Gyros Inconsistent after update to 3.6.8

To start with just repeat the calibration steps, and let me know of the result.

When I will have some free time I will recalibrate compass and try this firmware. Calibration of IMU is bigger problem because I will need much more time for disassembly copter.

Dear @sidbh is that possible that you will add code with better cold boot to 3.6.8 firmware ? not to 3.7 beta… It will be easier to resolve this problem.

Did the issue get resolved with the modified firmware?

With this 3.7 firmware cube behave differently each time I am starting it. Today I didn’t get any message “gyros inconsistent” I just get different messages like"PreArm: High GPS HDOP" maybe because I am testing my copter on balcony.

Yes, that’s the reason, you should either disconnect GPS while testing this or go to a more open area. The firmware image I shared is not stable, but any Prearm checks or faults it detects are real, and more correct than previous version.

I tried 3 times to start Cube after longer period (cold boot) and noticed “gyros inconsistent” when imu tempertaure was around 4372 in this 3.7 dev firmware.
Here you can download my last logs:
Maybe next week I will try to test this outside on open area.

The fix is not to fix it for cold boot. It fixes it so it won’t happen after reboot once temp is reached. Please go back and re read inter thread. You should never fly without reboot after temp reached.

You should never fly without reboot after temp reached

@hdtechk Where in the documentation is this procedure? I have never seen this. Yes it makes sense to do for the IMU, but it is never specified in the documentation that I have seen.

@hdtechk I am working on fixing that. @Norimbo is helping me test the modified firmware I created to solve the issue.

@Norimbo I looked at your logs. It seems the bias growth over temperature (which grew from 15deg to 45deg) for one of the sensor is just too much for even EKF to be able to estimate. Your best bet is to restart after cold boot. Since, the cube is heated, the issue won’t persist in flight. Other suggestion would be to reduce the target temperature (BRD_IMU_TARGTEMP), but ensure that you are not setting it below 15 degrees greater than Highest Operational Ambient Temperature.
Also, thanks for doing the tests! I will continue to evaluate this and see if there’s anything we can do to improve this in the future.

Also, for your reference. This is the PR which resolves the issue to certain extent:

@sidbh I had opportunity to check the copter on open field. First log is cold boot and at the end I calibrated the compass. Second log is just after reboot. Everything looks good I did also short low flight during second try.

Dear @sidbh is there any chance to repair this problem with Cube Black ? The problem still exist in Arducopter 4.06 I have to reset Cube every time. Can I help with development of the fix for this heating IMU issue ?