Bosch Sensortec Community

    Showing results for 
    Search instead for 
    Did you mean: 

    BME680 TVOC value maxed out

    BME680 TVOC value maxed out



    For a client, me and my team developed a product, based on BME680, controled with ESP32-C3 MCU using the latest version of the BSEC library.

    However we are experiencing a problem, that sometimes the value of the TVOC goes way above normal values and stays at maximum value, until the whole device is reset.

    Normal values of TVOC-equivalent (bVOCeq) is between 0.5 and 5, but in our case goes to 200 and even to 1000 (maximum value). We are not using the IAQ or CO2 estimates in the application - thus I don't know their values.

    So far most of the cases happens when the sensor is powered and working fine (read values every 8 seconds) and we restart the ESP32-C3 (soft reset), which is forcing the sensor to recalibrate and after the calibration process - the uncommon values starts ramping.

    Since the product is entering into mass production at the moment, the randomness of this event becomes unpredictable and we can't ask the end user to turn on and off the device. I don't know where to look for the issue, since it is hard to replicate on all of the demo boards.



    3 REPLIES 3

    Community Moderator
    Community Moderator

    Hi nibopet,

    Which BSEC version, configuration, power mode you used?

    Hi BSTRobin, 

    Thank you for the reply.

    The BSEC version is 2.4.0

    The power mode is: BSEC_SAMPLE_RATE_LP

    The configuration function is:

    bsecSensor sensorList[] =
        };    if (!envSensor.begin(BME68X_I2C_ADDR_LOW, Wire))
        }    if (!envSensor.updateSubscription(sensorList, ARRAY_LEN(sensorList), BSEC_SAMPLE_RATE_LP))
        }    envSensor.attachCallback(BSEC_DataCallback);



    We are comparing a prototype board with first series of production boards and the only difference we see in the performance is that on the production series board the QC sticker is in the range of 1cm of the BME680 sensor. Could it be possible that when the sensor starts operating, it heats little by little particles of the glue, which cause the maximum values?

    During the prototyping stage we got such values only once - when a sensor was soldered with excessive flux on it. When washed out with alcohol, the problem was resolved.