Maybe you have an old record player sitting somewhere?
If I had it, it would have been the simplest solution, but I don't.
I tried with the glass pane. It was more difficult than I thought.
The main problem are the surfaces, the glass is too slippery and the tested discs slip far too much. The tested ball instead seems not smooth enough, (the noise while it rolls tells it), so its trajectory, especially when it rolls slowly, is not accurate.
I could observe some stationary rolling.
The stationary rolling is not stable, when the ball slows down a bit, (rolling resistance), it starts to move in circles.
The circles tend to be larger and larger, until falling down from the pane.
The center of these circles tended to shift towards the center of the glass pane, making difficult to observe the ball circling at a side of the center of the glass pane.