Ardupilot Tailsitter

Hello, RC fixed wing VTOL Tailsitter renovation project.

It was made by remodeling the SONIC MODELL AR WING 900mm airframe.
The motor mount was designed and manufactured with a 3D printer.

The attachment position of the motor mount was located in the middle of the elevon and attached to both wings. I thought that only then, the rear flow of the propeller could have a big impact on the elevon.

I finished two flights with the plane that was built like that.

First of all, it’s a successful flight video.

The video above is a simple production process and a successful flight.

The first flight mode was Q_Stabilize mode. (First video 3:10)

When it first takes off, the aircraft tilts and waddles to go forward.

This seemed to be controlled so that the drone would not tilt further because it had achieved its maximum lean angle.

And here’s the next video. It was a very windy day.

This video fails to secure a stable attitude without maintaining a certain position due to strong winds while flying in Q_LOITER mode.

As you have seen in the above two videos, the tailsitter controls the drone’s position by the drone’s inclination, making it difficult to control the drone’s position and posture when the wind blows.

I’m thinking about how to make it strong in the wind.
Will simply increase the maximum tilt of the drone solve it? (Q_ANGLE_MAX: Angle Max)
If the maximum tilt angle parameter is set large, wouldn’t the tail-sitter drone be able to tilt the drone a lot to maintain a constant position even if it’s windy?

So I looked at the flight log about the roll, pitch, and yaw angle of the drone.
But I don’t know how to interpret the drone tilt angle in quad mode and airplane mode.

The pitch slope in quad mode must lie 90 degrees forward to 0 degrees of pitch slope in airplane mode.
The roll axis and yaw axis intersect each other in quad mode and airplane mode.

So how do I analyze the flight log when I interpret it?