You know, after thinking about this awhile, the week you've been having this problem with this system is the same amount of time since you added the 64 core monster, is it not? It's very possible that "just" removing the ~175 watt RTX 2060 got you back below the maximum that circuit can handle. Since you always use quality PSUs, and you swapped out one oversized PSU for another, the one thing we can be sure of is it cannot be the PSU. It can be the circuit supplying that PSU, though.* Also, with PSUs when two are fighting for the same current, the larger one will always win. The larger one also has more capacitor storage, meaning only the smaller 850 watt one would shut down. Maybe you'd be better off selling me that 2060.
*To the uninitiated, Mark has more computers in his house than the average office building. Large office building, that has multiple servers. He was already bumping up against the amount of AC electrical power he had available to use, before adding the 64 core monster. Since he only removed one of the older, less efficient 16 core Threadrippers, it is quite possible that with the more efficient, but 400% more cores 64 core/128 thread monster, he was a bit over the limit. It could of course also be a bad 2060, but that makes less sense, since you would expect a bad card to cause problems immediately, and not 2+ days later.
edit: I reread the OP, and you said it has been happening for a few weeks, meaning it was happening before you got the 64 core beast, so can't be from overloading the circuit.