For completeness, when you start talking about errors on temperatures measured by Marlin and RAMPS with a thermocouple, there's another substantial error component to consider. I ran into this...