For when power is lost and/or communication to the piepan is cut off we need an independent warmup routine in the temperature board. Steve Hope will flesh out the specification and attach it to this ticket. But roughly speaking there will be some criterion for the board to decide on its own that it needs to keep the detector safe (some combo of loss of power, loss of networking, loss of keepalive commands, etc), and a simple procedure to follow at that point (for example: turn off PID loops, slow down sampling, switch to maintaining a 1-2K diff between a detector sensor and some other one). As simple and robust as is possible.

This also requires the PCM to power things off, in order to minimize load on PFS and SM UPSes, so that heater power can me maintained as long as possible.