A Method of Measuring Energy dissipated in Condensers, 39 



acquainted, and cannot say whether it is something in the 

 method of manufacture or the nature of the dielectric which 

 makes the dissipation of energy so large ; we presume, how- 

 ever, that it is the latter. There were six condensers joined 

 together, each being a solid slab of about 11 centim. x 15 cen- 

 tim. x 1*5 centim., thus having a volume of about 250 cub. 

 centim., and a capacity of one third of a microfarad. 



The six slabs were placed on a table, joined together in 

 parallel, and in series with the resonance-coil (which was at a 

 distance from them and from the measuring instruments), 

 loosely covered with a woollen cloth, and coil and condenser 

 subjected to an alternating electromotive force of about 

 50 volts, and a frequency of 120. No effort was made to 

 secure the maximum degree of resonance, and the voltage on 

 the condensers was found to be about 900. In a short time 

 the temperature of the condensers had risen to 30°, as indi- 

 cated by a thermometer inserted between two of them, and 

 the first set of readings was taken. The loss of energy in 

 the condensers was greater than it had been at lower tem- 

 peratures, and continued to increase as the temperature rose. 



At the same time, owing to this increase in the equivalent 

 resistance of the condenser, the resonance ratio decreased and 

 the current and voltage on the condenser decreased. The 

 loss at 36° C. is 50 per cent, greater than at 30° C, and is 

 approaching a maximum. At 39° it is 9*5 per cent., and the 

 fourth reading at sensibly the same temperature (but which 

 doubtless was a little higher, at least in some of the conden- 

 sers) showed a slightly less loss. At 47° C. the loss had 

 decreased to 8*0 per cent., and at 49°'5 C. to 6*5 per cent., 

 only two-thirds its maximum value. The condensers were not 

 all at the same temperature, and the indicated temperatures are 

 therefore not exact. But they show unmistakably a maximum 

 value of the condenser loss, or energy converted into heat, at 

 about 39° C, and beyond that a very considerable diminution. 

 No further readings were taken until the condensers had 

 risen several degrees, when it was suddenly noticed that one 

 pair was hotter than the others and getting soft. The ther- 

 mometer in a cooler pair registered 59°, but the warmest pair 

 was considerably higher. The loss was astonishingly large, 

 but the condenser had not broken down. Moreover, the 

 " leakage current " had not greatly increased, for while 839 

 volts gave 1'50 amperes at 49°' 5, 433 volts gave '80 ampere 

 at 59° C. To be in exact proportion to the voltage the current 

 should have been *774 instead of *80 at the higher tempera- 

 ture, a comparatively small difference. 



To find so large a loss in commercial condensers of good 



