On 9/14/2010 10:25 AM, James Knott wrote:
Coax is much easier to build as a true constant-impedance line, but it has higher loss, especially at higher frequencies, which is the true limiting characteristic. If you can build a twisted pair to a very high standard of consistency, it will give greater ultimate range due to the lower signal loss. Actually, it's not quite that simple. Loss depends on physical dimensions and the dielectric constant of the insulator. However,
John Perry wrote: properly constructed coax generally has better shielding than Unshielded Twisted Pair (UTP).
And, since there's more dielectric in coax than in TP, there's more loss. And, since copper losses are greater than in TP due to the much smaller center conductor, there's more loss. And shielding has been known since the 1930's (demonstrated by Telco research, actually) to be useful only for near-field electric fields (generally a small part of the interference environment). The main benefit of the coax structure is forming the outer conductor around an easily-controlled internal insulator.
That's also why the ancient Ethernet cables were coax, but limited essentially to 10Mbps and short distances, and we had to go to twisted pair (cat-5, cat-5E, cat-6) to get higher bandwidth and signal fidelity over longer distances. Manufacturers are building Ethernet cables much more carefully than they do short-range, low-frequency twisted pair, and for longer distances the transceivers are more complex to compensate for the poorer signal fidelity,.
It was cost that drove the switch to twisted pair. ... With copper you're still limited to 100M, but with fibre, microwave or other, you can now go essentially unlimited distances, limited only by what the hardware can do.
So it appears I've gone astray trying to extend my instrumentation background where it's not appropriate. :-/ I'll now drop back into the background and continue watching you all argue this interesting topic. I'll have to ask my Russian in-laws if their home town is set up like Ilya's. irrelevant jp -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org