MAGNET STEELS AND PERMANENT MAGNETS 



399 



A further application of the idea that the demagnetization curve 

 can be represented by a hyperbola is made by Mr. E. A. Watson.* 

 Watson applies a graphical construction for a hyperbola to the demag- 

 netization curve and extends the construction into the region of positive 

 values of //, thus providing a means of extrapolating the hysteresis 

 loop to the saturation value of B. It has been the writer's experience 

 that Watson's graphical construction cannot be legitimately extended 

 into the first quadrant because the curve thus predetermined almost 

 always lies considerably under the true curve, and yields too low a value 

 for the saturation flux density. Fig. 13 illustrates the case in point. 



200 

 IN OERSTEDS 



Fig. 13 — Comparison of graphical construction of the descending portion of the 

 hysteresis loop with actual measured data. Note the lack of agreement for positive 

 values of //. 



If, instead of potting B vs. II for the points of the demagnetization 

 curve, we plot B/Br vs. H/IIc, it is found that all test specimens yield 

 points fitting fairly well the single curve represented by the equation 

 y = A{x -{- l)/(x + A), in which y = B/Br, x = H/Hc and A = ^l2 

 approximately. It is possible to derive from this equation the rela- 

 tionship that {BII)max = 0.423 Br He which checks very well with the 

 line drawn In Fig. 11. It can also be seen from the equation y = 

 A{x -\- \)l{x -\- A) that the representation of the demagnetization 

 curve by a hyperbola cannot be extrapolated to give the correct 



