Compactlogix and Ethernet I/O Fault

I have a 1769-L32E compactlogix processor with a number of modules connected through a managed switch in a star/ring combined topology. The PLC port faults cutting all connections to the I/O halting the machine unexpectedly, about 2-3 times/week. This so annoying... The HMI was configured to work through the PLC's serial port. I am out of ideas and tired of trial and error procedures. I contacted Rockwell's support and, as I had formerly an unmanaged switch, they recommended to replace it with a managed one. The chosen switch is a Stratix 5700 1783-BMS20CA which was given a static IP address to have the AB-global macro executed, per instructions from its manual.

The fault persists; the only thing that has made the faults less frequent was to increase RPIs in every node. The network is only composed of a few nodes (well below the published limit for the L32E). Please help! Customer is quite upset and not without a reason.


We have had a similar issue with, I believe, compactlogix processor in VFD communicating to another PLC and Rockwell's support finally fixed.
Basically any communications loss caused reset of parameters and before the Device(VFD)refreshed info in output registers bad information was passed to PLC causing shutdown from PLC level.

I might have timing off but bottom line, adding a time delay after communications loss before sending commands to device stopped problem.
Hope this helps conceptually even if not totally precise on the run.