958 THE BELL SYSTEM TECHNICAL JOURNAL, SEPTEMBER 1952 



too 



80 



d 50 



^40 

 o 

 o 

 o 

 - 30 



to 20 

 O 



o to 



I 2 3 4 5 6 8 10 20 30 40 50 60 80 100 



BIAS FIELD IN VOLTS PER CENTIMETER 



Fig. 6 — Variation of noise with dc bias in single crystal filaments. 



on the noise. Some of these processes also affect the hfetime of carriers 

 in the bridge to a large extent. However, there seems to be no direct 

 and simple relation between the two effects, since treatments have been 

 foinid which change the noise by a large factor with almost no effect on 

 lifetime, and vice versa. 



Fig. 8 shows measurements of noise voltage on several dozen bridges 

 at a uniform bias of 10 volts per centimeter, all having sandblasted siu^- 

 faces, mostly of n-type but a few of p-type germanium, and with widely 

 different values of resistivity, produced by varying impurity concentra- 

 tions. There is considerable scatter in the results, but there is a fairly 

 obvious tendency for noise voltage to increase in proportion to resis- 

 tivity. Since Johnson noise also increases in proportion to resistivity in 

 a structure of fixed dimensions, the conclusion is that with constant bias 

 voltage the ratio of current induced noise to Johnson noise tends to be 



