158 BELL SYSTEM TECHNICAL JOURNAL 



were designed and made by the Bell Telephone Laboratories es- 

 pecially for this work. The band filters used for the purpose of 

 selecting the carrier and side-band frequencies had a cutoff of 40 

 Transmission Units 250 cycles from the mid-band frequency. 



These cutoffs together with the position in the frequency range 

 of the pass bands of the filters preclude any troubles from cross modu- 

 lation of the radio carrier and side bands during the beating down 

 process. The products of such cross modulation would be fre- 

 quencies which are multiples of 250 cycles and these cannot pass the 

 filters. On the other hand the beaten down frequencies will pass 

 practically intact, since as has been shown by the previously de- 

 scribed single frequency tests, each of the three frequencies received 

 although subjected to amplitude modulation by fading, represents 

 only a very narrow band of frequencies for which the filter pass bands 

 were of adequate width. 



As the modulating tone was carefully calibrated to 250 cycles and 

 the filters adjusted to transmit the frequencies specified, it was only 

 necessary to transmit the carrier while adjusting the receiving beating 

 oscillator. The following procedure for this adjustment was found 

 to be very successful. A local audio frequency oscillator was set to 

 the reduced carrier frequency of 1,625 cycles, and its output con- 

 nected to a telephone receiver. The audio beat note from the radio 

 signal and local beating oscillator was reproduced by a loud speaker 

 and its frequency adjusted to zero beat the 1,625-cycle tone from 

 the telephone receiver. 



When this adjustment had been completed the carrier was modu- 

 lated with the 250-cycle tone, and the side-band signals automatically 

 pass through their respective filters. 



The signals from the outputs of the filters were amplified, and 

 recorded separately by the three oscillograph elements. The sample 

 records shown in Fig. 12 are representative. 



Strips 1, 2 and 3 are taken from a long record obtained May 7, 

 1925, 3 :22 a.m. The upper trace is a record of the upper side-band 

 signal, the center trace the carrier, and the lower trace the lower 

 side-band. Strips 4, 5 and 6 are from a section of a similar type of 

 record made May 23, 1925, 1 :06 a.m., where the carrier was modu- 

 lated with a 500-cycle tone and different filters were used. In this 

 record the upper trace is the lower side-band and the lower trace the 

 upper side-band. 



It will be noticed that the timing interruption appears only in the 

 side-band signals, as the tone was interrupted before modulation 

 took place, and that the amplitude of the carrier signal is not affected 



