[footloose]

I originally posted this in the metalworking forum, but got no response, so now I'm trying here.

I would like some feedback on a machine design.


I would like to make precise molds in protoboard, machinable wax, and aluminum.

So, cutting aluminum to .001 is my goal.

I have spent a lot of time reading up on stuff, here is my current design.

Once I get a design nailed down, I'll decide if I want to take the plunge (based primarily on price).

Current plans:
25mm hiwin rails from automation-overstock.com
c7 rolled ballscrews from linearbearings2008
2.2kw spindle from linearbearings2008
8020 80mm and 80mm/160mm framing
shars 18" x 8" cast iron t-slot table
3 axis servo kit from keling

Drawing is a bit rough, I didn't want to take much time to model parts. The ballscrews shown are the wrong size, they're just in there for rough placement.

All the framing is the "correct" size.

Drawing shows just a flat plate rather than the t-slot table.

Drawing is missing:
Mounting brackets for ballscrews.
Mounting bracket for spindle
Holes depicted in plates.
Nuts and bolts
servos

Working area is 450mm x 250mm x 200m, roughly 18" x 10" by 8", except for where the z axis ballscrew might interfere along the right hand side of the x axis. There is enough space to make the table a couple hundred mm deeper, this could be 450mm x 450mm area no problem, I think.


Things I'm proud of in this design:

The axes don't really depend much on each other spatially, so it's relatively easy to adjust the design to a new size. Previous iterations were a real pain in the ass if I decided the spindle anchor plate needed to be bigger, or whatever.

The mill is nice and small. Currently the two sets of 8016 are 750mm long, which comes out to just under 30 inches (without the servos which will stick off the sides). This small size is nice, because then I can build an enclosure and still fit comfortably on a 36" table.

I think it uses its space well. most of the available space is usable working volume.


Things I'm worried about:

Y axis deflection, if the spindle is right in the middle of the wide cross plate and fed in the Y direction. I tried to use deflection calculators with an estimated force of 1000N. 1000N because this paper (referenced elsewhere on these forums http://www.cadanda.com/CAD_4_1-4__35.PDF) suggested 700N for milling aluminum with a 1/2" end mill. I will be doing small work, and probably won't use larger than 1/4" end mill, but I have no idea how to estimate this cutting force. The 8020 deflection calculator was simple and easy to use, but I couldn't get a good design that 'felt' good with a 8020 crossbar, so I just went with 3/4" aluminum plate. Is this good enough?

The ballscrews are off center. Does this matter?
If it really doesn't matter at all, then I'll put the Z axis screw on the 'outside' and eliminate the issue that it takes up some working volume.

Bellows for ballscrews and rails. These seem hard to find, or expensive?

The Z axis is different from most other designs, which would fix a gantry without an axis, and then have Y and Z axes on the gantry crossbar. I put my Z rails directly on the gantry pillars, and I wonder if there's any issue I'm not aware of that makes this a bad idea.

In general:
What things should I be worried about that I'm unaware of?
Are there things in this current design that are problems waiting to happen?

[RomanLini]

...
I would like some feedback on a machine design.

...
Are there things in this current design that are problems waiting to happen?

It's a pretty neat design but I think there is a serious flaw.

I think you will need dual vertical leadscrews to lift the Y plate. With just one screw at the side you will get big problems with the Y plate twisting if the spindle is at the far left side of the Y plate.

Dual vertical screws AND dual motors will also help with the other flaw in the design, that the vertical motors will need to lift quite a lot of weight as you lift the entire Y plate and any strengthening beam it may have on the back.

[footloose]

Thanks for the feedback.

I think you will need dual vertical leadscrews to lift the Y plate. With just one screw at the side you will get big problems with the Y plate twisting if the spindle is at the far left side of the Y plate.

I thought the linear rails would prevent any twisting?

Dual vertical screws AND dual motors will also help with the other flaw in the design, that the vertical motors will need to lift quite a lot of weight as you lift the entire Y plate and any strengthening beam it may have on the back.

I had figured I could put one of those hydraulic cylinders they use for car trunks to make the effective weight small. Is that a bad idea?

http://www.speedwaymotors.com/Hydrau...ener,2427.html

[stevespo]

I've been mentally planning some type of "light duty" milling machine as well. Similar components, 25mm rails, possibly 8020 (3060, 3030, 1545, etc), 2.2kW spindle. I already have a moving gantry (8020 woodworking machine) which can take light cuts in aluminum, but want something more rigid, smaller footprint, and mist/coolant proof.

Anyway, I agree with RomanLini. You'll want 2 screws, or somehow move the single screw to the middle - which I suspect won't be possible. Even high quality linear rails are not rack proof. With a single screw in the middle, I bet you'd be able to move one side by yanking on it.

I haven't seen many DIY machines of that type (moving bridge?) I recall seeing at least one or two proposed designs, but I'm not sure if they ever got built. I'm sure there are some strengths to that design, but it appears to add a lot of complexity. Personally, I'm leaning towards something a bit more conventional with a moving XY table and a spindle mounted on a single central post of some sort. Maybe fill the 8020 with some type of e/g mixture for additional mass.

Here's a couple of interesting designs:

http://www.cnczone.com/forums/showthread.php?t=91225
http://www.cnczone.com/forums/showthread.php?t=91552
http://www.cnczone.com/forums/showthread.php?t=64974
http://www.cnczone.com/forums/showthread.php?t=96410

So, I don't have a lot of constructive feedback to offer, simply because I'm not real familiar with that design. My gut feeling is that it might be a fair amount of additional effort for fairly low return. What's your rationale for using that design, instead of a more traditional mill, or fixed (or moving) gantry?

Also, where will you purchase your 18x8" t-slot table? Shars?

Thanks,

Steve

[footloose]

I've been mentally planning some type of "light duty" milling machine as well. Similar components, 25mm rails, possibly 8020 (3060, 3030, 1545, etc), 2.2kW spindle. I already have a moving gantry (8020 woodworking machine) which can take light cuts in aluminum, but want something more rigid, smaller footprint, and mist/coolant proof.

Anyway, I agree with RomanLini. You'll want 2 screws, or somehow move the single screw to the middle - which I suspect won't be possible. Even high quality linear rails are not rack proof. With a single screw in the middle, I bet you'd be able to move one side by yanking on it.

I haven't seen many DIY machines of that type (moving bridge?) I recall seeing at least one or two proposed designs, but I'm not sure if they ever got built. I'm sure there are some strengths to that design, but it appears to add a lot of complexity. Personally, I'm leaning towards something a bit more conventional with a moving XY table and a spindle mounted on a single central post of some sort. Maybe fill the 8020 with some type of e/g mixture for additional mass.

Here's a couple of interesting designs:

http://www.cnczone.com/forums/showthread.php?t=91225
http://www.cnczone.com/forums/showthread.php?t=91552
http://www.cnczone.com/forums/showthread.php?t=64974
http://www.cnczone.com/forums/showthread.php?t=96410

So, I don't have a lot of constructive feedback to offer, simply because I'm not real familiar with that design. My gut feeling is that it might be a fair amount of additional effort for fairly low return. What's your rationale for using that design, instead of a more traditional mill, or fixed (or moving) gantry?

Basically, my process leading up to it went like this:

With an 18"x8" table and a traditional mill configuration, the spindle would have to 'stick out' 10-12 inches, and I was worried about the deflection with an all-aluminum frame, 1 fixed end with the load at the free end is the worst case scenario.

So then I thought, fixed gantry over both ends so that both ends are fixed.
(So, moving XY table, fixed gantry with Z axis).

I then discarded this, because I really like the looks of the 'enclosed' desktop mills: most of my projects are using tiny endmills, 1mm sometimes, in machinable wax or aluminum. The chips are almost powder at this point, so a full cabinet enclosure would be nice to keep the mess contained.

For the 'enclosed' mills, it makes a lot more sense to have 1 axis on the table, and 2 on the spindle, since the spindle takes up much less 'footprint' than the table from an overhead view, and the total enclosed size is basically footprint + travel. (Put another way, if I want the whole 18" width of the table in the working volume, then with a moving table I now have 36" width as an absolute minimum. With moving spindle, if the spindle assembly is 4" wide, then the "minimum" width of the machine is 22" instead of 36")

So, I think I still want 2 axes on the spindle.

But, I am now leaning back towards the more 'normal' design of having the bridge fixed and the spindle moving horizontally along the bridge. This is the way I started anyway. I like the looks of the first link you posted, I hadn't thought of having the horizontal rails on the bridge be parallel to the table.

Also, where will you purchase your 18x8" t-slot table? Shars?

Yeah, I already got that piece, before I wised up and thought, "look at how much the whole thing will cost and decide before you start burning money".

It was $90 on ebay (but $60 shipping to Cali, it's heavy).

I haven't even opened the box yet, hopefully I can get a good price if I decide not to build.

[stevespo]

Yes, I can see that reasoning. I alternate between a design like the first link, and something more like a DIY Datron (moving gantry). A small, dedicated mill is probably all I need to compliment my wood router, but then I think - why not one (enclosed) benchtop machine that can "do it all"?

For small diameter mills and softer materials, I'm sure there are several ways that would work nicely. If you can limit your Z travel somewhat, you'll avoid the long cantilevered arm that is probably the weak link of most gantry designs.

It sure is easy to stick with aluminum construction throughout, but there may be a few places where plate steel would add mass and strength. Epoxy/granite and a granite table might be interesting as well.

http://www.datron.com/products/cnc-m...g-machine.html

At some point (Z travel than 4-6") it might make sense to consider a moving table mill design. Also the tooling choices and cutting forces may move me in that direction. So, I'll be very interested in seeing how your design progresses. It might inspire me to do some actual drawings of my own and we can compare notes.

FYI: here's my current router
http://www.cnczone.com/forums/showthread.php?t=27527

Steve

[RomanLini]

...

But, I am now leaning back towards the more 'normal' design of having the bridge fixed and the spindle moving horizontally along the bridge. This is the way I started anyway. I like the looks of the first link you posted, I hadn't thought of having the horizontal rails on the bridge be parallel to the table.

...

I think the best setup for a small accurate machine is a moving table with fixed gantry, then the vert axis mounted on the gantry.

Check out my machine, it was built for accurate fine machining on plastics and some soft metals;
DIY hobby small plastics mill/router
It has a couple of high perf features like my "L shaped gantry plate" which gives huge gantry rigidity in one axis, also a large box beam on the rear of the gantry. Other high perf stuff like a really flat vertical axis, reduced vertical travel, preloaded polymer bearings etc all add up to practically zero play and flex in all axes.