_STP_188-1956.pdf

Presented at a Meeting ofASTM Committee D9Philadelphia,Pa.,February,1956Reg.U.S.Pat.Off.AS1M Special Technical Publication No.188Price$1.75,to members$1.35Published by theAMERICAN SOCIETY FOR TESTING MATERIALS1916 Race St.,Philadelphia 3,Pa.SYMPOSIUM ONMINIMUM PROPERTYVALUES OF ELECTRICALINSULATING MATERIALS Copyright,1957,by theAMERICAN SOCIETY FOR TESTING MATERIALSPrinted in Baltimore,Md.November,1956 FOREWORDThough factor of safety is sometimes facetiously called to as factor ofignorance,there is some basis for referring to it as such.In the design ofelectrical equipment,for example,the safety factor is based on the minimumexpected values for electrical and physical properties of the component ma-terials,as well as an estimate of the extremes of conditions which theequipment may encounter under adverse circumstances.Results of mate-rials tests should enable a prediction as to the lowest expected value for aproperty of a material so that the designer may take this into account hiestablishing the safety factor.Unfortunately this information is notreadily obtainable from many present test methods.Though there has beensome progress in establishing procedures for predicting minimum values,there is still great need for further work.This symposium may be consideredas a report of progress,indicating a base line for additional investigationstoward establishing safety factor on a basis of knowledge rather than igno-rance.This Symposium on Minimum Property Values of Electrical InsulatingMaterials was held in Philadelphia,Pa.,on February 14,1956,as a specialfeature of the winter meeting of Committee D-9 on Electrical InsulatingMaterials of the American Society for Testing Materials.The program wasarranged by C.L.Craig,Sperry Gyroscope Co.,with the assistance andcooperation of Subcommittee XI on Significance of Tests of which Mr.Craig is Chairman,and Subcommittee XV on Technical Papers and Re-search,L.B.Schofield,Chairman.Mr.Craig presided,assisted by Mr.C.J.Photiadis(Westinghouse).iii NOTEThe Society is not responsible,as a body,for the statementsand opinions advanced in this publication.CONTENTSPageSignificance of Minimum Property Values of Electrical Insulating MaterialsC.L.Craig.1Extremal Nature of Dielectric Breakdown-Effect of Sample Size.H.S.Endicottand K.H.Weber 5Discussion 11Variability of Dielectric Breakdown in Sheet InsulationH.K.Graves 12The Establishment of Minimum Quality Levels of Sheet InsulationC.J.Photiadis.17Low-Temperature Resistance Test Methods for Non-Rigid InsulationC.L.Craig.33Discussion 38Statistical Treatment of Heat-Aging Studies on Varnished Glass ClothL.E.Sieffertand E.W.Alexander 39Discussion 42Quality Level Cannot Be Determined By GoNo-Go Gage Lester J.Timm 43Discussion 47 This page intentionally left blank SIGNIFICANCE OF MINIMUM PROPERTY VALUES OFELECTRICAL INSULATING MATERIALSBY C.L.CRAIG1HISTORYMinimum property values are test re-sults which indicate the lowest expectedresult within a designated accuracy for adefinite property of a specific material.Test methods to determine minimumproperty values are basic requirementsfor design purposes and for anticipatingservice performance.Many tests areavailable to control properties of provenmaterials,but few methods allow pre-diction of suitability of materials beforetrial.The papers of this symposium are in-tended primarily to present the back-ground of a few attempts to correlategeneral material test methods with ex-pected service and design requirements.They are planned to set a base line forfurther work to be sponsored by Subcom-mittee XI on Nomenclature,Signifi-cance,and.Statistics of ASTM Com-mittee D-9 on Electrical InsulatingMaterialsthese covering the generalfield,while future discussions will delvedeeper into specific properties or materialtypes.The cooperation and assistance ofthose having interest or data on anyphase of the subject is greatly desired.A few of the papers present substanti-1 Senior Materials Engineer,Sperry Gyro-scope Co.,Div.of Sperry Rand Corp.,GreatNeck,N.Y.ated methods for filling a small portionof the void hi performance-evaluatingtest methods,a lack that is commentedon below.In general,we are attemptingonly to point out the need,and to offersome suggested remedies for considera-tion and possible further work.It mustbe emphasized that very limited datahave been obtained on some of the testsintroduced,and confirming evaluation isrequired before these can be put to gen-eral use.Many of the presently available testmethods are empirical and are limited inthen*possibilities for correlation withexpected service.As detailed duplicationof all service variables is obviously im-possible,present methods generally haveattempted to grade the material by al-lowing relative rating hi some charac-teristic.Frequently,test results cannotbe translated into expected field per-formance;in fact,at tunes the specificcharacteristic measured is foreign tonormal application requirements.To beuseful hi design analysis,this grading or