Some aspects of the dielectric loss measurement problem

This paper points out some of the difficulties of making accurate power measurements at very low power factors. It deals with the lack of reference standards of known and constant power factor, especially for moderate sized samples and at high voltages. It also points out that a method of measuring an added loss as a check on a given test outfit´s accuracy is not a check at all, although it has been used as such in several instances. Reference is made to calorimetrical methods of calibrating a testing outfit and to special forms of resistors that have very low time constant and are suitable for this class of test work.