Compass error results to nonGPS

[BudWalker]

Thanks to BudWalker for that little dance.

I consistently get "magnetic interference" and weak GPS. Even though my phones, GPS units have no trouble.
The compass will calibrate, now that I have it figured out.

I had a similar experience to Ekto. I had a 'brain fart' regarding the axis.

Could you be more specific please. Are you saying that you've changed your calibration procedure?

 

[Trill136]

Yes, I have changed my procedure. I am now doing it correctly!!!

I read, and thought I understood, the procedure from the manual. But, since I do tend to overthink things, I made it more complicated than it was.

I watched a couple of videos, posted by people here, slapped myself on the forehead, muttered a few 'choice' words, and then did it properly.

 

[BudWalker]

Yes, I have changed my procedure. I am now doing it correctly!!!

I read, and thought I understood, the procedure from the manual. But, since I do tend to overthink things, I made it more complicated than it was.

I watched a couple of videos, posted by people here, slapped myself on the forehead, muttered a few 'choice' words, and then did it properly.

OK, just to be clear the comment about my compass dance was in reference to it's entertainment value? It's not the method you now use when you're calibrating your compass. I'm curious - what was the method you were using that apparently caused a thawed calibration.

 

[Trill136]

When I first had the error, I placed the little guy on the ground and rotated it 360 degrees. As it says in the manual. I simply picked it up a bit and spun it around using the battery button as the axis point. Then the manual says to point it "nose down" and rotate it 360 degrees. I spun it 360 using the same axis point. The battery button.
In other words, the first part worked properly, the second failed.
The failure was due to me using the wrong axis to rotate on.I should have rotated it, when pointed nose down, on an axis from nose to tail.
And, I kind of liked the 'twist/turn' kind of thing. I realized it was a bit of entertainment.
It is not the method that I used.

 

[ekto]

When I first had the error, I placed the little guy on the ground and rotated it 360 degrees. As it says in the manual. I simply picked it up a bit and spun it around using the battery button as the axis point. Then the manual says to point it "nose down" and rotate it 360 degrees. I spun it 360 using the same axis point. The battery button.
In other words, the first part worked properly, the second failed.
The failure was due to me using the wrong axis to rotate on.I should have rotated it, when pointed nose down, on an axis from nose to tail.
And, I kind of liked the 'twist/turn' kind of thing. I realized it was a bit of entertainment.
It is not the method that I used.

So in the second stage, you ware rotating the drone around the power button, while that button was facing away from you? It’s good to know that Mavic did not accept that calibration. It could be monitoring the process with gyros to see, if you are doing it right.

 

[Trill136]

Yes, that is correct.

I read the manual and had the body in the correct position, but, in my haste I did not pay complete attention.

I pointed it downwards and then rotated it on the same axis as in the first step. Which was incorrect.

In the first step the Mavic wants to rotate on an axis through the centre of the drone. Roughly where the power button is.
In the second step, it wants the axis to be on a line from nose to tail. This is where I made the mistake. I did not change the axis and rotated the drone on the original axis.

The way I have been quite successful with now, is to hold the drone in front of me. In the normal flight orientation, legs down, arms out, powered on, basically horizontal and rotate myself 360 degrees.
Then I point the drone "nose down", hold it in that orientation and rotate myself 360 degrees.

 

[ekto]

Update from DJI service,

After testing and diagnosing your product, DJI will provide you with the free repair service. The repair is expected to be finished in 2 to 3 working days, then the product will be sent back to you. Thank you for choosing DJI.

GPS Module 42.98
Compass Board 2.48
Repair Service Fee 50.00

With these prices, I won’t hesitate servicing the drone with my own dime, if it breaks later on. I was expecting it to be much more expensive. Shipping cost are not known, but they can’t be more than 20 euros, with standard service. I’ll report back when I get the drone.

 

[ekto]

I got my drone back! The whole process took 20 days and did not cost me anything. I have to say, DJI pulled this one off nicely. There ware emails from their automated service every now and then, I was kept informed about whats happening. Did not feel the need to send any enquiries.

It was pretty windy today, so I wasn't going to test any mapping today. But I did do a controlled test flight, and the bird handled very nice and sharp. No wondering off even thought the winds was 5 - 12 meters per second. I had to pair the Mavic with my controller and then I did the compass calibration, since it showed 1/3 of interference. After the calibration, there ware basically nothing. And no warnings of any kind after that.

So it seems that there ware something hardware related earlier, and that it's now fixed. But one thing is odd, the firmware is jumped back couple of steps. I don't know if I dear to upgrade it. Should I?


Anyway, here's a pic from the log showing YAW from two sources:


It mostly disagrees about 5 decrees, and mostly in the same direction. I think that is good enough. It the beginning there is the compass calibration, something I was not able to record earlier, funny.