RELIABILITY OF RADIO TELEPHONE [CIRCUITS 



185 



-30 



-20 



6-A NOISE IN DECIBELS 



Fig. 2 — Percentage distribution of 6-A noise measurements made on London to 

 New York short-wave circuits during 1930 on (a) one 18-megacycle circuit and in 

 (&) each of four frequency ranges. 



type of distribution. The bend at the low noise end is apparently due 

 to noise transmitted from the distant terminal and that introduced 

 by the local receiver. These sources of noise are approximately con- 

 stant and, when the atmospheric noise is very low, become the limiting 

 factors. The bend at the high noise end of the curve is probably due 

 to the action of the automatic volume control in the receiver. This 

 control normally holds the speech volume approximately constant, 

 but when noise is exceedingly high the noise in itself reduces the re- 

 ceiver gain and depresses both the noise and signal output. Since at 

 such times speech volume cannot be accurately checked, the measure- 

 ment is no longer an accurate indication of noise-to-signal ratio and 

 the curve reaches a limiting value. Evidence confirming the inaccu- 

 racy of the 6-A noise readings at the high and low noise extremes will be 

 discussed later in connection with the observed distribution of high- 

 frequency signal intensity values. 



In Fig. 2 (b) are shown distribution curves for 6-A noise values 

 measured at New York during 1930 on several of the London to New 



