[joani]

Okey friend changed the orientation of the mag, but this is pointing forward like the controller of sea 0 degrees, because it should change the orientation of the mag ??, do not make sense

[TheOtherCliff]

Different board inside different brand GPS has magnetometer mounted different ways.  Some need 0,0,0 and some need 0,180,0 because mag chip is mounted differently.

Real question is "did it fix it for you?" 

[joani]

Okay but the module comes with an orientation arrow, in apm2.6, the gps module has an arrow and this must be in the same orientation as the controller. Forward, but good friend I'll see what happens

[TheOtherCliff]

All orientation arrows on all boards should point forward, but then the configuration needs to be told that THIS kind of GPS/mag has the mag chip mounted upside down, so you tell the configuration that the mag board rotation is 0,180,0.

[joani]

Hello, do not log the horizon in gps 13 mode, calibrate the sensors, but when we return to the screen of prisiod this does not stay, not so with the other modes, the complementary for example

[f5soh]

Looks normal, the complementary only uses the gravity measured from Accelerometers and INS13 as the name says, uses 13 inputs to compute the Attitude estimation.
https://librepilot.atlassian.net/wiki/display/LPDOC/INS13+-+GPS+Navigation

Some quotes from previous messages that answered your question:

If you are using an I2C aux mag, you must change mag settings board rotation to either pitch=180 (or to roll=180, yaw=180).  If you do not, your artificial horizon will twist and flip constantly.  Even with it set correctly, you may see 5 degrees or so of horizon tilt and some small shaking.  That is normal.

A little bit not level (5 degrees maybe) and a little bit of shake (2 degrees maybe) is normal.

[joani]

My friend, what I read in many comments that the calibration of the accelerometer has to be quiet, which I almost never do I hold with the hand and the later I record, but my hand shakes, that can be part of the problem too? , There is no magnetometer that one can have on hand and save

[joani]

What I am going to do in the last instance is to remove it from the drobne and calibrate it in a fixed position

[TheOtherCliff]

All calibrations (except magnetometer calibration) should be done with vehicle sitting level on table or ground, not in your hand.  You cannot hold it still enough and level enough.

Also: when Complementary (Basic) attitude estimation is enabled, it should be placed on table or ground immediately after plugging in the battery because the gyros get calibrated then (the fast blinking LEDs tell you it is doing gyro calibration).  It is smart enough to wait till it sees that it is not moving to do the gyro calibration.  It does not have to be level for gyro calobration, but it does have to be completely motionless.  Then again when you arm in complementary mode it must be a table or the ground to be completely motionless.

[TheOtherCliff]

I am trying to understand what the problem is that you are having.

Magnetometer calibrations should be done with it mounted in the vehicle so that vehicle mag fields are taken into account.  When using INS13 the first issue is that ATTI and STAB stay red until both GPS and MAG turn green.  After that, the horizon in the GCS Flight Data page begins to work.  Even when the vehicle is level and motionless you will see some small tilt (maybe 5 degrees) and some small motion (maybe 2 degrees).  If the horizon is flipping all over and upside down, then your mags are not configured correctly.  If horizon never works, it may be that your GPS or mag is not green.

If you give more details and keep it all in one thread, we can help you get it working. 

[joani]

Thanks for answering all my concerns. Now I am testing a walkera 250 runner with gps, to see if with this result it is easier the autonomous modes. However I will resume the Revo Mini and see what I can do
1. You can pretty much cc3d the frame and calibrate accelerometer, gyroscope and leveling on a fixed bench, keep them and mind lacc3d and calibrate the magnetometer as I said.
And I will see after the oscillation of the gsc when I finish calibrating the gps13 mode. Ahh happy holidays !!!

[joani]

Friend because it can not be configured rth and assisted gps that gives configuration error, however if I put rth and I do not put gps assisted, if it takes, that is due to this and gps13

[f5soh]

And I will see after the oscillation of the gsc when I finish calibrating the gps13 mode.

Little oscillations in PFD are normal using INS13

[TheOtherCliff]

Generally it is not recommended for a beginner to have RTH (we call it RTB) connected to failsafe.  It is OK to have it on the switch, but it is good to test everything with VelocityRoam first.  There is much to learn about failsafe and GPS build and configuration so that it is better for a beginner to have failsafe drop to the ground than risk crashing into something or flying away because he did not set it up and test it completely.

[andyp]

I built my quad in the last 2 weeks and it was my first one beyond the of the shelf ones. In this article I found a lot of useful information, because I had the same problems as described here and they caused a lot of pain and frustration. Could we attach this information to the GPS setup page or make at least a link to this article? Also the aux and onboard mag calibration video is really helpful and should be available in the user manual.

As an engineer I like the article about the extended kalman filter, but it should be placed also somewhere in the user manual, as it provides useful and important information. I also find a description of the attitude estimation algorythm very important, because this is not self explaning.

Just mentioning this because it took me quite a while to find this.