FLEXIBLE REPEATER DESIGN 85 



capacitor in the transatlantic system will not occur in less than 16 years 

 nor more than 600 years. 



There is, of course, the possibility of a catastrophic or early failure due 

 to mechanical or other defects not associated with normal deterioration 

 of the dielectric. Such potential failures are not always detected by the 

 commonly used short-time over-voltage test. Thus, for submarine cable 

 repeaters, all capacitors subjected to dc potentials in service are sub- 

 jected to at least 1| times the maximum operating voltage for a period 

 of four to six months before they are used in repeaters. Experience indi- 

 cates that this is adequate to detect potential early failures. The results 

 of this type of testing on submarine cable capacitors is an indication of 

 the care used in selecting materials and manufacturing the capacitors. 

 Only one failure has occurred in more than 3,000 capacitor-years of 

 testing. 



An important aspect of the control of quality of components is the 

 control of the raw materials used in their manufacture. For the trans- 

 atlantic project, this was accomplished by rigid specifications, thorough 

 inspection and testing, supplemented in some cases by a process of se- 

 lection. 



This can be illustrated by the procedure used for selecting the paper 

 used as the dielectric in capacitors. The Western Electric Company 

 normally inspects many lots of capacitor paper during each year. Those 

 lots which were outstanding in their ability to stand up under a highly 

 accelerated voltage test were selected from this regular inspection proc- 

 ess. These selected lots were then subjected to a somewhat less highly 

 accelerated life test. Paper which met the performance requirements of 

 this test was slit into the proper widths for use in capacitors. Sample 

 capacitors were then prepared with this paper and so selected that they 

 represented a uniform sampling of the lot at the rate of one sample for 

 approximately each three pounds of paper. These samples were impreg- 

 nated with the same lot of oil to be used in the final product. Satisfactory 

 completion of accelerated life and other tests on these samples consti- 

 tuted final qualification of the paper for production of capacitors. Rela- 

 tively few raw materials were adaptable to such tests or required such 

 detailed and exhaustive inspection as capacitor paper. But the attitude 

 in all cases was that the material be qualified not only as to its primary 

 constituents or characteristics but also as to its uniformitj^ and freedom 

 from unwanted properties. 



To a considerable extent, stability of components is assured by the 

 practice of using only those types of structures which have long records 

 of satisfactory field performance. However, in some cases, a product far 



