operation cnc

[WandrinAndy]

... in my experience though it gets harder the further you get into the nuts and bolts of the job... one small amendment can have a bitch of a knock on effect :( Its frustrating when you want to get building but try to design the knickers out of it before you commit because .....

Very true! That is precisely the design phase I have been wading through for the past many weeks, and it's very frustating but, with what I have discovered and had to correct in my little design, I'm very thankful to have found these problems before the build proper!

[kingcreaky]

thanks for the comments gents.

however, although id love to build a machine 3x3. I cant afford too!. Im simply make the biggest I can whilst maintaining my marriage :D

anyway, done a bit more on the design. but those bearings look mighty small for such a large spanning gantry....

might have to have a re-think...

what do you guys think?

[Jonathan]

I don't like the bed being unsupported over a 3m length. You can at the very least mount a piece of steel box section vertically at the back and the front, but the front one would have to be removed when you make the axis longer. A third piece of 3m long box section in the middle wouldn't go amiss either. Also I would make the bed bolt on to the 4 verticals on the other side, again to make it easier to extend the bed in the future. Currently the frame can 'skew' quite easily, since there are no triangles - so add some pieces to make triangles! You can do this in all 3 planes which is a good way of using up the left over box section since no doubt you'll be ordering in 7.5m lengths. Incidentally how much did Adey Steel quote you for the box section, presumably 60x60x3mm? Their pricing seems to be a bit random...

The way you have mounted the ballscrews is good since it allows adjustment of the height via moving the 4 pieces of box section, which is nice and strong. You can make the end mounts fasten to the box section via holes slotted horizontally to get adjustment in the other axis, which will help when aligning the ballscrews.

[irving2008]

Some simple calculations can give you a better feel.

By way of example (and i'm not suggesting this is how to do it necessarily) lets say your gantry is a couple of 60 x 60mm x 3mm 2m long steel box sections joined by 3 bits of 10mm ali plate 200mm x 200mm at each end and the middle. The overall weight is 15.5kg and the deflection in the middle under its own weight, (ignoring the extra stiffness from the ali plates) would be ~0.1mm. A 10Kg spindle/z-package would cause a further 0.1mm sag, for a total of ~0.2mm. The max load on the bearings would be 255N which is nothing for a 25mm SBR. Increasing that to 100 x 60 box section with thicker walls would stiffen it further. Box section has its own issues with resonance but there are ways round that. The Y-rails would add further stiffness. All calcs are approximate and a more thorough analysis could be done.

So its doable... just need to get the detail right...

[Jonathan]

I think if we included the torsional stiffness in the calculation then the results would be a lot worse. Once this 3m gantry is drawn I should be able to do a FE simulation to find how much it twists, rather like the one in this post.

[irving2008]

I think if we included the torsional stiffness in the calculation then the results would be a lot worse. Once this 3m gantry is drawn I should be able to do a FE simulation to find how much it twists, rather like the one in this post.

Don't disagree, my point was to illustrate that the bearing sizes were OK for a gantry of that scale. Of course it will need to be stiffer in torsion and I'd consider plates both side of the box section and some perpendicular as well.

[kingcreaky]

Gents.

first of all, thanks you all so much for your input. You are all clearly much more astute than I. To be quite honest Im still pulling myself back together after reading Jonathons post http://www.mycncuk.com/forums/linear...ign-ideas.html about how one day he decided to manufacture the driven ball screw nut housings. That is on a complete other orbit to anything I have ever done / made.

Dont get me wrong. I have a metal lathe / miller of sorts


but ive never made anything close to the type of engineering displayed by Jonathon in that post. To be quite honest its now making me doubt my ability to persue the static ballscrew idea.

So Im reverting back to the idea of spinning screw static nut.

Also, I can see where I went wrong in my blogs and I have mislead everybody. To be clear forget about 3 meters. the bed im planning to make will be at best 1700mm by around 1100mm

[Jonathan]

Whilst I did make the rotating ballnut housing with CNC, that didn't really gain anything except aesthetics. If you just don't bother with the rounded corners and change the pockets for the pulley/motor to rectangular instead of having rounded ends then the design is much easier to make manually (since it doesn't require a rotary table) and none of those things will affect its performance.


The difficult bit is boring the 70mm hole for the 7207 bearing since the lathe needs a lot of swing to accommodate the plate and also a large chuck or faceplate to hold it. However it seems your MF42B machine has 420mm swing, which is plenty. To avoid buying those you could machine somthing in the extisting chuck to hold them using the pattern of holes already present.You will of course have to run it slowly. The other option is to use a boring head in the milling machine. You could also make the bearing housing and motor mount separate parts. If you clamp the two halves together whilst boring them this will ensure the bearings are concentric when it is assembled.
I made the shaft for the first one on the manual lathe out of two pieces, which was a mistake as it's harder to fasten them together reliably and ensure the face upon which the ballnut mounts is normal to the axis of revolution. Instead make it from some 3" aluminium bar and take it slowly. When you get close to 35mm diameter for the bearing use very small cuts and keep test fitting the bearing as it needs to be very accurate. Also thread-cutting the shaft and using a locknut to pre-load the bearings turned out to be a much better solution.

To be clear forget about 3 meters. the bed im planning to make will be at best 1700mm by around 1100mm

So just to be clear, am I right in thinking you want 1700mm*1100mm working area, but the available space for the machine is 3*3m? If so then that's very close to my machine - 1700mm*740mm and you would be best just using rotating ballnuts on the 1700mm axis as the ballscrews will be about 2m long. It's possible to do it without, but you'll have to spend a lot more money on the motors and drivers to get similar performance. With two bearings at both ends of the ballscrew the critical speed of a 25mm ballscrew will be about 760rpm, so you could get 7.6m/min with an RM2510 ballscrew which is an acceptable feedrate but it is rather close to the limit so may not be very 'stable' compared to the rotating ballnut which allows at least twice that feedrate. The other option is to use RM2525 ballscrews, but again these are much more expensive.

[blackburn mark]

To be quite honest Im still pulling myself back together after reading Jonathons post

take it easy man :)
be assured that most of us start pretty much where you are
i think some posters on this thread are being a tad unfair in flexing their intellectual muscles without remembering what its like to wade up to their neck in information and the inevitable burnout that causes... reminding you to chill should be on their list :)

you have a lath/mill so its safe to assume you are a practical bloke... if you have the gumption and time and a little help, nothing will stop you designing building a machine that will do as you require... it doesn't have to be rocket science and most machines are built without the in-depth mathematical analysis.

it wont be long before your the one sharing the love and helping some crazy arsed dude on a mission to build a cnc (just try not to melt his brain lol)