Graves, H. K.
Electrical Engineer, New York Naval Shipyard, Brooklyn, N. Y.
Pages: 5 Published: Jan 1957
In addition to average values obtained from a number of measurements on a sample of material, users of dielectrics have long been interested in minimum property values for design and reliability purposes. Tensile strength, elongation, thickness and dielectric breakdown might be mentioned as some of the characteristics where minimum values have been of interest. Dielectric breakdown has perhaps been one of the most important of these and numerous attempts at control have been made over the years. Many specifications have carried a minimum average dielectric breakdown or dielectric strength requirement with a somewhat lower minimum value specified for any one reading. This method of control appears to give little protection to the user while at the same time requiring the manufacturer to take a very great risk if the minimum single value is set at all close to the average. If 1000 breakdowns were made on a sheet of varnished cambric, one or more of the values obtained would probably be quite low, lower at least than any limit which might reasonably be set for an individual breakdown. If 10 or as often happens only 5 readings are taken, it is a matter of chance whether one of these very low readings appears and the material must be rejected on the basis of a specified minimum individual value. Other specifications such as those for magnet wire, for example, have for years run a long specimen such as 100 ft through a conducting bath with a relatively low voltage perhaps 100 v or less impressed to pick out pin holes, or at best extremely weak spots. Such specifications permit some pinholes per hundred feet and it is probably safe to assume that magnet wire will not be tested by passing the entire lot through the electrode. High voltage wire on the other hand has been tested 100 per cent for many years at a relatively high voltage. Defects are cut out or repaired.
Paper ID: STP46275S