LOADED TELEGRAPH CABLES 391 



nautical mile in length, selected at intervals during manufacture, 

 the effect of strains and of superposed fields being estimated by means 

 of experiments on short lengths of cable. However, there are ordi- 

 narily small unavoidable variations in electrical characteristics from 

 point to point along the cable and it is not entirely certain that the 

 average inductance obtained from measurements on a fraction of the 

 core lengths entering into the cable structure will represent the 

 average inductance of the entire cable. The resistance of the laid 

 cable is likewise difftcult to estimate. This parameter comprises, in 

 addition to the copper resistance, the resistance of the return conductor 

 consisting of the armor wires and sea water in parallel, components 

 resulting from eddy current and hysteresis losses in the loading 

 material and other components of lesser importance, the nature of 

 which will be discussed later. The losses in the loading material 

 depend upon the average permeability obtained in the laid cable, 

 and their predetermination from factory measurements may be un- 

 certain for reasons that have been pointed out. As regards the sea 

 return resistance, rigorous methods of computation are available,'' 

 but there is some uncertainty regarding the conditions that should be 

 assumed as existing at the ocean bottom. 



Measurements on Laid Cables 



For the purpose of placing the design of loaded cables upon a 

 definite basis, it has appeared desirable to measure the parameters of 

 a number of cables of this type that have been laid, and to compare 

 the values so obtained with the estimates based on analytical methods 

 and upon factory measurements. In order to simplify the problem, 

 attention will be devoted mainly to determining the values of the 

 parameters corresponding to a very small value of current in the 

 cable conductor. Under these conditions the hysteresis component 

 of resistance is negligible and the inductance and eddy current re- 

 sistance can be considered constant at any frequency. This is entirely 

 consistent with the method employed in the design of loaded cables, 

 in which the attenuation constant is computed, first on the assumption 

 that the current is very small throughout the cable, and then corrected 

 for "head end losses" due to the effect of hysteresis losses which are 

 present under actual conditions of operation. 



The usual method of determining the parameters of a transmission 

 system consists in measuring the propagation constant, V, per unit 



^ J. R. Carson and T- J- Gilbert, Jour. Franklin Institute, Vol. 192, p. 705 1921- 

 Electrician, \'ol. 88, p. 499, 1922; B. S. T. J., Vol. 1, No. 1, p. 88. 



26 



» 



