In message <(E-Mail Removed)>, Pete Flood
<(E-Mail Removed)> writes
>
>Hi, I'm flummoxed, nice to meet you....- can anyone help?
>
>I live in a long house (40m) with three computers. The router is at one
>end with one computer, a pc. A self-made ethernet cable runs through the
>loft to another computer, a mac. It works fine. Recently my girlfriend
>moved in, bringing her laptop pc. Intending to eventually run a second
>lead through the loft, I first tried her laptop out on my long lead.
>The network connections window didn't recognise the connection,
>although all worked fine when I connected the laptop to the router with
>a 1m ethernet cable. Since then I have tried out the pcs with a variety
>of cables - a bought 25m one, a self-made 15m one and a self-made 10m
>one, all to no avail (however with the shortest cable, the pcs register
>its existence but can't establish IP addresses). The mac, of course,
>works on all three.
>
>So...what can I do to get my girlfriend's computer working 40m from the
>router? I can move the router halfway, but I still have the problem that
>neither pc recognises the existence of a 20m cable. Should I give up and
>move to a house that isn't absurdly long, force her to renounce the evil
>bill gates, or go the wireless route (although from what I understand,
>40m is way too long a distance for this option to be worth pursuing,
>and in any case none of the computers have wireless cards, making this
>option a triple headache). I'm hoping there's a magic button somewhere
>for the sole purpose of solving this very irritating problem.
>
>Let me know
>
>Pete
>
>
>
>
Have you tried coiling the long cables up and using them with the
devices fairly close together. Does this work? If it does then the
problem is probably a serious source of interference on the route.
What speed does the Mac connect at? 10M or 100M?
If you force the PC to connect at 10M does it work? WinPC's don't seem
to drop back to 10 if 100 doesn't work. I suspect the auto-negotiation
signals get through OK, but the real traffic (being faster?) fails. But
the link "thinks" it has agreed 100M.
I, like the others suspect the cables (not sure about the problem with
the bought cable ). I had a similar problem a few months ago trying to
cable a long run.
To minimise hole size in the walls they had run (good quality) bare Cat5
cable and then terminated once it was in situ. The link would work if I
forced it to 10M, but would fail at 100M. But they had terminated the
pairs to pins 1&2,3&4,5&6,7&8. They SHOULD be wired so that
pins goto twisted pair
1&2 1
3&6 2 <-Note
4&5 3 <-Note
7&8 4
Full details are at
http://www.datalinkcom.net/rj45_pinout_wiring.html
http://www.hw-server.com/ethernet#twisted_conn
If the pairs are not connected to the correct pins, the signal is in
effect travelling down untwisted wiring one wire from each of two pairs
- it will work at 10M , but not at 100 over any distance
Why the commercially made up cable doesn't work, I have no idea. It
could just be a bad cable, or as someone suggested its wired as a
crossover.
Regards
--
Peter R Cook