thermal calibration issues

[jbourne15]

I am trying to perform a [offboard thermal calibration on my pixhawk 3] with v1.9.0 firmware (https://docs.px4.io/en/advanced_config/sensor_thermal_calibration.html). And I am getting step like response for my bias compensation (see attached figure). Notice the z acceleration bias is very large. Looking at the flight review it seems that the raw data is fine and has temperature drift, but the estimate rpy angle and rate go crazy.

Why do I get step like responses for my bias curves? Is this a hardware or software problem?
Has anybody got this issues before and if so how can I resolve it? This problem exist and my 3 new pixhawk 3 systems.

log001.ulg.pdf

thanks

[dagar]

Interesting that it's only on the 2nd sensor (both gyro & accel). Is this reproducible?

Also note that your baro cal looks completely bogus.

[jbourne15]

Yes it is repeatable. Probably tried it 4-5 times now.

[jbourne15]

Does the pixhawk accelerometer and gyro need to be previously calibrated before I thermally calibrate? According to step 2. I need to set all CAL_GYRO* and CAL_ACC* parameters to defaults. But I am wondering if CAL_ACC*_ID and CAL_GYRO*_ID needs to be set up to perform a thermal calibration.

[jbourne15]

I forgot to mention that I am performing the offboard calibration procedure.

[jbourne15]

After talking with Drotek, it seems they have some production issues with their latest pixhawk 3 units (IMU/IMU temperature sensor). So this may be just a hardware issue.

[julianoes]

@jbourne15 That's strange but thanks for the note.