[TAG] Wiring a house with ethernet

[Benjamin A. Okopnik]

On Thu, Oct 20, 2005 at 10:36:00AM -0400, Jay R. Ashworth wrote:
> On Wed, Oct 19, 2005 at 08:22:30PM -0500, John Karns wrote:
> 
> > [1] As Ben mentioned in a reply, high freq signal characteristics are vastly
> > different than DC.  So although the cable may pass a (DC) continuity
> > test, it can fail to perform adequately at high frequencies.
> 
> Skin effect is your enemy.

IIRC - and I may be wrong, since it's been ages since I've done the math
- at 1GHz, the effective cross-section of a given conductor is around

0.00008" (skin depth @ 1GHz) x circumference

so that R(total) works out to 

bulk resistivity / cross-section * ( length / ( pi * diameter ) )

There's the impedance effect to be considered as well (this being the
factor that hoses you when you make sharp bends); in properly-handled
shielded wire it won't amount to _that_ much, but leaving off even an
inch of shielding or having another wire within coupling range can send
it through the roof. When I was at Hughes, I learned to make a perfect
(i.e., less than 3dB loss in a foot-long run) SMC pigtail - but it took
a bunch of practice and a lot of testing with an RF meter to get there.


* Ben Okopnik * Editor-in-Chief, Linux Gazette * http://linuxgazette.net *