OPEN-WIRE CROSSTALK 199 



A very long segment has another effect on near-end crosstalk not 

 indicated by the above discussion. If there were no deviations in 

 any of the segments, the near-end crosstalk would be the vector sum 

 of a number of current elements of various magnitudes and phase 

 angles and the sum would be small due to a proper choice of these 

 magnitudes and angles in designing the transpositions. If a segment 

 deviates from its normal length, the magnitude of the crosstalk due to 

 the segment changes and the phase angle also changes. The phase 

 angles of the crosstalk values due to succeeding segments are also 

 changed since they must be propagated through the segment in 

 question. For ordinary deviations in segment lengths these effects 

 on the phase angles may be neglected. 



Since transverse crosstalk is independent of transpositions occurring 

 in both circuits at the same point, it would appear from the above 

 discussion that the location of such transpositions need not be accurate. 

 This is not ordinarily a question of practical importance. If some 

 circuit combinations have both circuits transposed at a certain trans- 

 position pole there will usually be other combinations which have 

 relative transpositions at this pole. The transposition pole is of 

 importance, therefore, in connection with the latter combinations and 

 the same accuracy of location is required for all transposition poles. 

 A question of practical importance, however, is whether the above 

 rules for locating transposition poles properly limit the interaction 

 crosstalk. This is affected by transpositions in both circuits at the 

 same pole as well as by relative transpositions. In the following 

 discussion of this matter it is concluded that the effect of transposition 

 pole spacing irregularities on interaction crosstalk may be ignored at 

 frequencies now used for carrier operation. 



The effect of deviations in segment length on interaction crosstalk 

 is indicated by Fig. 18. This figure indicates a short part of a parallel 

 between two long circuits a and b. A representative tertiary circuit 

 c is also shown. The transposition arrangements are like those of 

 Fig. 9B. In connection with the latter figure it was shown that the 

 interaction crosstalk would be very small if all segments had the same 

 length d. On Fig. 18, D is used to indicate the normal segment length 

 and the deviation of two segments from D is indicated by d. Since 

 the length A C equals the length CF, these deviations have no effect 

 on the transverse crosstalk which is controlled by the transposition at 

 C. The deviations affect the interaction crosstalk between the length 

 CF and length A C. 



The circuit a has near-end crosstalk coupling with circuit c in the 

 length CF. This effect is normally practically suppressed by the 



