136 ALTERNATING CURRENTS 



encountered by it; these resistances are due to friction, hysteresis, 

 and eddies.* 



A sudden change of load is thus seen to start oscillations, which 

 become superposed on the uniform rotation of the motor. There are, 

 however, other ways in which such oscillations may be started. Let, 

 for instance, the load remain quite constant, but let the speed of the 

 generator supplying the p.d. undergo a sudden increase. This corre- 

 sponds, in the vector diagram of Fig. 97, to a sudden advance of 0V 

 towards OE (i.e. to a decrease of 0), and it is evident, from what has 

 already been said, that this will start oscillations. 



Again, a sudden change in the exciting current of either generator 

 or motor will have a similar effect. 



It must be clearly understood that the oscillations under considera- 

 tion are very slow in comparison with the frequency of the alternating 

 current which drives the motor. The vector diagram of Fig. 97 has 

 to revolve a considerable number of times before a single to-and-fro 

 oscillation of OE relatively to OY is completed. Hence it is that such 

 oscillations are readily observed on the ammeter, whose pointer sways 

 to and fro in time with the oscillations. 



As already mentioned, oscillations started by any sudden disturb- 

 ance, such as those we have considered, will gradually subside, their 

 energy becoming dissipated by friction, hysteresis, and eddy currents. 

 If, however, before the oscillations have been damped out a fresh 

 disturbance arises, of such a nature as to reinforce the already exist- 

 ing oscillations, then their amplitude may be considerably increased. 

 With a rapidly and suddenly fluctuating load on the motor, or a 

 generator whose speed fluctuates regularly during each revolution, the 

 disturbances will be repeated at intervals, a fresh disturbance starting 

 new oscillations before those due to the previous disturbance have 

 been damped out. Now, in general, the interval between two dis- 

 turbances will not bear any definite relation to the natural period of 

 the oscillations ; as a result, the disturbances will act in such a way 

 that sometimes they re-inforce the existing oscillations, while at other 

 times they weaken them. An ammeter in the circuit will show this 

 effect very clearly ; its pointer sometimes swinging violently showing 

 that the disturbances are re-inforcing the natural oscillations while 

 at other times it remains nearly stationary. 



A particularly troublesome condition arises when the disturbances 

 have a definite period not differing greatly from the natural period of 



* The Motional, hysteresis, and eddy-current losses here referred to are merely the 

 additional losses brought about by the oscillations of the rotor, and must be carefully 

 distinguished from the ordinary losses of this nature which are due to the rotation at 

 constant angular speed. Any change in 9 causes a shifting of the flux across the pole- 

 face, so that as 6 oscillates the magnetic flux sways to and fro across the pole-pieces, 

 giving rise to additional hysteresis and eddies, over and above the normal losses occurring 

 when 8 has a constant value. 



