Abridgment of some problems in dielectric loss measurements

The measurement of dielectric loss in high-voltage cables presents many problems which are not simple, and their solution is obtained only after considerable effort and research. This paper presents some of the problems that the authors have encountered in this type of work, together with the methods used for their solution. The bridge method, which is used for the measurement of dielectric loss, is described in some detail. Details of the auxiliary equipment, such as galvanometer, amplifier, air condenser, shielding and special methods of bringing the shielding to the proper potential are also given. If accurate results are to be had in dielectric-loss measurements, the entire bridge, including cable, air condenser, and leads must be properly shielded, and the capacitance between leads and shielding should be made small. It is entirely insufficient to shield the bridge and ground the shielding directly. At times, shielding cables may increase the measured losses. In measuring the capacitance of the air condenser, all stray capacitance to its guard rings and to surroundings must be eliminated. The voltage across the voltmeter coil of a high-voltage transformer is frequently taken as a measure of the voltage across its secondary. In some methods of measurement, this voltmeter-coil circuit is an integral part of the measuring circuits. Methods of measuring the ratio of primary to secondary voltage and the phase angle between them are described, and in the complete paper considerable quantitative data are given for typical testing transformers under various conditions of load. The accuracy of the bridge as a power-measuring device may be determined by inserting known resistances into its arms and comparing computed results with observed results. Due to unsuspected leakage and stray-capacitance currents, however, the introduction of such resistances may cause large errors. Methods of minimizing such currents are given, as are analyses of proper and improper methods of - aking such measurements and observed data and results. Harmonics probably do not introduce error in this type of measurement, but at times they are troublesome in that they produce anomalous effects in the measuring apparatus.