Identifying the source of network failures

I have a CompactLogix 1756-L72/B with only an EN2TR/C in a 4 slot chassis. Port 2 of the EN2TR is connected to two POINT I/O racks, while port 1 is connected to the local port of an 9300-ENA. The Uplink port then connects to an unmanaged N-Tron 108TX switch in our SCADA network. I should note that the run to the unmanaged switch is about 230 feet with shielded cable.

Since installation, this setup is plagued by connectivity issues. Our HMI loses connection every 5 seconds or so, and regularly goes in an out. Investigating, I found that port 1 of the EN2TR is slowly accumulating FCS errors and MAC receive errors (2 or so a minute). Documentation about these errors is sparse, but I have tried:

-Confirming auto-negotiate is on - both ports are in auto, and the switch is unmanaged.
-Ruling out EM interference by swapping the 230' cable with one that I ran directly between the panels, so a shorter run that isn't in the cable tray
-Bought a cable tester to ensure they were built properly. Both cables passed (this was just a small fluke meter, I don't know if it would catch more insidious wiring problems)
-Swapped ports 1 and 2 on the EN2TR, port 2 then began to see errors so I ruled out EN2TR hardware failure.
-Tried different ports on the switch. If I stare like a madman at the LED lights, I can see that while the other ports are all blinking continuously, the problem port is blinking, then going solid for a fraction of a second.
-I ran a wireshark capture of the traffic between the controller and the HMI server, which I attached a screenshot of errors that were caught. I do not have the expertise to decipher them unfortunately.

Does anyone have some insight as to where this might be going wrong? I suspect the NAT device since the FCS and MAC errors are only generated on the port connected to it, but I can't justify replacing it without some evidence that it is to blame.

I ran into a problem not unlike your about 2 years ago. Our in house Fluke cable tester said that the cable was wired correctly but we still had issues. We finally borrowed a tester that could test the data rate as well as the connections. The new tester showed us that we were dropping packets and therefore the data throughput was severely degraded.

After reterminating the ends and testing again we were hitting 100 Mbs and the connection problems were gone.

Can you post screenshots of the following pages from the EN2TR's web page?

-Diagnostic Overview
-Application Connections
-Bridged Connections

(I think those are the names of the pages, they should be close)

Here are the screenshots you requested. 192.168.1.21 and 22 are both the POINT I/O racks. 172.16.1.51 is the server which is running ignition.

I will look into sourcing a more robust cable tester. However, I figured even if I did miswire two cables, it seems unlikely that they were miswired in exactly the same way as to cause the same symptoms. It was more likely that it was the port shared by both cables (the NAT uplink) that was causing this behavior.

In the meantime, I can try wiring up another cable just to confirm. Unfortunately, since this run is so long, I don't have a prebuilt cable of that length on hand.

Hook up any switch you have lying around in place of the NTRON. Easy test with minimum effort.

I have some consumer grade switches I can try temporarily. The N-Tron switch also has the PLC for our brewhouse and our MCC connected to it, both of which are behaving normally. This brewhouse PLC is also an L72 with an EN2TR, but has older firmware. I can't remove the switch at the moment while the brewhouse is active.

I mention this because there is an extra port on the brewhouse EN2TR I could potentially connect to instead of the switch.

Problem PLC -> NAT -> Switch -> Server
Problem PLC -> NAT -> Brewhouse PLC -> Switch -> Server

It wouldn't remove the switch, but it would at least give me full control over the duplex settings of both sides. This should be possible without disrupting the brewhouse PLC right? Isn't the EN2TR basically a glorified switch?

I didn't pickup on the fact that you were talking about 2 different cables. Both cables being bad in the same way would certainly be a stretch.

I have had problems with an EN2TR acting strange before. My fix was to get rid of it and use an ENBT but I would recommend making sure you have the correct AOP add on profile for your module and Call Rockwell Tech Support. Press them for answers how to verify the module is working correctly.