COPPER LOSSES 



109 



.'E 



R 



The hysteresis and eddy-current losses as calculated by the 

 methods explained above are always found to be less than the losses 

 actually occurring in transformers. A little consideration will show 

 that such a discrepancy might be expected 

 to exist. For in calculating the losses we 

 have made the assumption that the induction 

 is uniformly distributed over the cross- 

 section of the core. Such, however, is by 

 no means the case, as a single glance at 

 Fig. 88, whicli represents a core stamping, 

 will show. For the length of path corre- 

 sponding to ABCD is very considerably 

 snorter than that corresponding to EFGH, 

 with the result that the flux will be crowded 

 towards the inner surface of the core. Now, 

 any departure from uniformity in flux dis- 

 tribution results in an increase of both the 



hysteresis and the eddy-current loss. An and Observed Core Loss, 

 upper limit to these losses may be obtained by 



calculating them on the assumption that the value of B is that 

 corresponding to its value close to the inner surface (ABCD) of 

 the core. 



55 Copper Losses. Best Dimensions of Core 



The copper losses are easily calculated from the resistances of 

 the coils and their full-load currents. In the case of windings con- 

 sisting of heavy copper conductors, however, additional losses come 

 into play, due to eddy currents in the conductors, and these may 

 bring about an apparent increase of resistance, amounting, in some 

 cases, to as much as 20 per cent. For this reason, most makers 

 prefer to avoid the use of such heavy windings, employing instead a 

 number of coils wound with smaller wire and connected in parallel. 



The total losses at full load depend on the output of the trans- 

 former. The following table gives the efficiencies which might 

 reasonably be demanded in modern transformers : 



Output, in k.w. 1 5 10 20 50 100 150 250 

 Efficiency ... 94% 95% 95'5% 96% 96'5% 97% 97'5% 98% 



Let us suppose that the average length of magnetic path and the 

 total volume of iron are given, but that it is permissible to vary 

 the relative cross-sections of the core-plates and yoke-plates. Any 

 departure from a uniform cross-section for the entire magnetic circuit 

 will increase the iron losses. On the other hand, however, a reduction 



