DESIGN OF SYSTEM — NORTH ATLANTIC LINK 35 



can be utilized then depends on repeater design considerations, includ- 

 ing gain-bandwidth limitations, signal power capacity and signal-to- 

 noise requirements. 



The early studies of this transatlantic system were based on "scaling 

 up" the Havana-Key West system/ In these studies, consideration was 

 given to extending the band upward as far as possible by using com- 

 pandors^ on the top channels, thus lightening the signal-to-noise re- 

 quirements on these channels by some 15 db provided they are re- 

 stricted to message telephone service. 



As the repeater design was worked out in detail, however, it became 

 evident that a rather sharp upper frequency limit existed. This resulted 

 from the parasitic capacitances imposed by the size and shape of the 

 flexible repeater, the degree of precision required in matching repeater 

 gain to cable loss in such a long system, and the feedback requirements 

 as related to the requirement of at least 20 years' life. 



These limitations resulted in the decision to develop a system with 36 

 channels of 4-kc carrier spacing, utilizing the frequency band from 20 to 

 164 kilocycles per second. 



Signal-to-Noise Design 



Scaling-up of Key West-Havana System 



The length of the North Atlantic cable was to be about 16 times that 

 of the Havana-Key West system. The number of channels was to be 

 increased as much as practicable. The length increase entailed an in- 

 creased power voltage to ground on the end repeaters. Increase in length 

 and in number of channels entailed increased precision in control of 

 variations in cable and repeaters. Work on these and other aspects was 

 carried on concurrently, to determine the basic parameters of the ex- 

 tended system. 



Increasing cable size decreases both the attenuation and the dc re- 

 sistance, and in turn the voltage to ground on the end repeaters. It was 

 soon decided that the largest cable size which could be safely adopted 

 was the one used in the Bahamas tests. This has a center-conductor sea- 

 bottom resistance of about 2.38 ohms per nautical mile. It consumes 

 about 28 per cent of the total potential drop in cable plus repeaters. 



Number of Repeaters 



As indicated earlier, the factor which emerged as controlling the 

 number of repeaters was the dc voltage to ground on the end repeaters. 

 Considerations which entered into this were: voltage which blocking 

 capacitors could safely withstand over a life of at least 20 years; volt- 



