322 BELL SYSTEM TECHNICAL JOURNAL 



feeds a relatively smooth wave to the line while under load this 

 cause feeds an irregular wave to the line. The fact that 1140 cycles 

 is about as strong as the fundamental and that its harmonics are 

 stronger than alternate ones which are overtones of 570 only, sug- 

 gests the likelihood of a fourth cause having a frequency of 1140 

 cycles. .Small irregularities at frequencies other than those already 

 mentioned occur in the record. These are more prominent under 

 load than at no load and indicate the presence of small, more or less 

 irregular pulses, which increase with load. All of the above fre- 

 quencies may be accounted for by a consideration of the construction 

 and operating condition of the machine. 



The generator was driven by a single-phase, 4-pok', ()0-c>ck' motor 

 which may give rise to torque fluctuations once per revolution, or 

 30 times per second. Under no load this may produce considerable 

 corresponding fluctuations in speed while under load conditions the 

 generator acts as a damper, eliminating these oscillations. 



The 60-cycle peak may be due to any one or some combination 

 of a number of causes, e. g., eccentricity of generator armature, non- 

 uniform winding, non-uniforni thickness of mica separators in com- 

 mutator, high mica between one or more pairs of segments, etc. 

 The records show that for this particualr machine in its present con- 

 dition (new) at normal speed the 60-c\cle voltage de\eloped increases 

 considerably with load indicating strongly that the cause is largely 

 influenced by an I R drop somewhere in the machine. The most 

 likely causes therefore appear to be commutator eccentricity, irri'gular 

 spacing of the segments, or high mica. 



The peak at 570 cycles may be accoimted for by c\clic \ariation 

 of flux entering the armature core as the teeth pass the pole faces. 

 At no load the speed is approximatcK- ISOO r. j). ni. The number 

 of teeth being 19, it is ob\ious that there will lie .')7() lliutu.itions of 

 air-gap reluctance per second. Under no-load conditions the record 

 shows a comparati\-ely pure wave form for this cause. This is to 

 be expected because of the comparatixeK- uniform distribution of 

 flux under the pole faces at no load. .\s the machine is loaded, 

 however, the field is distorted and shifted gi\ing rise to an irregular 

 wa\'c form of voltage which is responsible for at least a part of the 

 large harmonic content shown b\- the load record. 



The presence of 1140-cycU' pi'ak which is present only uii(kr the 

 load condition ma>' be due to the cyclic \-ariation of xoltage produced 

 by the commutator bars leaxing the brushes. Inasnuii-h as the speed 

 is roughly about 29 re\olulions |)er second the frequency with which 

 bars leave brushes is about 1100 cycles. This frequency is present 



