burrows: the koepsel permeameter 491 



PHYSICS. — An experimental study of the Koepsel permeameter. 

 Charles W. Burrows. ^ Bureau of Standards. 



A detailed experiinental study of the possibilities and limitations 

 of the Koepsel permeameter has led to the following conclusions 

 regarding the different factors which may affect the accuracy of 

 the readings : 



1. Readings on the two sides of the zero of the instrument may differ 

 considerably but the mean of the two values thus obtained shows 

 satisfactory consistency on repetition. 



2. Shearing curves for different grades of material show that the 

 correction to be applied to the observed magnetizing force is not con- 

 stant for a given induction, but depends upon the nature of the test 

 specimen. This correction is usually subtractive for points below the 

 Imee of the induction curve and additive for points above the knee. 



3. An increase in the cross section of the test specimen tends to in- 

 crease the observed values of the magnetizing force for points below 

 the knee of the induction curve, and to decrease the observed values 

 for points above the knee. 



4. The length of the specimen projecting bej'ond the j'okes produces 

 no noticeable effect for points below the knee of the induction curve. 

 For points above the Ivnee the projecting ends increase the observed 

 value of the magnetizing force. 



5. If the bushings are not pushed all the way into their proper posi- 

 tion, a higher apparent value of the magnetizing force is observed due 

 to the mcreased length of the portion of the bar under test. 



6. Hysteresis loops obtained by the Koepsel permeameter always 

 show a low observed residual induction and a high observed coercive 

 force. 



7. A theoretical and experimental stucty of the distribution of the 

 magnetic fluxes through different parts of the magnetic circuit shows 

 that shearing curves of the form observed are to be expected. 



Apparatus of this type if used without correction may yield 

 data greatly in error. The results of the present study show 

 that for small and moderate inductions the measured magnetiz- 

 ing force is usually in excess of its true value, sometimes by as 

 much as 100 per cent. At high inductions the measured magne- 

 tizing force is usually too low by an error which may be as great 

 as 25 per cent. With care, however, and the use of proper 

 correction curves, the apparatus is capable of yielding quanti- 



1 Detailed paper to appear in the Bulletin of the Bureau of Standards. 



