PLC Network analysis

Hi all.

Quick question, we currently have an issue with 1756-ENBT addressing to flex i/o. We keep getting issues with four machines linked together randomly getting flex i/o faults and losing comms to out site network. I'm just curious to see what type of software is used in the industrial world for ethernet network analysis such as for dual ip's and trafficking? We use wireshark currently.

Thanks in advance.

If you have wire shark you have one of the best tools but I also like JDSU network analyzer, Angry IP Scanner, Zenmap, PRTG and Network Stuff.

There are some more that I like better but they are Linux based. If you know Linux let me know and I will list those for you.

The first stop for me is always the embedded web page in the 1756-ENBT. Look for any kinds of packet errors, CPU utilization, total packets per second, etc.

Then if there's a managed switch, check it's built in web page for any kinds of errors.

Then port mirroring and Wireshark.

Thanks for the responses. We use 4 x 1769-L35E that communicate together through a central unmanged hub. What is happening is that suddenly all the 4 machines will stop displaying flex I/o faults losing communication between the hubs. Im on the diagnostic page for the 1769-l35e currently and utilization is maxed out (88% onwards). What would you be looking for in such a scenario? I'm noticing that the packet counter statistic's has a decent amount of rejected and conn timeouts on the CIP is quite high. Currently it is working, but we have changed one of the four from managed to unmanaged to get away with it.

Apologies, we're using 4 x Stratix 6000 that we are having the issues with.

Have you checked the module's status web page as suggested?

I've downloaded Rockwell's manual to fault diagnose these units, which has helped. Everything seems in order except the CPU usage near maxed out. I was thinking of setting the RPI time higher as currently each machine is set at 10ms interval. Another issue is the supplier has unfortunately passworded out the configuration of the Stratix cards, so the only way we can get online is a factory reset

88% is a problem. Certainly this will cause you grief.

Is there a lot of stuff on Ethernet on this system? I'd suggest setting the RPIs only as low as necessary. For example, there might be no need to have a VFD that controls a conveyor set at 20ms, maybe you up it to 100ms, or 200ms (whatever the system will allow).

Don't just go randomly adjusting RPIs though without thinking about the consequences. I/O for example may need to be fast.

edit: I should also mention that if your system wasn't working properly with a managed switch, you change it to an unmanaged switch and suddenly it "works" now, you just covered up the problem. The issue wasn't that managed switch was being a pain in the arse, it was indicating that there was an issue. Running any Ethernet module near it's max will certainly be unreliable. You will most likely need to decrease the traffic, or upgrade your system to handle the traffic.

AB Has a nice "Ethernet IP Capacity Tool". I used it to figure out a good balance for a system I have with 32 PowerFlex 525 VFD's, 22 Armor blocks and 6 remote 17 slot IO racks, all thru one EN2T.