RELIABILITY OF RADIO TELEPHONE CIRCUITS 189 



ances,* so that if this were the only effect to be considered the 6-A 

 noise which is dependent upon the noise-to-signal ratio would increase 

 slowly- But first circuit and tube noise in the receiver are probably 

 the real limitation during times of disturbance. If this high-frequency 

 first circuit noise remains constant and the field intensity decreases the 

 ()-A noise will increase in opposite proportion. Therefore it ma\' be 

 assumed that the slopes of the corrected 6-A noise curves in the high 

 noise region will correspond to the slopes of the field intensity distri- 

 bution curves. In a conservative estimate it is reasonable to assume 

 that this is the case and that although the field intensity falls during 

 times of magnetic disturbance the high-frequency noise will not 

 decrease. 



In correcting the less important low noise ends of the 6-A noise 

 curves, use of the field intensity distribution is not so easily justified. 

 It may be reasoned, however, that here atmospheric noise is again 

 low compared to receiver noise but due in this case to a scarcity of 

 electrical storms within favorable transmission distance from the point 

 of reception. Then the corrected 6-A noise distribution at the low 

 noise end would also correspond in shape to the field intensity distribu- 

 tion. For these reasons it is assumed in the absence of better data 

 that the field intensity distribution may be used to correct for the 

 bends that occur at both ends of the 6-A noise distribution curves. 

 Fairly dependable field intensity data are available over a much wider 

 decibel range than is accurately covered by the 6-A noise measurement. 



In Fig. 5 is shown by the full line e-c-d-f a form of noise-to-signal 

 distribution which is conservatively representative of that experienced 

 on several short-wave radio telephone circuits as described above. 

 The horizontal decibel scale in this figure is arbitrarily referred to the 

 midpoint of the distribution curve. The broken line extension d-b 

 represents the decibel distribution of the lowest 15 per cent of the field 

 intensity values as experienced during the years 1930 and 1932. The 

 broken line extension a-c similarly represents the distribution of the 

 highest 30 per cent. The reason for using the transatlantic data is 

 that there are many more measurements available in the low field 

 region than there are for transmission over less disturbed paths. The 

 available data indicate that if suitable frequencies are used at all times 

 of the day the distribution of field intensities within the lowest 15 

 per cent and the highest 30 per cent has roughly the same average 

 slope during different years and on different circuits. 



' R. K. Potter, "High Frequency Atmospheric Noise," Proc. I.R.E., Vol. 19, 

 pp. 1731-1765, October, 1931. 



