304 BELL SYSTEM TECHNICAL JOURNAL 



4. Demagnetizing Factor of Magnetizing Coil and Sample. 



The errors under this heading pertain to straight coils and samples 

 and include the open end correction of the magnetizing coil, the error 

 due to the non-uniformity of the magnetizing field of the open ended 

 coil, and the demagnetizing factor of the finite length of sample. 



Calculations by L. W. McKeehan and measurements by P. P. 

 Cioffi ^ on coils and samples of nearly the same dimensions as those 

 used in this apparatus, have shown that the errors in H and B which 

 result from neglecting the above factors amount to less than three 

 per cent. The exact correction to be applied for any particular 

 sample is not readily calculated, but the sample can be assumed to 

 be sufficiently long if displacing it a few centimeters in the coils 

 results in no appreciable change in the hysteresis loop. 



5. Eddy Currents and Magnetic Viscosity in the Sample. 



Currents induced in the sample flow in such a direction as to oppose 

 the applied magnetizing force, with the result that to reach any 

 given density of magnetization a higher applied field is required when 

 eddy currents are present. The loss of energy in eddy currents is 

 added directly to the hysteresis loss, so that the dynamic hysteresis 

 curve is different from the static one. Qualitatively this difference 

 manifests itself as a widening in the H direction of the recorded loop 

 at intermediate frequencies, and a shrinkage in the B direction when 

 the frequency is so high that skin effect prevents the induction from 

 reaching its full value before the applied field has receded appreciably 

 from its maximum. 



In dealing with the distortion introduced by eddy currents, the 

 procedure must be governed by the nature of the problem in hand. 

 If the total energy loss corresponding to a given magnetic cycle is 

 sought, no eddy current correction is required. If it is desired to 

 reproduce as nearly as possible the static hysteresis loop of a specimen, 

 then the frequency of magnetization should be chosen low enough so 

 that, with the aid of the condenser in parallel with the magnetizing 

 coil, the eddy current effect is not appreciable. If, finally, the object 

 of the work is to detect differences between the static and the dynamic 

 hysteresis curves, the presence of eddy currents must be reckoned with. 

 The following analysis, based upon a simplified static magnetization 

 cycle and an assumed sinusoidal time variation of the magnetizing 

 field, makes possible a rough estimation of the apparent increase in 

 the coercive force //,; introduced by the eddy currents.^ 



Let Fig. 12 represent the cross-section of a rectangular lamination 



6 P. P. Cioffi, Jl. Opt. Soc. Am. and Rev. Sc. Inst., 9, p. 53, 1924. 

 ^ A similar result has been obtained and used by Neumann, I.e. 



