That sounds odd, so not a specific gpu crash, but all gpus drop to 20 MH/s each instead? That's not the typical error scenario really, I wonder what's going on with the rig at that time. You don't see any excessive cpu consumption, and gpu temps are fine all the time, Rave is controlling fans?
Thanks for trying to help me! I use constant fan speeds with RaveOS (50%) and my temps across my cards vary from 46-49C. I just restarted RaveOS with TRM again so I can measure the power consumption of the rig with TRM and then compare it to Phoenix. I'm not sure how to check CPU utilization from within RaveOS.