14 



THE BELL SYSTEM TECHNICAL JOURNAL, JANUARY 1954 



The Measurement of Flux 



When attempts were first made to extend the speed performance of 

 general purpose relays much below 8 or 10 milliseconds functioning time, 

 it was found that the simplified eddy-current theory, which treated the 

 core like a single short-circuited turn, was no longer sufficiently accurate. 

 The subject has been clarified through the use of the dynamic fluxmeter 

 which permits a direct determination of how the relay flux changes with 

 time, under actual conditions. This fluxmeter, originally due to E. L. 

 Norton and recently refined by M. A. Logan, was described in a previous 

 issue, and has proven most useful for determining those constants of the 

 relay which the designer must understand in order to work much l)elow 

 8 milliseconds functioning time. The instrument is shown in Fig. 9. It 

 requires that the relay under test be pulsed, usually at a speed of about 

 10 to 20 cycles. When the main winding is alternately connected and 

 disconnected, the search coil which surrounds the Annding sends alter- 

 nate positive and negative voltage impulses to a dc ammeter connected 

 across its terminals. Now it can be shown that when the search coil and 

 meter are disconnected during the interval from time zero to the moment 

 of interest, then the resulting reading on the meter is exactly propor- 

 tional to the flux at that moment. This switching function is provided bj^ 

 a timing circuit comprising a 40-kc frequency source driving a 3 decade 

 counting ring, under control of s\\itches permitting selection of the num- 

 ber of cycles of delay time. In this way the flux maybe measured at inter- 

 vals of 25 millionths of a second, with better than 1 per cent accuracy. 



One interesting measurement has shown that in the short time inter- 

 vals of present-day relays, eddy currents have less effect on the initial 



0.4 0.6 0.6 



BILLIONS OF CYCLES 



Fig. 8 — Slide data on AF relay 



