Hi Ian,
The Mini-1 did this too, but obviously the capability of the Mini-2's greater pitch angle will bump into this condition more often.
I did a little video back in June with my Mini-1 demonstrating an odd issue. As the gimbal nears the mechanical limit of its range of motion, the gimbal will noticeably jump several degrees away from that bump stop and then remain fixed at that new displaced position until you decide to rotate it back to where you wanted it pointed. But it only behaves that way when the motors are running.
With the Mini powered up, but with the motors
not running, the gimbal behaves differently. If you tilt the Mini to such an extreme angle that it exceeds the limit of the gimbal's mechanical range of motion, obviously the camera cannot remain in its stabilized position but will be physically dragged out of position. However, when the Mini is returned to a more normal level position so that the gimbal is no longer pinned against its bump stop, the camera then resumes its same initial stabilized orientation.
If this (as when motors are not running) were to happen in flight the camera view would be displaced
only during the moments the Mini exceeds the gimbal's mechanical range of motion, but would then return to its previous stabilized position whenever the Mini levels off again. But the gimbal's actual behaviour with motors running forces the gimbal to
jump away from the bump stop and stick there and it does not automatically resume the initial position.
I actually would prefer the gimbal to behave as shown with motors not running.