Pushing forward but drifting to the left? Nearly lost the drone

[weirdorox]

Hi dear mavic pilots.

I was flying in denmark today, absolutely lovely weather here for taking the drone for a flight.
So i see a ship i want to shoot with my mate (we each have a Mavic drone)
After him returning with his DJI ZOOM i went for my flight, same route more or less (2km away above sea) no signal problems at all.
So i hit around 50% battery time and decided to return home, as im flying back home (straight forward) my drone decides to "turn its nose to the left" and drift where i more or less lose the control of it. The camera even "jaws" around 15 degrees to the left.
small of a panic here, i decide to try the RTH button, and luckily it worked and it flew home safe.

So my question is, what actually happened???

A link to the flight from airdata:


Dropbox - Oct 16 2018 8_20 AM.webm

 

[Chasing Light & Shadows]

First thing to do is an IMU calibration, you can do this from the app and is fairly straight forward.

 

[Robert Granata]

Is it possible you used the panning feature of the gimbal (long pressing the screen and panning just the gimbal) and forgot to re-center?

 

[weirdorox]

Good question, i need to check if this could have happened... Im gonna calibrate the IMU and see if that helps. Even the camera was also jawing to the left..you can see that on the second picture

 

[DanMan32]

Another possibility is VPS picking up water motion if you were low enough.

 

[51 Drones]

It is the IMU. The cause of most problems such as this. Calibrating should fix it.

 

[CanadaDrone]

You got multiple high wind warnings and attitude warnings (the drone had to tilt so far to fight the wind the obstacle sensors could not work properly), meaning your drone was being blown around pretty hard, and very likely more than it could handle which would cause it to drift. That would be my guess. Your altitude was almost 400ft at times, where wind can be wildly unpredictable and dramatically different than what you are experiencing on the ground.

My other thought was satellites, but you had a good number of them and a good signal.

Another possibility is VPS picking up water motion if you were low enough.

VPS is only used in a few flying modes and while really close to the ground - if he was in normal flight they should not have factored in.

 

[DanMan32]

VPS works in just about all modes, but does have a height limit. If you see VPS height listed in the telemetry, it is operating. When it is out of range it changes to RC distance (you see a person standing)

 

[CanadaDrone]

VPS works in just about all modes, but does have a height limit. If you see VPS height listed in the telemetry, it is operating. When it is out of range it changes to RC distance (you see a person standing)

But is it used to control altitude when the barometer is functioning properly in a normal flight mode? I think that is the important distinction - if the barometer reads X height, I don't see why the VPS would override that and cause a crash, but I am not 100% sure.