What would it take to extrude aluminum?

Has anyone tried extruding aluminum?

The obstacles I can think of are as follows:

1. A hot end material that will not melt itself
2. A temperature indicator that will not burn out
3. Obtaining over 660C
3b. Staying over 660C consistently without damaging the heat source
4. A build plate that will not melt or catch fire
5. Cooling
5b. Localizing the heat so that it does not creep up the aluminum and damage the machine
6. Limiting the size of the hot end so that it fits in the machine (i.e. not too bulky a heat sink or too long an extrusion path)
7. Oozing the molten metal out at just the right temperature and speed so that it bonds but does not run.
7b. I don't know if oxidation would be a problem, but if it's anything like welding... yeah.


Can anyone else think of some obstacles, or maybe some solutions?

Remember, the idea is to keep it inexpensive. Of course, if we threw enough money at it, all problems would disappear... but that's just no fun

i've taken one of my hotends to 590c which was interesting,

at those temperatures things like temperature control get fiddly and it would have to be made from brass or a stainless steel, the biggest problem however is that aluminum in it's molten state will react with materials like stainless steel and i think brass,

sodium silicate is probably a good starting point as it won't degrade as much and will take the temperature,

---

-=( blog )=- -=( thingiverse )=- -=( 3Dindustries )=- -=( Aluhotend - mostly metal hotend)=--=( Facebook )=-

Stainless reacts with aluminum? Hmm well what do they make MIG welding electrodes out of? I'd bet that would stand up. I know you can MIG aluminum. Or aluminum casting... what container do you keep it in?

the molten aluminium will react with the stainless, but only really when it's molten ( the aluminium)

---

-=( blog )=- -=( thingiverse )=- -=( 3Dindustries )=- -=( Aluhotend - mostly metal hotend)=--=( Facebook )=-

so... is it possible to soften aluminum to the point where it bonds with itself, but does not yet have the fluidity to screw up the stainless steel?

secondly, what do you mean by "react" ... is this a slow corrosion, over weeks, or is it abrupt?

thirdly, what effect would chromium have on stainless steel, with regards to molten aluminum? It supposedly drives up the corrosion resistance...

Wissing Wrote:
-------------------------------------------------------
> so... is it possible to soften aluminum to the
> point where it bonds with itself, but does not yet
> have the fluidity to screw up the stainless
> steel?
>
> secondly, what do you mean by "react" ... is this
> a slow corrosion, over weeks, or is it abrupt?
>
> thirdly, what effect would chromium have on
> stainless steel, with regards to molten aluminum?
> It supposedly drives up the corrosion
> resistance...

the problem with aluminium is that the window of softness is tiny, i'm not sure by how much it'll depend on the grade,

when i say react it's in a corrosive nature, it's been a problem for a lot of the diy backyard metal casting crowd who have been using stainless steel and mild steel containers to melt aluminium in, from some of the blogs i've read it most of them only get a few uses out of somthing with a wall thickness of 3mm of mild steel,

i have no idea about chromium but i dare say exposed to extreme heat like that it probably wouldn't last too long,

i've recently picked up a graphite crucible for melting aluminium in because i want to try out this lost pla casting, but my biggest issue is getting a blow torch that runs from bbq gas

---

-=( blog )=- -=( thingiverse )=- -=( 3Dindustries )=- -=( Aluhotend - mostly metal hotend)=--=( Facebook )=-

Aluminum will oxidize readily near the melting temperature. You might need something like an inert gas air assist or possibly a fully purged enclosure to help mitigate the oxidation.

There are probably numerous materials that you could use to extrude aluminum. You'll need to research binary phase diagrams and find the minimum temperatures at which intermetallics form. In the case of steel, it's an alloy of several different materials. At the melting temperature of aluminum, it's likely that one of them will form a low temperature eutectic, which means intermetallics will be created. A better solution would be to use single element materials which do not form intermetallic compounds until high above the melting temperature. Look up the binary phase diagram for readily available metals (Cu, Ni, Fe). At least with only two compounds, you will not have to worry about ternary eutectic points.

Alternatively, glass or other stable metal oxides (ceramics) will be unlikely to react with Al. Many kinds of glass (especially quartz) are inert to most metals and do not soften until well above 1000C. Also to consider -- the surface of Ti has a very stable oxide which may also provide a diffusion barrier against aluminum, even if the two have a low temperature intermetallic.

Well, alumina ceramic rod stock is readily available. Would that react with aluminum?

It would be nice, because ceramic apparently comes in pre-made tubing with inner diameters as small as .78mm; which could possibly mean no machining. I could just get a larger ID ceramic tube and a smaller ID ceramic tube, and put the smaller inside the larger halfway in, to make a reducer to push the aluminum through.

... I've got some 'pressured air resistors', made from brass with inset ruby-disks with different diameter bores from 0.1mm to 0.3mm ... AFAIK with prices around 1-3 Euros each.

The ruby-disks could be a good choice, but the original assembly won't withstand the pressure while printing ...

---

Viktor
--------
Aufruf zum Projekt "Müll-freie Meere" - [reprap.org] -- Deutsche Facebook-Gruppe - [www.facebook.com]

Call for the project "garbage-free seas" - [reprap.org]

Extruding is probably difficult. However printing objects in aluminium (or any other metal) should be childs play to do if you throw away the "extruding" mentality:

I would buy an off the shelf TIG welder (prices start at £150) and rig a Zirconium-tungsten electrode and wire guide to be moved by steppers in 3D. This would eliminate the problems with reacting/damaging the chamber or measuring temperatures, and would be much easier than trying to melt the aluminium inside a hot-end. TIG would be nice as you could use the tungsten electrode and just feed the aluminium in from an external drive (think Bowden extruder). Compared with MIG there would also be no worry with accidentally welding the aluminium wire to the build and shorting it out, and the weld quality is generally better with TIG. I would also buy off the shelf MIG welding wire (comonly available down to 0.5mm diameter), which would be simple to draw even thinner if there was a market/need for higher resolution.
Using the same system of drive unit as a MIG, you would have a metal drive wheel with an idler to press the wire against it, as there is no real back pressure to overcome (it's not being forced through a tiny hole) a simple NEMA17 ought to do the job fine providing your roller moved freely. The wire then runs inside a flexible cable to deliver it to the head (and we think bowden extruders are something unique)...
You'd probably want some sort of electrically insulating arc guide (eg ceramic with a hole drilled in it) to make sure the arc melted a small piece at the tip of the feed wire, rather than jumping around up the wire, or onto the printed object.
You'd also need to check how tolerant the TIG welder was of being switched on and off fairly rapidly (I think not likely a problem). The ability to instantly stop the heat would also get rid of the stupid whiskers of melted material which are an issue with plastic. Bridges may also be revolutionised as it may be possible to weld the feed wire to one end of the bridge and then feed it to the other end before welding it on there, thus no hot/melting wire in the middle.

Once de-bugged, a TIG system would also be easily be generalised for use with virtually any other metal (steel, titanium, chromium, aluminum, copper, brass, bronze and gold are examples), and wouldn't be fouled up by molten metal requiring head changes for each metal. It would be fairly easy to feed multiple metals into the melt zone to develop all sorts of novel items/properties.

TIG and MIG use inert gasses to avoid the problems of aluminium oxodizing/reacting with the other parts, so thats instantly eliminated, you also would have very little to worry about regarding heat transfer to the feed system which would be sitting well away from the head (probably on the bench beside the TIG welder. The head would only need to be a Zirconium-tungsten electrode and a guide for the incoming metal wire, and could be feasibly be lighter than most current plastic melting systems.

You would be sensible to enclose and do an initial flush of the build chamber during a build or the volume of gas would be prohibitively expensive. I'd use some kind of plastic bag inflated inside the chamber initially to prevent mixing of air and inert gas which would be deflated by the pressure of inert gas then a low level vacuum pump to suck it back out. With a bit of thoughtful design, the system could easily waste virtually no gas. However I think this would almost certainly also be the case if you used a traditional style hot-end else it would regularly oxidise and foul at the build point or extruder nozzle. Walls would also probably have to be metal to help radiate the heat build-up as I suspect without cooling the entire printed object would be a puddle of metal on the bed by the end of an extended print.

Given the weight of the printed object and possible issued with necessity of cooling it, the design would have a static build plate, with probably a small hole in the centre to insert a piece of wire as the 2ary electrode. This would be chopped off the build at the end of the process.

This system would solve:
1. A hot end material that will not melt itself (easily replaceable electrode, which is largely undamaged)
2. A temperature indicator that will not burn out (probably not needed)
3. Obtaining over 660C (easy)
3b. Staying over 660C consistently without damaging the heat source (some erosion of tungsten electrode, but this can be easily replaced as required).
4. A build plate that will not melt or catch fire (not sure on this one, probably a ceramic, but a heated build plate may allow for faster builds, or maybe a cooled build plate would decrease sagging of the object)
5. Cooling (basic fan, but I'm not sure on the heat capacity of the gas and transfer to the build chamber walls)
5b. Localizing the heat so that it does not creep up the aluminum and damage the machine (likely irrelevant)
6. Limiting the size of the hot end so that it fits in the machine (easy)
7. Oozing the molten metal out at just the right temperature and speed so that it bonds but does not run (fairly simple to establish with a few experiments).
7b. I don't know if oxidation would be a problem, but if it's anything like welding... yeah (solved).

PS I know infrared thermometers aren't generally very accurate, but if they are taking readings from a non-varying target (so emissivity isn't changing), they should be able to be calibrated accurately enough to provide useful readings without damage if you still wanted to create a do it the 'usual' way (ie a melt chamber in a hot-end)...

PPS I'm not a trained welder, so I would also find someone to check my theory against before I built it...

PPPS I think this artist is doing something like what I envisage, ie building up the metal to form the sculpture rather than welding bits together:
Stainless Steel Sculptures

Edited 3 time(s). Last edit at 08/12/2013 05:48PM by dgm3333.

Yeah I'll bet that would work. Something along those lines anyway. Although I don't know if I would start from TIG. If we're gonna have to mod something, it seems like there would be less modification required if the feed system and gas was already done. I've played with a MIG quite a bit, and as long as you get the speed/voltage ratio right, you can usually eliminate shorting.

The tough part, I think, would be getting it to be nice and clean. Maybe just using really small wire and low voltage with a (like you mentioned) arc guide would do the trick. But in my experience, welding is messy unless you're really skilled. So building a machine that was really skilled may be hard to do. Not impossible, but it would take quite a bit of calibrating.

I still stand by TIG.
Welding is messy because it's hard for a beginner to keep the spark gap constant. A robot (think printer) should manage this fine. TIG is inately cleaner than MIG.
TIG does already have a gas system, but I think it would be unnecessary extra weight on the carriage to have to drag the gas line around, thus my reason for separating it from the 1ary electrode electrical line and redirecting the gas to fill the entire box.
A basic Wades (or similar) extruder would work fine for feeding the line - there's nothing different to feeding a wire of metal vs a wire of plastic. only change would be (perhaps) needing to reinforce the "bowden cable" so friction with the wire didn't cut through it. (in my opinion the only useful part of a MIG would be this line).
Also if you wish to use MIG, you will have to develop some system of monitoring for the wire for when it attachs/welds itself to the print and shorts (an electrical monitor would be fairly easy though). Also some way of disconnecting it (often bending the wire breaks it, but perhaps not always close enough to the surface to be ideal, as you don't want it to be a point of shorting next time, so you may need a small grinder or clipper to cut it either of which would be unnecessary extra weight. Admittedly a robot would have less likelyhood of this happening due to constant height control, but I don't think it would be 0%, as random splatters can affect the layer height.
With a TIG system, you'll still need a system for disconnecting the wire, but it wouldn't have to be monitored. I'm not sure, but you should be able to use retraction on the wire and activate the spark to cut it, thus a much simpler system to institute.

The only other slight problems I thought of overnight would be reconfiguring the slicer software to
1. start at a given point (so the feed attaches to the 2nd electrode), and
2. making sure every part (including layer 0) was connected to the origin at the lowest layer, ie supports would still be required, but in this case for electrical conduction as well as the usual gravitational support.
Neither of these are particularly tricky, and I'm sure there would be a slic3r or Skeinforge guru who could do that, and it would be fairly easy to manually add supports until that was developed (and perhaps trick the slicer by having the initial layer a single point which was placed at the 2ary electrode origin.
David

Edited 3 time(s). Last edit at 08/14/2013 02:53AM by dgm3333.

Hi Folks! I'm another new convert to 3D! Just getting my bits & pieces together to build my first Prusa i3.
This question of printing metal interests me, because of my pre-existing interest in metal casting. I've also done some mig & tig welding.

I've been trying to visualise your tig solution, David, but the main problem I see is that the arc turns an area of the metal you've already deposited into a molten pool into which molten drops of the feed-wire drip and fuse. Whilst adding more metal may not be too much of a problem, the big challenge is how to keep the size of that molten pool so tiny that the already deposited layers do not flow, distort, and cause the print to lose definition.

I agree with several of your ideas, though. To prevent oxidation, certainly an argon shield is a must, and you are right - ways could be found to minimise wastage (cost). One possible solution we have not yet discussed here, is the idea of using thin alu wire - or, frankly, just about any other wire - and fucusing a laser beam on the point where it touches the print. We'll already know the distance from nozzle to print, so finding the focus point where the wire will touch should not be a problem. The beam will then melt both the wire and a tiny pool in the printed metal into which the droplets can fuse. The difference is that with a focused laser beam the size of this pool can be very finely controlled to a degree that I just cannot visualise with tig or mig.

Wissing, I can see why you started with aluminium when thinking of melting metal. Wire is readily available, and the melting point is not as high as the MP of many other metals. Also, the light weight of the printed product may be of advantage in some engineering applications. However, if suitable wire can be sourced, why not begin the experiments using tin? A lovely, shiny, corrosion-resisting metal whose melting point is only 232 degrees C! It's a bit soft, but then so is pure aluminium. And when you want to take a break from printing it, you can melt some on the kitchen stove and cast a few toy soldiers!
The tin metal can be bought - at a price - but in case anyone's interested, you can also recover it (little at a time) from the tin cans in your kitchen before you put those out for recycling! From a ceramic dish which can be heated, make an electrolysis bath ( a cheap slow-cooker from a car boot sale is perfect). An electrolyte of caustic soda solution, heated to about 85C will by itself strip the tin off your cans without attacking the steel! However, to make the process continuous, get/make a variable dc power supply and electroplate the tin out of solution again onto a metal (e.g. copper) strip. It can be melted off the strip later on your stove. My electrolysis setup consists of a 2amp variac I picked up on Ebay (mains in - zero to 270V ac out). This is feeding a simple car battery charger (not a fancy electronic one), so I can finely adjust my electroplating voltage between zero and 13 volts.

Forgive me, if I've strayed off topic a bit, but I get a wholesome feeling from reclaiming scrap materials to turn into works of art. If we can find a way do that by printing metal, that would be really neat!

Edited 1 time(s). Last edit at 08/15/2013 03:32AM by 3Diddy.

So tin.

I'll put the welding idea on hold for now, cause I think that deviates a bit too much from the RepRap - foraging into the area of "not tried and true" can take a lot of time that I don't have.

But I didn't realize tin melted so low. I guess I sorta knew, because it's used in solder.

Back to the original questions then: could I use a stainless steel nozzle? I've got one being built right now by a friend at work... 0.5mm hole with a 3mm feed tube. I would almost just try printing solder... any sort of noxious cancer/reproductive-disorder-causing problems therein I should be aware of?

Furthermore, would tin react with a glass build plate?

3Diddy: Agreed: My suggestion of an "arc guide" was aimed to reduce the melted area. I think lasers are used for the metal powder 3D printers, but quite expensive even with using high temp ovens to heat up the base material (so you only need to add a little extra energy with the laser to melt the metal. When someone comes up with a cheap powerful laser it will revolutionise the world! (and probably blind lots of school boys). Oh and if cost isn't a problem then plasma gives more precise control than TIG, but I would agree, the only ones I've seen were still more messy than a laser :-p

Another sl problem with my idea is welders would be generally rated for short 'burst' usage, so cooling would have to be beefed up considerably for continual operation.

Wissing:
Solder has been done (Youtube (not by me).
Depending on the solder, it melts between 200-400 deg (approx). You would need to get flux free "solid" solder.
Retail is about £10/500g, but if you were printing volumes you could probably get it quite a bit cheaper. Price varies by the components which are also important in controlling properties (eg hardness) (if I remember cheaper solders have higher melting point and are harder, but I could be wrong). Amongst other things, surface tension is also an issue with getting fine resolution with melted metals, so this also varies (you'd need to experiment to get an optimal alloy for printing)
There are lots of metals which melt at low(ish) temperatures: mercury, tin, lead, silver, gold are common, but there are others. The problem is they tend to be quite soft unless in an alloy (thus Woods metal). AFAIK you should be able to put them through a standard metal (ie brass, stainless, aluminium) hot-end, the problem would be isolating from the surrounding plastic components (as heat transfer along the filament would be greater). But there are lots of oxides formed, which might clog the nozzle.
I want something you can use at kitchen temperatures (ie up to 400-500deg) without melting or going soft, so aluminium is more of an interest...

I guess the solder is pretty goopy. Although that setup looks like it could be improved. Big old nozzle there. Noisy too.

Another thing is this: I'm wondering if the brass nozzle would jam up with solder (brass is an alloy containing copper - a solder wick). Or does something else in brass neutralize that effect?

dgm3333: I agree that aluminum would be more of an interest. For one thing, there's something about it that doesn't scream "health defects". For another, it's a metal that has a lot more uses. We could be printing coffee makers, toasters, electric motors... anything that requires a heat sink. Also, a lot more of the parts in a 3d printer could be self-replicated.