5-B] TEST BY LOSSES. J 73 



first in parallel on the appropriate normal voltage and then in series 

 on the same voltage (which will then be half normal) ; thus, two 

 55-volt coils in parallel on a 55-volt circuit give the reading for 

 normal voltage, while the two coils in series on the 55-volt circuit 

 give the reading for half normal voltage. For the same degree of 

 magnetization, the wattmeter will indicate the same core loss what- 

 ever coil is used; the ammeter will read twice as much for the 

 parallel as for the series arrangement (the ampere turns being the 

 same) and the ammeter reading must, accordingly, be divided (or 

 multiplied) by two to reduce all readings to common terms. The 

 voltmeter reading must be multiplied (or divided) by two. , 



47. In the short-circuit test, the matter of wave form is practi- 

 cally of no consequence ; the only result affected is the reactance drop 

 and only a very large change in wave form could materially affect 

 its value. For the short-circuit test, therefore, any means of adjust- 

 ment may be used which is found convenient. 



48. Normal Voltage and Current. In determining normal full- 

 load values of current and electromotive forces, the assumption is 

 commonly made that the efficiency is 100 per cent, and that currents 

 and voltages are transformed exactly in the ratio of turns. Thus 

 in a 2,000/100 volt, 2 K.W. transformer, the secondary current is 

 taken as 20 amperes and the primary current as I ampere (whereas 

 strictly the latter should be a trifle more) ; the secondary voltage is 

 taken as 100 volts and the primary 2,000 volts, neglecting the fact 

 that these no-load values of voltages do not strictly hold at full load. 

 Any change of rating changes the results of a test. If the voltages 

 are rated as 2,200 and no, the corresponding primary and secondary 

 currents are 0.909 and 18.2 amperes; the copper loss is less and the 

 core loss greater. In comparing transformers and their guarantees, 

 each transformer should be tested at its rating. In comparing trans- 

 formers for a specific service, the tests of all should be made at a 

 common voltage according to the conditions of the service. From the 

 curve sheets it is very easy to pick results for different ratings from 

 one set of data. In the laboratory, to facilitate comparison of data, 

 certain voltages should be adopted as standard, as 100-200/1,000-2,000, 



1 1 0-220/1, 100-2,200, etc. 



49. Core Losses and Their Variation. Eddy currents flow in local 

 short-circuited secondary circuits which are practically non-inductive. 



