616 



BEIJ. SYSTE\r TECHNICAL JOURNAL 



source ^2 displayed r.m.s. deviations of ±.088 db. These variations are 

 in fact comparable with the probable experimental error, and the proof that 

 they actually exist at all still remains to be demonstrated. 



Of thirty-two different lamps, including 10 different types of fluorescent 

 coatings such as used in the pink, red, gold, soft white, daylight, green, 

 white, 4500° white, black light and blue, thirty-one^ were all within ±0.25 

 db of each other as was also a germicidal lamp with no fluorescent coating. 

 Thus it appears that the source of the microwave noise energy lies chiefly 

 in the gaseous discharge rather than in the fluorescent coating. 



16.2 



16.0 



15.8 



'^15.6 

 |o 



^15.4 



O 



-* 15.2 



o 



15.0 

 14.8 



14.6 



28 30 32 34 3G 38 40 42 44 46 48 50 52 54 

 WAVEGUIDE TEMPERATURE IN DEGREES CENTIGRADE 



Fig. 6 — The microwave noise power depends slightly upon the temperature of the 

 waveguide circuit. 



If this noise is tied up with the electron temperature of the discharge, we 

 should expect the noise to be flat, or "white" noise. Corroborative evidence 

 of this was observed when the spectrum of the noise was examined over the 

 band from 3700 to 4500 mc at points 20 mc apart and no irregularities were 

 found. The nature of the experiment was such that frequency bands of ex- 

 cessive noise power would have been observed had they been present. 

 Further tests should indicate whether or not a gradual change in noise with 

 frequency exists. It appears, however, unlikely that such a slope exists at 

 4000 mc. 



Furthermore, since the level of the noise energy is so constant with respect 

 to time, reproducible from tube to tube, practically mdependent of the 

 current and only slightly affected by the ambient temperature, we might 

 expect that it is being controlled or limited by some invariant physical 

 property of the atoms and ions within the gaseous discharge. If this is the 

 case, an absolute measurement of the noise power might lead us to some 



* One of the 32 lamps llickcred erratically. At times its noise was ^ db higher than the 

 average. 



