How does a drone rotate about a vertical axis while in a stationary hover?

[LawnDart]

I am intrigued how my (Mavic Air 2) drone can rotate about a vertical axis while in a stationary hover in still air.

It is obvious how a drone moves in any direction in a straight line - apply forward stick and it tips down at the front to develop a forward thrust, and the stability modules provide the information to maintain stable flight and prevent the drone from climbing or losing altitude while flying in the desired direction.

However, I was hovering my drone indoors after a firmware update to check all was ok and was watching it as I rotated it to the left and to the right on its vertical axis. There is a change to the sound of the motors while it turns, but other than that there is no change in attitude. The drone remains horizontal and stationary. How is that? There does not seem to be any action or reaction to Newtonian forces!

An explanation would be very welcome - Thanks

Great question and interesting discussion. I've never really thought about it.

 

[DoomMeister]

I have a related question; I'm on my third Mavic 2 Pro (each replaced with a new one under warranty) and they all experience this undesirable rotation about the vertical axis when I hover for long periods shooting hyperlapses. Not all flights have this problem but many do. I've always thought the issue was wind but it happens even on what seem to be calm days (but who knows what's at altitude).

I just wish there was a way to counter it or even better, have the Mavic MAINTAIN its orientation. It does a great job of maintain POSITION but the rotation (yaw) is ruining many of my hyperlapses.

Oh and almost all of the time, the rotation is in the clockwise direction (more than 9 out of 10 occurrences).

I would suspect drift in the IMU. Per recommendation of @sar104 uploading the DAT from a Hyperlapse session that was affected will give a much better clue to what the cause may be.