Is there any reason why this won't work? On a long ballscrew turning at high rpm's we all know there is a problem with whipping of the screw. A rotating nut system is good but complicated. Why not a piggyback type of arrangement with a Delrin block riding on a linear wagon on the linear profile rail that hits a stop just shy of halfway and therefor stops the screw from whipping as the nut travels to the extreme end of travel. There would be such a setup on each side of the ballnut. In the video I only modeled one support but you get the idea hopefully.

https://youtu.be/kkcVWIsXGjk