490 BELL SYSTEM TECHNICAL JOURNAL 



transmitting a pilot frequency over the channel. For this purpose the 

 carrier frequency serves as well as, if not better than, any other fre- 

 quency since it is easily obtainable at the transmitter and is readily 

 utilized at the receiver. If a single-sideband transmitter is fully 

 loaded by two equal* side frequencies and a carrier of amplitude 10 db 

 below one of the side frequencies, the power in the carrier is only about 

 5 per cent of the power in the side frequencies. If the peak voltage in 

 the transmitter is kept the same, each side frequency could be 1.3 db 

 greater when no carrier is transmitted. Practically, it is found that 

 since distortion rather than peak voltage is the limiting factor, the 

 presence of the carrier 10 db down has no appreciable effect on the 

 permissible sideband amplitude. By using a very narrow filter at 

 the receiver to pass the carrier, the same carrier-to-noise ratio can be 

 obtained with the reduced carrier as is ordinarily obtained with a 

 common double-sideband receiver receiving a carrier of full strength. 

 After passing through this narrow filter the carrier may be used to 

 synchronize automatically a local carrier, or by amplifying to a 

 greater extent than the sideband and recombining with the sideband, 

 it may be used for direct demodulation of the sideband. When used 

 in the latter manner it will be called "reconditioned carrier." 



In 1928, after extensive tests of short-wave double-sideband trans- 

 mission had been conducted ^ and while the short-wave transatlantic 

 telephone channels between the United States and England were under 

 construction, some preliminary trials of a short-wave single-sideband 

 system were made under the direction of Mr. R. A. Heising between 

 Deal, New Jersey, and New Southgate, England, using a local carrier 

 supply at the receiver. The local carrier was produced by beating 

 the output of a variable-frequency tuned-circuit oscillator with that 

 of a crystal oscillator. It was necessary to adjust the oscillator 

 continuously in order to keep the oscillator frequency in the proper 

 relation to the incoming sideband. 



Notwithstanding the limitations of the equipment, encouraging 

 results were obtained and study of the problem was continued, al- 

 though along a slightly different line. Receivers were built which 

 were capable of separating the sidebands and carrier of an ordinary 

 double-sideband and carrier transmission in such a manner that single- 

 sideband and other types of reception could be simulated. The carrier 

 could be separately filtered and reconditioned so that even with 



' Reports of some of these tests were contained in the following articles: "Some 

 Measurements of Short Wave Transmission," R. A. Heising, J. C. Schelleng and 

 (i. C Southworth, Proc. /. R. El., October, 1926. "Transmission Characteristics of a 

 Short-Wave Telephone Circuit," R. K. I'otter, Proc. /. R. E., April, 1930. "The 

 Propagation of Short I'iadio Waves over the North Atlantic," C. R. Burrows, Proc. 

 J. R. E., September, 1931. 



