Mavic Air looks unstable in flight. Why? (demo video inside)

[Dimedrol]

Greetings All!

Recently I've noticed my Mavic Air looks like "unstable" at flight time: it's moving all the time - slightly to the right, to the left, up, down, etc.
I thought it was poor calibrated compass.
Exploring many Youtube videos I've realized that (looks like) I did incorrect compass calibration all the time:
I've turn my drone clockwise holding it almost still, stepping around by myself.
On Youtube videos everybody calibrating their drones holding them on outstretched arm.

OK.
Yesterday I took my "Mavic Air" outdoors, try to calibrate on outstretched arm, go to fly, but, it seems not helped a lot...
Check my video (total time: 1:10):

Typical behavior - between 0:10 and 0:20 sec. - see, it's moving all the time a little bit, not holding position still.
Another "place of interest" - between 0:50 and till the end - 1:10.
Same behavior - not holding position stable.
You can see, especially about 1:02 - check the trees branches, the day was a bit windy, but wind was not very strong (I think)

Is it normal?
Or... what I am doing wrong?
Any help/advice will be greatly appreciated.

Thanks.
Dmitrijs.

 

[steve_lyt]

it looks like wind, the drones are not very aerodynamic , especially when hovering , it is always trying to hold its position, so I would say theres not a problem

 

[Deleted member 94047]

also, uploading the flight log will be much more informative. A lot can be learned from it than the video.

 

[sar104]

It looks to me like it needs an IMU calibration.

 

[Dimedrol]

also, uploading the flight log will be much more informative. A lot can be learned from it than the video.

Flight log like this? -

 

[steve_lyt]

the wind speed looks on the high side, to keep a stable hover, from the log it looks like normal behavior

 

[Dimedrol]

the wind speed looks on the high side, to keep a stable hover

You mean - 10.7 mph ?

 

[steve_lyt]

your flight when moving was stable , it was just the hover. even my mavic 2 pro wont hover with that windspeed, it may do better than yours due to the weight , but it will still try to move

 

[sar104]

Flight log like this? -

Okay - it was a bit windy, but not enough to cause those deviations. Can you find and upload the mobile device DAT file ending FLY002.DAT?

 

[sar104]

your flight when moving was stable , it was just the hover. even my mavic 2 pro wont hover with that windspeed, it may do better than yours due to the weight , but it will still try to move

The M2P should be stable in that wind.

 

[steve_lyt]

it is fairly , their is a slight bit of movement . but it is their , the gps keeps correcting

 

[Dimedrol]

Okay - it was a bit windy, but not enough to cause those deviations. Can you find and upload the mobile device DAT file ending FLY002.DAT?

Shared here: 20-03-15-02-39-41_FLY002.DAT

 

[sar104]

Shared here: 20-03-15-02-39-41_FLY002.DAT

Well first of all, your compass needs to be calibrated - that alone could cause those movements.

 

[Dimedrol]

Hmm... as I wrote at my very first post - I have calibrated my compass.
Maybe I did it wrong?
How to calibrate compass correctly?


P.S.
and... how to get this picture, you attached above?
As I understood, these points in the top area of graph means - magnetic yaw is too "wide"?

 

[sar104]

Hmm... as I wrote at my very first post - I have calibrated my compass.
Maybe I did it wrong?
How to calibrate compass correctly?


P.S.
and... how to get this picture, you attached above?
As I understood, these points in the top area of graph means - magnetic yaw is too "wide"?

You may have calibrated it, but the calibration is bad. Removing the data from during the calibration makes the problem clearer, perhaps:



The red and green points represent the discrepancy between magnetic yaw and IMU yaw as a function of IMU yaw. They should lie on a horizontal line at -7° due to the local declination. Instead they show a much larger variation with a period of 360°, indicating an uncompensated magnetic field originating on the aircraft. The variation is large enough to cause directional errors when the FC corrects for wind drift.

If calibration continues to fail to remove that then the aircraft may need demagnetizing.

 

[Deleted member 94047]

Sar104 has already covered this but do another compass calibration in a clean environment (with no magnetic interference) and I would also do a IMU calibration (this one you need to do on a perfectly level flat surface ).

 

[Deleted member 94047]

@sar104 I have often wondered about the meaning of the estimation errors reported in the DAT event stream after a compass calibration. For this one it is reported as 21.8 but I have no idea what these numbers signify.

-26.451 : 3618 [L-COMPASS][scale cali(0)] estimation error:[21.8]

Does that mean anything to you?

 

[sar104]

@sar104 I have often wondered about the meaning of the estimation errors reported in the DAT event stream after a compass calibration. For this one it is reported as 21.8 but I have no idea what these numbers signify.


Does that mean anything to you?

The calibration likely produces a scale and bias value for each magnetometer axis, so that may be the estimated error in a scale value.

 

[Deleted member 94047]

The calibration likely produces a scale and bias value for each magnetometer axis, so that may be the estimated error in a scale value.

I see, thanks! So I suppose the lower the number the better the result? And what do you reckon is the range of the scale, 0 - 100?

 

[sar104]

I see, thanks! So I suppose the lower the number the better the result? And what do you reckon is the range of the scale, 0 - 100?

I'd have to take a look at some of my compass calibration logs to figure that out. Given the current situation I might have some extra time to do that.