M172PDG28SI (Ethernet port randomly turn to offline)

[a_shapoval_mers]

TM172PDG28SI (Ethernet port randomly turn to offline)

Dear developers,

There are  about 10 PLCs (TM172PDG28SI). All of them connected to switch (BMS network) via ethernet (ModBus-TCP).

Randomly some of ports turn to offline.

But!!! After reconnecting (off/on) RG45 connector,  - PLC turn to online with absolute guarantee.

So,

1) Length between each one PLC and switch less, than 50 m

2) Every of PLCs connected to switch with FTP cable

2) There is a plastic RJ45 connector at PLC side

3) There is a metal RJ45 connector at switch side

4) Switch is connected to the PE

!Also I had been trying to use the metal rj45 connector at both sides, but result is the same

Is anybody have an experience for solute my problem.
May be there are some system words or system bits at the PLC level, that's can to provide some extension information about ethernet port

BR

[Bastian_Schmitz]

Hi,

At the moment in never had a situation like explained.

Could you tell me which firmware version do you use?

Did you use static IPs or DHCP?

You could check the connection on PLC side to BMS via the targetfunction sysMbSTcp_SocketServerinfo.

[a_shapoval_mers]

1. Firmware version is 596.9

2. Static IPs

3. for now sysMbSTcp_SocketServerinfo is not implement in PLC application, so as soon as possible I'll download PLC application with sysMbSTcp_SocketServerinfo, and when PLC will turned to off, I'll try to monitor return value via the RS-485 interface. But I'm absolutely sure, that PLC has another behavior at socket overflow -  it is not accessible via ModBus TCP, but it has the IP ping. In my case, PLC has no any ping, but after RJ45 reconnecting, the ping is restore

[BevanWeiss]

1. Do you have duplicate IP addresses configured anywhere in your network?  Some L3 aware switches can administratively disable ports where duplicate IPs are detected
2. Do you have any duplicate MAC addresses (i.e. virtual machines) on any of the ports.  Some switches disable ports where they see duplicate MAC addresses
3. Do you have any of the PLCs with multiple ports connected into the same switch (or into other switches which may result in a network ring being formed without STP/RSTP being appropriately configured)?

It's most likely that it's the switch being responsible for the 'port down' action, so you should be able to look in the logs for the switches.

If you have lots of PLCs, and a cheap unmanaged switch, then it's possible that the large number of small packets/frames from the PLC communications is overwhelming the switch processing (if it's not able to offload at line rate).

I've definitely seen this behaviour on older 10/100Mbps switches, where whilst they could handle 1x 100Mbps stream, when given 16x 100Mbps aggregate traffic from thousands of frames per second (i.e. like 200 bytes per frame) they just locked up ports, and then rebooted the PHYs (obviously they got too confused).