Pixhawk 2.1 cube wandering within a large radius?

hi all,
Just finished my build and got it into the air for the first time, the aircraft seems to be “stable” (as in, doesn’t oscillate or twitch) BUT, it seems to want to wander in a large radius from its GPS point in modes like Loiter. I set it into autotune mode on chan 7 of the TX, and it began tuning, but then just randomly fell from the sky, so I am currently fixing the motor arms that came loose and such, luckily no major damage.

I have a fairly small log file to share, in the hopes that you guys can point me in the direction of why this guy is wandering so much.

the frame is a 350mm quad frame, with MRT 800kv motors, BLHeli_S 35a 4 in 1 esc, and controlled by a spektrum dx9 with a spektrum receiver.

I am running on APM 3.5

Any info is appreciated!

Log link: https://www.dropbox.com/s/njlfs81tmt3bm0b/2018-10-03%2014-20-22.log?dl=0

There’s some Z vibrations you might be able to make better, but it seems RCout 1 to 4 (outputs to ESC’s) are rising steadily during autotune and eventually hits maximum and then the FC cannot maintain control.
It’s like the ESC’s or Motors are overheating and losing thrust as time goes on.
There’s no battery voltage or current recorded in this log - that would be very handy.
Be sure you’ve run the RC calibration and then ESC calibration.

RCin 3 is up past centre as if the craft should be in a gentle climb, yet altitude is fairly steady - but I’m not sure if that’s pilot-commanded, or if it’s an effect of the FC learning the hover throttle.

For autotune, get some good altitude first, like 10 meters or so - this allows for plenty of maneuvering and gets you right away from any ground effect and up into good GPS signal. Any manual tuning you can do is highly beneficial and you may not need autotune. In my case autotune made things worse, or at least strange, and I didn’t feel like running autotune multiple times when I didnt really have any issues in the first place.

1 Like

Hey Shawn,
Thanks for your response, we were actually recording the aircraft earlier (I got it to fly a bit better by increasing the P and I main loops) but it was still trying to walk around a LOT and was kind of oscillating pretty bad. My friend recorded it and said he could see one of the motors (the left rear) seemed to look different in the camera like it was moving either slower or faster than the other 3.

What could possibly cause this do you think?
I have calibrated the RC and the ESC (it’s a 4 in 1) so I don’t think that should be the issue. But I honestly don’t know what could be causing it.

I was a little concerned that my motor mounting bolts might be a bit too close to my coil windings, and could cause interference, but is that a realistic worry that could cause this?

I just don’t know where to start.

Hi friend, first stabilize, then alt hold, then a GPS assisted mode . (reading might help) !
A full (sudden) deflection of the sticks in roll & pitch , and if IT is in air , it is ready for an auto tune !!!,

1 Like

hi Hari,
when you say stabilize, then alt hold, then GPS assisted… do you mean those are the modes I should tune it in to start with? thanks!

Update -> this is solved, GPS was getting interference, once it was raised up on a mast, and the pixhawk was reset to the stock tune, it flies beautifully!