DIFFUSED p-n JUNCTION SILICON RECTIFIERS 681 



tests, and load tests. The last tests are really the important tests; how- 

 ever, these require the dissipation of large quantities of power in the 

 load to test only a few devices. Therefore only a few units were tested 

 in this condition and the majority tested under other conditions. The 

 several units under load test have been operating for six months with 

 no noticeable change in their characteristics. These devices are the small 

 and medium size development units. The large rectifiers would require 

 about 10 kilowatts of dissipation each in a load to give them a fair load 

 test. 



The shelf tests at room temperature and at a temperature of i50°C 

 have been running for six months and have indicated that most of the 

 units remain practically constant. There have been some units that 

 improve on standing but there is no method of predicting which ones 

 will improve. Some units get worse on standing; however, most of these 

 can be predicted from the initial tests since these units usually have a 

 noisy reverse characteristic near the reverse breakdown voltage. The 

 units that change differ only in their reverse characteristic; the forward 

 characteristic changes are not detectable indicating that the contacts 

 are stable. The changes in the reverse characteristic are probably due 

 to the trapping of ions and vapors on the surface of the devices during 

 the packaging operation. Another source of these variations is due to 

 the non-hermeticity of the glass-to-metal seals allowing gases to diffuse 

 into the package where they may cause changes in the reverse charac- 

 teristic. These leaks have been found in many early units and new as- 

 semblies are being tried at present. 

 f The forward characteristic life test was considered a good test since 

 the device is subject to practically all the internal power dissipation 

 without reciuiring the relatively high load dissipation. It is tests of this 

 nature that allow one to rate the various size devices. The medium size 

 rectifiers that ran at 15 amperes in this test failed after three months 

 of testing; whereas no units running at 5 and 10 amperes have failed 

 during the six months since the tests have started although their re- 

 verse characteristics have changed slightly. It should be noted that 

 most of the change of reverse characteristic occurred during the first 

 test period of two weeks. These changes are probably due to the causes 

 mentioned in the above paragraph. 



Reverse characteristic tests have been running for several months on 

 a group of 10 small rectifiers which we feel have a better gas tight seal 

 than the other development units. The voltage has been adjusted on 

 these units such that they are pulsed into the breakdown region with a 



