148 THE TRANSFOTniKK. 



Iron Losses. The iron losses occurring in the core of the 

 transformer are similar to those which take place in the 

 armature core of a direct-current dynamo or motor, and are 

 due to the same two causes. 



The power spent in hysteresis increases with the intensity 

 of magnetisation of the core, but with a constant magnetic 

 flux is proportional to the number of reversals in the polarity, 

 i.e., it is proportional to the periodicity of the current. 



The eddy currents in the core, being induced in the con- 

 ducting plates of the core by the varying flux across them, 

 will also increase in direct proportion to the rate of alterna- 

 tion of the flux, obeying the usual law of induction in con- 

 ductors by magnetic lines. The watts spent in producing 

 the eddy currents will, however, vary as the square of these 

 currents (since watts = C 2 E). Hence the watts lost in 

 producing eddy currents will vary in proportion to the 

 square of the periodicity of the current supplying the 

 transformer. 



Separation of Iron Losses. The usual formulae employed 

 for the iron losses in the core of a transformer are as follows : 



Watts lost in hysteresis = K n B 16 . 10 7 per cub. cm. 

 of core. 



'K = constant between -002 and 004. 

 n = periodicity of current. 

 B = max. induction in lines per sq. cm. 

 t = thickness of core plates in mils, (thousandths of 

 an inch). 



Watts lost in eddy currents per cub. cm. of core 

 = w 2 f B\ 10" 1G 



From these expressions it appears that it would be 

 possible, but somewhat difficult, to separate the two losses from 

 observations made on a transformer at various magnetic 

 densities, since the eddy current losses increase as the square 

 of the induction B, while the hysteresis losses increase as 

 the l'6th power of B. An easier way of experimentally 

 separating them is to keep the induction constant, and to 

 measure the watts lost in the iron at different frequencies. 

 Thus if two measurements are made, one with a current of 

 double the frequency of the first, it will be found that the 

 power spent on the core has more than doubled, since the 

 eddy current losses increase as the square of the frequency. 



