BME680 too much variation over a short period of time

Hello all, I have setup and configured BME680 air quality sensor with hope that I will get to know the in-the-office air quality. I am using bme680_bsec library in esphome configuration. My concern is the readings even in a short period of time are changing too much with nothing else happening in the room. For folks who have this sensor installed. Do you also see the same? At this point, I feel any data coming from this sensor is inconclusive. I have even tried to totally enclose the room, but lot of variation is still seen.

Yes, I am aware that when the AQ is 25 its calibrating but the other variations don’t look right.

Any guidance is highly appreciated.

Part of bme680 config below

bme680_bsec:
  address: 0x77
  sample_rate: ulp

Do you have this sensor housed in plastic case that may be out-gassing?

Also does the ESPHome IAQ sensor have an IAQ accuracy attribute?

The IAQ accuracy value is defined as this: https://community.bosch-sensortec.com/t5/Question-and-answers/What-does-the-IAQ-accuracy-mean-in-BSEC/qaq-p/5935

Without an IAQ accuracy of 3 your readings will be kind of meaningless.

Thanks Tom for your reply.

  1. The sensor is really out in the open. Yet to put it in any case.
  2. I have seen these cycles where it says “Calibrating” and then “Calibrated”… ( values 2 and 3 ). I agree, we need to discount values when accuracy is 2 (Calibrating). The vast fluctuations are seen when it’s done with calibration and accuracy indicates “Calibrated”.

This was an expensive sensor so hoping to get some meaningful data out of it.

Regards,

-Sam.

There are a lot of posts about it on the Bosch support site. e.g. https://community.bosch-sensortec.com/t5/MEMS-sensors-forum/BME680-IAQ-and-eCO2-worse-after-ventilation/td-p/20634

I’ve never liked their temperature and humidity sensors (though I might have been getting factory seconds from China), they seem to suffer a lot from self heating.

I find temperature and humidity ok but IAQ is just a wild roller coaster ride. This is closed room but I haven’t lit the log burner and I need to do some proper testing

Thanks Julian for your reply. With my sensor, the variations are even worst. In a span of 6 hours, I have seen it going from 50 to 400 and back… ( in a closed room with nothing going on )

I wonder if these sensors are just not something we should use. Or I am unable to find right way to use them.

That was the impression I got from the Bosch forums. This is not specific to ESPHome.

SCD40 is the ceiling sensor and other sensors are on a side wall in the same room. Bottom graph is C02. I find the measurements react wildly different when there is activity in the room; I came to the conclusion that the VELUX sensor is accurately calibrated and dampened vs my DIY effort.