7-1 THE BELL SYSTEM TECHNICAL JOURNAL, JANUARY 1954 



volves application of the same relations that apply to the operate per- 

 formance. 



The procedures outlined for the estimation of sensiti^'ity and work 

 capacity, or of the magnetic parameters required to give a desired per- 

 formance in these respects, have their primary use in preliminary design. 

 They make it possible to determine from the performance requirements 

 definite dimensional criteria which must be satisfied by the design. 

 After models have been made and measured, the load characteristics 

 may be compared directly with the observed pull. Study of the mag- 

 netization relations, and comparison of obser\-ed and estimated magnetic 

 circuit constants, however, is of ^'alue through the whole course of 

 development, and even in the engineering use of an established design. 

 Such study relates the performance directly to the dimensions and ma- 

 terials used. In particular it permits extrapolation of the performance 

 observed on particular models, providing estimates of the range of per- 

 formance variation corresponding to the range of tolerances in dimen- 

 sions and material properties. An essential phase of such studies is the 

 experimental determination of the magnetic circuit constants by the 

 methods discussed in a companion article. 



Validity 



The solutions to the static field equations based on the magnetic cir- 

 cuit treatment are inherently approximations. The discussion of this sub- 

 ject in many engineering texts implies that the approximation is rather 

 crude. In principle, as shown above, the magnetic circuit method may 

 be applied to as close an approximation as desired, sul)ject to the ac- 

 curacy with which the pattern of the field can be recognized. 



The criterion of satisfactory estimation of the magnetic circuit parame- 

 ters is the ability to predict the magnetization relations actually ob- 

 served. Analysis of observed magnetization relations by the method 

 described in the companion article has shown that these relations can 

 be satisfactorily represented by the magnetic circuit relations given 

 here. Satisfactory agreement has been found between obser^'ed and 

 estimated values of the magnetic circuit parameters when the latter are 

 determined by the procedures described above. These conform in large 

 measure to those initiated by Evershed, and are distinguished princi- 

 pally by explicitly recognizing the existence of leakage paths shunting 

 any air gaps, and by recognizing that these cannot be identified by vsearch 

 coil measurements, so that thc^ field between two members l)ounding a 

 gap must })e regarded as the \'ector sum of a constant and a variable 

 field. 



