Why thermocouple readings don’t compare with camera results


Question received by email, slightly edited;

We are busy working with a foundry that mainly does High chrome white irons. The problem that we are encountering is that there seems to be a variance of about 80°C between our thermal camera and the thermocouple that they are using (which is supposed to be very accurate). Could the problem be that the thermography camera is only reading the surface temperature of the molten metal whereas the Thermocouple is slightly immersed and is therefore reading a slightly higher temperature due to the fact that molten metal is not yet cooled by air on the surface?


The problem that you are encountering where your temperature measurements don’t correspond with a contact measuring device is NORMAL. As thermographers we come across this often, and the longer that you work in thermography, the more you will come across it.

Firstly, you are correct, they are not measuring the surface, they are measuring below the surface (based on your description), and the temperature at the surface and below the surface is different. This type of error, normally causes the thermocouple to measure higher than the camera, as the surface is usually cooling quicker. The other issue, is that the thermocouple and the camera will not really display the same result, as they are measuring different things. The camera intercepts radiation and calculates this to a temperature based on user inputs (emissivity, Trefl etc), and assuming that you have all those inputs correct, it should display the correct temperature from the surface. If a thermocouple is used for surface measurement it will not usually display the same value as the camera. The reason for this is simple. Thermocouples do NOT display the temperature of the object they are measuring, they display their own temperature. The thermocouple will only display the temperature of the object that they are in contact with if they are in thermal equilibrium with that object. The thermocouple will only be in equilibrium with the object that it is measuring if its mass is significantly less than the object, there is good contact, the object is not in transient and it is in contact for long enough to reach equilibrium. Only when these conditions are met will it display the correct temperature. I would suspect that they are measuring an object that is cooling down, this means that it is in a transient state, and the measurement that are achieving is not accurate. Measuring something in a cooling transient may give a higher measurement as often the object cools faster than the thermocouple responds.

One Comment

  • Interesting article Bob.
    Is that thermogram by any chance a certain power station we are both connected with ? Looks familiar 🙂

Comments are closed.