1903.] The " Hunting " of Alternating-Current Machines. 245 



J A /V Hence for a fixed field excitation the damping term is about 

 proportional to the square of the load; the more the machines are 

 loaded (within limits) the better they run in parallel. This is in accord 

 with experience. Furthermore, with constant load, /3 is inversely 

 proportional to A, and a reduction of A results in better damping. 

 This I have also found to be true on the machine with which I experi- 

 mented, viz. : that if with a constant load you diminish the field 

 current you get more stable running. 



(3) The co-efficient of instability, or y, will in most cases (for which p is 

 small compared with Lp) be proportional to />A 2 and inversely propor- 

 tional toL 2 ^ 2 . 



(4) If ft be increased from zero, in other words if the resistance to the 

 induced currents in and about the pole pieces be diminished, the 

 damping effect corresponding to y' first increases to a maximum and 

 then diminishes. It is possible therefore to carry the application of 

 " Amortisseurs " too far. That this is so is obvious when one considers 

 that a coil of no resistance round the pole piece would completely 

 destroy all variation of induction and all the damping effect which 

 depends on such variations. It would in fact correspond to the case 

 first investigated of a motor with a fixed field, or without armature 

 reaction, which as shown is usually unstable. 



(5) Referring to the general expression for y (equation 9) in 

 which the resistance is taken into account, it appears that if 



p 2 < .P'.f-" \ N ~ a )2 , y becomes negative, and the clamping coils, 

 {L(L + A)p +/>*}* 



instead of reducing the oscillations, actually tend to increase them. 

 Jy ~ y P. ; is about equal to C( N /2), where C is the current 



in the machine when standing and connected to the mains. The limits 

 outside which /? must lie in order that the damping coils may operate 

 as such, and not as additional causes of instability, are therefore about 



J C( >/2), and are very narrow in machines of any size. The 

 Lp 



matter only becomes practically important in the case of small machines, 

 and of motors on long transmission lines. The behaviour of the small 

 motor with which I experimented is considerably influenced in 

 this way. 



Experimental Confirmation. 



I have in a general way confirmed the results here obtained by 

 experiment on a small alternating current motor. This machine is a 

 4-pole generator, made by the Westinghouse Company, and intended 

 to give an output of about 10 amperes continuous current at 

 110 volts. It was converted into a synchronous alternating current 

 machine by fitting slip rings on to it. 



