Abstract
The underlying causes of the losses occurring in dielectrics when subjected to alternating stress have long been a subject for speculation. There have been a number of proposals of methods for linking such losses with the dielectric behavior under continuous or d‐c. electrical stress. To investigate the validity of the suggestions for correlating the d‐c. properties of dielectrics with their a‐c. losses, the continuous currents flowing must be measured as soon as possible after the d‐c. voltage is applied to or removed from the specimen. The development of the amplifier‐oscillograph permits a much improved and satisfactory technique. With the apparatus and method described here, observations could be begun within one or two thousandths of a second after the specimen had had the d‐c. voltage applied, or had been short‐circuited and they could be continued as long as desired. By this method, a connected series of observations was obtained with a single cycle of voltage application, removal, and short circuiting of the specimen, greatly expediting the work and eliminating inaccuracies arising from changes in the dielectric sample during the long periods of time heretofore required for such studies. The apparatus employed consisted principally of a quick acting, spring‐actuated switch for operating the specimen circuits rapidly, and a sensitive‐amplifier‐oscillograph for recording the dielectric currents. On operating, the quick acting switch shunted the geometric charge of the specimen to ground for a brief interval after applying voltage to the specimen, or short‐circuiting it, and then connected the amplifier‐oscillograph to the measuring circuit. The observations could be continued as long as desired with a D'Arsonval galvanometer of high sensitivity. The amplifier‐oscillograph attained a sensitivity of 1.6×10−8 ampere per mm deflection, which is about 300,000 times that of the oscillograph alone. The vacuum tube amplifier which accomplished this extension of range was in three stages, pure resistance coupled for the accurate reproduction of continuous currents. Many precautions were taken and tests made to ensure the reliability of this instrument and its connecting circuit apparatus.

This publication has 10 references indexed in Scilit: