Sphere-gap calibration

The following report contains an account of experiments carried out at East London College upon spark-over between spheres and the standardization of the sphere-gap as a voltmeter. The origin and objects of the research are described. A detailed account is given of the methods employed in measuring the high voltages, which include the following:? (i) Measurement on primary. (ii) Measurement on secondary tap. (iii) Measurement of current through a high-voltage air condenser. (iv) Use of capacitative potential divider. (v) High-voltage electrostatic voltmeter. (vi) Measurement of frequency and wave-form. The accuracy and convenience of these are discussed. Special apparatus constructed is described, including a high-voltage air condenser and gaps with spheres 62.5 mm and 20mm diam.respectively. A number of experiments on the effect of various conditions upon the spark-over voltage are described. The existing standard recommendations with respect to the proximity of neighbouring objects were confirmed, but it is shown that some of the precautions appear too stringent, whereas some factors particularly the size of earth plate, should be taken into account. It is shown that a strong source of ultra-violet light decreases the spark-over voltage, radiations of wavelength at least less than 3200 ?.U. being necessary. The decrease varies with the intensity of the radiation and does not reach a limit under normal circumstances. Since a weak source of ionization appears adequate for ensuring regularity of results, illumination with ultra-violet light is not recommended, but it is desirable that some degree of initial ionization should be ensured, such as that due to weak corona, although this ionization appears usually to be present without special procedure. The effect of the surface conditions was investigated and found to be very complicated; it is concluded, as in a previous investigation, that a clean polished surface, but not a mirror polish, is desirable. The effect of - humidity and different types of contaminated atmospheres was investigated. The effect of humidity alone was a small decrease in spark-over voltage. Smoke of different types produced different effects and varied also according to the humidity. A large decrease of spark-over voltage was produced only by a very thick cloud of smoke. The spark-over voltage was found to be irregular in an atmosphere the humidity of which was changing. The behaviour of spheres of brass, phosphor bronze and mild steel was studied and is compared. In the authors´ experiments phosphor bronze was found to be most satisfactory for a number of reasons, but definite consistent differences between the metals were not established. Calibrations are put forward for 20-mm diameter and 62.5-mm diameter spheres up to 80 kV (R.M.S.). The calibrations for 20-mm diameter spheres agree substasntially with those previously put forward as provisional standards by the E.R.A. based on experiments at the British Thomson-Houston C0., Ltd., They also do not show a very great variation from the recent figures adopted by the American Institute of Electrical Engineers. The calibrations for 62.5-mm diameter spheres agree with the revised American I.E.E. figures up to about 50 kV (R.M.S.), but are low compared with the latter for voltages above this figure. The accuracy of the calibrations is discussed and it is concluded that these are limited by the inherent variations of spark-over voltage. It is considered that a probable error up to about 0.5 per cent in a mean of 4 or 5 readings must be admitted for which it has not been found possible satisfactorily to account. Thus the sphere-gap is not considered to be a measuring instrument of the highest grade such as, for example, a standard voltmeter at lower voltages, although it is more reliable than many high-voltge voltmeters at present commonly used in commercial practice.