STUDIES IN RADIO BROADCAST TRANSMISSION 



157 



with a single frequency differing from them by a small amount and 

 detected may thereby be reduced to audible frequencies without 

 having their interrelations of phase, amplitude or difference frequency 

 composition, changed in any respect. For instance if the frequencies 

 expressed above are beaten with a local constant frequency, 



B cos {qt-\-\p) 



the resultant lower or difference frequencies will be 



+ ^^ cos [(p-^v-q)t + 4>-H 



-\-kBA sin [(p-q)t-^p] 



2~ cos [{p-v-q)t-<p-yp]. 



Each one of the three components has been changed in amplitude 

 by the same factor kB representing the efficiency of detection. Each 



Monitoring 

 Loud Speaker^ 



phone 

 cciver 



Fig. 11 — Diagram of system used for three-frequency tests 



one has been reduced in frequency by exactly the same amount ^ 



and each has had its instantaneous phase shifted by an angle — lA- 

 Relative to each other they remain unchanged. 



P 

 In our actual case the carrier frequency ^ was 610 kc. The modu- 



Ztt 



V Q 



lating frequency ^r- was 250 cycles and the beating frequency ^ 

 Ztt Zt 



was 608,375 kc. so that the resulting three audio frequencies were 



1,875 cycles, 1,625 cycles and 1,375 cycles. 



As indicated in Fig. 11 in order to make a record of these signals 



they are separated at the receiver by means of band filters. These 



filters and others similar in type for other modulating frequencies 



