EN CN
Horizon drift problems with latest firmware (2.56b7)
  • I have 2 gimbals with 32bit controllers with encoders in one of them that I use in a multirotor.
    Since the upgrade for 2.56b7 I am unable to control the horizon drift when flying. The roll moves to one side and I did try every trick I thought of to minimize. I did the normal 6 point gyro/acc calibration, the temperature calibration also on both IMUs meanwhile tried lots of "gyro trust" values.
    Nothing seems to minimize the roll drift.
    But there is something that puzzles me, that when the drone is tilted because of the wind and stationary (since no accelerations are being made) the horizon is still tilted. And this I can not reproduce in the ground.
    Is there any new "gyro trust" variable I should know about?
    Is this a side effect of the improved positioning algorithm?
    Or is this only in my gimbals?
    Thanks
  • Have you used the same version of the GUI than FW. not doing so can cause problems.

    Try reseting ACC and Gyro and recalibrating, if that does not help save your datas and erase EEPROM and set everything up again.

    Before doing that you could verify your IMU readings using the real time data. Sometimes IMUs can go broken and need replacement.
  • Hello Garug,
    Thanks for your input.
    FW mismatch was checked already, so that I avoid using the basecam AP for android because it wasn't update for the last FW version.
    EEPROM reset was one of the first "radical" solutions I tried and I restored everything manually (didn't use the settings file).
    IMU's seems ok in the real time data...
    I have also replaced the main control board for a new one and it is all the same with version of the firmware.
    Do you think I should try anything else?
    Thanks
  • " when the drone is tilted because of the wind and stationary (since no accelerations are being made) the horizon is still tilted. And this I can not reproduce in the ground."

    This is very weird. It is normal that on lateral accelerations the roll and tilt drift, but just tilting the gimbal should not cause it, airborne or not.

    Make sure that you have no weird stuff going on with RC control, mixing on transmitter, etc.

    Other than that it could be vibrations, but that severe vibrations would show also on video. There could also be electrical interference. How do you power the gimbal? Try with separate battery or at least main battery balancer port for powering the gimbal. Electrical interference could cause I2C errors when airborne and they could cause problems.

    Try different IMU, Do you have frame IMU? is it also for sure ok and well calibrated.
  • Hello,
    I did a full reset 3 more times and on the last one didn't calibrate with temp comp. and it seems to be working a lot better.
    Thanks for your help Garud
  • My experience is also the same that everything is working better in new firmware without any temperature calibration.
  • The temperature calibration has been working good for me, but needs to be carefully performed.
  • Hi there!
    Same problem here. I was flying withthe normal encoder enabled board. Yesterday i installed the "extended" version. I carefully perfomed temp calibartion and Gyro and Acc, 6-axis... everything.
    The horizon is worse than before. Then I deleted everthing (calibrationwise) and made a sngle Z+ Axis calibration. It was almost better than before... strange.
    Now there is another problem I observed: when I tilt down 70-90 deg. the gimbal also moves the roll axis and when I tilt back up again the horizon is 10-15 deg off.

    Thanks for your help!
  • How did you perform the calibration, I assume IMU disconnected from gimbal?

    How have you installed the IMU? screws can easily pend the IMU and totally destroy any performed calibration. I mostly use thin two sided tape, or IMU installed on separate block.

    Also, it is important that IMU angles are correct related to the frame.