I think I have this licked, with experimentation!
The problem is again CABLING!
I’m supplying sufficient voltage - but the USB cable that I used, from a MacBookPro 13, is insufficient to deliver consistent power to the VIM4 under load.
Who’d have thought a premium laptop labeled “pro” would have such a problem with USB-C power?
I am now unable to repro, using the OC3.0 port on my power supply, and a shielded USB-A → USB-C power cord from a noname Amazon supplier. I’m also down to drawing 5.2 volts, which tells me the Apple cord introduces significant resistance - probably result very thin copper wiring.
Lesson learned again: Power, power, power. The VIM4 tolerances for supplying voltage are very finicky, and fail in ways that resemble overheating, etc.
Ours are running on 6V and doing well. Just idling after boot the VIM4 with ubuntu server is drawing 560 ma.
Been cutting off connectors so the VIM can be connected to lab power supplies and would you believe some of the cables are using the small data wires for power. Unless the cable specifically states 5 amps it might be not very good.