334 BELL SYSTEM TECHNICAL JOURNAL 



a single antenna. The improvement in signal-to-noise ratio caused 

 by the use of 16 antennas should average 12 db, but instantaneous 

 improvements might vary from large negative values, if the equipment 

 were not kept properly adjusted, to values of 25 or 30 db, which might 

 be expected when the noise came from the direction of a null in the 

 musa antenna directive diagram. 



In the operation of radio telephone circuits there is a minimum 

 signal-to-noise ratio below which commercial service cannot be given. 

 As the signal-to-noise ratio is increased a value is reached where 

 further increases give little benefit. The range between these two 

 values is about 25 db. Transmitters and receivers generally are 

 designed so that their maximum signal-to-set-noise ratio is somewhat 

 greater than the maximum beneficial circuit value in order that set- 

 noise shall not degrade the circuit. The maximum signal-to-noise 

 ratio obtainable with a musa receiver and a single-antenna receiver 

 should be approximately the same. 



The 12 db average improvement which the musa receivers should 

 give should make it possible to obtain, on the average, commercial 

 circuits with signal field strengths 12 db less than those usable with a 

 single-antenna receiver. This in turn will decrease the amount of 

 time in which commercial service cannot be given. ^ On the other 

 hand the musa receiver should produce its maximum signal-to-noise 

 ratio for field strengths 12 db lower than a single-antenna receiver and 

 at field strengths 12 db higher the musa receiver would show no 

 improvement over the single-antenna receiver. The net improvement 

 in signal-to-noise ratio therefore should be expected to average from 

 12 db at the lowest usable signal-to-noise ratios to db at fields 25 

 or 30 db higher, with fairly wide variations with time from the average. 



The results of a comparison between the musa system and a single 

 sideband receiver operating from one of the same antennas confirm 

 the theoretical expectations to a fair degree. The fraction of the time 

 that given improvements in decibels are obtained follows approxi- 

 mately a normal probability curve. 



The reduction in selective fading effected by the musa receivers is 

 difficult to state numerically. Most of the objectionable selective 

 fading is removed. There are times when waves that have traveled 

 over distinctly different paths arrive at so nearly the same angle that 

 they cannot be resolved. Fortunately these times are fairly rare. 

 When waves of closely adjacent angle are present the monitoring 

 system does not give a true indication of reception angles, as can be 



9 R. K. Potter and A. C. Peterson, Jr., "Reliability of Short-Wave Radio Tele- 

 phone Circuits," Bell Sys. Tech. Jour., Vol. XV, pp. 181-196, April 1936. 



