It would be interesting to try a test with the thermal runaway protection in the Marlin firmware, just to confirm it works. I'm trusting it at the moment, but it would be good to have documented confirmation.