Attitude bias and high drift in APM 2.5

I am in the process of setting up my 3DR Hexacopter and I am having some problems with my APM. I tested the sensors as it says in the tutorial a few days ago and everything looked fine as you can see in the link below. This week I finally got my tranmitter and I decided to work on setting up the RC control/Transmiter/APM working. I was able to get it working however I noticed that my HUD for the attitude looked weird as seen in the second video.

APM 2.5 a few days ago

https://www.youtube.com/watch?v=Wrs4PzPwmhQ

APM 2.5 now

https://www.youtube.com/watch?v=qpIAZHHrlZ4

Any ideas what this can be? Any other data that could help?

This is my first post, I am hoping I posted on the correct place :)

Regards,

I fixed the problem, more than likely it was me committing a newbie mistake. I am unsure what I did that corrected the misalignment but this is what I did in case anybody gets my same problem.

1. Using APM Planner I ran the calibration for the magnetometer.

2. I disabled the magnetometer

4. Disconnected the apm from the usb then reconnected it.

3. Reloaded the firmware (I most have done this like 20 times in the process of trying to fix it but go figure why this time it fixed it).

4. Connected back to the APM through the usb.

5. To my surprise then the virtual horizon started working properly again.

It is worth mentioning that before this I couldn't find any weird drift on the accels or gyros. However the nav computing for roll would compute a noisy constant slope even when the APM was static on my desk. My best guess would be that my magnetometer was acting up? I was able to get the apm working before going in to detail on the magnetometer measurements but I didn't see anything when i gave the measurements a quick look. It is worth mentioning that I am not very familiar with the outputs for the magnetometer at this point. I am looking forward with playing around with sensor blending AFTER I get the thing flying :) Again it is highly likely that it was my fault and not the apm (I am rather new to it), but it could also be some random bug I stumbled upon?

Thanks to those who made comments on the video.

Regards,

Heimdall

Replies

Elektro October 20, 2012 at 8:20pm

I fixed the problem, more than likely it was me committing a newbie mistake. I am unsure what I did that corrected the misalignment but this is what I did in case anybody gets my same problem.

1. Using APM Planner I ran the calibration for the magnetometer.

2. I disabled the magnetometer

4. Disconnected the apm from the usb then reconnected it.

3. Reloaded the firmware (I most have done this like 20 times in the process of trying to fix it but go figure why this time it fixed it).

4. Connected back to the APM through the usb.

5. To my surprise then the virtual horizon started working properly again.

It is worth mentioning that before this I couldn't find any weird drift on the accels or gyros. However the nav computing for roll would compute a noisy constant slope even when the APM was static on my desk. My best guess would be that my magnetometer was acting up? I was able to get the apm working before going in to detail on the magnetometer measurements but I didn't see anything when i gave the measurements a quick look. It is worth mentioning that I am not very familiar with the outputs for the magnetometer at this point. I am looking forward with playing around with sensor blending AFTER I get the thing flying :) Again it is highly likely that it was my fault and not the apm (I am rather new to it), but it could also be some random bug I stumbled upon?

Thanks to those who made comments on the video.

Regards,

Heimdall

This reply was deleted.

Discussions

Attitude bias and high drift in APM 2.5

Replies

Activity