When I built the last small plasma table I would have liked to spend an extra couple grand for some better screws that would have zero backlash but cash was a bit short back then... still is and the screws I got on there now have decent accuracy and backlash for their cost, the actual backlash may be a bit less than I mentioned in the earlier post as that was tested with the full machine torch and air marker onboard.

The main reason I was trying faster speeds was for a welding experiment of fast deposition wire feed for a custom project but I had too much trouble with the feed mechanics on the wire system so it is on the back burner for a while until I can redesign/build a better feeder.

If you switch to screws you need a reason for it, and they are more expensive without being more accurate [Screws also have backlash.] They also have rotational inertia that adds a lot and prevens quick movement changes, which 3D printing is all about.

They're totally suitable for Plasma CNC, though.

It will let you know alot!

Screws are overrated for low-power situations. They can make sense with traditional CNC machines, because of the ease of developing thrust, but this is not very helpful for a 3D printer.

They have many drawbacks, too.

But if you want to do further research, look into the "Lisa Simpson screws" but I'm remembering that off the top of my head. If I am not mistaken, it is the most recent public screw attempt that I have seen, with a few advantages over belts (in a delta configuration.) but there are still drawbacks.

Belts give you a lot of performance for the buck and have much lower friction than standard screws, I run MS screws on my plasma rigs and they work pretty good but they do have their issues and while I run antibacklash nuts on them they still have measurable amounts which for plasma work is ok but if you start worrying about ~.5mm for aesthetics in 3dPrinting then it will cost you