210 



BELL SYSTEM TECHNICAL JOURNAL 





II it 



500 1000 



FREQUENCY IN CYCLES PER SECOND 



5000 



Fig. 10 — The frequency distribution of contact noise in a standard carbon trans- 

 mitter for normal operating conditions. The ordinates give the mean square contact 

 noise voltage in a twenty-cycle band the midpoints of which are given by the abscissae. 



resistance carbon grid leak. It is seen that the noise has precisely the 

 same frequency distribution in both the carbon transmitter and in the 

 grid leak, which further supports our belief that the noise mechanism is 

 the same in each case. 



Otto ^ reported similar measurements of noise as a function of 

 frequency in carbon transmitters, single contacts of carbon, carbon grid 

 leaks and copper oxide resistances. Whereas we find an almost exact 

 inverse relationship between noise and frequency for all types of 

 elements tested he shows curves with negative slopes ranging from 1.0 

 to 1.4. Meyer and Thiede ^ in measurements on thin carbon films 

 obtained negative slopes having values between 1.0 and 2.0. 



Contact Noise as a Function of Temperature and 

 Surrounding Medium 



For the complete elucidation of contact noise the knowledge of its 

 dependence upon temperature is important. However, the difficulties 

 involved in such a measurement are so great that we have been unable, 



