208 



BELL SYSTEM TECHNICAL JOURNAL 



pairs in a short length of Hne which, with point transpositions, would 

 have very low crosstalk. At points B and E both circuits are trans- 

 posed alike. With point transpositions the near-end crosstalk in the 

 two spans adjacent to one of these pairs of transpositions would be 

 NK2d, where d is the span length, N the near-end crosstalk coefhcient 

 and K the frequency in kilocycles. For drop bracket transposition 

 the crosstalk would be K{Ni + iVz)^ or a change of K{Ni -f A^2 

 - 2N)d. 



The transpositions are so arranged that the crosstalk in the two 

 spans at B tends to add to that in the two spans at E. With drop 

 brackets at B and E the major crosstalk in this length of line would 

 be twice the above change since the crosstalk with point transpositions 

 is very small. 



ALSO TRANSPOSITIONS IN BOTH 

 PAIRS AT THIS POINT AND 

 /"each Va f^lLE THEREAFTER 



J^MILE 



5 10 15 20 25 30 35 40 45 50 



FREQUENCY IN KILOCYCLES PER SECOND 



Fig. 24 — Near-end crosstalk with and without drop brackets. 



If the arrangement of Fig. 23-B is reiterated in a long line, the total 

 increase in the crosstalk due to drop brackets at such points as B and 

 E may be marked. It may be noted that the crosstalk in the two 

 spans at A tends to cancel the crosstalk in the two spans at C and 

 likewise there is cancellation at D and F. Drop brackets may, 

 therefore, be used at points ^, C, D and F without a consistent increase 

 in crosstalk. Arrangements like those at B and E of Fig. 23B should 

 be avoided in transposition design involving drop brackets. 



The change in the crosstalk due to drop brackets is not necessarily 

 an increase. Fig. 24 shows an arrangement of transpositions in an 

 eight-mile line and three crosstalk frequency curves. Curve A shows 



