BME688 totally inaccurate Sensor Data

Breakout boards, sensors, other Adafruit kits, etc.

Moderators: adafruit_support_bill, adafruit

Please be positive and constructive with your questions and comments.
Locked
User avatar
SomeNix
 
Posts: 4
Joined: Sat Mar 25, 2023 6:20 am

BME688 totally inaccurate Sensor Data

Post by SomeNix »

Hi,
i purchased a Adafruit BME688 Sensor at BerryBase and would like to refund that product.
It does produce totally inaccurrate Co2 and Volatile Organic Compound Data.

As you can see in the attached image, i do have an Awair Element, which has a dedicated Co2 Sensor that should quite be accurate and does produce plausible data. e.g. When i open the window the Co2 value of the Awair Elements drops and the value of the BME goes UP?
Screenshot 2023-03-25 112821.jpg
Screenshot 2023-03-25 112821.jpg (161.48 KiB) Viewed 422 times
The BME Accuracy Level was always at 3 in the shown Messaurements. See also attached.
Screenshot 2023-03-25 113451.jpg
Screenshot 2023-03-25 113451.jpg (153.39 KiB) Viewed 421 times
If you need further data or information, im happy to provide that for you

User avatar
mikeysklar
 
Posts: 13936
Joined: Mon Aug 01, 2016 8:10 pm

Re: BME688 totally inaccurate Sensor Data

Post by mikeysklar »

Thank you for the graphs.

Is it fair to compare a dedicated CO2 sensor to a VOC sensor? I would expect the data to be different if the VOC is including other organic compounds.

Adafruit does sell TRUE CO2 sensors and they cost 3x the price.


We can start the refund process for you if this sensor was purchased in the last month. What was the date of purchase?

User avatar
SomeNix
 
Posts: 4
Joined: Sat Mar 25, 2023 6:20 am

Re: BME688 totally inaccurate Sensor Data

Post by SomeNix »

The package was delivered on 15.02.2023. I did not try the sensor at first and then waited a few days to ensure that the burn in and accuarcy can build up.

To be honest, i did not expect readings that are as accuare as dedicated CO2 Sensors, allthough Bosch claims very accuarate messurements for that sensor...
I did expect, that these readings were in a matter, that is plausible to the cause.

As i stated before you see in attached pictures, that the sensor shows a rapid increase in C02 Levels, when i open a window, and naturally the dedicated sensor shows a decrease in CO2 Levels, as it should be.
Its not like the sensor is a little bit inaccurate.

For example:
28/02 8am: BME688 ~800ppm; Awair ~2000ppm = Diff >~1200ppm
25/02 8am: BME688 ~500ppm; Awair ~2250ppm = Diff >~1750ppm
While the Awair Readings seem to be quite plausible (e.g. door opens, door closes, humans in room, window open etc.), the BME Sensor seems very random.


The messurements are completly useless as such and do not show in any particular way how the CO2 Levels are in reality.

User avatar
mikeysklar
 
Posts: 13936
Joined: Mon Aug 01, 2016 8:10 pm

Re: BME688 totally inaccurate Sensor Data

Post by mikeysklar »

You are outside the Adafruit warranty period for replacement.

It is not appropriate for you to continue to compare VOC to CO2.

User avatar
SomeNix
 
Posts: 4
Joined: Sat Mar 25, 2023 6:20 am

Re: BME688 totally inaccurate Sensor Data

Post by SomeNix »

As i stated i do not compare VOC to CO2, iam Comparing CO2 to CO2 equivalent. The sensor creates a CO2 Equivalentmessurement based on some freaky AI thing by Bosch.
I can surely can give you the VOC data which is as inaccurate as the CO2 Data, when you need me to.


So do i see that correctly, that you refuse to refund a broken sensor, which gives me crappy output data?

User avatar
SomeNix
 
Posts: 4
Joined: Sat Mar 25, 2023 6:20 am

Re: BME688 totally inaccurate Sensor Data

Post by SomeNix »

?

User avatar
millercommamatt
 
Posts: 831
Joined: Tue Jul 31, 2018 4:57 pm

Re: BME688 totally inaccurate Sensor Data

Post by millercommamatt »

Based on everything you've shown, your BME688 appears to be working as it should.

This sensor does not measure CO2. It says so in the data sheet [https://www.bosch-sensortec.com/media/b ... -ds000.pdf] (e.g. pages 16, 31). What it does do is estimate CO2 by measuring the VOCs in the air and then assuming that the VOC source is human exhalation and using built-in correlations of VOCs to CO2 in human breath to estimate the CO2 concentrations. If the source of VOCs measured by the sensor isn't human breath or something equivalent, it will be wrong.

In your case, when you open your windows the primary source of VOCs measured by your sensor is from outdoor air pollution whose VOC levels correlate to CO2 differently than human breath. Thus, the sensor says estimated CO2 goes up when VOCs go up, which is incorrect in this case, because the underlying assumption behind the estimate isn't valid.

It's important that you understand how your sensor work and understand what they actually measure. Read the datasheets! In your case, your sensor works fine. You just didn't understand what it actually measured. Hopefully you now understand and have the proper expectations for what your sensor can do and for which applications it's best suited.

The CO2 estimates from this sensor will be most accurate in an indoor setting where the dominant source of VOCs and CO2 are people. For example, this would be suitable for monitoring a conference room.

Locked
Please be positive and constructive with your questions and comments.

Return to “Other Products from Adafruit”