How does AQ know when it is absolutely leveled?

[LPR]

I think it would be better for you to stay with the firmware Version v6.6.1 stable (b1036).
The newest firmware is for developers to test. The main control features are for the most part the same in the newer firmware.
If your parameter file is good and the AQ board is level in your frame then there maybe a problem with your board.
A log should help to understand is it's a motor, ESC or other problem.

Have you flown the hexa with a different flight controller?

Larry

[LordPikaChu]

[chschmid]

Looking at your LOG 140, the ACC is still far off, assuming the hexa was standing more or less level before liftoff. Furthermore I guess your MAG values are kind of wild. Do you have some powerlines or Tx modules close by the board?

Your vibs look fine to me. They can always be better though

Cheers
Christof

[Max]

Have you verified your calibration using any of the info available here: http://autoquad.org/wiki/wiki/autoquad- ... on-result/ ? Especially the ACC calibration check. Check the ACC values as recommended -- though you can do a rough check by leaving the board mounted as well, and using live telemetry view is pretty quick.

Also, using telemetry view, check the ACC Magnitude and MAG Magnitude values. The former should be close to 9.8 (1G) when the AQ board is in ANY orientation and not moving. The latter should be close to 2 (Earth's magnetic field). You can also compare AQ_ROLL/PITCH with ACC ROLL/PITCH -- the former is derived using all available sensor data while the latter is derived purely from ACC data. This can tell you how much the MAG sensors are influencing the final solution.

-Max

[Max]

I added the magnitude check info to that docs page, screenshot and everything -- http://autoquad.org/wiki/wiki/autoquad- ... -influence

-Max

[LPR]

Pika

Max and Christof have given better advice then I can.

Photos of your Static and dynamic Logs curves and of your Hexa may help us find more of that is causing the level problems.

Larry

[LordPikaChu]

[srinath]

One cause for the drift problem may be that the board is not mounted perfectly level. Your calibration may not be too bad but if your board is tilted in some direction by 5 deg, then a sticks neutral command from the radio will try to level the AQ board, which obviously means the craft will tilt and hence move.

This is always a problem when mounting with screws and standoffs. A small difference in height or not screwing a standoff fully into another could cause such problems.

How about trimming it away using the radio or rechecking whether the board is actually level

[LPR]

Mateusz

The other problem I see is that you have the wires carrying power way too close to the AQ board. Twisting the power wires would help but I would make a new wire harness that has twisted wires and that follows around the bottom outside of the carbon fiber center plates. The only power wires that need to be close to your AQ is the twisted pair that power the AQ board.

Having your power wires twisted and a few inches away from the AQ board should get the mag working very well when you power the motors. The power wires from the battery should also be twisted.

With a good parameter file, a level AQ board and the wire harness away from the AQ board your copter should fly quite level. To get it to be very level you can fly it outside with GPS and after a minuet or two it will autotune it's level even better.

Larry

[kylekyle]

my hex has the same problem. it drifts on manual mode. when i check mag--magnitude, it varies alot depending on orientation (it suposed to be constant 2, right?)
thus when i am trying new dynamic calibration, step 3 mag calibration gives me trouble. after about 5rerun, MAE does not decrease and 6th run never ends. what MAE error do you finally get?

i am womdering if this gets better if i remove gps sheild. my gps sheil is very big and it might interfere mag calibration?