I spent some time this morning reading up on CNC design and spec’ing ball screws and linear rails.
I started a spread sheet to keep track of parts that can be purchased. There’s a company that makes fully assembled z-axis that you can buy that all you need to do is toss your stepper motor on it. Looks like that takes a lot of the more tricky part of the design out of the equation.
With using one of those for the Z axis, I’m at $2000 for stepper motors, drivers, ball screws, and linear rails. This gives a 59"x39" (1500mm x 1000mm) linear rails and ball screw setup, which should net me close to the 3’x4’ cut size I’d like. The pre-built Z axis is about $300 of that and a 120v 2.2kw water cooled spindle is almost $400.
Reading on design lead me to 2010 ball screws. I’m currently considering HGR20 linear rails for the X-axis (mounted horizontally on the gantry) and 20UU rails for the Y-axis (mounted vertically).
There is a little concern that NEMA23 are powerful enough for the materials I’m considering using, so I upped to NEMA34 on the spreadsheet for the X and Y axis.
The current thought is to use 1.5" square steel tubing for mounting the linear rails with mounting brackets on each end welded on for the X axis. I can weld braces between the two tubes for additional bracing through the span if needed.
I’d probably weld a steel frame for the Y axis rails to mount to. I’m considering an under-carriage for the gantry to ride on that will go under the table. Then I have to design a table that’s only mounted on the two ends… again… something else that’d probably be welded. What I like about the under carriage design is that it would allow the rails and ball screws for the Y axis to be somewhat hidden. It also turns the entire gantry into a box design. Not sure what I’d make the end plates out of.
In my head, this seems to be fairly simple, so I’m sure I’m missing something. I think purchasing the rails is what helps simplify the overall design.
There would be a lot of machining I’d have to do to make sure all the rails are parallel. My mill should make that easier. It only has a 36" long bed, so I’d have to reposition the work piece at some point during drilling the holes, but I can use a previously drilled hole to re-index the part.
Add in the build materials and I’m probably looking at a $3000 machine. Quick google show Avid’s standard maching is $2799 with NEMA23, but it uses lead screws and bearings. That makes this idea pretty tempting.
It’s been a fun thought process, but I really should be focused on current projects 
