steel/copper etc printing with arc welding method

[22:25] hmmzzzz.... i just had an idea how to do steel/copper etc 3d printing
[22:26] coulndt it be done with arc welding method
[22:26] would be easy mechanically aswell
[22:27] has anyone tried this yet ?
[22:27] any reasons why it wouldnt work
[22:27] etc ?
[22:29] all you need is a more heat resistant printing base
[22:29] ceramics?
[22:29] metal wire
[22:30] would be nice to try it in oxygen free environment aswell
[22:30] welding psu
[22:30] and some simple extruder mechanism to feed the wire
[22:31] should work like a charm

penny for your thoughts about the idea

Hi r2kordma,

... there are two actual methods for direct-printing objects from metall.

Both of them base on selective-laser-sintering.

In the powder-bed-method you cover a complete surface with a special powder, where you have a plastic-hull and a metall-core.
Then you scan with a laser (weak, only some 10 to 100 Watts) the area which should become solid, so the plastic-coated particles melts together.
When repeating this for every slice, you build your 3D-area. Then you can pull out the solidified object from the powder and bake it in an oven to a nearly solid metall-object (with some shrinkage too!)
In some variants the small cavities between the matall-particles will be filled in a third step with solder or another fluid metall, so the object would be perfectly solid ...

In another direct-laser-method the laserspot (of a very strong laser, many kilowatts) melts a point of the topmost surface and some metall-powder is blown into the hot spot, so it melts, adds and forms a metall-blob. With drawing this 'blobs' over the surface you can add pure metall until you finish your object.
Here you receive perfect and solid metall bodies in only one step ...

Viktor

10-100 watts a weak laser? hardly 100W laser is already smth that cuts steel, anyway i don't really care if that solution is used in commercial 3d printers. the thing is that it can be used for 3d printing. anyway I'm planning to look into that idea, gotta ask from some professional welder if the idea is feasible.
if someone has extensive welding experience now is the time to shine with your knowledge.

Edited 1 time(s). Last edit at 04/12/2008 04:15PM by r2kordmaa.

I'm no professional by any means, but I have arc welded a few times in a friend's shop (building a go-cart), and understand how the process works.

This is an interesting idea and definitely worth looking into and testing. However, in my experience with this technology, if you don't keep moving the weld rapidly, the heat in a particular area builds up VERY quickly and within a few seconds the whole region can melt down and start dripping molten metal (undesirable). You'd have to find a good way of getting heat out of the area fast to keep the lower layers stable while building on top of them.

yeah well thats just controlling the robot movements, software issue. im rather afraid there might be some harder difficulties to face like extreme inaccuracy due to molten steel freely running around for a while before cooling down or smth like that one must remember that in printing material isnt always deposited on level surface but edges aswell where it might drip off

Hi r2kordmaa,

... it's a balance between accuracy and time, what makes a 100Watt-laser for this application really 'weak' - besides, a 100 Watt-laser is good for cuting plastic ... for cutting steelsheets of some millimeters thickness commonly lasers in the higher kilowatt-range are used ...

When you want a good accuracy at a fair speed (minutes, not days per centimeter heights with 0,2mm accuracy), then you have to scan your area with very high speed, so every point of the object is only some microseconds in the laserspot - so you have to calculate with energy per time, what makes the absorbed input-energy per spot something in the milliwatt-range.

Try to sinter the steelpowder with a laserpointer - no chance ...

So you have to overcome the minimum energy-level for melting and with a common scanning speed of some meters per second and a 'fine' spot of 0,2mm diameter you should expect the needed laser-power up to some kilowatts!

With lasers in the 10 to 100Watt-range you can only sinter plastic-powder in a usable time/accuracy-balance, so the 'cheap' SLS-application uses plastic-coated metall-dust and performs a second curing/sintering-step in the oven for the completed object ...

Viktor

well 1kw+ lasers are smth you wont build yourself and if you buy one you'll prolly end up poor. around 100W co2 laser can be built at home and you dont really need such scanning speed. anyway thats hardly this threads discussion, rather the idea is if you could use arc welding to print, note that pauses in printing to cool down the detail are probably necessary.

I have dicussed this at length with a welder friend of mine and the general consensus we reached was that it would be possible MIG would be the most likely best option. Warping would be a major issue (as in day to day welding) getting a good resolution might also pose a problem. The mechanics of moving the weld head is very well understood just look at almost any production factory the welding robots are very impressive.

A good welder can easily fill a hole in, surely this is 3D additive fabrication.

---

Ian
[www.bitsfrombytes.com]

fabulous putting darwin in a box and filling it with mig isnt hard eighter. it will take some time till i have my own darwin and i probably wont get my hands on welding psu in no time soon so im afraid unless someone takes the lead and tries it out it will just remain a theory. btw i think the troube is: filling the hole is one thing, building a object layer by layer is smth else im especially concerned about putting metal on the edges as it might drip down from there but in some time next week i'll discuss it with a professional welder.

r2k

Hi,

Here's a study about the rapid manufacturing of complex parts of stainless steel and titanium (pdf): [files.erikdebruijn.nl]

The're (indeed) all based on sintering-like techniques, as Vic pointed out.

Regards,

Erik de Bruijn
[blog.erikdebruijn.nl]

The metal won't drip welds only do that if you don't know what you are doing but they do form interesting shapes that would take some calculating and you would get massive thermal stresses, if you take a thick piece of steel and weld one side it will bend and this might make it impossible for all but the crudest items

---

Ian
[www.bitsfrombytes.com]

Plus you could burn a whole... I think...maybe..idk...just a suggestion

---

Jay

found an article about printing in aluminum in an argon atmosphere using similar techniques to inkjet printing. Tiny droplets sprayed across a small gap to a surface. Layer takes place form there.

However, overhangs of any sort are just not doable. What kind of support structure would you use? You can't even do slight overhangs like with Darwin as this stuff is fully liquid at application. SLS, as pointed out above, is how metal is done. I'm pretty sure it is done that way for exactly these reasons.

The other way of printing metal just doesn't appear doable to me.

On the topic of welding robots...very good industrial robots exist to make defined passes. It isn't a dynamic process by any means. It's a controlled--trial and error produced--manufacturing process that only seems (to me) to be worth while if you are going to make thousands of similar welds.

Just my opinion. I'll be looking into SLS in the future.

Demented

This is a shameless regurgitation of information previously posted in the forum by others. But I thought it might be interesting based on the discussion here.

MIG Weld Deposition 3D Printing:
[repositorium.sdum.uminho.pt]

Another related link, more along the lines of piezo injet technology for aluminum:
(I think this is what Demented referred to above)
[gram.eng.uci.edu]

That aluminum article is indeed the one I was referring too.

Your MIG welding article is fantastic though! Still, no overhangs > 45 but they definitely made it work within those constraints.

Demented

you can make overhangs, read the entire article, you need to turn the object during the process tho
anywho looking at this article it seems like welding robot type reprap would be much more useful:

it can print objects larger than itself ergo an itself in full eliminating some or all parts of manual construction process
LOTS of work area, enough room for holding different toolheads, drill bits etc for automated changing
you can turn the toolhead in different positions, it doesnt have to be vertical
because of abundance or space you could have multiple robot hands working simultaneously - for example one could hold and turn the object while other deposits material

compared to a cartesian bot a welding robot is simply uber
so how about investing that build possibility for next gen reprap(or the generation after that)

[edit]
i think the robotic arm design needs a whole separate thread

Edited 1 time(s). Last edit at 04/17/2008 01:03PM by r2kordmaa.

Hi r2kordmaa,

when you want high accuracy and more rigidity with 'weaker' material (as plastic) and motors, then other designs as planar-knematics or scara-robots shouold be more interesting, than the single-arm-design - go through the links int this article: [forums.reprap.org] or search the forum with the keywords "parallel-kinematics", "scara" or "string-bot" ...

Viktor

scara bot is a good idea but doesnt have the merits of:
360 dergee work radius
ability to access details from side

if the arc welding 3d printing could be implemented stronger matwerials than plastic could be used. with steel frames there wouldnt be a need to worry about arm strenght. solution to move the joints remains. anywho theres a thread under mechanics section for the robot arm design

Hi r2kordmaa,

... for 360

i kinda see the flexibility of the arm as an uberbonus, these things arent used in car industry without a good reason, you can even access within structures etc. the difficulties to build it and program it are simply something to overcome. its next gen stuff one way or another. about the hydraulic structures im really suspicious about the positioning accuracy, although its really interesting idea and useful in many applications i doubt you can do high precision automation tasks with it. robot arm positioning on the other hand is just a software job

I disagree about the accuracy capability of a hydraulic system. Think of putting some linear encoders on there. Then, positioning is--as you said for the robot arm--a software job.

Demented

... i'm more for the bionic approach - hydraulic actuators and pressure-sensors all over the machanic (read through the posts, ther it's described)

Viktor

It occurs to me that if you had a good clay extruder head and a powdered metal extruder head, you could make cast metal objects that you just print and fire. You could make anything from cast iron frying pans to jewelery with a primitive backyard forge that way.

Or instead of going through all the trouble to extrude clay, you could just extrude wax and do lost wax casting with plaster of paris.

I like the fact that the first printed metal object Mr Ribeiro shows in his study is a 3kg steel pint glass - obviously, great minds think alike.

hi my name is chang vang
Are u guy have any postion opnnig
Because right now i have no job so i want to find a job right now


Chang Vang