462 Mr. W. E. Sumpner on the Variation 



sufficiently ballistic galvanometer is to be had, it is not easy 

 to obtain L in absolute measure. 



At the suggestion of Prof. Ayrton some experiments were 

 made at the commencement of 1887 to determine the variation 

 of the self-induction of the field-magnet coils of a Ferranti 

 dynamo. The method* used was the modification of Maxwell's 

 method already alluded to. (The coefficient has been since 

 compared with the capacity of a standard condenser by the 

 swing method ; the value obtained for a current of O'Ol 

 ampere was 0*61 secohm, and for larger currents larger 

 values were obtained.) When a strong current was flowing 

 through the coils the resistance of the circuit was only a few 

 ohms, and the coefficient was probably larger than 0*6. The 

 time-constant was therefore generally larger than one tenth 

 of a second, and the discharge could never be considered 

 completed in less than a second. Probably no galvanometer 

 would have been sufficiently ballistic under these circumstances 

 to give good results in absolute measure. As, however, the 

 only galvanometer conveniently situated with reference to the 

 dynamos was one of the D'Arsonval type, with a short period 

 and large logarithmic decrement even when on open circuit, 

 the hope of any but comparative measurements was abandoned. 



10. It is possible to calibrate any galvanometer for ballistic 

 purposes by charging a condenser to a standard potential and 

 discharging it through a resistance in series with the galvano- 

 meter to be calibrated. By suitably altering the resistance or 

 the capacity the time-constant of discharge may be made to 

 have any value. This was done with the D'Arsonval galva- 

 nometer, in order to interpret the results obtained. A con- 

 denser of 18*6 microfarads capacity was charged by a Latimer- 

 Clark's cell, and then discharged through the galvanometer 

 (whose resistance was 700 ohms) in series with a variable 

 resistance, R ohms. The time-constant of discharge was 

 therefore 18-6(R+700)H)- 6 seconds, 



and the ratio of this to 1*66 was the ratio of the time-constant 

 to the period of the galvanometer, and is denoted in the 

 following table by the letter p. Observations were made 

 of the value of the first swing of the galvanometer-needle for 

 many different values of R. A few of these are given below. 



* This method, due to Prof. Ayrton, will be found more fully- 

 described, together with several otber methods, in a paper by the 

 present writer on "The Measurement of Self-induction, Mutual In- 

 duction, and Capacity," Journ. Soc. Tel. Engineers, May 1887. 



