MEASUREMENT OF POTENTIAL DIFFERENCE 293 



until 50 is again registered. After this the dynamometer should 

 remain exactly at 50, while the secondary is turned through 

 180; if it does not so remain the process is repeated until it 

 does remain at 50 for all positions of the secondary. 



In measuring currents or e.m.fs. it is not necessary to split the 

 phase with great accuracy, although it is more convenient to do 

 so; but the utmost care must be taken in doing this when vector 

 diagrams are being constructed or when an accurate knowledge 

 of phase angles is required. 



Above all things, it is necessary that the supply voltage and 

 frequency remain perfectly steady. 



After the adjustment of the phase splitter, the first step in 

 using the instrument is to calibrate the electrodynamometer at 

 the reading corresponding to the standard potentiometer current. 

 To do this the battery is used as a source and the adjustment 

 made as with the ordinary potentiometer. The Tinsley instru- 

 ment has no separate standard-cell tap, so it is necessary to set 

 the slides at the voltage of the cell and then adjust the rheostat; 

 this process must be repeated whenever the standard potentiom- 

 eter current is checked. When the galvanometer is in balance 

 the reading of the dynamometer is taken; if the instrument is 

 not astatic, reversed readings must be taken and the two results 

 averaged. By means of the change-over switch, alternating is 

 substituted for direct current and the reading brought to the same 

 value and held there. Then the graduation in volts on the 

 potentiometer scale gives r.m.s. values of the P.D. First 

 the phase of the potentiometer current is roughly adjusted; then 

 the unknown P.D. is balanced as nearly as possible by the po- 

 tentiometer slides. The balance is then improved by shifting 

 the phase of the potentiometer current and still further im- 

 proved by resetting the slides; thus, by a process of double 

 adjustment, the vibration galvanometer which is used as the 

 detector is brought to rest. As the vibration galvanometer is 

 a tuned instrument which responds freely to currents of only 

 one frequency, the periodicity of the supply current must be kept 

 constant if the sensitivity of the potentiometer is to be main- 

 tained. With any potentiometer the deflection of the detector 

 is dependent on the difference of the two potential differences 

 which are being balanced. As the vibration galvanometer which 



