Cnc lathe using thk rails possible?

[dberndt]

I've been looking to get myself a lathe for some time and CNC it. But it seems that finding an old cheap machine with good ways in my price range (cheap). So I've been wondering about the practically of building my own. I've really only given thought to the ways. Would THK linear bearings work ok for ways? I'm kind of at a loss to figure out why they wouldnt. But I've never seen it done so maybe theres a reason.

I'm thinking about a machine something with something like a 10" swing and 40" between centers.

[RotarySMP]

That would be a pretty huge machine. Do you have a Mill that can machine the bed to accept the linear was?

Linear ways are used in most of the modern CNC lathes and other machines.

Have a look at Swede's www.5bears.com, for ideas, as he built a nice mill with linear ways.

[MIKE JEFFERS]

have a look at this
http://www.compucutters.com/machines...ompucutter.htm
it is a lathe ,a small one, but a lathe non the less.
as the basis for a larger machine (but maybe not 40") it might be an interesting
project or might give you some ideas.
mike.

[dberndt]

Well in terms of the bed. I''ve been thinking just a decent piece of tooling plate, mic6 or whatever. Should be reasonable enough in terms of flatness if properly supported no?
some sort of adjustment could be built in to remove any warp that might be in it, though there should be none, but we're probably talking .0001"s here.

Maybe I'm crazy, as I've never mounted a thk rail. But it seems to me that you need 2 rails, possible more. Mount your first to the plate base with milled holes. Possibly in a channel with a shoulder for alignment on one rail. much like how swede has done his mill. Then the second can just be dialed in with an indicator, pretty easily I would think.

I'm not so much interested in continuing a discussion of how it could be built but more so if it is practical. The loading in a lathe must be different then a mill. In a mill you are always exerting force directly over the same spot (provided you dont rotate the head). Meaning that all the forces center in the middle of the rails. Even if the loseness of the balls in the trucks on the thk rail was enough to let the axis twist slightly or be shoved to one side it doesnt really matter that much in a mill, it would just show as backlash and that'd more or less be that. In a lathe I'm wondering if the cutting forces can be absorbed by the rails, will they allow more chatter then a conventional design. Should there be some sort of other guide system used instead or in addition?

Does the machine really need to be a 2 tons beast? Why cant a piece of tooling plate, some rails, a carriage, cross slide, and headstock be enough? I'd really like to make a light(ish) machine on the order of no more then 500lbs if thats possible. I'm willing to deviate from my spec's in terms of swing over the crossslide, but 30" or more between centers seems pretty reasonable. And I figure convential machine swinger over crossslide and cnc machine swing over crossslide dont really have to be looked at in the same way. With a manual machine you have a compound eating up a good 2" and a generally really unoptimized design imo.

Any data on what sort of forces are actually imparted by/on the tool to make reasonable cuts in something like stainless steel would probably be helpful. I've done enough turning on a manual lathe to call myself proficient at it. But I am still woefully ignorant of just how much force might be applied at the tool.

Anywho, I could be crazy.

[RotarySMP]

There is a reason the bed of a 40" turning lathe bed is a few hundred Kg of cross braced cast iron. I can't see how a peice of guage plate with THK railed screwed to it could ever resist the forces, but that is just a gut feeling.

What are you aiming to make with it?

What would you use as a headstock? It would be hard to come close to the utility of a chinese import 10x lathe for the prices these things go for.

[Mcgyver]

Even if the plate is within tenths over 40", how you are you going to support it such that it stays that way under varying forces? By the time you are done you may end up fabricating some cross section that’s as heavy and complex a typical cast iron bed.

How are going to get said superstructure flat to begin with – ie if you’re attaching a 40” piece to the superstructure, it will bend to the superstructure. Either scraping or a big mill as Mark suggested. Then you've got all kinds of internal stressed that will keep pulling it out shape and you have only 1/2 the dampening effect of cast iron.

I don’t mean to discourage, only to point out that presumably the 2000 pound monster was also engineered and therefore it’s not going to be easy to maintain performance while reducing weight AND going with a less desirable material (rolled vs. cast) simply through better design. My lathe is 1000 lbs on a cast iron stand and there are lots of times I wish for something beefier.

What about the approach seen on some high end wood lathes? A large cylinder with brackets and ways welded to it then filled with sand to dampen vibration (or my fav, epoxy and granite which would contribute strength as well as dampening).

Strictly based on personal experience, 30” – 40” between centres would be good, although 99.9% of time you are working within a foot of the headstock. I would definitely would want more swing on a machine this size,. Space over the compound has not been an issue – you might turn 3’ lengths and 12” dia, but at the same time?

I am faced with the same questions on my mental (both a noun and adjective) designs. Ultimately approaching it like an engineer and figuring out in forces, vibration, deflection would provide the answer vs. anecdotes which it sounds like you want to do. It would be great to find more resources on engineering for non engineers, but here’s a couple of links…
Here’s some basic stuff on cutting forces: http://www.hytec.com/Designin.pdf
Here’s lectures from MIT machine design course: http://pergatory.mit.edu/2.75/index.htm

[ESjaavik]

Here you have a guy that did it.

http://mitglied.lycos.de/dieter096/photoalbum.html

It's hosted on Lycos, so there will be a storm of commercials, but you get a glimpse in between all the s**t.

[dberndt]

Flatness, flatness, flatness, and bed support is all anyone wants to talk about. Well It seems to me that something like a concrete base could be poured and leveled. If that isnt flat enough I wonder if there is some sort of top coating liquid that could be applied that has a low viscosity that will harden level. If not one could spend hours or days leveling the surface after the pour. Either way a the plate could be secured to something reasonably solid.

I know a concrete base sort of throws out my idea about light and portable, but who cares about the concrete base, if I move I can always pour another one. And frankly, with a cnc machine for hobby use, why do I need a 36" high slab, why not just bolt the plate right on the floor after leveling it nicely. Sure it might suck for access, but if I made one part a day on it average over its lifetime I'd be highly suprised.

Even with an 80/20 subframe option or something similar I dont see why a mill is going to help. So you've attached your rails to the frame and now the plate is bent. How's a mill going to help, you cant take a pass with the rails installed, and without them the plate returns to normal. Scraping the thk rails really doesnt seem like a practical solution to me. So how would this operation work?

[Chris D]

If the ways are not supported, they will flex under load (while cutting). This will lead to very low accuracy and cutting problems like crazy. Linear ways by themselves have very little support - they would sag if you suspended them at the ends only.

A person could build the entire lathe out of wood and use linear ways, but that doesn't mean it will perform good.

I believe I saw that you were going to cut cast iron with this lathe. If that is the case, then you will surely want to isolate the linear ways from the cutting area. Cast iron will create plenty of very fine "Dust" which will get into the bearing packs and ruin them in pretty short order.

It all comes down to what you are willing to accept for performance in the end. If you want performance and accuracy, you will have to build the machine to provide those attributes.

I am in the process of thinking through either a CNC conversion or building a lathe from scratch. The biggest problem with making a machine tool from scratch is machining all of the parts properly to provide the strength and rigidity that is needed for a precise lathe.

[Mcgyver]

Originally Posted by dberndt So you've attached your rails to the frame and now the plate is bent. How's a mill going to help, you cant take a pass with the rails installed, and without them the plate returns to normal. Scraping the thk rails really doesnt seem like a practical solution to me. So how would this operation work?

you don't scrape or mill the rails, you scrape or mill whatever they are being fixed to. obviusly consistent distance between centres and bed is important for a lathe and therefore you need a flat surface on which to bolt the rails to maintain this distance. I know i'm i stating the obvious here, but that little literally underlying fact will determine the performace of the machine no matter how much money you plough into it up top.

the concrete base has merit, but i would want to build it regular height - may with a hollow core? I don't know how much movement if any concrete has over time but that would be a consideration. steel strips on the concrete that you scrape in?

[mxtras]

Linear bearings and rails are made for accurate guiding - so - Heck YES! I say go for it.

The rails will not offer any support - as previous posts emphasize, so I also suggest that you make sure you have a very stout, accurate base structure to attach them to.

IDEA - I have used Moglice (epoxy) to "pot" in a flat surface on smaller projects with excellent results. If you have a stiff structure that's relatively flat - even within .250" or so, you can pot it upside down on a known flat surface (granite table) to duplicate the flatness of the master surface onto your structure. Then mount your rails directly to the duplicated surface in the normal manner. The epoxy surface is extremely strong and stable.

I don't like the concrete idea, but it's an idea. I would expect you to run into a host of issues - especially with regards to attaching other mechanical elements in the future. I would also expect moisture retention to create issues with corrosion, among other issues.

Good luck - and post your progress!

[ViperTX]

Since you probably won't save any money building it yourself....the question is "do you just want to have the satisfaction of knowing you built it?"