614 



BELL SYSTEM TECHNICAL JOURNAL 



on a hot plate and allowed to return to room temperature gradually, then 

 cooled with an air stream and allowed to warm up gradually while the output 

 noise and the temperature of the waveguide were being recorded. This re- 

 vealed the temperature coefficient of — .055 db per degree centigrade. The 

 data (plotted in Fig. 6) show an r.m.s. deviation of ±.114 db from this 

 coefficient. 



^,e. ALONG LINE /^ 



Fig. 3 — Admittance diagram of microwave noise generator. 



The ambient temperature of the waveguide circuit had very little efifect 

 on the admittance of the gaseous discharge. 



As a check on variability with respect to time, two of these noise sources 

 were compared, one against the other, at tive-minute intervals for 65 min- 

 utes. During this time the waveguide temperature of source # 1 rose from 

 34° to 35.2° C and that of source # 2 rose from 33.7° to 36.1°. Each compari- 

 son was corrected, according to the coefficient of — .055 db per degree C 



