Building CNC machines from Epoxy+Granite

There's a very long thread on cnczone, indexed here:
[www.cnczone.com]
about building cnc machines using a mixture of granite or quartz aggreate and small amounts of epoxy.

This has applications for making RepRaps and for making CNC mills and routers using RepRaps.

Hi Sebastien,

in the german CNC-forums they show their selfmade 'stone'-bases and mill-beads and discuss different mixtures and methods of mixing and finishing.

The best mixture seems to be small pebbles, coarse sand, fine sand and some glass-dust or cutted glass-fibres with 6% to 10% epoxy.

For exctracting air-bubbles you have to compact the mixture with heavy vibrating, but then you have a perfect stiff structure ...

Another methode for getting feasible building-blocks is a mixture of sweet beer (malt-beer, "Malzbier" in german), some gelatine and ordinary sand.

Or pour waterglass into the sandfilled form, when you didn't want to have a water-soluble material ...

Viktor

This is getting a bit freaky.

On of the many attractions to CNC milling and FDM the the ability to be able to make molds for anything you like almost at will.

I have had a kinky urge to make some Escher style tiles (Flying Books for a library, originally for a house we were buying) for a bit and spent some time reading up on and researching Terrazzo. As a tile material I could use in differing colors and shades but needed to make molds for them.

Terrazzo is pretty much what you are describing here.

Originally terrazzo was made with lime cement then later Portland cement as the matrix. Current trends are towards using an epoxy matrix.....

Many differing aggregates are used, quite often marble chip for it's coloring properties but more durable aggregates are routinely used ie Granite.

Overall, observing methods for binding within the cast parts could be advisable due to the thickness. ie Horse Hair, Mineral fibers and steel reinforcement or mesh. It is after all concrete of one form or another and has predictably similar properties. ie Great in compression, feeble under tensile stresses without reinforcement or pre tensioning.

The prospect of achieving glass smoothness/flatness though by casting on sheet glass is a clever idea.

Hey it looks ultra cool too.

Terrazo suppliers could be a useful source of materials for this activity..

Cheers

aka47

---

Necessity hopefully becomes the absentee parent of successfully invented children.

Here's a photo from the cnczone index-thread which I liked.

Viktor, you may want to include links when you mention the german cnc forums. While I'm not deutscheophonic, some people are.

Edited 1 time(s). Last edit at 02/15/2008 04:12AM by SebastienBailard.

Hi Sebastien,

... i can post some images, but you have to login to read the posts - it's mainly the Forum "CNC-Ecke" [5128.rapidforum.com] (or www.cncecke.de )

Viktor












Edited 1 time(s). Last edit at 02/15/2008 01:18AM by Viktor Dirks.

Ummm I think I could be missing something here.

The arrangement appears to be savagely counterintuitive when you take into account gravity, materials and the center of gravity.

The frame, base, etc appears to be light weight aluminum extrusion and the z axis is Granite/Epoxy.

In an idealised design:-

Surely the heaviest components are better placed lower down (enhanced stability & low center of gravity) with the lighter components anchored firmly on to or preferably in the high mass base. ie Granite & Epoxy Base.

The Table could be a compromise. ie Granite & Epoxy with an indent to take a slab of MDF as a sacrificial surface. (It can also be milled smooth by the tool head) I guess you could also drop in a billet of aluminium and mill in T Slots for fixing. Higher mass for the Table would be beneficial as it would damp the vibration when milling.

The bits that wave the tool head should be v high stiffness but very low mass. ie Aluminum Extrusion or space frame. (Space frame cnc milled out of aluminum sheet etc would be cool)

Low mass gives low inertia (less likely to loose steps) and quicker moves of the tool head for the same power input.

The technique/s etc though in the piccys and from the reading I have done on t'internet are ubber cool and the end effects more so. (Do I sound like an anorak here)

cheers

aka47

---

Necessity hopefully becomes the absentee parent of successfully invented children.

Hi Andy,

... the images are out from the building process (and this are two different machines) - in the end the complete machine should be embedded in virtual stone ...

The framework is made from steel, not from aluminium - the people building this targets mainly rigidity and stiffness, not leightweight systems - so some hundred kilograms per component aren't bad and some complete assembled systems have weights of some tons ...

Viktor

I think that listening to Kraftwerk is imperative whilst building one of these machines...

Wowzie! If you ever have to move you'll need a crane to move that thing.

Kraftwerk were very good.

TransX (Living on video) were better.

League Unlimited Orchestra were the best.

Given that it's Germany though it's more likely to be Tangerine Dream.



I can definitely see a lot of mileage in the Stone base (Y Axis), Stainless steel gantry (X Axis) and Aluminium Z axes though. With the Gantry set into the base. All shiny and custom.

I wonder if my wife would let me sneak it into the lounge.........><

aka47

---

Necessity hopefully becomes the absentee parent of successfully invented children.

When moving I recommend Einst

Viktor

Which epoxy is the favored one and are there some useful example recipes somewhere ??

cheers

aka47

PS Ripping apart I would have to go for Ramstein.

---

Necessity hopefully becomes the absentee parent of successfully invented children.

Hi Andy,

... it's every type of 2K-epoxy you can find, the people didn't fix on a special type, it's simply the 'glue' ...

For the best results you have to mix coarser and finer sand with some sort of cutted fibres (glass, carbon or such) with 6% to 12% of epoxy.

The mixture didn't flow, so you have to insert it with some force and compress by vibration to exhaust air-bubbles.

Viktor

Wait just a minute here, a cnc made of beer? Could you give more information on that? Could you potentially substitute out gelatin for alginate? Alginate is a binder and gelling agent that can be made by bacteria. Sand is available everywhere, and most beer is made with microorganisms, and you can grow bacteria, so it might be possible to achieve grey goo potential with RepRap.
more on alginate: [en.wikipedia.org]

It definitely has some use in deserts, which are made of sand and don't receive much water. Deserts also receive a lot of sun light too, so you put a solar-powered total RepRap out in the desert, and solar power becomes the cheapest source of energy in two years or less.

And any end of the world scenarios are prevented by the fact that it dissolves in water. What's the fun in that?


BTW isn't asbestos a mineral fiber?(asbestos is very very heat resistant)

OK

What looks like a useful supplier of Resins and molding compounds in the UK (As well as other bits and pieces including carbon fiber) Note I have never tried pricing or buying form these folk I just found them on t'internet but they do appear to retail.

[www.mcmc-uk.com]

Their sr-ca-85 and sr-1690 sd-7160 seem to pretty much fit the bill, and they do silicone molding stuff too. They have good data sheets online so even if they are no good for you their data is a good starting point.

Pressure and vibration. If I was going at this low/intermediate tech, I think I would be starting out with a casting bed mounted on old car springs, with an electric motor and off center weight attached to give vibration. Pressure could be vacuum bag with ye olde Hoover to give suck. Used in combination with a slow cure time and low viscosity casting Resin to get best effect (Reasonably Long settling time).

For smaller components, a Hot water bath could give the necessary post cure backing at aprox 40 deg C the data sheets suggest. For larger items probably greater ingenuity would be required. (Hot tub and crane anybody)

Use of asbestos in the UK (and I think pretty much in Europe is very much a no no, classed as v dangerous hazchem/carcinogen and provokes much hysteria from local authorities) cheap fibrous material is likely to be rock wool glass fiber insulation, the yellow fluffy itchy stuff you put in your loft. (Ferrodo use this sort of stuff in their friction material these days instead of the Asbestos which is now a complete no no, cuts down on the litigation) Horse hair has proved functional over centuries of lime plaster walling too.

Most builders merchants in the UK stock Granite Aggregate (Usually Grey) and of course the usual colours of sand. For colorful chippings find a monumental mason they have all sorts for topping off grave plots (Including very colorful glass like stuff). (Leave the cemetaries alone)

A clean concrete mixer has got to be good if you are doing any quantity, mixing heavy material by hand for the period required to get good dispersal is back breaking work. (I have fitted floor safes in my younger days using a 3-2-1 granite concrete mix made in a wheelbarrow and can testify to this)

For recipes I think I will go look at some of the references I found for Terrazzo they give ratios and grades which should be pretty proven if only for highly durable flooring and can be used as a start point to tweak onwards.

Hope all this is useful to someone.

Cheers

aka47

---

Necessity hopefully becomes the absentee parent of successfully invented children.

Terrazzo

[en.wikipedia.org]

[www.premfloor.co.uk]

[www.ntma.com]

NB this is all about marble chip flooring, however if you substitute Granite for Marble you won't be far out. When casting V Large pieces including larger aggregate sizes in the core of the piece will substantially reduce cost. (Possibly using a two stage process) plain old concrete could even be used in a core slab. drastically reducing cost but getting the same finish results.

Note the NMTA references above I think they have got their mesh table mixed up and the retains on mesh versus drops though mesh should be the other way around.

The NMTA though in their specifications section give a useful amount of technical details that could be of value if you selectively ignore the flooring and cementatious references.

Cheers

aka47

---

Necessity hopefully becomes the absentee parent of successfully invented children.

ive made epoxy granite fo 24 years what you wanna know

For starters...are you related to anyone from Oklahoma?

Are you in Oklahoma Sean? I live near OKC.

An about the granite.

Do you have recipes etc proportions, ratios and grades of components ??

aka47

---

Necessity hopefully becomes the absentee parent of successfully invented children.

Two questions:

Are these materials pretty straightforward to use in a small workshop? Is ventilation an issue?

Are people interested in using this to build, say, a laser cutter and a large CNC router?

Sebastien

I can't actually speak for anyone else on this one but my interest as well as making granite tiles stems from a method to add mass to a V. Small (Minature realy) desktop design of rapid manufacture machine.

The sort of thing that will do PCB work on a standard eurocard size workspace, Mill out board shapes and copper, drill holes, draw on etch resists and component idents, deposit solder paste, pick and place SM components etc.

As well as be able to mill out enclosure openings (LCD Display windows, Potentiometer cutouts etc)

Making something small is great but it will generaly lack enough mass & stiffness.

If we then consider that such a small workspace could also with the addition of an extruder toolhead make just about all of the Darwin Parts. (For bigger fabbing) Then it could be a repstrap.

I'm thinking here of something here that is to RepStrap and Electronic prototyping what Arduino is to Embeded Systems.

It's Open Source/Creative Commons, looks cool and sexy and does the job and more besides.

My current thoughts are something along the principles of Ianus (My test bed) but much much smaller with the base being cast from Granite Epoxy Concrete. The cast base would include either by shape or by casting in, the rails for the lower table.

The Gantry and Bracing (Something I need to add to Ianus) would be cast into the base in the same molding operation. These would likely be made from Pipe/Tube for ease of construction.

The Gantry Rails would be made from square hollow section (possibly steel of some variety for strength and cos it's cheap) fish mouthed at each end and bolted through the gantry frame using inexpensive studding.

Bearings and trucks would be created using M8 x 20 bolts, skate bearings and a M8 nut each screwed into holes drilled and tapped into equal angle section for the same square dimensions as the square hollow section making high quality linear slides.

Multipart molds for the Granite Base should be something easily printed out from an RP machine. Or constructed as a one shot mold using standard ish materials and molding techniques.

Thoughts for what they are worth. (No serious intent here just day dreaming)

aka47

---

Necessity hopefully becomes the absentee parent of successfully invented children.

Andy,

What size workspace are you envisioning?

Edit:
Something this size, 24"x15",
[www.k2cnc.com]
would be useful in the small workshop and for what you're doing, but that's moving gantry rather than moving table.

Personally, I want something big, 4'x8' ideally. That's a different project, you wouldn't want to build a huge cnc route to do pcbs.

We should pick Viktor's brain regarding the laser cutter stuff.

Edited 1 time(s). Last edit at 03/02/2008 01:56AM by SebastienBailard.

Sebastien

Re 4' by 8' (2440mm x 1220mm) I assume here you are talking the standard sizes for sheet materials. ie ply wood, plastics & some metals etc. (Steel is'nt typical for those dimensions but I think can be purchased after being processed such)

The 4x8 is typical for materials that are used to clad something (2440x1220 goes very well with a modular distance/grid of 600 ie that used in building and kitchens plus some cutting)

A guy I know has a sign making business and he has just such a router with a vaccum bed. Definitely moving gantry, moving something of that size about as a moving table at speed is guaranteed to be problematic and something of a safety hazzard.

BTW noise is a problem for the sign maker, hearing defenders are a must and the unit lives in an industrial unit away from residential areas as it can often be running all night. 8x4 works great as a sound board even on a vacuum table. He also has issues with some of his employees mis setting the max plunge depth for the router, it makes a mess of the vacuum bed. Not cheep.

The only real exception to the moving bed being problematic at these dimensions is where the material can be loaded on a roll and the machine operates like a roll plotter/router.

Again though this imposes constraints on what you can cut out and how you keep it in it's place relative to the rest as it gets thrown around. No vacuum bed.......

I come at all this from an electronics engineering/production point of view and in seeking to bring the means to manufacture to the many.

To actually make a given electronic product (Embeded or otherwise) there is just so many different things an individual has to know how to do, be able to do and ultimately tool up for.

Making low volume production is very cost prohibitive. Especially if you want the end product to have that professional look you need to be able to sell it.

Incidentally standard eurocard size is approx 10cm x 16cm. In terms of electronic module size this is good because when you want a system on a board size greater than this you break it down into eurocard sized chunks and mount it all in 19" racks or multi card enclosures A lot of Low volume production Lab Electronics and Instrumentation is made this way (Nothing like confusing measurement systems is there, fact of life in the UK)

Most case manufacturers make a good range of eurocard compatible enclosures with compatible mounting systems.

The Range of 19" Rack manufacturers is quite surprising and their production ranges resemble something from Meccano.

The eurocard foot print is I think a size that is sufficiently large to empower it as a workspace dimension for Rep Strapping. (I don't think any of the parts is larger than this.

I agree with you that moving gantry is the optimum solution for big work area.

The only quibble I have with moving gantry is cumulative error which could arguably be engineered to a figure your application could live with in most applications. For sheet sizes 8'x4' cutting out large items (signboard letters) a precision of 1mm is plenty.

On cumulative error, I see it this way, Where you mounting one axis upon the moving part of another you have at best added the positional error of one to the other. In a worst case (Articulated Arm Robots) you have multiplied them. Where you then mount those two axes on the moving part of a third you have done the same again.

In a moving gantry system your Z is mounted entirely on your X which in turn is mounted entirely on your Y.

For the Ianus test bed I am working on this is why I chose to go for a fixed gantry and moving bed. The Y and X are mounted on a stiff frame that arguably yeilds little only the Z is mounted on the X minimizing cumulative error at the expense of having a machine size somewhat greater than the bed/workspace. But as I want to be able to mill some metal components as well (maybe aluminum molding) I went for this configuration.

Are you perchance looking to use such a machine for sculpture ??

Cheers

aka47

---

Necessity hopefully becomes the absentee parent of successfully invented children.

"Are you perchance looking to use such a machine for sculpture ??"

Maybe eventually; if I start working in a large scale, I could use it to route out the foam for a large relief work or the foam slices to make up the body of a larger statue. I'm still in the 'boy that would be really cool' stage.

I think it's worth highlighting the fact that we probably want to design up three CNC routers,

small,
14" x 25" (356 mm x 635 mm)

medium,
2' x 6' (610 mm x 1828 mm)

large.
4' by 8' (1220mm x 2440mm)

(And similarly for laser cutters)

Edited 1 time(s). Last edit at 03/02/2008 10:07PM by SebastienBailard.

Cool

Personally if I were looking at these sizes I think anything above 1 Meter in either X or Y (as they are effectively transposable in machine terms moving Gantry would have to be it.

Moving tables waste too much space in that your machine size is a multiple of your table size in at least one axis. (X & Y).

Up to 1 Meter and it's going to be a compromise between preference and efficient machine size versus allowable error.

That's my take if it is worth anything....

Cheers

aka47

PS I am still "Boy who would be cool" and will probably be so until my grans age or decomposition gets me. Unfortunately the more I stick at it and try again the more the majority of the population (my kids included) call me an anorak.

---

Necessity hopefully becomes the absentee parent of successfully invented children.

Andy wrote:
PS I am still "Boy who would be cool" and will probably be so until my grans age or decomposition gets me. Unfortunately the more I stick at it and try again the more the majority of the population (my kids included) call me an anorak.

At least the kids say it with affection (well i think mine do )

---

Ian
[www.bitsfrombytes.com]

On making additional machinery using rep rap

There is a fundamental limitation (Which Darwin cleverly takes into account) in that you can not make a part for a larger (Or same size) machine which exceeds the size of your work area.

Your machine at which ever stage must always be larger than your work space in order to reach all of it's workspace.

Darwin works around this by ensuring that each individual component fits within it's workspace and then uses a semi space frame approach to expand the space between the components giving you a machine of the same size or larger (Just use longer rods etc)

The amount you can expand the basic Darwin model and still have usable machine is governed by the diameters of the interconnecting rods.

The only way top go beyond this fundamental limitation and keep the rod sizes the same is to make the leap to using structural members that are a full space frame.

For me personally Darwin is great but has too many different components in it's makeup. for the next generation we ideally need fewer components of more universal application. Think of the Balls and Sticks approach to making molecular models or the Lego Approach to building houses. (I think that the attachments for the cross bracing could do to be built into the corner blocks as blind holes, this would mean making the corners blocks bigger, but leave less components to print)

A modular space frame system that works in this way can produce machinery larger than itself by more orders of magnitude than our current most excellent design.

Not only that but maybe also other things too. Automobiles, Housing, Workshops etc

Interestingly enough, a space frame bed that is then filled in with a composite like Epoxy + Granite becomes a re-inforced composite bed with very desirable machine characteristics and overcomes the limitations of the inherent weaknesses of Concrete Mixes .

In this respect I am very interested in looking at making larger machinery from the smaller and Darwin is a very able enabler to produce the right designs of universal connectors etc that would be needed.

Cheers

aka47

---

Necessity hopefully becomes the absentee parent of successfully invented children.

At the practical limit, assuming the darwin printing space is a cube,
((1+1)^.5)+1)^.5 = 1.553774
The darwin, assuming it could print everything, including the rods, could print the parts of a unit up to 1.5 times the dimension of the printing volume.
Rods, being the most difficult, would be printed at a diagonal.

Unfortunately even printing the rods on the diagonal would not allow you to be able to print a rod that would make a diagonal brace that went corner to corner on the frame (just as so many of them do on Darwin) as again it would be larger than the diagonal available for the workspace.

Good try though.

However utilizing space frame methods, no one rod would have to be bigger than the workspace for structures as big as you could fund.

It's all down to the size of the triangles.

The difficult bit would be devising a smooth carriage as space frames are inherently bumpy (It comes of having all those nodes joining together the short rods). But not I think Impossible with the right additional components.

aka47

---

Necessity hopefully becomes the absentee parent of successfully invented children.