Quadcopter attitude unstable when stopping against wind while position control moving with v1.7.3

Problem:
The attitude of a quadcopter is unstable and almost falling when it stops against wind while horizontally moving in position control mode when wind blows about 7 - 8 m/s, although it is stable even in any more abrupt attitude maneuvering in manual mode with the same wind (so, it is not hardware problem).

Additional Description:
The symptom does not happens in no wind.
It happens when the copter stops oppositely to the wind.
When it happens its heading mostly changes by itself about a few tens of degrees and restored.
Attitude and position parameters tuning cannot avoid it.

Environment:
PX4 v1.7.3, Pixhawk2.1 + Here GNSS, EKF2, F450 sized quadcopter frame

Let me know if it is my fault or comment please.

Do you have the log files? Can you please post on the https://review.px4.io/

Also please send a picture of the drone if you have one?

Thanks Ronildo!

This is the log:

Flight log

The picture I cannot give sorry, because the drone is not mine but its shape is not special. It’s a DJI F450 sized quadcopter. Its shape is more long in x axis of body the bar shape of battery is aligned with it.

Explanation of log graphs:

  1. Around the time about 17:27 sec, attitude and altitude are unstable in several graphs.
  2. Related to this period, from 17:26.45 sec to 17:26.58 sec logging dropouts happened in Sampling Regularity graph.
  3. In this period, barometer altitude also seems not output if we zoom in Altitude Estimate graph.
  4. Around this period, thrust and magnetic field are correlated in Thrust and Magnetic Field graph.

Although thrust and magnetic field are correlated, is it a root cause of all symptoms?

More:
Pitch seems to exceed -90 degree when stopping against wind. So it falls during some times and its heading rotates by itself about 36 degrees clockwise and then returns counter-clockwise.

Question:

  1. Can wind cause attitude overshoot quite bigger when position stopping?
  2. Does the gyroscopic effect appears significantly while falling (no thrust)?