>Twisting always increases capacitance between pairs.
Hmmm. Twisting WILL, of course, increase the LENGTH of the
conductors a bit, which will increase the capacitance by that
proportion. But the question is, HOW MUCH does it increase? I
recently untwisted some #14-2 and 16-2 twisted pair to use as
radials. The difference in length between twisted and untwisted
was less than 10%. Consider CAT5. It is VERY tightly twisted with
a very short lay (a lot of twists per unit length), but its
capacitance approaches the lowest capacitance cable I know of.
Conductor size also contributes to capacitance, with smaller
conductors having less capacitance. So making conductors larger
increases capacitance. Does that mean we should use only #40
This fixation on cable capacitance reminds me of the late Dick
Heyser's comment that trying to describe an audio system with only
one parameter -- its frequency response -- was like trying to
write poetry with only one word in your vocabulary. Heyser was a
very bright guy whose day job at JPL was designing communications
for the space program. He was the inventor of Time Delay
Spectrometry, and held a dozen or so patents in fields ranging
from medical electronics to underwater sound to audio.
TWISTING is a powerful rejector of interference in any system, and
especially in a balanced system. Twisting is good for at least 20-
30 dB. If it adds a few hundred pF to a circuit that already has
0.1uF (or even 0.005uF), who cares?
TowerTalk mailing list