Designing DIY CNC Milling Machine

Thanks for the suggestion, I'll consider that. The 'trough' for ballscrew had two goals:
- I wanted to avoid additional spacers between the rails and saddle not to amplify any play/bending any more than necessary.
- It actually makes protecting the ball screw from chips super easy with two slightly overlapping pieces of 1mm rubber in between which a connector block between Y and X drives (similar to what I had in my last build on X axis).

I might also add another set ot tensioning bolts on the other direction so there is *something* stopping the casting from breaking in half:
 

Another thing to consider is that linear rails are not self-supporting.  i.e. they need to be fastened to precision flat surface to work properly.

If you cast a base and then try to grout the linear rails directly to the cast base, the rails will not be straight or flat and you'll wind up with binding issues.

Wide-series rails need a flat surface, but they don't require a straight edge on the mating surface.  Normal-width rails require both a flat surface as well as a precision edge to butt up against to establish (and maintain) rail straightness.

That's why I modeled machined steel plates on top of the cast surface.

Both HIWIN and THK have excellent brochures and technical guides available online.  Suggest you read them over, especially the sections on mounting linear rails and various configurations.

I was planning to use straight edge and some kind of epoxy to establish flat surface between on both rails, similarly to:


But first I am lurking for the straight edge, but seems like this is going to be very expensive, even second hand - especially for around 1000mm length that will be necessary for the X axis.

Do you have any ideas for different tools that are straight enough that might serve as a _stamp_ for shimming compound?

Obviously dependent on local availability, but monument fabricators usually have granite-cutting and polishing equipment.  (cemetery tombstone/headstone makers)

They may be able to make you something locally from a chunk of granite they can't otherwise sell to a normal client.

I wouldn't expect them to be able to make a granite square or parallel with any real precision, but a 1m long rectangular chunk of granite with at least one flat surface shouldn't be too hard.

Yeah, I may try that.

One thing I found: firma-pro.com/en/poziomnica-pro900-heavy-duty-black-2xfrez-uchw

This is milled on both sides, maybe there is a chance it will be reasonably flat for the purpose of my project? What do you think?

It may work, but I wouldn't trust it to remain flat over that distance for multiple uses over a period of time.

Box levels are made from aluminum extrusions, and extrusions have residual stresses from the manufacturing process.  Machining two parallel faces helps eliminate some of the compression forces remaining in the outer skin, but I'd be afraid of thermal movement while handling it or using it on a hot day vs. a cold evening.  Plus the fact that you may need some force to impress a flat surface in to some epoxy and the aluminum would flex more than a steel, cast iron, or granite device.

You are contemplating a significant investment in time and money building a CNC machine from scratch.  The project will likely take 2-3 times as much time & money as you think.  Whatever you think - double it.

Keeping that caution in mind... there are some places where you can save money and achieve something 'good enough.'  And there are other places where you will likely regret shortcuts.  Struggling to achieve reasonable flatness & perpendicularity or parallelism so your linear rails don't bind up would be high on my list of places to avoid problems.

Cheap ballscrews can be compensated.  Cheap(ish) linear rails can be managed.  Cheap steppers instead of servos.  Homemade electrical enclosure and used electronics...But an un-flat, un-square frame or base will be a nightmare to deal with later.

Have you considered fabricating something completely out of granite?  You mention that steel is expensive - whether in plate form or tubular/box sections.  There's a gent on YT who documented fabricating a rather nice CNC router exclusively (nearly) out of cut granite pieces.  He drilled and epoxied the sections together, and used the base pieces to establish a flat surface... everything else was referenced off that main section.

I can't post a link now, but do a search for "granite cnc" and you'll likely find the series I'm referring to.

If you can find a local monument fabricator you might find it's economical to work with them and come up with a design that doesn't require you to create quite as many precision surfaces or geometries from scratch.  Even better if you can get something fabricated from a scrapped headstone with a mis-spelled name on one side. 

spumco wrote:

It may work, but I wouldn't trust it to remain flat over that distance for multiple uses over a period of time.

Box levels are made from aluminum extrusions, and extrusions have residual stresses from the manufacturing process.  Machining two parallel faces helps eliminate some of the compression forces remaining in the outer skin, but I'd be afraid of thermal movement while handling it or using it on a hot day vs. a cold evening.  Plus the fact that you may need some force to impress a flat surface in to some epoxy and the aluminum would flex more than a steel, cast iron, or granite device.

 

Oh, that's very valid argument. I haven't considered that this will quite a lot of surface and possibly require some significant force. Definitely won't go that route.

I'd like to keep the costs down, since this is purely a hobby but I already built un-square, un-flat and bendy CNC router that's barely usable - and whole goal of this exercise is to make something that's slightly better this time.

Thank you for not letting me make dumb mistakes!

I'll reconsider my options, start looking for real steel straight edge and also call around to some cemetery-stone guys.

If I'll be able to grab something thats, let's say, 1000x300mm - what do you think about using such granite chunk to imprint the shimming for both rails at the same time? I'd feel much better doing both rails at a time, with guarantee that these will be flat (at least as the slab is). Doing each rail separately with straight edge leaves a lot room for error.

One last question, when it comes to granite tombstones - will _any_ shop be good enough? Like, is the granite processing quite universally good enough or should I be looking for someone with some specific equipment? In other words, how to look for a tombstone that's flat enough?

And thank you again for all the words of caution!​​

I think your idea of using a single slab to imprint both rail surfaces at once is excellent.  They would certainly be co-planar, and getting them aligned & parallel would be much simpler from there on.  Just do a test to make sure whatever you use as a mold release will keep you from permanently bonding your casting to the granite.

As far as what tombstone makers can do or what to look for... I've no idea.  But I suspect there are tombstone fabricators as well as re-sellers.  I'd stay away from re-sellers and try to find someone who actually cuts & polishes the things in-house.  Look for someplace dirty - not a flowery funeral home.

Look for somone close, tell them what you're up to and want to achieve, and see if they're interested in something slightly unusual.

Maybe watch a few YT videos on tombstone-fabrication to get an idea of what a shop should look like and what equipment to expect in a full-service outfit?

Another topic to consider - in the video, guy used Diamant 310FL as a shimming compound: www.diamant-polymer.de/en/shop/dwh/dwh-3...egroessen=DWB+250g#0

I assume it was specifically with steel filler to make it expand in similar rate as steel underneath?

However, in my case I won't be shimming steel but cement casting instead. What kind of shimming compound should I look for? Some concrete filling putty?

---

Update!

I just got back from neighbour tombstone guy. He had thin (30mm) granite plate cutoffs. While talking to him it dawned on me that instead of casting concrete and then imprinting the shims I could try use the granite plate directly as a reference surface for rails, mold concrete over it and bolt it with through-rods.
This will require slight re-design because ball-screw supports and rails will be mounted at the same level, but that's not a problem.

Only trimming will be necessary when stacking one part on top of another to ensure parallelism between X an Y and perpendicularity between Y and Z, but I think that should give far better chance of success than any imprint (in the end, best case scenario imprint won't be worse than original plate).

So for 3 granite plates I paid less than what it would cost me to buy this aluminium box level
This stuff is heavy! I am starting to worry whether the 3Nm NEMA23 steppers I was planning to use will be strong enough...

Good score!

You've got a lot of work ahead of you, but I think using one of the plates as a built-in reference surface will pay off in the end.

I suspect you won't be happy with NEMA23 motors.  Quick way to check would be to make a rough model in Fusion 360 and set the materials to granite.  That should give you a ballpark weight for each of the moving components.  From there, most of the name-brand ball screw (and servo) manufacturers have on-line calculators (or formula in their catalogs) to determine how much motor power you need to achieve desired accel values.