>This has nothing whatsoever to do with the problems of interference with 
>USB cables, which carry serial data streams, not packets, and are 
>subject to external electromagnetic interference.

I think you are mixing logical layer concepts (packets) with physical 
layer concepts (stream of bits). At the physical level, both Ethernet and 
USB are performing serial communication: that is, sending streams of 
bits. Both will be subject to electrical noise.

Staying in the physical world, USB cables are surrounded by a metal 
shield. This is a costly, but very effective means to block electrical 
noise. Ethernet is not shielded so must rely on more sophisticated 
techniques. Ethernet signals are sent using a differential circuit 
(actually, several of them). A differential circuit uses a pair of wires 
and puts an opposite and equal signal across the pair. It depends on 
noise putting an equal signal across the pair. So a differential circuit 
looks only at the difference between each wire of the pair, thus it is 
able to ignore the noise. I think Ethernet twists per foot has more to do 
with crosstalk than noise rejection. I will leave that to someone else to 
look up.


*************************************************************************
**  List info, subscription management, list rules, archives, privacy  **
**  policy, calmness, a member map, and more at http://www.cguys.org/  **
*************************************************************************

Reply via email to