Inserting a dummy value is better than risking an exception with
failure to write a thermistor table at all. Happens when for one
of the usual thermistors accidently a nominal resistance of
1000k instead of 100k is used.
Two thermistor tables included:
ThermistorTable.h.dist.old, which is the table we used to have, included so those already using it don't have to recalibrate.
ThermistorTable.h.dist, which was generated with 50 entries, and trimmed down in lower temperature ranges where we don't care as much about accuracy and there's less deviation from a straight line anyway.
Corrected default temp sensor entry in config.h.dist to thermistor instead of intercom. (Now matches earlier definitions which by default say we're using a thermistor)
Added noheater sensor example to config.h.dist
Copied CreateTemperatureLook.py from "official" firmware, and modified it for 14.2 fixed point, no negative temperatures (we're using uint16_t's), and PROGMEM.
Since I simply copied the example linear interpolation formula from wikipedia, I'm certian that it could be more efficient. The code that was there wouldn't work with 14.2 table values, and my understanding of it was too shallow to fix it. This works, and upgading the calculations to uint32_t actually takes less code space than leaving them at uint16_t. I assume it's calling a library routine that's already being linked in for the 32-bit math.
Markus Hitter
Phil Hord
David Forrest
Andy Kirby
John Gilmore (none)