings 

 friction). 



The losses (a) can be calculated for any load by multi- 

 plying the armature resistance by the square of the current 

 at that load, care being taken in the case of polyphase 

 alternators to take the value of the current actually flowing 

 in the windings, which is not always the same as that in the 

 external circuit (e.g., in a delta-connected 3-phase winding). 



