CNC Design Help

[bprzybyl]

Hello,
I found this picture of a CNC machine on here, but can't find any more info about it. Personally, I think it's an awesome design. Really nice looking, anyway.



My question is, would it be rigid enough to mill steel or any other metal. I would like to build a very versatile mill, with hopes of adding a 4th axis soon after completion of the 3-axis. To me, it looks like that would foot the bill as far as machine design.

Also, does anyone know where this picture came from, and if there are any more? Models? Drawings? Anything? any help would be greatly appreciated.

Thanks,
Brooks

[fkaCarel]

To me it looks like an unfinished CAD-drawing of a gear driven router. Skinny table, design of fighting rotational momentum at the base of the gantry could be much improved. Not heavy(enough). Just my thought.

Carel

[2muchstuff]

I would question the use of a single rail on both of the x and y axis.

[bprzybyl]

I'm sorry, I have had very limited sleep the past few days. I was and have been completely oblivious to the way that it's drawn, while making changes in my head. I figured it was just an unfinished drawing and that those racks were just there for show. As for the racks, they would be Linear Slides (DEFINATELY). And as for drive, I have decided on double nut roller screws, custom made by me. I saw the design on here, and fell in love. I didn't even know how this one was supposed to work. I just figured the designer was lazy and just put those in instead of modeling the slides. I learned a good deal when you helped me before Carel.

Now that my intentions are a little clearer, do you think this design would be rigid enough for metal milling? To me, it looked like it would. I would design use square aluminum tube, probably 3" wide, with the thickest walls possible. No welding will be performed as I don't know diddly about welding. Everything will be drilled, tapped, and countersunk with allen screws. I love the way it looks when you can just see the head of the screw flush with the surface.

If anyone has any more questions, ask away. You're helping me out so I should be as helpfull as possible.

Thanks,
Brooks

[ger21]

I think what you guys are calling racks are actually linear rail covers.

[bprzybyl]

Yeah, like bellows over them. Has anyone seen where these originate? a model would help greatly.

[ger21]

Yeah, like bellows over them. Has anyone seen where these originate? a model would help greatly.

http://www.gortite.com/bellows.htm

[bprzybyl]

No, I mean the machine. I haven't even begun planning how to protect the rails. Large, table width bellows were what I originally figured on, I would just have to find a source.

So do you guys think this machine is rigid enough now?

[fkaCarel]

If you want to design, you need to know the basics of design and the battlefield. The teeth of a mill must remove metal. This creates a momentum against the spindle bearings. The workpiece wants to go in the opposite direction. This creates flex and you want to minimize this as much as possible. This can be done with maximizing rigidity, so the force gives only a minimal movement. This movement is always there, how unmeasurable it may be, it is the law of action/reaction. The next is inertia, which means that the resulting movement of an intermittant applied force (the mill) is smaller when the to energize mass is heavier. A hammer against a nail drives the nail in, the same hammer against a building does nothing. And yes for the ant-****ers, as we call them here, there is still movement and absorbtion, because you can hear it. So you want a one piece unmeasurable heavy thing. The rest is a compromise to allow the thing to move, and have access. If you ignore these rules, you will get vibration and lose control over quality and shortened life of everything. If you have vibration, you can also have resonance which is plain destructive.

Now we apply this to your proposed design. It looks nice, but... The table. The stiffness is governed by the material thickness left after milling the T-slots. It is not much, so the table warps more then necessary and spoils the alignment of the slides, if you attach a workpiece to it. More thickness, more weight, more stiffnes, more inertia. The gantry is sufficiently supported for reaction in the width direction, but if you look in the length direction, you see that the table is attached to two pipes. These two are connected with two pipes, and on one of these are the upward pipes fixed. If you look at the pipes as an arm, creating momentum, they can and will rotate the low pipe and they are 1 pipe-width too long. Following the laws of momentum, the underside should also be wider then the upside. In general if you look at things, ignore the weigth (it may still be lightweigth although you can't lift it) and imagine it made out of elastic material.

This may or may not sound all boringly theoretical, but I find it easier to break things on paper with a calculator, then a never ending evolution on a basically flawed design. If you goto into the effort of making parts, you may as well be proud on them.

Carel

[dkowalcz]

It's not a bad design, there are worse ones... I've milled steel on a production basis with a converted $50 drill press and made money doing it, so there's no doubt that this would be at least as good, and probably a whole lot better. I'm not sure it'd be a better effort than to buy a used/blown control NC mill and fix it up though.

No dimensions, so I'm going by eye here, but as others have noted I'd thicken up the table quite a bit. It's unclear what the base tubes are for exactly - if the whole assy was a solid plate under the X rails and a concrete bed under that I would think you'd be better off. That would shorten the uprights a bit, put the flanges closer to the loads, etc.. Not to mention make it more likely to actually be flat.

Looks like the Z isn't designed, but you knew that already.

Oh, and definitely use steel! It's 3x stiffer than aluminum, cheaper, and you can weld structural sizes with a $200 welder.

[bprzybyl]

Can welding be accurate with steel? I remember hearing that the steel will deform once that much heat is applied to it. But, I also know NOTHING about welding. I should be joining a club in fall, and would have the opportunity to learn. I also have a friend that knows how, maybe he can give me a lesson.

Going off of that design, I used a program called "ImageJ" to make measurements. Assuming the table is a 12x12" table, it is made with 3x3" stock for all uprights and supports. Through this design, I was going to use 3x3 aluminum tube with .25" wall thickness. Maybe I'll sit down and figure out some deformation with given loads. Which brings me to another problem-

How much force is incurred when milling? I haven't found this information anywhere.

As for mounting a large slab of aluminum/steel as a base, I thought about that a little while ago. The only problem I can see is that it might not be very accurate (thickness wise), and would therefore call for some MIC-6 Plate. Now, I understand that MIC-6 plate isn't very expensive, and I would agree that it isn't. But for a good 1/2" to 1" thick slab of the stuff, 18x24" long (guestimate), I think it would end up being fairly expensive for my purposes.

Personally, I do dislike the bars that run parallel to the y axis on the bottom of the machine. I think it's just to crude of a base when you'd want it to be as stable as possible.

Maybe I'll try to get some designs up by the end of the weekend. Now for another difficult decision- Design in Rhino 3.0, or Pro/e WF 2.0. Decisions, decisions, decisions...

Thanks,
Brooks

P.S.- "Boringly Theoretical"? Try "Very helpful"

[fkaCarel]

Welding is'nt accurate, as raw material is not. The very fast cooling spot of molten metal induces stresses in the material. They can be removed with heat treatment. For an accurate machine, you need to have all contact area's machined, so that they all match flawlessly. A perfect rail fitted on a wobbly surface will give a wobbly finished surface. You don't have to be a fundamentalist to prefer grinding in this case.

I checked the electro motor data and a 0.75kw motor gives a torque of 2.53Nm. You can extrapolate these data through gears and the mill diameter. This will give stall data. Then you create a worst case-scenario, and multiply with the classification.
There is an upgoing classtable, starting with aeroplanes(1.05)- consumer goods- smooth industrial machinery - heavy used idem - heavy used, heavy vibration etc. I think I would multiply with 20. Mounting a motor with more torque would offcourse mess everything up. If you are limited in machining options, don't rule out 80/20. Here you can buy it upto the incredible size of 320 * 160 mm (~13"*6.5") with direct mount of linear slides. With proper design you can make exceptional things out of it and if you compare it with welding/machining or ordering/building it is cheap. I threw out my welder after discovering it.

Carel