Hello all.
Extruder puller drive, PowerFlex 755, 75 hp, on EtherNet/IP to a CompactLogix 1769-L30ER. The drive drops off the network now and then and the controller shows 16#0204 on it. The drive doesn't fault, it just loses comms and the puller coasts on the last command until it comes back. Twice a shift maybe.
Switch is a Stratix 5700 in the main panel. The drive's port, gi1/5, has CRC and alignment errors climbing, a couple of hundred a day, and every other port sits at zero.
I swapped the drive's comm card, the 20-750-ENETR, with the one out of the other puller and the errors stayed on gi1/5, then moved the drive to gi1/6 and the errors went with it. So the switch is fine and the drive's port is bad? The errors followed the drive both times, so is it that drive's own cable I want rather than the card?
Errors move with the cable, not with the port and not with the card. That leaves the cable and whatever's next to it. What is the cable, and where does it run between the switch and the drive?
Grey patch lead, Cat5e, one of the ones out of the IT cupboard. Goes from the main panel across in the tray, into the drive panel through the bottom gland, up the side of the 755 to the card. The motor leads come out of that same gland and run up the same side to the drive.
There's your antenna. An unshielded office lead lying against the motor leads of a 75 hp drive. The drive chops the 650 VDC bus at a few kHz on those leads, the patch lead has no shield, so it picks up a bit of every edge and the switch reads it as a bad frame. CRC. The comm card was never the problem.
Fix is shielded industrial Ethernet, Cat5e SF/UTP or better, in through the other side of the panel from the motor leads, 8 inches of separation at least, cross at right angles where you must. Bond the shield at the panel entry with a shield clamp, not a pigtail. Then reset the counters and watch it.
Why does the shield help when the noise is on the motor leads and not on the Ethernet pair? I'd have thought the twist cancels it.
The twist cancels what both wires see equally. Motor leads a finger away from one side of the pair don't hit both wires equally. The shield takes it all, so the pair sees nothing. And those motor leads want their own shield bonded both ends, which I'd bet these haven't got.
Ran a shielded industrial Cat5e in through the top gland on the other side of the panel, shield clamp at the panel entry, 8 in off the motor leads. Reset the counters on gi1/5 Wednesday afternoon. Zero CRC, zero alignment by Friday. No 16#0204 on the drive and the puller hasn't coasted since.
So it came down to the unshielded office patch lead lying against the motor leads. Shielded cable, separation, shield bonded at the panel, done. The comm card was fine all along and the original 20-750-ENETR is back in.
Odd thing, the other puller has the same grey lead on the same route and zero errors. Different gland, maybe. Not touching it until it acts up. Thanks edj, boris_h.
Touch it now. Zero errors means nobody has looked yet.