BME680 how to consider the gas sensing values

[LeonardoPapale]

Hi!

i am using the BME680 sensor for a while now, i like it.

I am getting good values in Temperature,  Relative Humidity and Pressure. Now I am considering the Gas values. In the first graph there is the Iaq accuracy in time, 250000 seconds so around 70 hours. In the second graph the iaq values. Now i woud like to understand when i can consider accurate the iaq values. Can i consider correct the values when iaq accuracy is 2/3 and not accurate when iaq accuracy is 0/1? another question is about the Calibration, at first i tried the sensor for  some hours and get always iaq accuracy=1, then i red in the forum that it needs some stimuli to get a correct calibration, so now i used a bit of paper wet of alchol. This time the iaq accuracy in few minutes reaches value=2 and then 3 but after few days it returned to 1. What caused this behaviour? And how can i avoid it to have always an accuracy=2/3 and a valid iaq value? Values obtain are in the range 0-250 and values higher than 250 seems to be clipped to that value, but from datasheet the possible range is 0-500, is it normal? am i doing something wrong? Values of CO2eq and Bvocs are always clipped to 0, and i know for sure in that 70 hours CO2 increased a lot, maybe there is a enable parameter that i am not considering? I am using the C driver and the example from Bosch without differences in the code. Last thing i am not using the save/state load state functions because i am using the Lite Version, can that functionality solve some of these problems?  i would like to avoid the alchol thing every time i power on the sensor, could be possible with these functions? 

sorry for the ammount of questions, but i would like to use this sensor in best way possible, and i am not finding info about these arguments.

Here an example of CO2eq and Bvocs values:

BME680:
Time: [50:27] IAQ accuracy: (3)
Temperature: (22.60310 C), RHumidity: (81.26712)
Raw Temperature: (23.63362 C),Raw RHumidity: (76.59995)
IAQ: (243.50282), Pressure: (101148 Pa)
C02eq: (0.0 ppm) , BVocsEq: (0.0 ppb)
Static IAQ: (155.51780) , BSEC status: (0)

[handytech · Accepted solution]

---

@LeonardoPapale wrote:

Now i woud like to understand when i can consider accurate the iaq values. Can i consider correct the values when iaq accuracy is 2/3 and not accurate when iaq accuracy is 0/1? another question is about the Calibration, at first i tried the sensor for  some hours and get always iaq accuracy=1, then i red in the forum that it needs some stimuli to get a correct calibration, so now i used a bit of paper wet of alchol. This time the iaq accuracy in few minutes reaches value=2 and then 3 but after few days it returned to 1. What caused this behaviour? And how can i avoid it to have always an accuracy=2/3 and a valid iaq value? Values obtain are in the range 0-250 and values higher than 250 seems to be clipped to that value, but from datasheet the possible range is 0-500, is it normal? am i doing something wrong?  i would like to avoid the alchol thing every time i power on the sensor, could be possible with these functions? 

---

• The IAQ accuracy definition was indeed added to the Q&A section in this post. Hopefully that can help clarify the meaning of each possible value.
• The alcohol trick can be used to validate that BSEC is running properly, but shouldn't be used as a method to calibrate the sensor and could explain some of the described behavior. Since BSEC's calibration process is self-calibrating based on its background history, the performance of the IAQ output will be influenced by such a calibration method.

---

@LeonardoPapale wrote:

Last thing i am not using the save/state load state functions because i am using the Lite Version, can that functionality solve some of these problems? 

As BSEC seems to be running without any error, the described behavior doesn't seem to be linked to the usage of the Lite version. Note that using the Lite version does prevent you from loading any config string, and therefore from possibly increasing the history period considered for the background calibration.

---

@LeonardoPapale wrote:

Values of CO2eq and Bvocs are always clipped to 0, and i know for sure in that 70 hours CO2 increased a lot, maybe there is a enable parameter that i am not considering? I am using the C driver and the example from Bosch without differences in the code.

---

Based on your overall behavior, it seems most likely that these outputs were not enabled in your applications, and you should look in the direction pointed out in this reply. In the meantime, we have now enabled these two outputs by default in the sample code of the current (and upcoming) BSEC release(s).

[LeonardoPapale]

Thank you for the valid answer!