A Rostock Welding 3D Printer?

Tired of printing in boring old plastic? Why not try metal? Researchers at Michigan Tech have come up with an open source reprap style design of a 3D printer that can print metal for only $1200.

The paper was published in IEEE Access a few weeks ago that it outlines the design and testing of this printer, which is basically an upside down Rostock with a MIG welder used as the extruder. As you can imagine, the quality and resolution of the parts isn’t that amazing (hang around after the break to see an example), but this is an exciting step forward for 3D printing. Equipped with this and a mill and the possibilities are quite endless!

Did we mention how cheap welding wire is? A cost that could add up is the shielding gas though, but as a user on Reddit points out, an upgrade for this machine could be an enclosed build chamber which could then just be flooded with the gas. Alternatively, would flux-core welding wire work?


As you can see, the printed part is rather rough — but it is solid carbon steel (ER70S-6). Combine this with a quick machining pass on a CNC and you’ll have a prototyped metal sprocket that barely wasted any material! We can’t find the video right now, but there was a commercial machine that is basically the same concept, using a modified 6-axis robotic welder. Its main purpose was for the production of large mostly hollow parts, like a wing casing on a plane. Instead of machining a giant chunk of metal, it could be printed and then cleaned up with a single machining pass — considerably less material processed.

So who wants to try building one?

[via Reddit]

65 thoughts on “A Rostock Welding 3D Printer?

      1. This is metal parts which are easily detectable with metaldetectors. And production of this type of weapon with this kinds of tools already regulated by existing laws. All screaming was about deadly scaryfull plastic guns ))) This commentator just missed topic :)

        1. What laws? You can make guns if you want too. I’ve looked into this. The only thing I’ve found so far is you cannot manufacture firearms with the intent to sell them unless you possess a firearms dealer license. Although it is perfectly legal to sell a firearm you’ve made after you’ve made it regardless of your licensing status. Just so long as no one can prove you made the firearm with the express purpose of selling it you’re legal.

          1. I missed this story when it first got posted but I’ll drop this here for anyone that comes after me. My understanding of US gun law (and this might only apply to IL) is that while you CAN fabricate a gun yourself legally, you are required to report it’s existence and register it with the authorities as soon as you manufacture the specific component that is, legally, considered to be “the gun”. For example, I believe that part is the “lower receiver” in the AR-15.

          2. There are people working to get the unconstitutional gun control laws either repealed or nullified. All gun control laws are unconstitutional, it’s just that the ruling party want to shit all over the Constitution and the opposition party haven’t got a ball between them.

    1. Fluxcore also requires cleaning off the slag after every single pass. This could be accomplished using a robotic arm with a needle gun mounted to it and a vacuum attachment for getting the ditrious out of the weld zone and redirect the exhaust from the vacuum back into the work chamber.

    1. Unlike plastic extruders. the MIG gun can push out bare wire, so bridging could be fairly easy if the axis motors have enough force to bend the wire ;-)
      (or if it was a 5-axis machine to tilt the gun)
      Adding a a shearing tool to cut the wire is all that is needed :-D

    2. I bridge all the time when welding by hand. After blowing a hole in sheet metal it is easy enough to slowly reduce the size of the hole by using a filler around the edge. A similar process could be used to bridge.

  1. Well, looks like that by means of such combined technology (additive + subtractive) it is quite possible to create machine at the price within 10000-15000 dollars (if to use delta-robotic arms and ball-screw actuators) which will be able quickly to make not really difficult metal details of any form. There one manipulator which will be able to change a welding toolhead for an toolhead with the milling spindle is required. Or it is better to make two manipulators – one with a welding toolhead, and another with milling. Cool. Fully automated production closer and closer.

    1. BTW, probably it will be possible to combine welding and milling in one process, for example when you welding fewlayers, than milling surfaces, which will be not available after complete welding, then welding one more layer, milling it borders and so on. In this case the parts could be more difficult, than if to use separate welding and milling.

        1. Yes, but this was done with metal smelted parts, which is goes in one piece, and need melting forms and so on. With combination of welding and milling in one process it is possible to mill surfaces which is unavailable if you first welding and then – milling. Not 100% of course, but more than 50%. An you also can split your design in smaller parts. As it is happens now.

      1. And, probably, for this task it will be better to use 6DOF manipulator to be able to create parts without support, like here – 3ders(dot)org(slash)articles(slash)20130517-mataerial-3d-printer-allows-you-to-create-objects-in-the-air.html

    1. With Microplasma you can weld wires with less than a hairs diameter. It’s kind of a TIG welding, so you can feed cold metal to the arc.
      Oh, and i reported you. Because stupid buttons are stupid. Sorry.

  2. I’d like to see it cut in half and take a look at what the actual weld quality is like. Just because it looks flawless on the outside doesn’t make it flawless on the inside :D

    Very cool idea though. Flux core wouldn’t work well though because of the slag and also the much higher heat requirements that would result in a flatter and lower weld passes.

    Mark what do you mean by thinnest? The biggest problem with stacking weld passes vertically is controlling the amount of heat placed into the object. Also you can only go so cold with your weld before you start introducing weld faults. Different wires stack differently, depending on your alloy requirements I would use the thinnest wire I could that required the least amount of amperage to perform a quality weld.

  3. There isa way to do this without a shielding gas by going with a modified sub arc machine. The process uses a sand like media to act as the flux negating the need for a gas. For it to work with this setup the platform would have to adjust its height for each layer and the media would have to be replenished as needed. There was an artist who built a solar powered printer recently that used a Fresnel lense to melt sand and an adjustable platform to do the same thing.

    1. That artist is Marcus Kaiser, and he did create glass-like objects, not metal ones. basically he melted/sintered sand under a very large magnifier. No welding and the fluxes were purely incidental (no added soda or calcium oxide added, that would make much better glass).
      Also keep in mind that his was mostly an art installation, he was not interested in the best result, the process and recording of it were just as or more important.

  4. No one mentions part strength either. I assume the point of this would be making stronger parts. But being as that sprocket is made entirely of weld I can’t imagine it would handle a very heavy load. Also in the quest for a thinner weld (allowing for better resolution) I can see untrained welders losing all the penetration between layers. That’s going to leave you with an even weaker part. Plus, like they said, welding wire IS cheap, but steel isn’t terribly expensive. You can print a weak part on an overly complicated machine and have it fine tuned on a mill, or you can buy a cheap block of steel and have it machined on a mill. Either way a mill is going to have to get in on the action, if you want a quality part.

    Still neat to see, I just don’t see it being feasible. As a species we’ve been working metal for centuries (without computer aid). By now we’re pretty good at it.

  5. “Did we mention how cheap welding wire is?” By volume? Because metal is somewhat heavier than plastic is.

    As far as what shield gas costs goes that varies by supplier. But the ones I’ve dealt with and the gasses I buy it seems to me that gas itself is not the expense, filling the bottle is. I’ve had little bottles, and big bottles, and the big bottles never cost that much more to fill than the little ones did. Buying compressed gas isn’t like buying gasoline. I wish it was!

    1. Inert gas cost has some relation to the gender but it also has to do with supply and demand and in the instance of helium its subsidies and a dwindling supply and a lack of production. If a shielding gas is used then the work area could be enclosed fully and a low pressure release valve used to allow the system to vent excess pressure without wasting an excessive amount of gas that you would find with an open work area and even an open top unit.

      1. Helium is not used with mig welding, it is only used with tig and even then it is limited to situations where better penetration is needed.

        Many shielding gas combinations are used for MIG, the simplest being straight CO2. Most people use a 75/25 mix of argon and co2 to minimize splatter. Also most MIG machines now made are designed with a V/A output curve for 75/25.

  6. I just did a back of the envelope calculation on operating cost. In one hour, the printer should be able to deposit ~5kg (10lb) of metal for a total cost of around €20 ($25, or 17mBTC :D ). For scale, that’s about a 10x10x10 cm (3″) cube. The cost depends largely on feed-rate, since the cost of gas is ~$15 of that $25. Here’s my calculations:
    It looks like a typical #2 argon tank (1133 l, or 40 ft^3) is enough for about 2 continuous hours of welding (regardless of feed-rate), and costs €22 ($30) to refill.
    Metal wire is cheap, at ~$1/pound (~€1.6/kg).
    I don’t see the welder settings listed on the wiki, but ~6kW is about right for a MIG. At 10 pence per kWh, electric works out to $1/hr.
    Their welder uses 0.8mm wire at anywhere between 15 and 420 in/min. If we assume ~200 in/min (5m/min), the volume deposited is 3.14*(0.08cm)^2*508cm = 10 cm^3 per minute. Density of iron is 8g/cm^3, so deposition rate ~5 kg/hr.

      1. That’s an interesting idea. Air is composed mostly of N2 (78%), O2 (21%), Ar (1%), CO2 (0.04%). Molecular weights are N2: 28, O2: 32, Ar: 40, and CO2: 44 AMU. The air in a room doesn’t separate out because random convection currents keep it well mixed.
        Convection currents are very slow, though. You’d need to add more Ar occasionally, or keep very slight positive pressure in the box, but an enclosure could substantially lower the AR consumption, especially for long prints.
        Another possibility would tea candles. Instead of displacing oxygen, why not just remove it? Enclose the printer with a couple candles burning somewhere in the enclosure, then start the print when they no longer had enough oxygen to burn. We might have to experiment to find something that would stay lit long enough to sufficiently low O2 levels.

        1. Ignore my candle idea above. Removing O2 will help, but water vapor and nitrogen also pose problems at welding temperatures. (Props to “AKA the A” for pointing this out below.)
          Also (as I currently understand it) pure argon is only used for welding aluminium, stainless, etc. For welding mild steel, 75% Ar, 25% CO2 is used. The CO2 helps clean up the weld. (Although Macona wasn’t 100% positive that the CO2 is actually consumed)
          If the CO2 is consumed slowly, the gas pumped in to maintain positive pressure would probably be more than enough to keep an acceptable CO2 level. The intake rate could be upped to maintain this, or a gas with more CO2 could be selected.
          If the CO2 is consumed fast enough for these measures to be impractical, the concentration could be monitored and adjusted from a separate CO2 tank. This has a number of advantages. High CO2 levels introduce carbon to the weld, making it harder but more brittle. This is good for printing wear-resistant, high-carbon tools (files, saws, drill-bits, router bits, etc) but bad for things that should give before they break. For the best of both worlds, the perimeter of a part could be printed in high carbon steel to case harden it, while using softer but stronger steel for the center.

  7. Why not go the other way and stick it in a vacuum chamber? All you would need is a roughing pump and then add some argon, pump it down again and add a bit more argon to make it arc instead of make a corona. If you did that I’d think the parts would end up looking much cleaner and you could use basically any metal wire. You might need do the necessary anealing and heat treating to get it to proper strength, though.

    1. Argon would dissipate in some kind of low pressure. Argon in welding is also producing the Arc beam, transporting the energie, melting your electrode and the material your trying to weld on. If the Argon leaving the gun began to flutter, also would your beam do.

        1. Actually i do welding construction, programming of industrial welding robots and am qualified as International Welding Specialist (IIW 1170), among some other stuff.
          But no, don’t do much handiwork.

    2. Straight argon will not work for mig. You need some CO2 for cleaning and I believe it is consumed or broken down during welding so you would need some way to monitor CO2 levels and inject CO2 back into the chamber.

    1. I think it’s common usage for professional 3D printers now, at least for plastic injection molded forms production. But if you have such metal printer you can get (theoretically for now) the same result with less equipment, and it is more flexible in usage, but lower in speed (relatively).

    2. You can. Pick your favorite paste extruder (there are a bunch of designs) and use it to print the mod from plaster of paris or a ceramic. Alternatively, you can print the part you want, and then do lost-PLA instead of lost-wax casting. Basically, you buy plaster of paris (it’s cheap!) at a hardware store, mix it with water, and then pour it around your piece. Once it hardens, you heat the whole thing up and melt out the PLA.
      The problem with casting is bubbles. DIY casts tend not to look as good as professional ones, although it’s an artform that anyone can master with practice.

  8. It seems like wire-feed TIG would work better for this process. Wire feed tig is sort of like a bowden tube attached to a TIG torch. This way the arc is not generated by the wire itself vaporizing, but between the tungsten and the work surface. It would provide more leeway to decouple the power settings from the size of the wire. And the arc itself and puddle zone can be a lot smaller.
    With MIG, there is a delicate balance between the diameter of the wire and the voltage setting. Wire-feed TIG would give you the ability to fiddle around with more variables:
    pulsing / duty cycle (where the arc is modulated at some number of Hz), wire feed speed, wire diameter, voltage, amperage, and polarity. It should be possible to create “taller” weld beads that much more closely resemble the profiles of plastic extrusion in a typical 3d printer, perhaps even using a pinpoint infrared detector aimed at the puddle zone to dynamically adjust the feed rate and/or power settings so that it has the desired size.
    It looks like this MIG sprocket they made had way too much power! Perhaps also a heated bed would help (maybe 600-900 degrees F) so stress warping would be less of a problem on longer parts. Although, the built-up stresses for this kind of process could create some neat opportunities to optimize part strength.

    1. No. At the temperature the metal melts, nitrogen is a problem (there is a reason they use fairly expensive argon instead of cheap nitrogen, which is a waste product of air distillation). Moisture is even worse, since is dissociates into oxygen and hydrogen (both make the metal brittle and weak)…

    2. probably not as they are not using nitrogen for welding anyway. But on a side note: an oxygen generator/concentrator does have the downside that your nitrogen comes out at the low pressure. But if you use another material for your column (activated carbon instead of zeolite) you can use the same process to concentrate nitrogen while getting rid of the oxygen. I’m looking at this to use as gas-assist in a laser cutter: it makes sure that you blow out any flames above the cut. Perhaps you might also be able to cut metals by flowing pure(-ish) oxygen into the focal point. But that is a whole other story/machine.

  9. Another thing to watch out for is duty cycle. Usually only the very large machines are rated for a 100% output at rated current. Most little cheap machines in the $600 range are rated ay about 20% duty cycle at 90A output. If you build this you need to size your welding power source appropriately, if not you will either trip your thermal overloads, fry your diodes, or worse case, fry your transformer.

Leave a Reply

Please be kind and respectful to help make the comments section excellent. (Comment Policy)

This site uses Akismet to reduce spam. Learn how your comment data is processed.