Electric Arc Furnace Closes The Loop

When we think of an Electric Arc Furnace (EAF), the image that comes to mind is one of a huge machine devouring megawatts of electricity while turning recycled metal into liquid. [Gregory Hildstrom] did some work to shrink one of those machines down to a practical home version. [Greg] is building on work done by [Grant Thompson], aka “The King of Random” and AvE. Industrial EAFs are computer controlled devices, carefully lowering a consumable carbon electrode into the steel melt. This machine brings those features to the home gamer.

[Greg] started by TIG welding up an aluminum frame. There isn’t a whole lot of force on the Z-axis of the arc furnace, so he used a stepper and lead screw arrangement similar to those used in 3D printers. An Adafruit stepper motor shield sits on an Arduino Uno to control the beast. The Arduino reads the voltage across the arc and adjusts the electrode height accordingly.

The arc behind this arc furnace comes from a 240 volt welder. That’s where [Greg] ran into some trouble. Welders are rated by their duty cycle. Duty cycle is the percentage of time they can continuously weld during a ten minute period. A 30% duty cycle welder can only weld for three minutes before needing seven minutes of cooling time. An electric arc furnace requires a 100% duty cycle welder, as melting a few pounds of steel takes time. [Greg] went through a few different welder models before he found one which could handle the stress.

In the end [Greg] was able to melt and boil a few pounds of steel before the main 240 V breaker on his house overheated and popped. The arc furnace might be asking a bit much of household grade electrical equipment.


21 thoughts on “Electric Arc Furnace Closes The Loop

  1. EAF might be a bit much to ask, I’ve had pretty good luck melting and heat treating with a resistive heating coil, ~1000W gets a few lbs of salt up to 1600F (molten salt heat treating for O1) in about 20 minutes. I’m using resistive wire from E-cigs, and controlling it with an arduino. IIRC the max temp of my bricks is ~2200-2300F, which is about an order of magnitude less than molten steel, however I think my graphite crucibles will take that temp without issue, but I also think at the 3000F+ temp for molten steel you need to use a pyrometer instead of a K-type thermocouple. Either way, very cool project. The real trick for something like this is to control the amount of heat lost, so even if you have a duty cycle or power problem you can work within that constraint by making sure your material doesn’t cool down too quickly. Shield gas can also help quite a bit, 95/5 nitrogen hydrogen is common in industrial settings (the hydrogen burns off any oxygen present).

    Someday soon the home EAF foundry will be more within reach to the home experimenter.

  2. I wonder if you can make an arc induction hybrid where the induction coil gets the metal to a given point and is then lifted off before the arc comes into play? With induction you can keep the load very well thermally insulated too, no hole at the top for all that heat to convect out of.

  3. Please, Please, Please do not do anything with molten metals over a concrete floor. One accidental spill and the water moisture in the concrete will instantly vaporise, with no where to go the concrete will spall (explode), throwing molten metal all over place.

    There is a reason why foundrys have sand floors, sand is porous so that any vaporised water will escape in a safer manner. Put down a bed of sand or soil if you are going to do this sort of thing.

    1. Yeah…when I was in high school two of my buddies where melting pennies on a drive way. I heard the scream, turned around and saw one of them trying to claw the molten metal off his cheek and the other guy just staring while his hair/scalp where smoking. Ah the glory days…

    2. For anyone attempting to build or “do” something of this nature. Sand with a heavy dished plate of iron on top inside a box with moderately high sides. All of your melting and pouring should occur within this box, preferably over the iron. You shouldn’t be within arm’s reach of any of this, as molten metal can pop when it contacts any moisture, even just uncovering it to pour it can cause it. One piece of molten iron or even Al will burn right through skin and muscle like it wasn’t even there. Molten iron will cause bones to explode on contact. If you survive something like that, you will wish you hadn’t. No flammable materials within spitting distance is the best work environment for casting metals.

      Use a chain and gantry style pulley system and a load balancer to control the tilt of the crucible. As always, only work with this sort of thing in protective gear, and with plenty of ventilation.

      What he’s doing int he video is the result of years of practice, and not something a newb should attempt.

      1. Years of idiocy would be more accurate. What he’s doing should NEVER be done.

        If you want to do foundry work, read a proper book on the subject. Don’t use something like this as a guide. Reinventing serious accidents is not healthy.

        A perfectly proper setup would be a bed of dry sand in a wooden box large enough to contain the entire operation. The depth of sand needed depends upon the amount of metal being poured. The “the plate of iron” in this comment is daft.

        Here’s a valid source of information:


        1. > What he’s doing should NEVER be done.

          Are you implying that plastic mesh Nike joggers aren’t adequate PPE for hot metal?!

          (At least in the video he seems to be wearing leather boots)

    3. I think what’s happening is that the extreme heat reverses the process of the concrete setting, driving out CO2 and moisture, un-slaking the lime part… but yah, sudden superheated steam with nowhere to expand, same effect.

    4. Thank you, I work at a blast furnace and haven’t been able to figure out why the floors explode when they spill material on the concrete sections. I never thought about the water absorbed into the floor.

  4. Odd, didn’t see him heat the mold before pouring. I’ve seen even slightly damp molds explode before on contact with molten metal. Overall, this looks like a great clinic on how to burn down your house. At least move the setup into the driveway. Glad this gent didn’t slip.

  5. I dunno if he actually managed to boil any iron. But for the duty cycle, maybe a second welder? Swap them in and out of circuit? With something like the mother-of-all-relays, made from a servo motor and some decent-sized contacts.

    The MOAR could alternately switch the output electrode between the two welders. There’d be no worry about contacts welding, since another switch could disconnect the welder from the mains first, so no power.

    And then, i dunno, that gets you up to 60%. You could use a third welder. Or maybe rig up a decent cooling system, investigate the welder, and see if there’s any components you might up-rate, or attach bigger heatsinks to. 30% is probably a good enough duty cycle for their intended use, so why would the manufacturer bother trying to extend it? But I bet there’s a few things you could do, if you needed to.

    This is all assuming he wanted it as a permanent thing, rather than just something fun to do, then scrap afterwards.

      1. Yeah I spose, but can even a good welder deal with being shorted for long enough to reduce steel to a puddle?

        Talking of welders, if I had a friend whose MIG welder needed fixing, because the wire dispenser no longer advances automatically, what might be a thing to fix that? I’m gonna have a look and a tinker anyway, but are there any caveats or obvious things to look at first?

  6. Reads like needs a better workshop to perform the foundry work with it’s own generator and high power source. I did a quick peak at what the mid scale mills require and… wwwooo-aaahhh!!!

    “A mid-sized modern steelmaking furnace would have a transformer rated about 60,000,000 volt-amperes (60 MVA), with a secondary voltage between 400 and 900 volts and a secondary current in excess of 44,000 amperes. ”

    I never did look at the energy use for the cupola’s storage tank part (induction I think… might have been arc) or the smelter cupola (gas) operation when I worked in the Bosch foundry.

    Man, that’s huge amounts of energy that I’m trying to grasp operations sources. Smaller scale doesn’t seem so challenging and a 10KW generator might help with duty cycle regarding issues from the power source. The http://www.electrode.ca/ link is interesting also. I’m not going to scope creep… I’m not going to scope creep… I’m not going to scope creep…

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.