Semi-newbie looking for some design advice

[galacticroot]

Note: This is not a wood router, but this seemed to be the best category, since it is basically a typical CNC gantry router with a very different type of tool attached.

I'm hoping to start building a modular CNC "wax/resin fabricator". It applies droplets of material to build up an object, layer by layer. While there are commercial machines that do this, I haven't found any home built versions, so I'm going to try to design one.

This machine will hopefully be a prototype for a machine that I would like to eventually sell in a pre-built or kit form for (hopefully)less than $1000 each, after various improvements. The plans and software will probably be open-sourced(GPL probably).

The first part of the machine is just a simple 3-axis gantry machine. The theoretical maximum object size will be around 2'x2'x4'.

Here are some of the basic design considerations so far. Since this is also my first machine of this type, I'm looking for any design suggestions. Also, I want to make this design as cheap as possible, but is there anything where it would be worth going with something more expensive?:

-Most of the machine will be made from wood to cut down on cost. This includes the table(which will have tin foil placed on it prior to starting a build) and most of the gantry mechanism.
-The machine will use cheap standard v-screws. These are a small fraction of the price of ball screws, although they do wear out sooner. I already have the screws, so I'll at least try them at first. They don't seem to have a big problem with backlash. Bolts will be attached at each end to hold the screw in place.
-The machine will use a single center-mounted screw(like in the zoltar router) to move the gantry frame across the table. The sides of the gantry will also be similar to those on the zoltar router, but will be triangular and made of thick plywood.
-Another single screw will move the slider along the top of the gantry.
-The Y axis slider will move along two 2x4s, and use small wheels as bearings. I will need to figure out the wheel adjustment mechanism later.
-The Y axis screw will go through the middle of the 2x4s.
-The Z axis will be similar to the Y axis design, but mounted vertically.
-The screws will be connected to stepper motors using high-pressure tubing and hose clamps a couplers.
-My stepper motor controller will be a microcontroller and 4 10-amp logic-level mosfets. The mosfets are IRLZ14, and I already have them in stock. I've tested them, and they drive unipolar stepper motors very well.

I'll probably draw the design in ProE and render some pictures of it later this week.

[craftech]

Hey galacticroot this seems like a realy cool project I would not advise using 2x4 or plywood these materials will warp and shrink after time ,if you want to keep the cost down use MDF is not that expensive and you can laminate and cut out all of your parts after with your CNC for production of your kits , this material is easy to work with and there is a wide variaty of attachements and hardware that you can purchase in any home depot for assembly.
OK
craftech

[Graham S]

The hardest part of this idea is the wax application and the software, any ideas for those?

[ViperTX]

Cool....if I help with the design...do we share the revenue?

[galacticroot]

Originally Posted by Graham S The hardest part of this idea is the wax application and the software, any ideas for those?

Yep. I built this thing yesterday:


Its a stepper motor driven mini screw pump. I don't have a tip attached to it yet, but it will hopefully be able to pump wax in small, precise, amounts. The tip will be a blunt syringe needle or other very small nozzle. I still need to design a heating jacket for it.

The software shouldn't be too hard to write. My first program will probably just use depth maps for the surface data. Writing a similar program for my mill took about an hour. I suspect the biggest problem would be calibration of the amount of wax pumped.

Cool....if I help with the design...do we share the revenue?

Heh, well, it will be mostly, if not entirely, open source(I may eventually make some seperate proprietary add-ons if I actually sell these), so you can use the design however you want, just as long as you also keep it open source.


If the wax idea fails, I am also considering another process that builds with metal instead of wax, based on electroplating, but I will need to experiment with it before I will know if it would even work. The base machine would still be the same for either process, except the metal machine would have an electrolyte tank instead of a table. Other possible methods of building 3D objects include a thermal wax transfer system, or various systems working with epoxies.

Since the basic CNC machine would be modular, tools could be easily changed out to do all sorts of processes, also including routing, drilling, pick&place assembly, vinyl cutting, plotting/painting. I suppose it could even have two or three tools mounted at once for things like circuit board assembly, although that would reduce the travel.

[Graham S]

You say you have a mill, why not make a wax add-on. Routers are easy enough to make and there are already free plans available for those. If you can get a decent applicator working then you have really done something. A screw type does seem a good idea, much as is done in injection moulding. I wonder about temperature control, I assume the wax they apply is not totally molten. I also wonder about nozzle design, it is a real shame you never see decent pictures of such things. Where did your screw pump come from?

A prototyping process based on electroplating would be increadibly slow, I can imagine it could work though, you could have a wire electrode that could be fed out of the end of a tube as it was dissolved. The tip of this tube would have to be held almost in contact with the metal substrate so as to concentrate the current as much as possible to provide some resolution. It is very much like ECM in reverse, you could make increadible parts I am sure.

Graham

p.s. I did see something about a homemade wax printer based on a HP inkjet once but I couldn't ever get any decent info.

[galacticroot]

Well, my current mill is a Max-NC 10, which is too small to do anything like that with.

As for the electroplating version, that was basically the idea, although I was going to use a large plastic tube with a tiny plastic capillary tube at the end. The machine could be fed with bronze/copper/whatever. The metal would be oxidized in the large section of the tube, flow down through the capillary tube, and be reduced on the workpiece at the end of the tube. The capillary tube would limit the plating area. It would be very simple, although limiting the range of the plating might be tricky.

I'm probably going to do some experimentation with both wax printing and metal plating methods and go with whichever seems better. The wax printing wouldn't be that fast either, considering that it would only have a one "pixel" print head.

[cbass]

Hey Graham,

Awesome idea! I'm a sculptor who has been considering how to use electronics and computers to make work for the past few years and an additive process makes perfect sense.

I would say this is VERY marketable. However, if you really want to make money at this, I would focus on this as being an add-on to a pre-existing cnc system. Something that will replace a router. This may keep the costs down and will let you concentrate on one area (rather than having to get into machine design as well). I realise this may not be ideal, as a RP system has specific requirements, but its just a thought.

A question that arises for me is how you will support the wax as it is being applied. The areas of the model that contain just Z movement shouldn't be a problem. Its when there are areas that combine X, Y and Z that problems may occur (in a curved surface for instance). Gravity is not your friend, especially since you are using a liquid or semi-liquid wax. From what I've read, a sacrificial powder is sometimes used. Maybe this is not necessary as the wax will cool almost instantly and will be able to support itself.

Anyways, this does sound like a great idea. Bring Rapid Prototyping to the masses. Let me know if you are looking for a partner(s). I make a mean espresso and could use the royalties too!

Carlo

[ViperTX]

Okay, it seems to me that when dispensing wax you must have some method to control the amount of slump (cool air jet?).

Also it seems that it would be better to machine a piece of wax using a 4th axis ....any thoughts?

We have one company in Austin that was using a laser beam to solidify some special resin to generate these quick prototypes.

[Graham S]

Galacticroot (any chance of a real name), I have a Proxxon which is smaller than a Max and I would still consider it, it didn't stop me going to 5-axis Remember that the basic depositing system could even be tested on a single axis or better yet two.

Your idea for the tube is good however you need to remember that the process is about current flow, the current is what can be used to localize the plating. Actually you can't really localize the plating as such but you can produce an area with a higher plating rate than the surrounding. One thought, if you got a solution such as copper sulphate that already contains a salt of copper it might be possible to produce a multipronged electrode and scan this across the surface energizing only the electrodes that need to be and keeping them very close to the surface, perhaps having only the tip uninsulated, I am not sure if this will work or if the tips would just dissolve, I suspect stainless steel or Ti ones wouldn't.

Sorry should get back to wax.

Viper, you can't make one piece hollow objects with 4-axis.

There are lots of rapid prototyping methods, some harden resins with UV lasers, others melt powered plastic or metal with a laser. Some add wax or plastic with a print head and some harden starch powders with a binder. There is also a type that builds objects up from layers of sticky back paper which are laser cut during the process.

Graham

[galacticroot]

Well, I am going to consider both the wax and plating methods, and possibly a laser powder fusing method, although that would be expensive.

I'll probably experiment with the plating process this weekend if I can find a place to buy sulfuric acid to make the electrolyte with.

Titanium is oxidized by hydrogen ions, but silver isn't so it might work. Of course once you add power it might dissolve anyway. I could easily make an array of about 25 individually controllable electrode/tube sets to form the "print" head, though.

[ViperTX]

Okay, so what problem are you attempting to solve. Seems that if you're making stuff out of wax and it's not for rapid prototyping...then you'll probably thinking of casting it or you could also be thinking of plating over the was....you know like the Baby Shoes that are bronzed. As far as machining wax using a fourth axis you could start out with a hollow cylinder....