We have a distributed PLC system consisting of 105 PLCs out in the field. There are 14 RS-485 runs and master SCADA is polling each PLC through etehrnet-serial gateways. Communication is Modbus.

6-7 months ago a lightning hit some part of the field and some PLCs stopped communicating. We have surge protectors at the communication line and panel power inlet. The PLCs were all good, there was nothing wrong with the ports, but they were just not communicating. After trying everything I could for a month or so, I isolated the PLC rack from ground and used a DC/DC convertor just to feed the PLC. That could only solve the problem. So I did this for each panel with this problem.
Last month again a lightning hit the field, this time different area. And same thing happenned. I solved the problem the same way.
Any idea how this solution works and what is happenning to my system. Is it possible that the modbus cable may change empedance with the effect of the voltage peak?

 

One possible explanation is that the surge protector was damaged by the lightning, partially shorting one or both of the communications lines to ground. This is a typical failure mode I have seen with MOVs. The shorting resistance would appear to be in parallel with the input resistance of the PLC communications port, and would reduce the received voltage. When you isolated the rack supply common from ground, you isolated the input resistance of the PLC from the surge protector input resistance.

 

IF there's one thing I've learned about electronics and lightning, It's that lightning does as it damn well pleases and the results are anything but predictable. I've seen a ttl chip get zapped in the middle of a MOS memory array that was unharmed. And electronics that survived when the cables looked like a snake eating eggs. In your case, I would check for multiple grounds where you don't expect them and isolated things that are no longer isolated. That might explain why removing a ground fixes the problem. And although we think of cable as pretty low tech and immune to damage, lightning provides the exception to the rule. And lightning can damage things you would never expect could be damaged. I had to fix up a computer installation in FL where nearly every cable into a particular building was working but damaged in some way causing problems only once or twice a day. A TDR showed impedence "lumps" at one of two points on nearly every run where the foil shields had been vaporized, drain wires opened, or insulation melted inside. And from even six inches away, you'ld never see the damage. And we had to replace "66" punchdown blocks. Now, to look at a punchdown block, there's simply no way for it to go bad. But they were, with shorts where they'd flashed over when the humidity rose.. None of this was ohmmeter or test light stuff, they checked out fine without the TDR,several times. And worse, after replacing everything that we knew was bad, there was "walking wounded" that waited till after we left to fail. My point is that you didn't fix anything and you are running with equipment known to be degraded. It's wonderful that you can work around it, but I would schedule time to find and fix the damage. When this happens, a claim should be started with your insurer so that the cost of fixing this is covered rather than kludging around it and having unscheduled bizarre problems and uncompensated downtime at a later date. But that's just my opinion. For all I know that site in FL may still be having problems. But they aren't suing because they understand that they are residual damage from an "Act of God" and not shoddy practices on our part. And the repairs are covered.

Regards

cww