Hello.

I have a FireFLY6 as a Y6 copter using a PixHawkY6B configuration with Arducopter 3.2 release

After minutes of flying level, the Y6 starts to roll and pitch gently out of level.  This gets worse over time.  After about a minute, full stick will no longer hold the craft level so it gets landed.  I have shown my logs to the mfg.  The mfg did not feel there was any problem with the PixHawk so will offer me no more help.

The tests were done with no wind.  And toward the end of the flight, I am keeping the craft level even though the actual roll value does not show that.  When I can no longer keep it level, I lower the throttle and land.

The flight mode was always just hover level assist.

I am asking if someone can please look at the logs and help me understand why late in the flight the copter no longer is stable.  I have charted the variables I think show the problem, but the mfg just shows me back charts of the accel cal being ok.  I would like to use 3.2 but don't know if it is stable for the PixHawk.  Thank you.

The logs are also attached.  3691205103?profile=original

Thank you very much.  

62.BIN

63.BIN

61.BIN

You need to be a member of diydrones to add comments!

Join diydrones

Email me when people reply –

Replies

  • I had a quick look at just 63, and it seems you are having a Gyro failure (not accelerometer) on the MPU6K.  It picks up a heavy bias when the trouble starts.  

    • Thank you for the info.  I appreciate your help.  I will study those variables.

    • @Rob: what causes it to fail like that ?

  • - usually such problem would be caused by too much vibrations.

    Yes, I've seen similar degraded AHRS performance on recent ArduPlane too, but not on ArduCopter

    I am not sure what kind of AHRS related processing was removed(?) from the code.  - but the EKF just gets better.

    if you plot log 63 , ATT.Pitch and EKF1.Pitch  you can see they disagree around line 89300   AHRS_EKF_USE would have saved the day for you.

    • Are you saying that EKF would have held the frame level when the gyro bias drifted?

      • Yes it definitely looks like that.
This reply was deleted.

Activity