POWER MEASUREMENT 



325 



Source^ of Error. 4 When measuring small amounts of 

 power the resistance R must be large. This produces two re- 

 sults; the condensers formed by quadrant 2 and one-half of the 

 needle must be charged through this resistance, while that formed 

 by the quadrant 1 and the other half of the needle is charged 

 directly from the source. Consequently the potential of the 

 condenser formed by the needle and quadrant 2 is a trifle lower 

 than it should be, the time-phase relation of the potentials on the 

 two condensers is not quite correct and the result is that with 

 the main circuit open there will be a small deflection. A high 

 resistance also alters the power factor of the circuit which in 

 testing insulating materials may be very low. The current must 

 be relatively high, and the correction term in (23) becomes large. 



Fio. 190. Illustrating Rayner guard -ring electrode. 



The net error is practically proportional to R. When in- 

 vestigating dielectric losses a correction can be made with 

 sufficient accuracy by observing the deflection with two known 

 values of R and extrapolating for the watts expended when 

 R = 0. 



If the needle is not tapped into the transformer but is en- 

 ergized from a potential divider, an error may be introduced be- 

 cause of the alteration of the potential of the point b, Fig. 187, 

 due to the charging current necessary for the needle. Errors 

 may also arise from capacity effects in very high resistances of 

 the ordinary construction. In electrostatic wattmeters for use 

 at high voltage it is essential that there be ample separation be- 

 tween the quadrants and the needle so that brush discharges 

 will be avoided. 



