CROSSTALK BETWEEN COAXIAL CONDUCTORS 



375 



turbing circuits are interchanged. The direct crosstalk component 

 would not exhibit this effect. 



The results of tests on 6000 and 12,000-foot lengths are given in 

 Fig. 7. Again, the trend is in the same direction as in Figs. 5 and 6 

 except that in this case {Fi + F//) is nearly equal to Fi and has an 

 appreciable influence when the two components are combined to give 

 far-end crosstalk in 12,000 feet. This indicates that the first term of 

 Fi in equation (6) is becoming more important as / is increased as 

 would be expected. 



105 



no 



115 



uj 120 



$ 



O 



S 125 



z 

 130 



O 



f^ 140 



150 



155 



40 50 60 



80 100 150 200 300 400 500 600 800 1000 



FREQUENCY IN KILOCYCLES PER SECOND 



Fig. 7 — Crosstalk components in 6000-foot and 12,000-foot lengtiis. 



It may be noted here that curve 7^/ in Fig. 7 differs considerably from 

 the AB and BA curves of Fi + Fnn + (Fi + F/f) in Fig. 6 although 

 all represent far-end crosstalk in 6000-foot sections. These differences 

 in magnitude must be due to differences in the construction of the two 

 cable sections. The difference between the curves varies from 3 to 8 

 db in the frequency range above 200 kilocyles. However, up to about 

 150 kilocycles the differences are not greater than 1 db. At the higher 

 frequencies such differences naturally will introduce difficulties in any 

 analysis since they superpose sizeable random effects on the major 

 component of crosstalk which is systematic. 



The curves in Fig. 8 present far-end crosstalk tests on 12,000 and 

 24,000-foot lengths. Here Fi and (Fi -f F/f) are of the same order of 

 magnitude and combine in such a way that the crosstalk in 24,000 

 feet is from 3 to 6 db higher than that measured in 12,000 feet. Com- 



