[schiavelo]

So i was wondering if anyone has any insight on the concept of making your moving gantry CNC predominately out of epoxy granite filled steel bar?

From everything i've read I'm under the assumption that heavy and rigid = precision. even in moving parts. is this the case?



As far as bearings go, I'm under the impression that fully supported guide rails are the way to go. Would i receive a measurable benefit from having over my guide rails being over 20mm smaller machine approx 800x500mm?


Also, would I receive a measurable benefit from doubling up on rails on my largest axis? (dont get hung up on that particular design, just the double rail on each side)



Where can i afford to be a little bit cheap to start with? Can i buy cheaper chinese stepper motors as long as i get a something like a gecko G540 controller?



Leadscrews. What diameter should i look for to be on the right safe side? is 12mm a decent size or should I go larger?



Thanks heaps for all the info I've already gathered from you guys on this forum and for any I may receive now!

[ger21]

1) Filling steel tube with epoxy granite is more beneficial for reducing vibration than increasing rigidity.

2) Just my opinion, but if high precision is your goal, than don't use round linear rails. They're not nearly as rigid as THK type profile rails and bearings. Single profile rails would be better than dual round rails. And aligning 4 rails on one axis would be nearly impossible.

3) Just about everyone uses Chinese steppers. Just make sure you get the right ones for the application.

4) Leadscrew choice depends on a lot of factors.
Ballscrew, or acme?
How long? How much weight are they moving?
How fast do you plan on going?

[louieatienza]

Lin Engineering has a plant that makes steppers here in the US. Sanyo Denki, Vexta (Oriental Motor) are made in Japan. Kollmorgen (Danaher, Pacific Scientific) are made in Malaysia or Japan. Way better quality, they can be pushed harder, and in reality not much more than the bargain Chinese steppers.

As Gerry says the epoxy granite fill contributes more to vibration damping than stiffness, though there should be some increased stiffness. Heavy and ridgid only heps if everything else is designed as such; and the machine will only be as strong as its weakest part.

Also I agree with Gerry that there really is no benefit from doubling up on round supported shaft. Surplus profile rail can be bought for not much more and have 5-10 times the load capacity of the round rail mounted in its best orientation (standing up.) Four-way profile rail (like THK SHS, HSR, etc...) can be mounted in any direction and have the same load rating.

[wizard]

You have already gotten some good responses with respect to your questions but I have to suggest that you actually tell us what you expect to be doing with the machine and what precision means to you.

So i was wondering if anyone has any insight on the concept of making your moving gantry CNC predominately out of epoxy granite filled steel bar?

I'm not sure if you know or not but there is a massive epoxy granite thread on this site that is well worth looking into. If you want to get into it real deep there is a paper that float about the internet written by a guy named Bamberg, called: PRINCIPLES OF RAPID MACHINE DESIGN which you might want to read to bone up on the subject a bit.

From everything i've read I'm under the assumption that heavy and rigid = precision. even in moving parts. is this the case?

Not at all. Precision is something that comes about for a well engineered and put together machine. A machine can be precise but light weight a classic example here is a laser that needs little mass to do its job.

Rigidity is extremely important once you start to try to machine materials that are relatively harder to machine. Without rigidity cutting forces can cause a machine to deflect or vibrate. If you can imagine cutting a piece of balsa wood with a saw and compare the effort required to say cutting a hard maple, you should see that you have to push harder to cut the tougher wood. In the same way a router has to push the tool harder the more difficult the material is to cut. This pushing twists the frame and is why rigidity is important to maintain machine alignment, so knowing what you intend to machine and the resultant forces will help determine how rigid the machine should be.

Mass comes into the equation because mass can dampen the vibration and if that mass is in the right place contribute to rigidity. The key here is that mass has to be in the right place and has to be balanced against what your controls are capable of driving. In low cost machines (if that is your goal) excessive mass may increase you controls costs.

As far as bearings go, I'm under the impression that fully supported guide rails are the way to go. Would i receive a measurable benefit from having over my guide rails being over 20mm smaller machine approx 800x500mm?

That depends upon what you are trying to accomplish. It has been mentioned that profile rails are much stiffer and in many cases a better choice. You have to balance that against what the machine will be doing. For example if you are machining dusty materials, You have to be concerned about that dust gumming up what ever bearing you may be using.

Also, would I receive a measurable benefit from doubling up on rails on my largest axis? (dont get hung up on that particular design, just the double rail on each side)

Two rails are plenty. You should spend some time with the various manufactures web sites, there are a huge number of designs for linear rails literally a solution for any DIY machine design out there.

Where can i afford to be a little bit cheap to start with? Can i buy cheaper chinese stepper motors as long as i get a something like a gecko G540 controller?

Without an idea as to what the machine will be doing it is hard to say. I would say this though, the machine frame is like the foundation of a building, done right the foundation should support whatever you build on top of it and support reasonable changes in the future. So I would not short change myself with respect to the frame.

Leadscrews. What diameter should i look for to be on the right safe side? is 12mm a decent size or should I go larger?

Depends upon the machine and the axis. There is no good reason to keep the lead screw diameters the same.

Thanks heaps for all the info I've already gathered from you guys on this forum and for any I may receive now!

No problem! To get better help though you need to be more specific about the machine. Here I'm thinking size and usage.

[awerby]

Good post, Wizard! I'd just add that while making the moving parts of a CNC machine heavier can help damp vibrations, there's also a downside to that. As well as being rigid, these machines also have to be somewhat nimble. When cutting wood, especially, the machine can be moving at speeds up to 1000 inches per minute, and the inertia that a heavy steel gantry filled with granite can develop at that speed is considerable. A stepper motor that's suitable for a lighter machine will be overpowered when it attempts to quickly change direction, and lost steps will result. A driver that can power a lighter machine will likewise be unable to produce enough torque at these speeds. If you really want to build a very heavy steel frame, go to a fixed bridge/ moving table design, where the weight of the gantry won't be an issue. Keep the table fairly light, and drive it from both sides instead of using one screw in the middle. Think about providing bellows to protect the rails and screws, so dust doesn't clog everything up.

Andrew Werby
www.computersculpture.com

[schiavelo]

3) Just about everyone uses Chinese steppers. Just make sure you get the right ones for the application.

4) Leadscrew choice depends on a lot of factors.

Ballscrew, or acme?

Imperial or metric leadscrew, not ballscrew.

How long? How much weight are they moving?
How fast do you plan on going?

The machine will be about 800x600 (thats millimeters, Imperial scum!) I dont need amazingly fast production time but lets work on it not being painfully slow. Unsure about weight. lets say its decently heavy though. If im pretty cool with over engineering what would you suggest? Im sure ill end up reusing hardware on a better machine once ive made the first one and learn a thing or two from that first hand experience.

You have already gotten some good responses with respect to your questions but I have to suggest that you actually tell us what you expect to be doing with the machine and what precision means to you.

the toughest material i'd like to machine is aluminium. I'd like to be within 0.5mm accuracy.

I'm not sure if you know or not but there is a massive epoxy granite thread on this site that is well worth looking into. If you want to get into it real deep there is a paper that float about the internet written by a guy named Bamberg, called: PRINCIPLES OF RAPID MACHINE DESIGN which you might want to read to bone up on the subject a bit.

ill check it out. thanks dude.

[LeeWay]

How much space do you have available in the shop? A moving table design might be your best shot at making a lower cost machine that will do aluminum okay. Probably still cost you around $4000. You need the Z axis to be pretty rigid for that. What kind of spindle solution will you be wanting?
I can tell you that a router designed for woodworking is not ideal for cutting aluminum. It can handle a little bit of it, but it cools itself through the machines windings. That means it can suck chips into itself. They also get very hot when machining aluminum. Not for extended use. A belt driven or purpose built spindle would be a better choice.

[wizard]

Imperial or metric leadscrew, not ballscrew.

On a CNC machine it makes no difference really.

The machine will be about 800x600 (thats millimeters, Imperial scum!) I dont need amazingly fast production time but lets work on it not being painfully slow. Unsure about weight. lets say its decently heavy though. If im pretty cool with over engineering what would you suggest? Im sure ill end up reusing hardware on a better machine once ive made the first one and learn a thing or two from that first hand experience.



the toughest material i'd like to machine is aluminium. I'd like to be within 0.5mm accuracy.

0.5mm is nothing really. For us imperial scum that is about 20 thousands of an inch.

ill check it out. thanks dude.

[schiavelo]

0.5mm is nothing really. For us imperial scum that is about 20 thousands of an inch.

Its "nothing" as in its not difficult to achieve?

[ger21]

Correct. .5mm is very easy to achieve.

As for screw diameter, 12mm-16mm should be fine.

[wizard]

Its "nothing" as in its not difficult to achieve?

Yes that is what was meant. Half a millimeter is pretty easy to achieve positioning wise. Of course the rest of the machine needs to be stiff enough not to deflect significantly and loose that positioning accuracy. The point is it is pretty easy to get a lead screw to position a nut within a half millimeter.

[LeeWay]

Lead screws are not optimal for a heavier machine though. Your money would be better spent on ball screws. You can get ball screws with fairly decent accuracy, at least better than what you are looking for, for a comparable price to lead screws. I see lead screws as something that might get the job done short term, but I would not put money in them.

[LeeWay]

As a matter of fact, belt drive or rack and pinion might even be a better choice for this application.
I used belt drive on my Prusa printer and use R&P on my plasma. One thing they accel at is speed and pretty good torque with smaller motors. They still have better accuracy than you sited. Cheaper to buy them and easier to maintain.

[schiavelo]

Ok so after some further reading and seeing a lot of "don't dream too big on your first build" stuff I had a rethink. I thought about doing what almost everyone was advising and starting small and even using that machine to build your second. Anyway, i'm looking at changing my working area (envelope?) to 350 x 250 x 200mm and I want decent precision, accuracy and repeatability with aluminium (to a level where I can make things like motorcycle triple clamps etc.)...

- Is it even feasible without converting an old mill?

and if so

- Is fixed bridge definitely the way to go or is that a small enough machine to be cool either way?

- ballscrews, these cheap chinese C7 ones..from reading the "ballscrew basics" thread i'm left with the impression that they're fine to use, especially at short lengths?
1 Antibacklash Ball Screw 1605 L400MM C7 BK BF12 2pcs 6 35 10mm Couplings | eBay

[ger21]

to a level where I can make things like motorcycle triple clamps etc.

If that's what you intend to do, then you should be buying a mill, and not even thinking about building your own machine.

[wizard]

Ok so after some further reading and seeing a lot of "don't dream too big on your first build" stuff I had a rethink. I thought about doing what almost everyone was advising and starting small and even using that machine to build your second.

Size is important here, especially if you don't have access to a machine shop. There is nothing wrong with building a big machine as your first if you are confident in what you intend to build and how to go about the build. The big problem though is making a large machine stiff enough to machine aluminum. It gets to the point that you need to put a lot of engineering time into it.

Anyway, i'm looking at changing my working area (envelope?) to 350 x 250 x 200mm and I want decent precision, accuracy and repeatability with aluminium (to a level where I can make things like motorcycle triple clamps etc.)...

Then you are on the wrong track. For one machining a triple clamp implies more precision than you specified. The second issue is that building a router machine stiff enough to do this reliably, even on a smaller scale machine, is not trivial.

To be honest I have little background in motor cycles but I believe I know what a triple clamp is. As such I would suspect that some of the tolerances are pretty tight on these parts. Further I suspect the materials would be rather thick.

- Is it even feasible without converting an old mill?

You would need to do a mill conversion or buy a new CNC mill. Frankly if I was doing motor cycle work, a mill would be more useful than the vast majority of router type machines out there. A bridge mill suitable for the type of work you are planning is a very significant project and likely a very expensive one.

and if so

- Is fixed bridge definitely the way to go or is that a small enough machine to be cool either way?

The only way that you would even come close to building a good enough router type machine for this task is to build a very stiff fixed bridge machine. Again though it would likely be cheaper to convert a mill or buy a new CNC machine.

- ballscrews, these cheap chinese C7 ones..from reading the "ballscrew basics" thread i'm left with the impression that they're fine to use, especially at short lengths?
1 Antibacklash Ball Screw 1605 L400MM C7 BK BF12 2pcs 6 35 10mm Couplings | eBay

It all depends upon the reliability & accuracy you want to achieve.

For example profiling triple clamps will require good repeatability and rigidity to machine parts with a minimal of post processing required. Part of the equation would be zero backlash ball screws and nuts. If extreme accuracy is required then you will need ground ball screws and good thrust bearings.

In any event to sum it up you really need to get a mill. It will likely be cheaper to buy a mill than to build a one off router type machine good enough to machine motorcycle parts.