1903.] Tke" Hunting" of Alteniatiny-Current Machines. 249 



with a brake until the applied power was about 1800 watts that the 

 " hunting " could be stopped. 



[That the hunting in this case was principally due to the negative 

 value of y was shown by the fact that it was not possible to stop the 

 hunting by the introduction of external resistance. Even with a 

 resistance amounting to 2 -5 ohms, with which p would certainly be 

 greater than (L + *) p, the motion was still unstable. The im- 

 possibility of making the motion stable by adding resistance puzzled 

 me considerably at first, as I had not noticed the importance 

 of the negative term in y', but had treated it as a negligible quantity 

 June 28th.] 



By inserting self-induction in series with the motor the violence of 

 the " hunting " could be gradually diminished, and at the same time its 

 period could be continuously increased from about 0'35 second to 0'45 

 second. When the self-induction was such that L}> amounted to 

 about 6 ohms the motion became stable, the oscillations when once 

 started rapidly dying away. This is in accordance with conclusion (1) 

 above, and the continuous variation of the period though the hunting 

 continues is proof of the essential instability of the motion. 



With external self-induction giving a total reactance of about 3 ohms 

 and resistance of T2 ohms the stability depended on the load and on 

 the exciting current. The following observations were made : 



(a) Exciting current ll'O, motor unloaded. The motion was stable, 

 oscillations if started dying slowly away. Period of oscillation 

 0*42 second. 



(//) Exciting current was increased to 18'5 when the motion became 

 unstable. By loading the motor slightly with a brake the " hunting " 

 could be stopped. Period 0'38 second. 



(f') Exciting current increased to 30 '5. The motion with the motor 

 unloaded became violently unstable. By loading the motor with a 

 brake until the power delivered to motor (excluding external self-induc- 

 tion) amounted to 1000 watts, the " hunting " could be stopped. It was 

 curious to watch the effect of the load on the current taken by the 

 machine. The ammeter had such a slow period that it could not 

 follow the variations in current due to the oscillations, but remained 

 steady, giving the average of the square of the current. When the 

 machine was "hunting" the ammeter showed a current of about 

 27 amperes, and the wattmeter an applied power of about 710 watts, 

 two-thirds of which were, of course, accounted for by the C 2 R losses. 

 On putting the load on, the current as shown by the ammeter 

 gradually diminished until when the motion was steady it was 

 13'5 amperes. The power indicated by the wattmeter was now again 

 about 1000 watts, but this time only about one-eighth was going in 

 C-R losses. 



The last series of experiments is a good confirmation of conclusions 



