372 BELL SYSTEM TECHNICAL JOURNAL 



to about five miles. Primarily the tests were made to indicate the 

 effect of sheath and quads upon the summation of crosstalk with 

 length as a check on theoretical considerations and were the first 

 extensive tests made on a coaxial cable with this end in view. The 

 layup of the cable is shown in Fig. 3. 



SHEATH 



COAXIAL UNIT ^;^^ ^^^^tV PAPER WRAPPING 



CENTER CONDUCTOR :— ,y//i, ,, w xv//i. 



I3AWGAUGE, SEMI-HARD Iv/// // ^7?\ \r~~\Y7A STEEL TAPES 



DRAWN COPPER CONDUCTOR 

 INSULATION: — 



HARD-RUBBER DISCS 

 OUTER CONDUCTOR :- 



TUBE FORMED FROM COPPER ,//..,k/., ,KXi, i^^ ^-^ l^yx. l^Z^L-aJ'^^ r,»r,.rr, ,K,e,„ .T^,, 



AND STEEL TAPES \//V^ *^/^ ^^ l^Wf^-PAPER-lNSULATED 



QUAD ■ 



■INSULATION 

 PAPER WRAPPING 



CONDUCTOR 



„ ' CENTER CONDUCTOR 



PAPER INSULATED, 19 AW GAUGE, 



SPIRAL-FOUR QUADS \//?>i>^ ^\ /f~7i^/^A/ OUTER CONDUCTOR 



Fig. 3 — Cross-section of twin coaxial cable. 



As indicated in the latter portion of Part I the general procedure was 

 to measure crosstalk in available sections of equal length, /, with the 

 tertiary circuits terminated in approximately characteristic impedance. 

 Interaction crosstalk between these sections was then measured and 

 finally the two sections were combined to find the total crosstalk in 

 length 11. This process was repeated until a total length of about five 

 miles was built up. 



Far-End Crosstalk Summation 



The results of crosstalk tests on 73 and 146-foot lengths are shown 

 in Fig. 4. The letters on the curves correspond to the crosstalk 

 components discussed in Part I. Only far-end far-end interaction 

 crosstalk was measured but for such short lengths the near-end near- 

 end crosstalk would be nearly the same. 



Remembering from the discussion in Part I that the total crosstalk 

 Fii in length 21 is equal tolFi -\- Fnn ■\- Fff it is evident that since in 

 this case the measured components Fnn or Fff are quite small the cross- 

 talk in 146 feet should be approximately 2Fi. That this is the case 

 may be seen from the measured crosstalk in 146 feet which is about 6 db 

 higher than for 73 feet. These lengths are apparently short enough 

 for equations (15) and (16) to hold reasonably well at the lower 

 frequencies. 



