"Don Pearce" wrote in message
...
I'm really not sure what you are saying here. Capacitance
increases
because the cable is shortened? No.

Yes.

Take a 1 foot long untwisted cable with zero conductor OD spacing
(i.e. the insulation of both wires is touching) you will have
some inductance and some capacitance.

Now take that same 1 foot section and twist the wires. Total
capacitance will remain unchanged since the spacing between
conductors did not change. However, the capacitance per foot
increases because twisting shortened the cable to something less
than a foot. However, even with heavy twisting the length is only
shortened by maybe 10% or so. So the capacitance per foot
increases by the same amount (say 10%).

Also, study the equation relating conductor area and separation
and
you will find that the best way to increase capacitance between
two
conductors is to bring the closer to each other. There is no
question
of "may or may not go up".

Reread what I said. In the case above the total capacitance is
the same but the capacitance/length increases. If the twisting
method brings the wires closer together then capacitance
increases. However, it's just as easy to take two parallel
conductors, twist them loosely, and get lower capacitance (both
total and per unit length)...because the separation distance
increases.

So capacitance per unit length can go up or down depending on
changes in separation distance and amount of twisting.

With the question of twisted vs untwisted pair, neither
inductance nor
capacitance changes. What does happen with a twisted pair is
that
coupling to external fields cancels at each half twist, but
that is
another matter entirely.

With twisted pair not only does the external coupling cancel but
the mutual inductance between the two wires cancels.

Inductance is defined as the algebraic sum of self inductance and
mutual inductance. Your analysis neglects the effect of mutual
inductance.

As you point out self inductance does not change with twisting
(assuming conductor diameter and separation do not change).
However, mutual inductance drops dramatically and in the ideal
case becomes zero.

The inductance of untwisted parallel wires is proportional to
x*(ln(s/r)+log(s/r)) where x is length, s is separation, and r is
conductor radius. For twisted pair the log term goes to zero
(ideal case). If you don't believe me take a run of zip cord put
it on an LCR meter and measure inductance before and after
twisting.

Capacitance can go up or down depending on conductor separation
before and after twisting and the increased capacitance/length
due to length shortening.