BMI088 sensor data discontinuity when lower the temperature

[Brucelgn]

We are using BMI088 by SPI to get the accelerometer and gyroscope data. The ODR is 500Hz. When the temperature goes down from above 65℃, all the sensor data smoothing by 1 sencond seems to be discontinuous at some moment. The figure below shows the case. The raw data is normal. Can anyone explain what happens?

[FAE_CA1]

Hi,

Thanks for your inquiry.

In your message you said that BMI088 accelerometer and gyroscope raw data look normal when temperature goes down from above 65C. However, when you plot accel and gyro data with 1 second smoothing, you saw some sudden jumps.The issue could be overflow in your smoothing formula.

Thanks.

[Brucelgn]

Hi,

Thanks for your reply.

The smoothing method we are using is examined and works well. The raw data looks normal just because the noise is too high.

The sudden jumps will disappear when the temperature continues going down. In the above case, the smoothing data shows large to 4mg change in accelerometers and 0.1°/s change in gyroscopes.

[FAE_CA1]

Hi,

Thanks for your BMI088 gyro raw data plot and temperature plot. I assume that the vertical axis unit is dps for gyro x/y/z raw data and C for temperature data.

For BMI088 gyro x/y/z raw data there are a few large jumps when temperature goes down from 20C to -20C. Maybe there are some vibrations created by the temperature chamber that causes gyro data jumps. Or maybe you are using 2000Hz ODR with 523Hz bandwidth unfiltered data for gyro. Please try to set BMI088 gyro to 100Hz ODR with 12Hz BW as shown below to see if there are still some sudden jumps or not when temperature goes down. The typical BMI088 gyro zero-rate offset is within +/-1dps and min/max +/-3dps. So if gyro x or y or z axis raw data goes beyond +/-3dps range, then this BMI088 is a failure part and we need to get it back from you to do failure analysis.

Thanks.

[Brucelgn]

Hi,

Thanks for your advice. The vertical axis units are as you point out. Maybe the temperature chamber vibrations cause gyro data jumps. But it is strange that gyro X/Y data will go back to the proper trend after a short time.

The current internal ODR is 1000Hz with 116Hz BW. We will try to set  gyro to 100Hz ODR with 12Hz BW and repeat the previous process.