222 THE BELL SYSTEM TECHNICAL JOURNAL, JAXUARY 1957 



similar t3^pe, but the available data applied only to frequencies up to 

 about 180 kc, whereas the Clarenville-Sydney Mines link was to operate 

 at frequencies up to 552 kc; extensive extrapolation was therefore in- 

 volved. As soon as the first production lengths of cable became available 

 in February, 1955, lajang trials were carried out off Gibraltar, and it was 

 found that there were serious changes of attenuation on la3dng, over and 

 above those directly attributable to temperature and pressure effects, 

 and that the assumed characteristics were inaccurate. Although the 

 attenuation in the factor}' tanks had been in reasonable agreement with 

 that of the earlier cable, there were changes on transfer to the ship's 

 tanks and again on lajdng, amounting in all to a reduction of about 

 1.5 per cent at 180 kc. This would have been comparatively miimportant 

 had the discrepanc}' been of 'cable shape', i.e., the same fraction of the 

 cable attenuation at all frequencies and therefore exactly compensated 

 bj^ a length adjustment of the repeater sections. As this was not so, and 

 as the cable-equalizing networks in the repeaters were settled by this 

 time, it was clear that precise information must be obtained in order that 

 suitable additional equalizers could be provided for insertion in the cable 

 on laying. There are a number of factors which can lead to small changes 

 of attenuation on laying, but most of these tend to increase the losses. 

 The primary reason for the observed changes appears to be contact 

 variations between the various elements of the inner and outer con- 

 ductors, i.e. the wire and three helical tapes forming the centre conduc- 

 tor and the six helical tapes forming the return conductor. These contact 

 resistances tend to change with handling, and as a result of a slight 

 degree of 'bird caging' when coiled, it seems that the attenuation de- 

 creases as the coiling radius increases, and vice versa. Also, the effect of 

 sea pressure is to consolidate the conductors and thereb}^ further reduce 

 the attenuation — an effect which appears to continue on a diminish- 

 ing basis for a long time after laying. 



To obtain reliable data for the Clarenville-Sydney Klines link, 10 

 nautical miles of cable with A-type armour was laid at about the mean 

 depth of the system (120 fathoms), off the Isle of Skye. The attenua- 

 tions, coiled and laid, are shown in Fig. 4, due allowance ha^dng been 

 made for temperature and pressure. The ordinates — attenuation 

 versus frequencj^ — are such that the value should be approximately con- 

 stant at high frequencies. 



In making a final determination of the cutting lengths for the re- 

 peater sections, it was assumed that the factory measurements of at- 

 tenuation would be reduced by 1.42 per cent at 552 kc on laying, that 

 the temperature coefficient of attenuation would be +0.16 per cent 



