AQ M4 compass/yaw problems

[saihv]

Hello all,

As part of the M4 setup, I did the DIMU tare and the magnetic calibration (moving around all axes for about 1-2 minutes until the lights become solid and QGC says finished calibration). The problem I'm having is that even after the magnetic calibration, the readings are 180 degrees off : It thinks S is N and such in the HUD. I tried the calibration twice with the same results. When I fly it in manual mode, the quad takes off and in a second or two, tries to do a full 180 turn, or sometimes drifts back and forth in the yaw axis. The IMU orientations and flip values are set correctly. It'd be great if someone can let me know if this is a known problem/ of any possible solutions.

Also, when I look at the Pilot window of QGC and the "horizontal situation indicator", only the values POS and GPS are green. RAT, ATT, PXY, PZ and YAW are gray and say off , and everything else is crossed out and yellow. Not sure if this is an issue and might be related..

Thanks,
Sai

[Max]

Hello Sai,

I wouldn't put too much faith into what the HUD reads in terms of compass heading, at least until you're actually flying with a decent GPS lock (the MAG sensor is notoriously inaccurate especially on small frames). But what you're describing in flight doesn't sound right at all. My guess would be that it is not related to the MAG sensor... especially since manual mode is just that -- manual.

You could look around the forums for troubleshooting ideas, and/or post more details on your setup, like saved params file and photos, a log from a short flight if you have one.

Ditto on the HSI indicators, AQ doesn't really send that info to the GCS (yea we should probably just remove them, then).

Cheers,
-Max

[saihv]

[Max]

I should mention that you can check/verify your mag calibration -- http://autoquad.org/wiki/wiki/configuri ... -influence If it is wildly off, I suppose that could have a negative influence overall. I haven't personally seen that in a long time (the firmware code used to rely more on the mag sensor but that was several versions ago), but who knows.

Good luck with the troubleshooting!

-Max

[saihv]

Hi Max,

I did check those values after calibration: it was close to 2.0 as it should've been.

The first thing I wanted to try today was to remove the GPS (the GPS module has a magnet inside it); and when I fly it after removing it, manual mode was perfect again. I have a different set of questions now, it'd be great if you could give me some of your views: so even manual mode does become worse with magnetic interference? Also, when I did the mag calibration yesterday, the GPS was on the quad, so shouldn't the offsets take this close-up magnetic influence into account as well?

Thanks again for your time.

- Sai

[Max]

Sai, what do you mean by "remove the GPS?" The antenna? Maybe the difference is the weight? What exactly are you using?

You really need to post more details on your build and settings before anyone can help you in a meaningful manner. There's dozens (hundreds?) of variables that can influence any setup.

-Max

[saihv]

Hi Max,

Yes, by removing the GPS antenna, the problem seems to be solved. This is the antenna I have on the setup.

https://www.sparkfun.com/products/464

I just powered on the quad and rechecked the magnetometer values, and it looks like I was mistaken when I said the calibration values were okay: the values are between -2 and 2 when the GPS antenna is not on the quad, but some of them are going down to almost -4 when I put the antenna back in. I am linking a few pictures of how the setup looks like, the IMU.MAG value graphs and also the view of the frame setup in Qgroundcontrol. Most of the settings apart from the frame/motor setup are just the defaults. I am using the M4 with the motor pwm lines attached to the N4 expansion board, and with an external ESC. The frame is a QAV 250mm frame with 5 inch propellers.

http://imgur.com/a/KkQE3

Please let me know if any more information would be useful.

Thanks,
Sai

[aBUGSworstnightmare]

[Max]

Well you put a VERY strong magnet right next to the sensor... what did you expect? The 5m(!) of cable aren't doing you any favors either.

You're also upsetting the balance of the whole quad by mounting that heavy thing in the front like that, and all the cable on one side. Not to mention adding a bunch of weight. On an overweight or out of balance MR, the yaw is typically the first control to suffer.

Bottom line, use one of the recommended GPS antennas and mount it on a shield, as recommended. Redo your Mag calibration with the final result (or in the meantime redo it w/out antenna and fly like that).

BTW to check the mag calibration you want to plot the MAG Magnitude in various orientations (check the instructions again), which is essentially like an average of the 3 axes. But your plot with the antenna mounted is obviously way off anyway.

Cheers,
-Max