1082 THE BELL SYSTEM TECHNICAL JOURNAL, NOVEMBER 1952 



median value of noise that the latter is equi\'alent in noise effect to a 

 steady random noise of the same \^aliie. In the FAI receiver, the level of 

 speech is essentially not affected by the strength of RF signal and so a 

 measurement of the output noise is directly related to the speech-to- 

 noise ratio. The speech-to-noise ratios given here are computed from 

 noise measurements by assuming that speech of —14 vu level is applied 

 to the system at a point where one milliwatt of 1,000 cycles tone would 

 produce a 10-kc frequency deviation. The strength of the speech signal 

 at the receiver output is expressed in the same units as are used for 

 the noise. 



As might have been expected the median speech-to-noise ratios cor- 

 relate strongly with the amounts of r-f signal received at the various 

 locations. This correlation has been e^■aluated in order that the most 

 likely relationship between speech-to-noise ratio and received r-f signal 

 may be known for the different frequencies. These are shown in Fig. 12, 

 where each circle represents the median speech-to-noise value measured 

 at one test location plotted against the median r-f signal received at that 

 location. The solid lines have been drawn in to show the trend. The 

 bending at the top of the curve is inconsequential. It only represents 

 the limit imposed in the test setup by tube microphonic noise, vibrator 

 noise, etc. The curves show, for example, that in order to produce a 

 commercial grade of transmission, which requires a 12 db speech-to-noise 

 ratio, the median r-f signal must be 122.5 db below one watt at 150 mc. 



The data given in Fig. 12 pertain only to the suburban locations. 

 Measurements in Manhattan have not been included, even though they 

 indicate that larger signals are required, because the limit of system 

 coverage is to be found in the suburbs. The data on the solid curves of 

 Fig. 12 have been used to derive the curve of Fig. 2 which plots the value 

 of r-f signal required at the mobile receiver for a commercial grade of 

 transmission. The dotted curve of Fig. 2, which shows the median signal 

 required in locations where noise picked up by the antenna is less than 

 set noise, is based on the assumption of an 8 db noise figure for a practical 

 150-mc receiver, 11 db at 450 mc, and 12 db at 900 mc and higher. 



Measurements have been made of the effect of noise picked up by the 

 antenna at land receiver stations. These are expressed here in terms of 

 the carrier strength required for just-commercial grade of transmission 

 (12 db speech-to-noise ratio) as compared with the value required when 

 there is no antenna noise and only receiver noise is present. These com- 

 parison measurements were made by injecting a steady carrier into the 

 receiver with an antenna connected normally, and again with a dumjny 

 antenna connected. Although these tests were made mth a steady rather 



