THE ROTARY CONVERTER. 319 



If it is not possible to drive the converter mechanically, 

 it must be driven electrically from the direct-current side. 

 The experiment is then carried out as follows : 



Connect the direct-current side of the machine to a 

 source of direct current in series with a variable resistance. 

 Supply the field windings with direct current, also through 

 a variable resistance, and through an ammeter. Connect 

 a voltmeter to the alternating side of the armature. 



The connections will then be as shown in Fig. 158. 



Begin by giving the magnets their full excitation, and 

 start the machine, gradually cutting out the armature 

 resistance until full speed is attained. Read the excitation 

 and alternating and direct-current voltage. Slightly 

 decrease the field, and increase the resistance in the arma- 

 ture circuit, so as to obtain the same speed as before. 

 Eepeat the readings and take a succession of such obser- 

 vations with decreasing values of the excitation. 



When plotted to give the magnetisation curve, the 

 terminal voltage on the direct-current side will be higher 

 than the true voltage, which should be considered to be the 

 voltage transformed into alternating volts on the A.C. side, 

 by the amount lost in armature drop and armature reactions. 

 If the armature drop ( = current x armature resistance) 

 be subtracted from each D.C. ordinate of the curve, a curve 

 which is practically the true magnetisation curve of the 

 machine will be obtained. 



Fig. 159 shows the magnetisation curve or no-load 

 characteristic of a 2 kw. rotary converter obtained by 

 driving the machine by a separate motor and varying the 

 excitation. Curve I. shows the readings of a direct- 

 current voltmeter connected to the direct-current side, 

 and Curve II. shows similar readings obtained on the 

 alternating side. 



Curve III. shows the relation between the direct and 



alternating voltages. This ratio is seen to be constant 



over the whole range of readings taken, and to be 



alternating volts ar7K , . 



ft- . volts = This value is less than the 



theoretical value of '707, which would have been found if 

 the wave form of the alternating voltage generated had been 

 truly sinusoidal. The normal wave form in a rotary con- 

 verter is affected only by the shape of the poles, since the 



