Using thermocouple with Arduino MKR zero & Therm shield

Dear all,

I'm currently trying to use a K thermocouple connected to an Arduino Therm shield. Using official Therm shield library, temperature values display in serial interface, but values aren't accurate for this kind of sensor, fluctuating from 25°C to 35 °C. Current ambient temperature is around 30°C, but it doesn't stabilize, I've tested my K thermocouple with a dedicated multimeter, confirming thermocouple was stable and OK. I've also tested another Therm shield, that gives the same results, and finally tested with another K thermocouple, which isn't the same as the first one (new reference), with the same result.

Looking for some information on the internet, I've read that MAX31855 component stability could be improved by adding a capacitor betwwen mesuring entries, but it doesn't improve the quality of my displayed values.

Does anyone have encountered similar troubles?

Best regards,


The voltages over the thermocouple are very low. So if any source of electrical noise is in reach the results will be affected.Is your complete setup within a shielded case?

BTW, what accuracy do you expect from your setup?

I Pylon, thank you for replying. Actually I did not test in a shielded case, I'll try it soon. I've also read on Google that some max31855 ship encounter problems of accuracy. I've bought some new ones to have a try, I'll keep you in touch. I'm quite surprised because I thought that the shield would be more easy to stabilize and more accurate. In terms of accuracy, using a thermocouple should bring to something closer to 1 degree Celsius either than 5 degrees Celsius. The max ship has its own measurement mistake, but I'm quite sure I can improve final result quality.

I'll keep you in touch



I've received my new MAX31855 chips, but it was exactly the same serie as the original ones. Hence, I did not make any hardware modification, and tested the shield again... And it is much better !!

There are 2 differences between my last test and today : The place where I'm testing from, and the computer I'm using to monitor the shield. I though troubles would come from EMI noise, but I've noticed that accuracy drops down when my computer increases its computing tasks and starts ventilating much more, so the quality of USB powering could be a good reason to get a decreased measuring accuracy.

I'll try it again in the same place where I performed my first tries, and also try again with the original computer to identify the problem source. But globally the problem might come from the quality of EMI shield or powering noise.

Thank you for helping !