312 THE BELL SYSTEM TECHNICAL JOURNAL, JANUARY 1957 



The program had three specific purposes; (1) to detect immediately 

 any fault which might develop during laying; (2) to perinit the design 

 and execution of corrective system adjustments while en route so that 

 transmission performance of the completed link would fall within speci- 

 fied objectives; and (3) to gather data on system characteristics at 

 intermediate points for eventual use in fault location or in aging studies. 

 The need for parts (1) and (3) of the program is more or less self evident, 

 but part (2) merits some further discussion. 



In an ideal submarine cable system in an average environment, the 

 attenuation of the cable from one repeater to the next would be offset 

 exactly across the frequency band by the gain of the following repeater. 

 Such a result is never achieved in practice, as the temperature and pres- 

 sure environments (which affect cable attenuation) cannot be known 

 precisely in advance, and the cable structure itself cannot be manufac- 

 tured for mile after mile without variation in transmission characteristic. 

 Additionally, the mechanics of the laying process induce minor changes 

 in the physical structure of the cable which reflect in attenuation 

 changes. 



If such deviations from desired characteristic produced only differ- 

 ences from the specified system gain objective, compensation could be 

 readily applied at the ends of the submarine link. Unfortunatel}^, this 

 is onlj^ partl}^ the case. Their more important effect is the resulting 

 misalignment of operating levels of individual repeaters from design 

 objective. 



Misalignment magnitudes must be watched carefully, for at best 

 misalignment narrows the system latitude for seasonal temperature 

 changes and for aging, and at worst it can result in intolerable sj^stem 

 noise. If a repeater is preceded by too much cable the signal to noise 

 ratio at the repeater input will be less than desired because of thermal 

 noise. In the opposite case of too little cable, the signal level will be too 

 high and the resulting overloading in the repeater will also affect the 

 signal-noise adversely. Once present on the signal, the noise cannot be 

 removed, and so the cure for excessive misalignment must be applied 

 before the misalignment has developed. Adjustments at intermediate 

 points along the route must therefore be contemplated. 



Testing Program — The program which was evolved to meet the three 

 objectives outlined was meticulously reviewed and practiced before the 

 start of laying, and various forms were prepared for entering data and 

 plotting and evaluating results. This was essential to avoid wasting 

 effort or missing \'aluable data. The wisdom of this was fully apparent 

 to all involved after experience with the close time schedules and the 

 mental tensions which developed during the actual laying. 



