BME680/688 IAQ - results of the index are counterintuitive

[G6EJDdavid]

I'm using the BME680 or 688 together with the BSEC library to obtain an IAQ value, everything appears to work as I expect, the index starts low at a steady 25/27 and then after some time starts to report a varying IAQ, apparently with the changing enviroment.

However, I noticed that when an SCD41 measuring CO2 besides the BME680 starts to detect an increased value of CO2, the BSEC IAQ falls, this does not make any sense to me, surely when in an enviroment with high or increasing CO2 levels, should the IAQ increase and not decrease as it does at present?

My logic is the IAQ should climb as CO2 concentrations climb...

The graph on the left is the BME680 CO2 ppm, on the right is BME680 IAQ and at the bottom is the SCD41 CO2 variations in the same environment.

 

You can observe that as the SCD41 reports rising CO2, the IQA falls as does the BME680 CO2 equivalent?

What am I missing here?

Has the BSEC library been validated and are there any results of the validation?

[BSTRobin]

Hi G6EJDdavid,

From the graph you sent, it can be seen that the CO2 values of BME680 and SCD41 have different variations trends, and it seems that they are not in the same environment. But from the variations trend of changes in CO2 and IAQ of BME680, their changes are synchronous, with CO2 increasing or decreasing and IAQ also increasing or decreasing.

[G6EJDdavid]

Actually the SCD41and BME680 are next to each other, it did that for this trial. You will see that the readings are time coincident, albeit as the SCD41 CO2 readings drop the IAQ rises. I also have an SCD40 running which validates the SCD41 values. I have not seen any validation results that demonstrate the IAQ results, which is why I’m querying the IAQ results!

[BSTRobin]

Hi G6EJDdavid,

If SCD41 and BME680 are are next to each other, but the trend of CO2 value change is inconsistent, it is necessary to check what causes the difference.

If the CO2 change trends of SCD41 and BME680 are consistent, it would be more meaningful to look at the changes in IAQ.

[G6EJDdavid]

That is why I uploaded the variation in IAQ in the three graphs above, you can see that as CO2 levels drop the IAQ rises, why is that?

It seems there is something strange with the BSEC algorithm, which is why I am asking if the BSEC has been validated with various test cases, for example when CO2 levels are high, low, or that the results are consistent with other sensors.

I have not seen any material published to determine the validity of the results...

SAGP40 VOC results

BME IAQ Results

Both sensors are in close proximity with the SCD41 refrerenced above, so two sensors are showing simialr results, but in all scenarioes the IAQ value response is the inverse of poor air quality.

This is the SGP40 result versus the BME680 BSEC result for IAQ, note how as the SGCP40 sensor detects a lowering VOC, the BSEC results start to increase, this cannot be correct and it looks like there is an inversion in the BSEC library.