680 THE BELL SYSTEM TECHNICAL JOURNAL, MAY 1956 



base of the package. This temperature difference depends mostly on 

 the material of the base and its geometry. In the devices presented 

 this drop is not more than 15°C at the maximum rated current. Thus 

 the largest drop in temperature occurs between the cooling fin and the 

 ambient which means that the design of the cooling fin is the controlling 

 factor in the operating junction temperature of the rectifier. 



6.3 It is possible to use the devices without an attached cooling fin. 

 In this case, the maximum current is limited essentially by the size of 

 the package. The small rectifier package is designed for 3^ watt dissipa- 

 tion and therefore the maximum current that should be rectified is about 

 500 milliamperes. The medium size unit will comfortably rectify 1 am- 

 pere without any additional cooling and the large rectifier unit will 

 conduct 3 amperes under the same conditions. 



7.0 RELIABILITY AND LIFE MEASUREMENTS 



7.1 One of the desired properties of any device is that it should op- 

 erate satisfactorily at its rating for a long period of time. The above 

 general statement contains many implications which should be made 

 specific for the devices under consideration in this paper. By stating 

 that these devices should operate satisfactorily we mean that they 

 should not age during operation; that is, the forward and reverse char- 

 acteristics at any temperature should not change with time. The state- 

 ment implies that a rating has been established for the units. Further- 

 more, a "long period of time" has to be defined. There are applications 

 where a few hours is considered a long time as in some military appli- 

 cations. However, in most Bell System applications, a long period of 

 time may be 20 years or approximately 200,000 hours. Clearly, in the 

 short time since these rectifiers have been developed, it is impossible 

 to make a fair statement as to their reliability and their life expectancy. , 

 However, it is possible to present some results of some early experi- | 

 ments and describe where and how the units have lived and died. It is ] 

 this information that we will present in this section. It is a common ex- I 

 perience that during the early development of any new component, i 

 there are many units that do not satisfy all the requirements of the de- | 

 sired end product. These units will generally deteriorate very rapidly | 

 on life testing due to some electrical or mechanical instability. The 

 units used for life testing have been screened to remove the above men- 

 tioned unstable devices. 



7.2 The life tests consist of four types; shelf tests at room tempera- 

 ture and at 150°C, forward characteristic tests, reverse characteristic 



I 



