We have a product using the BME680 and I was comparing the temperature and humidity reported from some recently built units (a couple of months old) vs. a group of units that are over two years old. The new group all report temperature and humidty that roughly agree with each other. The same for the old group. But the values for new vs. old are quite different. Based on what I'm seeing it appears that as the sensors have aged they report higher temperature and lower humdity.
Is there a known expected change in the values reported as the BME680 sensor ages? If so we can adjust the reported values accordingly.
Yes that is exactly what I did I had a group of 9 units total (4 new, 5 old) and had them all placed within a 2 square foot area with no heat sources around and next to them a couple of 3rd party temp/humidity sensors. The New units all reported a lower temperature and a higher humidity than the 3rd party devices reported. The Old group was the opposite they all reported a higher temperature and a lower humidity than the 3rd party devices.
Usually, we set temperature and humidity values in laboratory equipment, then test the temperature and humidity values output by the sensor and compare them.
Firstly, it is necessary to ensure hardware compliance with the HSMI specification(https://www.bosch-sensortec.com/media/boschsensortec/downloads/handling_soldering_mounting_instructi...) for design. Secondly, if you only need temperature and humidity data without using BSEC, you can refer to the sensor API and example code(https://github.com/boschsensortec/BME68x-Sensor-API) to ensure that you use official software.
If there are no issues with the above checks, you can record the sensor data and timestamp during testing for inspection purposes.
Thank you for the reply. But going back to my original question does Bosch have any data or guidelines about expected changes in the output of the BME680 as it ages? Just talking about the sensor itself and removing our design from the equation. Or is the expectation that the sensor output on day 1000 should match day 1?