The challenge of Thermocouples are two:
- Measuring their very small voltage
- Converting that voltage into an actual temperate
As mentioned here: I chose to use a 2nd INA-220 for sampling the K-type thermocouple This seemed a good match, as the K-type gives voltages around 10-20mV for the temperate range I am concerned with, and the INA-220 has lots of data smoothing and averaging capabilities.
The next challenge is how to convert the voltage into a temperate. This turns out to be rather difficult, as thermocouples are highly non-linear
And the result formula is rather computation intensive:
However, in looking at a simple Volts / temp data for the K-type, I have come up with this simplified formula:
Temp in F = (Volts + 0.78277775) * 43.73611852
(Volts as read by the INA-220's are in units of 10uV)
Over the temp range I am concerned about (300-800f) the error of this ranges from +7f to -4f. More then sufficient for this project.
Update 4/11/2013: I have revised the formula above to:
Temp in F = 800 + ((Volts - 1752.5)/2.4)
Where i is the mV's being read from the INA-220's. Should give better accuracy (+/- 1f) centered around 800f