MEASUREMENTS IN MULTIPAIRED CABLES 



655 



Expressed as a percentage of L per length, the correction term varies as 

 the square of length. It is only 0.5% for 750' of CNB but is 20% for 5000' of 

 the same type of cable. The single correction term given is not sufficiently 

 accurate for lengths of CNB greater than about 5000'. The comparable 

 Hmiting length for smaller gauges would be less; it is about 1000' for 26-gauge 

 pairs. 



It is also necessary to allow for stranding takeup effects when converting 

 from per length to per mile values. Stranding takeup is defined as the incre- 

 ment in length of a given pair compared to the cable length, due to the helix 

 introduced in the pair during the cabling operation. It depends upon the 

 size of cable and stranding lay and is negligible for cables of 50 small gauge 



Table I 

 Comparison of Measured 1000 CPS Inductance With That Calculated From the 



Theoretical Formula 



pairs or less but causes an increase of 2.7% in pair length over cable length 

 for pairs in the outside layer of a full size CNB cable (2|" diameter over 

 the paper wrapped core). 



Some short lengths of pair with wires parallel and approximately straight 

 and with accurately known dimensions were made up in the laboratory. 

 Inductance measurements at 1000 cps on these pairs agree very closely with 

 inductances calculated from formula (1), for d/2S ratios from about 0.25, 

 which is close to the nominal value for cable pairs, up to nearly 0.50, which 

 is the highest possible value. These results are shown in Table I. The 19-gauge 

 pair with a d/2S ratio of 0.261 was twisted fixed carriage style under constant 

 tension. There was no measurable change in inductance for twist lengths as 

 short as 2", compared with the straight, parallel condition. Measurement 

 accuracy was aboutdzO.25%, 



