1342 THE BELL SYSTEM TECHNICAL JOURNAL, NOVEMBER 1956 



20 



cr 18 



O 



^ 16 



u. 



o 



(0 14 

 Q 



Z 



^ 12 



O 



f 10 



UJ 



o 



z 



I- 

 < 



a 

 o 



Fig. 47 — Life test results. The open bars indicate tubes that have failed; the 

 solid bars tubes that were operating as of May 1, 1956. These tubes were operated 

 with cathode temperatures between 720° and 760°C. 



Figs. 23(c) and 23(e). The latter curves are repeated as Figs, 45 and 

 46 with the experimental points calculated from Si and S2 shown. It is 

 seen that the results of the two types of measurements compare remark- 

 ably well considering that the calculations of c and kp both require the 

 subtraction of nearly equal quantities. Thus we may conclude that our 

 method of considering the intermodulation is substantially correct and 

 that we can obtain compression and AM-to-PM conversion from an 

 intermodulation measurement . 



V. LIFE TESTS 



We feel that sufficient data have been accumulated to indicate that 

 tube life in excess of 10,000 hours can be expected. Fig. 47 summarizes 

 our life test experience. All tube failures were caused by cathode failure 

 and these were evidently the result of exhaustion of coating. End of life 

 for these tubes comes comparatively suddenly i.e., in a few hundred 

 hours after the cathode current begins to drop. At this time the emission 

 becomes non-uniform over the cathode surface with consequent beam 

 defocusing and helix interception. This in turn causes gas to be released 

 into the tube which then accelerates the cathode failure through cathode 

 poisoning. The rf performance remained good over the tube life — the 

 gain and output power actually increasing slightly near the end of life 

 as the beam started to defocus. 



