Type K thermocouple calibration techniques

[gnx]

Hello experts,
I am developing firmware for a TypeK thermocouple which uses ADC AD7193 for TC voltage. Another cold junction temperature sensor is used. I use the lookup table to convert CJ to voltage, add it to ADC measurement and then use NIST polynomials(10 coeff) to get temperature. On measurement using a calibrated TC simulator, the ADC was producing offset and gain errors leading to difference of 2-3 degrees in worst cases. I did a simple two point (zero and maximum voltage) calibration(zero offset removal and gain multiplication) and was able to get more accuracy. I simulated a few points and typical accuracy was about 1 degrees. My question is although there is no 'one size fits all', what is the standard procedure for doing TC calibrations? How many points over the whole range (-200 to 1372) would be tested and calibrated against?

[retrolefty]

Retired from instrumentation department of a large oil refinery, but our procedure was to do a five point verify/calibration check but not over the whole possible range of a K couple but rather over the range the sensor is going to be used in it's specific control loop, say 0-200 for a oil storage tank, etc. Actually we never considered thermocouples to be suitable for tight temperature control as the more accurate RTD sensors had come down in price over time Vs thermocouples and the total installed cost wasn't much different then a J  or K couple.