28 



BELL SYSTEM TECHNICAL JOURNAL 



is a lag ill the tield current behind the lield registered on the drum, due 

 to the self inductance of the coils. Added to this there is a lag in the 

 secondary due to its self inductance, and another lag due to the time 

 required for the fluxmeter to act. The effect of these is to widen the 

 loop. In Fig. 9 is shown a curve traced with no magnetic sample in 

 the field coil, and with dlljdt so great that the lag is appreciable. 



Fig. 9 — Loop made with an air core mutual inductance at a very high d Il/dt. 



Fig. 10 shows two loops, the outer one representing a loop as taken on 

 the apparatus, and the inner one the true loop corresponding thereto. 

 Let B be some induction near zero, on the traced loop. B will be 

 incorrect for the indicated value of // by an amount Bo — B, such 

 that if the field were held constant at that point while the drum con- 

 tinued to rotate the cur\e would approach Bo as an asymptote, as 

 indicated by the dotted curve. If dll/dl is not zero, B may be regarded 

 as momentarily approaching Bo as an asymptote. The equation for 

 B at any instant is: 



\,~ + B = Bo, (3) 



where Xj is the time constant of the circuit and Bo is not a constant 

 but a function of // and /. If we assume that dB/dll is constant for a 

 small region in the neighborhood oi B = Q, we have, putting Allc 

 equal to the error in coercive force lie, 



Bo- B =^j-^JJc 



