Printing In Metal with a MIG Welder

Whenever the question of metal 3D printers comes up, someone always chimes in that a MIG welder connected to a normal 3D printer would work great. A bit of research would tell this person that’s already been done, but some confirmation and replication is nice. A few students at TU Delft University strapped a welder to a normal, off-the-shelf 3D printer and made a few simple shapes.

This project builds on the work of [Joshua Pearce] et al. at Michigan Tech where an MIG welder and delta bot was used to lay down rather complex shapes on a metal plate substrate. The team at TU Delft used a cartesian bot – a Prusa i3 – for their replication because of the sheer mass of moving a metal build plate, firebricks, and welder around.

In the first few prints on their machine, the team was able to lay down enough metal to build a vertical wall. It’s not much, and to turn this into a finished part would require some machining, but these are only the beginning steps of what could become a legitimate way of creating metal parts. Video below.

71 thoughts on “Printing In Metal with a MIG Welder

  1. Not sure I’d want them machining anything for me with their ham handed hammer method of part extraction.

    Brings up what is the best surface for the welder to begin printing. metal? A higher melting point metal that would allow easier breakaway? I can see something like this creating blanks of odd shaped equipment that are then sent to a CNC for final cleanup and part inclusion.

      1. Actually, it didn’t look “promising”, it looked damned successful already. “3D printing” includes not only “hobbyist 3D printing on homemade and/or cheap (say, below $5000) equipment”, which has only become successful lately, but also “commercial 3D printing starting at 5-digit prices, with patents to keep cheap competition from existing”, which has been doing quite well for a long time.

        1. There’s a lot of youtube comment style negativity creeping into the discussions around here. This is hackaday. it’s about trying interesting stuff and showing the results, not evaluating the commercial viability of every random thought that passes by.

          Slapping a welder onto a 3D printer and seeing what happens is exactly what this community is about, and long may it stay that way.

        1. I have no 3d printing experience. I have limited CNC machine experience. I have significant experience with metal work and if they had scored the bottom of the print with a grinder prior to the hammer hits, I’d understand, but I see no evidence of that in the video.

          1. On their blog they note fiddling around with the welder on the ‘wrong setting’: it would create a bead of molten metal laying on top of the base substrate. One would assume that would be their starting setting. So instead of scoring the base, they start with a crappy weld. Sounds good enough to me!

        1. Of course actual testing is bit more involved, almost anything is going to break if you beat at it long and hard enough enough. Practically every major oil producer had a crew in our small town back in the day. To work for a major a welder has to be certified. I happen to be in a machine shop when the welders where being certified. Their samples where put in fixture that test how much force it actually took to make the weld fail.

      1. Actually, for several of my printers (With metal beds, or build surfaces) that is my preferred method, using a rubber mallet. No distortion from prying it up.

        Works well for both plastic and resin prints.

      1. Indeed! I assume you meant to suggest that milling leaves a much better finish and accuracy than 3D welding, and any 3D welder that does not include milling capability is rather silly.

      2. Okay, that was totally cool to see. I like the ability to switch tips to add/remove/finish the metal to/from/on the piece. My question is what is the scale that this can work at? Also, what was the piece that they was making? I swear I need to learn German… Or is that Dutch?

    1. some sort of thick gauge steel screen would work nicely. You would have to cut it out with a grinder, making the screen a waste product, but it could be produced cheaply, and steel recycling is common.

    2. Because I worked in the oil field I was in machine shops waiting for my work to be complete I was in machines shops quite a bit, hammers are used a lot. Unless there is evidence the hammer damage the part, it erroneous to call the use of the hammer “ham Handed”.

    1. Yeah this is probably the way to go. Much cheaper (In both hardware required and material), more precise and stronger. Probably wouldn’t need CNC machining either in most cases. I could see one benefit for mig printing being speed, but it’ll have to become a lot more precise to negate the need for the CNC step…

      Still a cool project though, it has potential.

    2. With metal deposition don’t think as normal methods…What can you do with this machine?….you can fix a complex mould by adding material and later a CNC process to extend life of the mould. if you have two MIG welder with different wires you can generated a lattice form with different composition from the outside to the inner part.

    3. I used to work at a company that rebuilt turbine blades by using a robot to do exactly this to the edge of the blade. I imagine that if its done right, its perfectly fine.

      Honestly these guys should talk to some people in industry because better versions of this type of robot already exist and some of the problems they are facing may have already been overcome.

      1. I’ll second that. I was doing that type of laser welding on gas/steam turbine shafts and blades 7 years ago. For prototyping it may make sense to combine the machines, but if you’re running a repair shop I’d want two different machines, one to weld, one to machine so that I could have multiple parts being repaired at once.

    1. A laser with the required power output is somewhat expensive and fairly NOT eye-safe…
      Also, fine metal powders tend to be quite flammable, so you need an inert atmosphere, otherwise things become a little too spectacular.

      1. i dont think a mig welder is very eye safe either. no matter how you do it, printing in metal is going to need to be done under a hood of sorts if you want something that you can use in a shop without blinding everyone else working there.

        1. A busy shop will have several welders going while machinists are using their machines, and other employees are going about what it is they do at the business. I’m not seeing why this would be more of a hazard. Also I’m not seeing how this is going to create as fine dust.

    1. I personally am impressed at how well it worked. I would have thought that each new layer deposition would melt the layer below it and you’d just have a mess. Some of this might still be going on, I’d suggest trying to cool the piece as it’s being printed (be VERY careful if you use water!). Also what current setting are you using? I’d expect that a thin wire and low current would produce the best resolution.

          1. Sorry, I was unclear. We used pressurised air to cool the part down between switching layers, so no welding takes place. But you are indeed right that something like argon would probably be better (and more expensive), but further research need to be done on that part.

  2. Use CNC to cut out a bottom piece the shape you need, attach it to the print bed surface, then weld onto it. Easier to remove, and no need to “print” out the bottom solid layers.

      1. I’ve seen the layered cnc idea many times before. One company did it with wax paper and a laser cutter but went out of business after their end users just bought wax paper instead of their expensive material. It was big enough to do engine blocks for casting. I’ve also seen another doing it with aluminium and steel foil and ultrasonic welding.

    1. Cant wire weld steel to aluminum. The difference in melting points alone makes this physically impossible AFAIK. kind of like building a solder statue on a wax substrate. it would punch right through immediately.

  3. OK i’ll be the negative guy here. horrible idea horrible execution, just horrible, plus this is not autonomous. people have to stop thinking “hey just print it, great idea” this is not the best idea in fact is is often not even a good idea. if you want a plate or a part, print out a positive in plastic or whatever, finish sanding it, then cast it in the material you want. or machine it then make a cast to finish off that way. maybe someday we will have metal 3d printing that has some type of tolerances (no way this has any at all, maybe +- 1″) but until that day comes this type of idea/project IMHO actually detracts from real innovation. and yes i did just say that, this is nothing more than a generic 3d printer with a mig welder duct taped to it.

  4. few things, an old trick when plugging a hole in a steel panel with a MIG is to hold a block of copper up to one side of the hole and then weld. the weld then fills the hole, but naturally doesnt stick to the copper. perhaps try this as a base?

    if sucessive beads are stacking well as-per normal 3d printing, then theres a fair chance the joints are not fused correctly, a proper weld profile looks like a smooth speed bump on the surface. using too low a current allows beads to stack nicely with a resulting weak weld.

    Last thing (im a self taught mig/tig/gas welder so appologies if this is drivel) but the metal transfer from the wire feed to the weld pool is by ion transfer/magic, if you look real close at a weld while laying a bead through a good quality welding mask the tip of the wire never quite touches the pool. so the accuracy of where metal is deposited varies with the stability of the arc, and they can wander a fair bit.

    Mig welding is used all over industry to build up surfaces, its a good process for firing a bulk of metal down in a reasonably orderly fashion, as a previous poster has mentioned i’d be very surprised if there isnt some literature available to academics on the subject, may help the project progress nicely.

  5. CNC plasma cutters are commonplace. CNC GMAW welders are not. Being able to do low cost production quality CNC welds is much more useful than doing expensive, poor quality 3D metal printing.

    The use of weldments in place of castings revolutionized manufacturing processes in the latter part of the 20th century.
    Things routinely made by casting in the 50’s and 60’s are now made by cutting, pressing and welding rolled plate. The quality is more consistent and the cost lower.

    There’s a lot more to 3D printing metal than just putting down metal. Surface finish is just the beginning. There are many technical reasons for the high cost of 3D metal printers.

    1. actually cnc gmaw is more common than most believe it to be. centroid a large manufacturer of retrofitting equipment makes just such a thing. very popular especially among pipe shops.

  6. Nice project, I hope we will see more of it. I would love to see this combined with a milling head of some sort. The surface finish you get right now requires special glasses (rose-tinted) to call it beautiful, hopefully a milling/cooling pass will improve that.
    There is quite some overlap in requirements with a CNC milling machine (all working with metal precautions plus gas and UV shielding), so it is not a big plunge to combine them.

  7. There’s tons of research being done into this in industry right now, it’s called Wire Arc Additive Manufacturing. There are some really cool niche applications that don’t work efficiently with regular CNC or casting (large hollow structures, varying metallurgical properties as you go along…) but you still need to do some finishing operations later. In answer to how you remove it from the build area, normally they weld to a base plate and then cut it off afterwards. There’s some good information here:

  8. The (not insurmountable) problem that I see is that metal isn’t nearly as nice a material for printing as thermoplastics like ABS. Liquid metals at welding temperatures tend to be relatively low viscosity compared to plastic, and they have a much shorter cooling/hardening time. This makes printing complex shapes very difficult, and makes the print quality of simple shapes relatively bad. This is in addition to the issue of thermal stresses in welds. Thermal stresses in standard welding (welding two pieces of metal together) can be large enough to frequently cause concern. When your entire piece is made out of weld metal, I can’t imagine what sort of stresses you will get. This will cause warpage or even breakage of the part. The weld stresses can be annealed away post-completion of the print, but the weld stresses during printing may cause the part to break or warp significantly before the part is finished printing.

    As i mentioned, these obstacles are not insurmountable. Higher viscosity alloys can be used for welding, or techniques can be developed for dealing with the undesirable properties of the liquid metal. These compromises may result in printing parts significantly thicker than the desired end result that require considerable post-print machining to reach the desired dimensions. I don’t think additive metal printing will ever be quite as ubiquitous as additive plastic printing, but I certainly see some applications for complex parts that would normally be cast and then machined down to final spec.

  9. Nice article on your techniques used to optimize 3d printing and welding. It was interesting to read your take on combining both welding and 3d printing for the product. Thank you for the info and the film posted as well. Take care and have a great day! Durandana

  10. Is this 3D Printer on the market right now ? It is awesome to see a 3D printer combining with a welder to make a real products although I think it is hard to observe the process without a lens filter.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s