Vertical X-axis PCB Mill carriage

Slightly self-promotion for the Vert X-axis competition, mostly an area for discussion so the thingiverse comments don't get too overwhelmed with my ramblings!

PCB manufacture needs to be easier for reprap; countless numbers of printed parts can be made, but I see electronics as a real bottleneck. Just like 3D printing, the more accessible it is, the better the take up. To this end, I have started designing a milling carriage for the Vertical X-axis standard: [www.thingiverse.com]
The idea is that the normal x-carriage can be unclipped from the x-axis, leaving LM8UU bearings on the linear rods, then this one can be clipped on to mill PCBs or other small parts.

References
Basic reprap PCB milling page here: [reprap.org]
Linear mounts are 70mm apart, as per the standard [reprap.org].
The LM8UU Linear bearing mounts were taken from Josef Průša/Greg Frost's X-end holders, which I think have the best grip on the bearings: [github.com]. I guess my idea is that it could be swapped, not that you'd be doing it every day. I have found Greg's other design [www.thingiverse.com] doesn't have as strong a hold, as is more difficult to print reliably (lots of small towers).
Spindle layout is based on David Carr's Mantis 9 Low Cost Spindle v2.0 [makeyourbot.org].
The carriage will incorporate the belt doubling system from Madkite [www.thingiverse.com] that Emmanuel is building into the Vertical X-axis he has designed [reprap.org] and [www.thingiverse.com].

Problems using a reprap for milling:
1. Using belts to drive something that should really be driven by leadscrews - Madkite's belt doubling technique I hope will improve the belt situation
2. Wobble in the drill - David Carr's spindle design should eliminate problems at the motor/spindle that others have seen using various Dremel-style drills
3. Wobble in a X-axis - hopefully the change to a vertical X-axis should mean this is much more tightly constrained
4. Wobble in the frame - The larger problem is the X-axis pushing left and right. I'm designing a simple cross-brace to triangulate this area with threaded rod. It's not ideal for 3D printing as it will interfere with the build area, but will not get in the way of milling, and will be easily removed to convert back to printing.

Any observations, comments and/or suggestions gratefully received!

Regarding wobble in the frame, maybe we should look at alternative and stronger frame constructions? Preferably not wood as nophead has reported inaccuracies due to wood being influenced by weather conditions.

devijvers Wrote:
-------------------------------------------------------
> Regarding wobble in the frame, maybe we should look at alternative and stronger frame constructions?

I agree, and evidence from CNC machines shows that MDF and other wood-based machines expand and contract to a measurable degree depending on humidity. Those built from thick plastic sheet or extruded metal seem not to suffer these problems.

However, the thing I'm designing is for small-scale hobbyist work, where perhaps there is only space/budget for one machine - that's my main constraint on the design. It still has to work accurately, but the idea is to have a swappable tool head on a Mendel/Prusa (albeit with many changes, eg linear bearings, vertical x-axis) that allows for milling to 'hobbyist' accuracy; I'm not looking to design a whole machine! I think there is sufficient scope for stiffening the frame specifically for milling. Still, don't know until we try...

Fair enough. What I understood so far is that a standard Mendel/Prusa frame is not stiff enough for the minimal accuracy that is needed. That means that to get from a standard Mendel/Prusa to a machine that can mill for example the Gen7 PCB one would need to upgrade her machine.

Those upgrades could possibly include (needs to be determined through experimentation):

- vertical X axis,
- bowden cable extruder,
- changes to the frame to improve stiffness.

Obviously once you have a working 3d printer you're set to start printing the required parts.

Can you elaborate on the spindle drawing on the mantis website? [makeyourbot.org]

I've looked at it a couple of times but I can't spot what's novel about this design.

devijvers Wrote:
-------------------------------------------------------
> Those upgrades could possibly include:
> - vertical X axis,

That's what this is designed for!

> - bowden cable extruder

Errr... what? For a mill?! Don't understand this! There are vertical X axis designs that include the extruder, no need to mess around with Bowden cables and their problems (I have a repstrap with one, direct drive is better!)

> - changes to the frame to improve stiffness.

Absolutely. The two areas on the Mendel/Prusa that are not triangulated are the square parts of the frame - the base and the front and back. The base sits flat and generally shouldn't need bracing, but there is possible movement in the frame 'squares' front and back. A cross-brace is needed, which I'm working on. For FDM this isn't needed, and would get in the way, but for milling I think it would very advantageous.

> Can you elaborate on the spindle drawing on the mantis website?
> I've looked at it a couple of times but I can't spot what's novel about this design.

I don't think it is particularly novel, except maybe the use of PVC tubing in the drive shaft; it's just a low cost way of making an accurate spindle that doesn't wander. I haven't built one yet, I'm using it as a starting point to understand the problems. I'm trying to spec a metric version, sourcing in the UK, so it will change a bit.

Regarding accuracy, yes it needs to be accurate. However, if the PCB is set up with Voronoi areas rather than straight tracks, the milling doesn't need to be quite so accurate, I think. [groups.csail.mit.edu]

>> - bowden cable extruder

> Errr... what? For a mill?! Don't understand this! There are vertical X axis designs that include the extruder, no need to mess around with Bowden
> cables and their problems (I have a repstrap with one, direct drive is better!)

Sorry, got confused

I've read elsewhere on this forum that the copper dust released during PCB milling is a problem for the reprap, namely for the bushings and electronics. Apparently the solution for this is to submerge the PCB in water. Any thoughts on that?

Quote

evidence from CNC machines shows that MDF and other wood-based machines expand and contract to a measurable degree depending on humidity.

This is true, but not a problem, as these measurement changes happen very slowly, like within a day. Within the time of milling a PCB, size changes are neglibile.

To some extents I still think using wood for the large parts is better than using steel. A Mendel with only printed and wooden parts, anyone?

---

Generation 7 Electronics	Teacup Firmware	RepRap DIY

Quote

A Mendel with only printed and wooden parts, anyone?

Yes that is what I am working on ATM, see [forums.reprap.org]. All the studding frame is replaced by a few sheets of 12mm MDF, much stiffer in the directions it needs to be. I doubt it would be much good for milling due to the use of belts, but it would be a lot better than a normal Mendel.

---

[www.hydraraptor.blogspot.com]

Quote
nophead
Yes that is what I am working on ATM, see [forums.reprap.org]. All the studding frame is replaced by a few sheets of 12mm MDF, much stiffer in the directions it needs to be. I doubt it would be much good for milling due to the use of belts, but it would be a lot better than a normal Mendel.

Hmm, very nice. I'd try to get rid of the continuous belt though. The woodwork seems simple enough too.

The belts aren't really continuous, I just hadn't got round to modelling the gap and the belt clamps on that snapshot, which is a couple of weeks old.

The advantage of wood is you can buy it cut to size with perfect right angles, so you only have to make the big cut out in the middle, or make that from three rectangles screwed and glued together.

---

[www.hydraraptor.blogspot.com]

Quit cost estimation:

- cut wood pieces: €15
- metalware to build frame: €15
- printed parts: €35 + shipping
- electronics + motors: €100 + €55 + shipping
- belts: €15 + shipping
- linear bushings: €15 + shipping
- hotend: €40 + shipping
- extruder montage kit: €15 + shipping

totals: €305 + shipping

Which is about €75 less than a prusa + you've got a platform to mill. Very interesting.

How about a vertical X-axis though?

devijvers Wrote:
-------------------------------------------------------
> How about a vertical X-axis though?

Ooooh, don't get nophead on that subject! I don't think he's keen, or rather, doesn't see any particular advantage in it. I think it's a fair comment, but only a bit of experimentation will show whether there are advantages in FDM, or rather making sure there are no disadvantages. For milling, I think there are distinct advantages, mainly in that the support of the spindle is tightly constrained in the vertical axis; I don't think I've seen a CNC router that doesn't have a vertical X-axis. Getting back on topic (please! ), I'm at least prepared to try milling on a Prusa frame. If it can be demonstrated to be workable, what's the problem? At the moment it's just people saying "it can't be done"! Okay, based on previous experience of standard frames using Dremel extensions (poor/loose spindles and slow mill rotation speed seem to be two of the main problems reported), but I'm proposing changes to address many of the issues. Feed speed will not be high, and drill tip resistance should not be great either.

devijvers Wrote:
-------------------------------------------------------
> I've read elsewhere on this forum that the copper dust released during PCB milling is a problem for the reprap, namely for the bushings and electronics. Apparently the solution for this is to submerge the PCB in water. Any thoughts on that?

I suppose you could do that; I read a film of light oil over the PCB is enough to trap swarf and dust and produces a nice cut too. Or rig up a vacuum cleaner. Again, CNC machines have survived for years. But then they use linear bearings with dust shields. Oh, hang on, so does my Prusa! And so does my mill design! Electronics and motors may need shielding, though.

droftarts Wrote:
-------------------------------------------------------
> devijvers Wrote:
> --------------------------------------------------
> -----
> > How about a vertical X-axis though?
>
> Ooooh, don't get nophead on that subject! I don't
> think he's keen, or rather, doesn't see any
> particular advantage in it.

I believe Mr Nophead is (as usually) correct. And i would say, pay attention and meter the length of his reaction in order to asses how bad situation is. Me for one, i would also like to yell more on the topic, except you know, nobody would pay attention to me ever, so ... i will have to save my breath


> I think it's a fair
> comment, but only a bit of experimentation will
> show whether there are advantages in FDM, or
> rather making sure there are no disadvantages. For
> milling, I think there are distinct advantages,
> mainly in that the support of the spindle is
> tightly constrained in the vertical axis;

There are no mechanical benefits in vertical bars except less working area is "wasted" on Y. There are disadvantages as mentioned, but cnc can "afford" those - dunno if its proper way to put it but thats what i think. Different world. Their bars are hugely over sized, and sometimes bars are also are part of the structure, which accounts for something extra and puts the weight of the carriage itself in the shadows. It is not our case though, in both terms.

I would say bar positions and dremels vibrating are not linked. I think the spindle issue might be a confusion. Trying to use a dremel by gripping it to its case is not appropriate and its not what the tool was designed for. Any number of its internal parts might vibrate like church bells. The proper milling motors have dedicated mount point for that, Euro 43 mm mount, and all internal parts and everything makes a common body with that support point. Which makes it a different game, even some datasheets (like suhner) might spec the vibration ranges. I have a 1.1kw/vac motor and a 250w/vdc motor both with proper E43 mounts, and in comparison my dremel (~ish) tool is like a toothpick. In comparison dremel case is not a support point, its ... just a cover, not much more.

> think I've seen a CNC router that doesn't have a
> vertical X-axis.

Attached a picture of horizontal X bars design cnc router.

As bottom line, i think vertical bars are doable, except there are a few of things that needs to be cleverly dodged. And also to be worth doing it, the benefits should surpass the disadvantages it brings.

Nothing stops us to try mill with our current horizontal setup. Might make 2 carriages on same bars (maybe a little more spaced so a small motor fits between the bars, and then print with one carriage while the other is parked to opposite side. Park the extruder to the other side and then dismount its belts, mount them to the other carriage, and use second one while first one is parked. This wastes some working area though, which could require an overall machine size increase to remain the same as before. Cant change between without human intervention. Maybe it sux. But well, its a thought, and maybe even a variant if all you want is a weak mill setup and jobs are separated from one another.

Edited 2 time(s). Last edit at 09/16/2011 07:53AM by NoobMan.

Quote
NoobMan
Trying to use a dremel by gripping it to its case is not appropriate and its not what the tool was designed for. Any number of its internal parts might vibrate like church bells.

Good point.