AN ELECTRICAL FREQUENCY ANALYZER 321 



Inn-n iihtaituti on a small niachinc direct-driNi-n li\' a ,'i-li- P-i <)0-ryrlc 

 siiijilo-phaso motor. Dat.i of importance rciatini; to the generator 

 tested are as follows: • 



Capacity of (ienerator '.i k\v. 



Number of Poles 2 



Speetl . . 172")- ISIM) r. p. m. 



Voltage r2o 



Field Shunt-connected 



Diameter of Commutator 2.75 in. 



Number of Commutator Bars 38 



Number of Armature Slots 19 



Size of Brush ^8 in. square 



Yoke Ring type 



Records obtained from this machine when operating under no- 

 load and half-load conditions are shown in Figs. 16 and 17, respec- 

 tiveK-. The corresponding speeds are approximately 1800 and 

 1750 r. p. m. In order to show what frequencies the machine gi\'es 

 out over the entire range 20 to 5000 cycles each figure is made up of 

 two parts: a portion of a 20-1250 record and a complete record over 

 the range 80-5000 cycles. On each figure is drawn the circuit con- 

 necting the d-c. generator to the analyzer. It will be noted that 

 a large condenser is inserted to prevent the passage of hea\\' direct 

 current through the analyzer. 



The consideration of these records leads to the conclusion that 

 there are at least three independent major causes of alternating 

 voltage operating in this d-c. machine. The fundamental frequencies 

 due to these causes are 30, 60 and 570 cycles. It will be noted that 

 the 30-cycle peak occurs only on the no-load record under which 

 condition the average speed is practically 30 revolutions per second. 

 •Sixty cycles and a series of its harmonic overtones are seen to be 

 present under both conditions of load. Under load the 60 cycles is 

 augmentetl whereas its harmonics are reduced. No harmonic o\'er- 

 tones of 30 cycles except such as might cf)incide with the harmonics 

 of 60 cycles are found in either case. This indicates the existence of 

 independent causes of the 30 and 60-cycle frequencies, that the 30- 

 cycle cause protluces an almost sinusoidal \oltage, and that the 60- 

 cycle cause unfler no load produces an irregidar wave which becomes 

 smoother as the machine is loaded. 



The no-load record. Fig. 16, show.s 570 cycles with no harnu)nics 

 while the load record. Fig. 17, shows 570 cycles with a complete series 

 of harmonics. This indicates that at no load the cause of 570 cycles 



