Not sure what section to put this in... move if appropriate.
OK, so I had a fun "RF tech" site visit yesterday. Theatre teching a new long-running show, with some electric winches pulling video screens up and down. They were having problems with the screens losing signal/sync when the motors were in operation. So, since somebody mentioned "RF interference", they called for somebody to come and sniff around.
I brought a TTi PSA 2740 RF spectrum analyzer and my usual case of tools. We went up to the grid and I sat around trying to find interference whenever a motor moved. The spectrum analyzer didn't show anything, and in fact these hoists (not chain motors, but double cable reels and a big motor on a chassis about 3 feet long) were completely encased in metal, etc. They'd tried building a Faraday cage with screening, which hadn't helped. They'd gotten some improvement by having an electrician connect an extra ground to the motors, using what looked like #6 back to a power panel.
OK, so I had plan B to look for magnetic fields and lower frequency stuff. I pulled out my Fluke scopemeter 120, and hooked up just a loose coil of wire. Sure enough, when the motors moved I saw some activity. But there was also some similar stuff happening just with lighting moves.
Since they were doing tech rehearsals and not sitting around moving the motors at my request, I had an hour or so sitting there staring at the setup and thinking about it. Finally, I measured voltage between the winch chassis and a huge beam running a few inches away, and saw up to half a volt of activity (nasty stepper motor sorts of waveforms) when that motor moved. Ohm reading between the same points was only 3-4 ohms.
Back to the video screens. Because of the movement, they were only able to feed the screens with something really small. They were using DVI over Cat5, and mentioned that the backup option would be fiber. No conduit around the cabling, of course. So it's plausible that there's a big common mode signal hitting the wire and swamping the balun, or even getting the balanced wiring a bit unbalanced. Symptoms were things like "unable to sync" type messages on the screens.
Thinking about the grounding, they had the extra ground running some 20-30 feet after it left the motors and going off to a panel. The control/power cables ran off a different way. Remember that I measured a half-volt of hashy motor stuff even with that grounding, compared to the giant beam. So that's current flow over the grounding wire, and current flow produces magnetic fields around the wire. Enough to cause the problems? I don't know.
My recommendation was
a) you've got a solution with the fiber optic, that's a pretty safe choice.
b) grounds should be as short and heavy as possible - go right from the winch chassis to that beam a few inches away, possibly a ground at each end of the chassis.
EDIT:They were in tech rehearsals, so "maybe we can try that later tonight". I don't know if the grounding helped, or even if it got done.