Ethernet Question

[bryan372002 0]

Need help please. I have to machines on different networks one being 192.168.20.xxx and the other being 191.1.15.xxx have them both hooked into a fiberoptic unmanaged switch and can get online with either network thru the fiber optic. The problem I may be having and I am not sure if this problem started after this was set up, is that some of My powerflex drives are getting DPI adapter faults. could this configuration with the 2 networks being tied together be causing network collisions ?

[J. B. Jennings 4]

If this "unmanaged switch" is actually a switch, you should be getting very few network collisions, if any. If it is actually a hub, then yes your problem is caused by network collisions. Since you are having issues with the PowerFlex drives, I would monitor input power and see if it is stable. If input power (voltage) is stable, then check the communications cables themselves, it may be a noise related issue. Keep us up to date on this and post which PowerFlex drives are having the faults. J.B.

[BobLfoot 301]

Not to argue with the answers already given but the last batch of Powerlfex Ethernet Drives we received had a rash of DPI Adapter Faults as virgin failures. The temporary solution was to disconnect the DPi port 1 HIM and leave only yhe DPi port 5 Ethernet adapter active. Our Engineers and RA Tech Support are still muddling thru the problem. I'll post more when the "mushroom club" gets told what happened.

[bryan372002 0]

Thanks Bob, thats interesting, I had contacted AB once about this issue or actually how I might be able to reset the drive without having to enter the electrical cabinet, so he first told me that Managed switches should have been used instead of unmanaged. anyway I may try disconnecting a couple of the HIMS and see if that makes a difference

[BobLfoot 301]

We have three ethernet drive systems with ControlLogix PLCs and Safety Modules IO designed by RA. So our application is "by the book" "to the letter" and still has issues. I'll post when and ifI learn or hear anything.

[bryan372002 0]

J.B. intended to thank You the other day for Your reply, but got hit with a breakdown (another issue) anyway thanks for Your reply, I will look into that switch to make sure it is indeed a switch and will post anyway findings that I may fall upon...lol

[Ken Roach 194]

It takes quite a bit of unmanaged Ethernet traffic to cause a 20-COMM-E to lose its connection with the controller. It's not impossible, it's just that I've never seen it happen. The easy and quick check to see if there's any IGMP management being done at all is to load Wireshark and plug into an open port on one of your switches. If you see a lot of UDP traffic with source IP addresses of your RA devices and destination addresses in multicast groups, the multicast management features of your switches are misconfigured or absent. I'd vote for absent, since you mention unmanaged switches. Another simple test to see if there are Ethernet capacity problems is to check the I/O connection list on the ControlLogix EtherNet/IP bridge module to see if there are missed packets being counted for the drive connections. Onboard the 20-COMM-E there are diagnostic objects that will allow you to check for DPI (the peripheral network inside the drive) noise problems. I usually use Drive Executive and select the Peripheral -> Display Diagnostic Items from the top level menu. The Diagnostic objects are not part of the ordinary parameter set. You're looking for the DPI Rx Errors and DPI Tx Errors counters.

[bryan372002 0]

Thanks Ken That's Great Info. I did have one of the PF 700 drives that seemed to be the worst, in fact it got to the point where it would not reconnect even after a Cold Boot (power shut down and restart) so I changed that 20-comm-e out and I dont think it has lost cooms yet, I will check on the RX and Tx error counters though. Thanks Guys