206 BELL SYSTEM TECHNICAL JOURNAL 



and all wires could be kept under the same tension, the effect of drop 

 brackets on crosstalk would be consistent and could, theoretically, be 

 made negligible by a suitable transposition design. 



There is, however, an accidental crosstalk effect. This effect is 

 partly due to the fact that it is more difficult to avoid deviations from 

 normal sag in the spans adjacent to drop brackets than in normal spans. 

 The main effect, however, is thought to be due to inequalities in the 

 lengths of the spans adjacent to drop brackets. 



The crosstalk in such a span is very nearly proportional to the 

 length of the span times a constant or "equivalent crosstalk coeffi- 

 cient." The usual crosstalk coefficient can not be used because the 

 wires are not parallel. 



Fig. 23-A indicates two long circuits, one circuit being transposed 

 on drop brackets at the first and third quarter points of the short 

 length D. The lengths of the spans adjacent to the drop bracket 

 transpositions are indicated by di to d^. The equivalent far-end 

 crosstalk coefficient for the span preceding a transposition bracket is 

 Fi and that for the span following the bracket is F2. (Fi and F2 are 

 usually quite different.) The total far-end crosstalk (output-to- 

 output) due to the four spans is (very nearly) : 



K{Fidi - F^di - Fidi + Fidi), 



where K is the frequency in kilocycles. 



If the four spans were equal the crosstalk would be zero (very 

 nearly). The actual value of the crosstalk is a matter of chance 

 since the deviations of the four spans from the normal length are a 

 matter of chance. These deviations cause a chance increase in the 

 near-end crosstalk as well as in the far-end crosstalk. 



This effect has been studied experimentally by using transposition 

 designs which suppressed the consistent effect. The pole spacing 

 effect was minimized by using very accurate spacing. The wire sag 

 effect was allowed for by comparing similar pair combinations trans- 

 posed alike except that dead-ended point transpositions were compared 

 with drop bracket transposition. Due to the great number of trans- 

 positions necessary at carrier frequencies it was found that the acci- 

 dental drop bracket effect was important at these frequencies. In 

 recent years, point-type transpositions have been extensively used on 

 lines transposed for long-haul carrier systems. 



When, for economic reasons, a transposition system is designed for 

 use with drop bracket transpositions, the consistent crosstalk effect 

 must be considered in the transposition design. The equivalent 

 crosstalk per mile for a span adjacent to a drop bracket must be 



