A new hope: Epoxy granite and steel.

[JAZZCNC]

I can't see how the stiffness loss with a slot is worth it.

The below shows two identical columns, except one has a slot cut in it. There is 2.36x as much deflection with the slot.


Bellows don't compromise rigidity.

Like I said the first time, it depends on the application if it's worth it or not. This machine has a specific usage (which I can't mention because of customer confidentiality) that doesn't require high strength so deflection won't be any concern even if it deflected 10x.
This is not my first rodeo, I wouldn't build something that wasn't more than up to the task it's been designed to do by a large factor.

If I was building a machine for cutting more aggressively and I will be very soon, then you'll see it's designed very different with much more bracing and thicker wall tubes, but it will still have the slots because they very much protect the screws. Also, you are making a big assumption that there is nothing inside those tubes.!!

I'd also be interested in the forces you applied to those tubes.? Are they realistic or would we need 30Kw servos with 2mm pitch 30mm ball-screws to repeat...

Who cares what it looks like?

More than you think do care, also usually IMO when something looks wrong, it usually works wrong or awkwardly.

Racking is a consideration. Depends on width between rails and bearing spacing along the rails.

Racking can't NOT happen IMO when it's plunging into hard materials for operations like drilling, it's just by how much.?

The extra expense of making it more massive, etc to counteract this far outweighs buying another ball-screw as it doesn't stop at just the width, spacing, etc. As it gets heavier, wider, etc, it requires larger and more costly everything.
Doubling up means you share the loads so can use a smaller size and less expensive components which cancels out any cost offsets but gives a much better-balanced machine.!. . . Which looks like it was designed by someone who knows what there doing and not a tight-arse who's trying to save a few ££.

I agree, ball screw expense is not the biggest consideration (though starts to get much more when you use high quality bearing blocks etc). I'm thinking about screw mapping and linear encoders - much harder with two screws.
Belt stretch is a concern for a long enough belt to link those two ball screws.

Well first I wouldn't use belts on something any wider than that little machine I did and I wouldn't use belts if I was chasing accuracy to the degree of mapping ball-screws as it defeats the point and like you say makes it very difficult.

Regards Screw mapping then very few do that at the DIY level and if using linear encoders rather than rotary encoders then you shouldn't need to map the screw. Your only concern then is keeping the two sides in sync and I'd assume if using linear encoders the control system will be fully closed-loop so will handle that side.!