1 5 6 TRANSFORMERS. 



[Exp. 



showing the core loss, W Q , for different voltages, as in Fig. 5. 

 Locate by heavy black dots the two points accurately deter- 

 mined at about half and full voltage. Draw a straight line 

 through these points and note that at normal and higher fre- 

 quencies this straight line gives the curve accurately through 

 - . 92 the working range, so tjiat 



W Q for normal voltage can 



**if 



be readily obtained from it. 

 At frequencies much below 



.-40 



3 



! 



. 32 normal, and at high flux- 

 densities, this straight line 

 relation may not hold. 



If it was impossible to get 

 data for any curve up to 



20 40 60 80 100 120 140 



FREQUENCY: CYCLES PER SECOND normal valtage, extend the 



FIG. 4. Watts core loss for different curve that far as a dotted 

 frequencies at normal voltage: 2 K. W. ,. rr-., . 



. line. This extension is 



transformer. 



quite accurate at frequencies 



r^ear normal or higher, but can not be depended upon at fre- 

 quencies way below normal. The curves, however, can be more 

 readily and more accurately extended on logarithmic paper than 

 on ordinary coordinate paper. 



10. The slope of these curves (the actual tangent with the 

 horizontal) is the exponent (a) of E or B in the formula show- 

 ing the law of core-loss variation for different voltages and 

 flux densities at a constant frequency; 



W oc oc B a . 



This exponent (a) should be determined and interpreted; see 



49-51- 



ii. Variation of Core Loss with Frequency. On the same 

 logarithmic sheet, see Fig. 5, plot a derived curve showing the 

 core loss, W Q , at normal voltage for different frequencies. If 



