On Thu, 28 Jun 2018 at 18:45, Eric Smith via cctalk <cctalk@classiccmp.org> wrote: > > In case it may not be obvious to some readers, the reason you should NEVER > ground an Ethernet cable (of any kind) at two points is that the ground > potential at two different points is unlikely to be the same, so that will > cause a DC current flow through the cable.
I installed the first LAN at the biggest builders' merchants in the Isle of Man, some 28 or 29 years ago in my first job. Normally, we'd have wired it too, but they said no -- we have staff electricians, we'll put in the cabling, you just connect it. So we gave them the specs, they fitted benches, ran in lots of 13A 220V power, thinwire, plus plentiful power sockets & breaks in the cable for every position. We hooked up the PCs, installed the server, installed the DOS networking client, and started testing. It worked. We left. Within days, I was back. Intermittent dropouts -- basically, at more or less any given time, one machine couldn't see the rest of the LAN. But which one changed every few minutes. Much testing. At software level because we were a small company, in a small island nation, and Ethernet testers were _way_ over anyone's budget. Basic continuity worked. I started testing each node. All worked individually. So I started bisecting the thinwire and checking each half. I got a shock. Off a thinwire cable. I had a voltmeter, at least, and it was registering a 3-digit number of volts. I think I actually did say the full unexpurgated WTAF. The Ethernet wasn't grounded -- most weren't -- but it *was* lying in the same conduit as the new mains cable, many dozens of metres of it. AFAICT the mains cables were _inducing_ current in the Ethernet. Quite a lot of it. The client had to rip out and relay most of the cabling. They were ignorant enough to be overconfident: wires are wires, we do wires, we'll do them like any other wires and it'll be fine. What amazed me is that none of the NICs blew, none of the machines failed or died. Once the cabling was sorted, it was OK. Who knew that BNC Ethernet ports could handle 100V or more flowing through them and mostly work? -- Liam Proven - Profile: https://about.me/liamproven Email: lpro...@cix.co.uk - Google Mail/Hangouts/Plus: lpro...@gmail.com Twitter/Facebook/Flickr: lproven - Skype/LinkedIn: liamproven UK: +44 7939-087884 - ČR (+ WhatsApp/Telegram/Signal): +420 702 829 053