Temperature-drift over time on single sensors

[Rudolph_Aschmoneit]

Hello everyone,

I installed 15 Ruuvitags in a laboratory for room monitoring at the end of October, this was the first stage of the project, we will extend the project even further, our customer uses the data to further optimize his production process. Now I have observed that a Ruuvitag since about November 13th shows a temperature drift of about one degree compared to the other Ruuvis. See also the diagram, the yellow line is the Ruuvi with the temperature drift, the Ruuvi with the red line is about 30 cm away and the rest of the room has a relatively equal temperature curve, except the purple one, which sits on a refrigerator and can be ignored here. Interestingly, the sensor only drifts with temperature and not with air pressure or humidity (see the other diagrams). All Ruuvis have the same firmware, 2.5.9. After a short review of the specifications of Ruuvitag and Bosch BME280, this drift is apparently still fully within the tolerances, although I didn’t find any exact information about the temperature drift over the lifetime of the sensor, only the general statement “Minimum/Maximum values of drift, offset and temperature coefficients are +/-3 delta over livetime” I would just like to know if there is an explanation for this behaviour, especially if the air pressure and humidity stay the same (because I read that the temperature value is used for the correction of the other values when I fly over it), if someone has already observed something similar or if someone knows the datasheet of BME280 better than me .

Many thanks in advance. Many greetings
Rudolph

[otso]

Hello,

That drift is very interesting, I’ve not seen temperature drift so quickly ever before. It almost makes me wonder if there’s something else going on, for example some sort of conductive contaminant onboard which would drain a little current and that would cause self-heating. If there is some current leakage, you’ll know in a few weeks or months as the battery will drain.

On the other hand, brand new RuuviTags often show a little lower temperature than ones from 2017 kickstarter times, so maybe most sensors will drift eventually? I’ve seen humidity having hysteresis, a tag coming from moist environment will show higher than a tag coming from dry environment when put side to side.

Pressure also drifts so much that tags cannot be used as differential pressure sensors for measuring home air conditioning without weekly corrections for drift.

[Rudolph_Aschmoneit]

Hey, thanks for your fast reply. And thank you for your thought’s. I will check from which production cycle the Ruuvi is, cause I think this is a mixed set up, from newer and older ones, but Im not sure at the moment, I will answer when I know more. The tip with the conductive contaminant is very interesting, but do you think is it better to wait and see what happens with the battery or should I just clean the board and see if it changes? also thank’s for the input with the drift correction in pressure management, we maybe want to measure air conditioning in this project in future, so I will read more about the drift correction and I think I have to go a little bit deeper in to the BME280 datasheet.

[otso]

@Rudolph_Aschmoneit

The tip with the conductive contaminant is very interesting, but do you think is it better to wait and see what happens with the battery or should I just clean the board and see if it changes?

I would just wait, chances of damaging the board while cleaning are pretty high and the problem could be something else.

[Rudolph_Aschmoneit]

Hello everyone,
sorry for the late feedback, but we had to continue to observe the process and draw conclusions, also because the new insights and their depth into our customer’s production line are new to him, it all took a little longer. It turned out that it was by no means a malfunction of the Ruuvi’s, but that the processes during bacterial growth created this jump in temperature and that this jump was not visible in the other Ruuvi’s, including the directly neighbouring Ruuvi, because there near by batches are in a different phase of the process. Through long-term analysis, we were now able to validate this reliably, since the temperature jump also occurs again over time and is repeated when the bacterial growth cycle is run through again, at the same time as the first observation. Our customer is very grateful for these new findings and the more detailed insights into his processes and we are happy that we were able to make it possible for him thanks to Ruuvi’s and are curious what we will learn together in the future, especially if we expand and improve the setup. We are very much looking forward to the Ruuvi Gateway and even more Ruuvi’s in our customer’s labs.
Many greetings
Rudolph

[otso]

Thank you for sharing the find, people getting new insights into their world through our sensors is what inspires us to do our work