5-B] TEST BY LOSSES. 161 



In this test, when full-load current flows in the primary, full- 

 load current flows in the secondary; when half-load current 

 flows in the primary, half-load current flows in the secondary, 

 etc. The flux density is very low, so that there is practically 

 no core loss. The wattmeter reading gives, therefore, the total 

 copper losses both primary and secondary for any particular 

 current. Included with the copper losses are the load losses. 



18 Load Losses. Load losses are due chiefly to eddy cur- 

 rents in the copper and are greatest, therefore, in large solid 

 conductors. They have the effect of causing a greater loss in a 

 conductor when traversed by alternating current than when 

 traversed by direct current, the resistance being apparently in- 

 creased. The term load losses includes all losses* which in- 

 crease with load and depend upon current, over and above the 

 copper losses as determined by direct current. Evidently such 

 losses should be taken into consideration in calculating efficiency ,f 

 and the Standardization Rules of the A. I. E. E. so specify. 



outside of the transformer. An ammeter and its leads in the secondary, 

 sometimes used, tends to give the transformer a poorer regulation and 

 efficiency. It is instructive, however, before taking readings, to insert an 

 ammeter in the secondary, as well as the primary, and to note that the 

 ratio of currents is practically equal to the ratio of turns. 



If there are two secondaries, it makes no difference whether they be 

 put in parallel or in series ; two primaries should be put in parallel or 

 series to suit the range of instruments. No coil should be left idle. 



* For example, eddy-current loss in the core due to local flux set up by 

 the current in a loaded transformer in addition to the normal core loss. 



t(i8a).Very commonly, however, this is not done, copper losses 

 (neglecting load losses) being determined by direct current measurement 

 of resistance. This tends to favor the transformer. In justification of 

 this, it may be said that it has not been fully established that the load 

 losses under actual load conditions are the same as those obtained on 

 short circuit it being held that they may be less. The two methods serve 

 as a check. If the losses by the wattmeter are only slightly greater than 

 by direct current, the result is satisfactory for the transformer. Any con- 

 siderable difference, however, shows the existence of load losses. For an 

 accurate comparison, great care is necessary in regard to temperature 

 conditions and the calibration of instruments. 



12 



