Abstract :
Considerable discrepancies between the results of different laboratories when measuring the dielectric loss of the same insulating oils has been reported in co-operative tests for C.I.G.R.É., Second Series, 1949¿50. These can arise either from contamination of the oil sample with impurities or from the instability of the oil. Errors due to the first cause have been rendered negligible by evolving a suitable technique for cleaning and handling the commercially available test cells. Instability of an oil, during the course of a measurement, is not so readily controlled, depending as it does on the composition of the oil. It is shown that, with an oil complying with B.S. 148: 1951, oxidation can cause the dielectric loss to fall to a small fraction of its initial value during the time required by some test cells to attain thermal equilibrium at 109° C. This factor can render meaningless exact values of the dielectric loss of an oil at high temperatures in the neighbourhood of 100° C.
