Linear-approximation model. Cold-junction temperature is always included — a real transmitter/DAQ measures mV relative to its own terminal temperature, not 0°C, so both directions need it to be accurate. For calibration-grade work, use NIST ITS-90 polynomial tables.
Standard method — stated scope limits apply
Inputs
mV → Temperature
Temperature → mV
The transmitter/DAQ terminal temperature — typically close to ambient. Set to 0 to see the uncompensated (0°C reference) value.
Enter values and calculate.
About This Calculation
A thermocouple generates a small voltage from the temperature difference between two junctions of dissimilar metals — the Seebeck effect. That voltage is only meaningful relative to a known reference (cold junction) temperature, which is why every thermocouple measurement needs cold junction compensation, either from an ice bath (historically) or an electronic reference built into the transmitter. Different metal-pair combinations (Type K, J, T, R, S, and others) trade off temperature range, accuracy, atmosphere compatibility, and cost — there’s no single best type for every application.
Frequently Asked Questions
Which thermocouple type should I use?
Type K (chromel-alumel) is the general-purpose industrial default, good to about 1260°C in oxidizing atmospheres. Type J is common for older equipment and lower temperatures. Type T suits low/cryogenic temperatures. Type R and S (platinum-rhodium) are used for high-temperature, high-accuracy applications like furnaces.
Why does cold junction compensation matter so much?
The thermocouple only measures a temperature DIFFERENCE. Without knowing the reference junction’s actual temperature, the millivolt reading alone can’t be converted to an absolute temperature — a CJC error shows up as a constant offset error across the whole reading.
Why is the millivolt-to-temperature relationship non-linear?
The Seebeck coefficient itself varies with temperature for any real metal pair, so the standard reference tables (and the polynomial curve-fits used in transmitters) are inherently non-linear — a simple linear mV/°C conversion introduces real error, growing at the extremes of the type’s range.
How accurate is a thermocouple compared to an RTD?
Standard thermocouples are typically ±1.5°C or ±0.4% of reading (whichever is greater) for common types — noticeably wider tolerance than a Class B RTD’s ±0.3°C. Special limits-of-error grades narrow this, at extra cost.