The thing is that the ballscrew is dead straight, it rolls on a surface plate with no whipping and I checked it on two V blocks, there's practically no runout. It was checked on a lathe after machining and the runout was 3um total, nothing even remotely close to what's visible on the video. The housing is bolted to a flat, machined surface, the bearing sticks out from the housing by around 0.1mm so when the housing is bolted down, the bearing is definitely pressed against the flat surface. I even tried it on the reverse side of the plate it bolts to (it's also machine) and it does the same wobble.

The strange thing is that it tilts in one direction, but if you go sideways it doesn't tilt by as much. If it was bearing clearance you'd expect to see that in any direction right? The annoying thing is that I have a similar problem on a large 25mm ID double row angular contact bearing, see here:



Same issue, tilts in one direction, but if I was to push on the end sideways, it wouldn't move as much. On that bigger bearing everything is tight, just like on the ballscrew, yet there's loads of play.

Regards,
dsc.