cable ratings Re: [TowerTalk] Using Belden 9913 on a crankup?

[Jim Lux]

>
> Ratings of cables use very obscure rating processes when it
> comes to voltage and power. I'm not even sure how they
> determine the ratings, but I know I can grossly violate
> ratings without any sign of failure.
The rating process probably isn't all that obscure.. it's probably laid out
in all it's glory in MIL-STD-C17 (I think..)

There are two limiting aspects for power handling on a cable: voltage and
current.  For voltage, it's breakdown, at the rated highest altitude, with
the worst case bend, at the end of rated life (and they probably use
something like a partial discharge tester to look for incipient corona, as
opposed to a pass/fail hi-pot breakdown type tesst). For current, it's a bit
tricky, because it's probably limited by the dissipation in the center
conductor, which will be frequency dependent.  So, they'd have to rate at a
particular frequency, and further, have to have some way of defining when it
had failed. They wouldn't go to actual melting, but probably, more likely,
that the cable doesn't meet the specifications (loss or characteristic
impedance) after having the test current applied.  Since it's a thermal
limit, duty cycle and ambient temperature would be important (so, for the
test, it's probably at the maximum specified temperature)

If they rate cable like we do other RF components, I would expect that they
rate for a given power, assuming a reflection of all of the power.. That is,
the voltage would be twice the nominal at the nominal impedance, and the
current would be twice the nominal, and then, they'd add a bit of margin to
make sure that it will pass the test.

A voltage rating is probably derived from a basic power rating, converted
using the nominal impedance.

For what it's worth, things like velocity factor and characteristic
impedance aren't all that well controlled in most coax.  If you're within 5%
(0.2dB), that's probably all you can expect.  Changes of a similar magnitude
can result from flexing and aging. Heck, you can get a 0.5 dB change in a
SMA connector (at 10GHz) just by going from finger tight to torqued to spec.
Even precision connectors like APC-7s are probably only repeatable to a few
tenths of a dB.


An interesting question.