DIY Spot Welder Can Join Anything Together, Even Copper


Hackaday reader [David] was looking for a cheap and easy way to spot weld copper tabs together. As he notes in his writeup, the properties of copper which are most enticing, such as high thermal capacity, make welding it all that more difficult. His home-brew method of spot welding is admittedly quick and dirty, but it does get the job done quite well.

He started off with an array of four 2.5V @ 2600 Farad ultra capacitors, which provide the high current required to do copper spot welding properly. They are wired in series and connected to his electrodes using heavy gauge wire. The graphite-tipped electrodes were an interesting DIY job themselves, cleverly constructed using copper tubing and a graphite block. The most simple/dangerous/clever part of the whole rig is his trigger mechanism, which consists of a pair of copper blocks that he bangs together manually to complete the circuit.

[David] is well aware that the setup is just a touch rough, but according to him it makes great welds, and it’s only a proof of concept at this point. He has a hefty list of improvements to make for the final version, including a different switching method among a few other safety precautions.

35 thoughts on “DIY Spot Welder Can Join Anything Together, Even Copper

    1. That’s not a bad idea. You could add a little solid state control with a 555 timer and an FET/IGBT/SCR. If you set up the 555 in monostable mode, you could precisely control the pulse width and initiate the weld with the press of a button.

  1. Banging two copper blocks together as a trigger mechanism – mental!

    I’m midway through making my own capacitance discharge spot welder for building battery packs, using a 1 farad and 1.5 farad capacitor in parallel, a Picaxe microcontroller to control the timing of the dual pulses, a bank of MOSFETs as the ‘switch’. thick copper cables and copper rods as the probes.

    Currently waiting for some IRFP2907 MOSFETs (209A continuous, 840A pulsed) to arrive because I tried a set of 10 IRFZ46N’s (53A continous, 180A pulsed) in parallel and it alsmost worked but during a test one died and two blew into pieces! and that was just using the 1 farad capacitor.

    1. make sure to pull the gates up fast enough (slew rate) and high enough (voltage).
      i have made the mistake of driving the gates too slow and making the mosfets linger in the 1-10 ohm range when they should be in the milliohm range.
      this instantly kills the mosfets and sends molten package plastic towards your eyes.

  2. why dangerous? 4 x 2.5 volts is only 10 volts. 10 volts can only be felt through the tongue or something else that sensitive…

    maybe hundreds of amps seems a lot, but what’s the risk?

    a 1000 volt source with only a few milliamps is a lot more dangerous than this…

    1. It’s not the volts/amps to be wary of, it’s if the wrong bits of metal accidentally touch & spark (put it in a decent case to prevent that) and when the thin/small metal being welded explodes from too-high voltage/ampage being forced through it.

      I’ve worn eye protection from the very start of my CD welder build, I’m not having a piece of metal in my eye and ruining my day.

    2. The ampre potential between the terminals is insane. Same goes for a 12v lead acid battery. Sure, you can touch them. But good luck unwelding a wrench that touches the two terminals!!

      1. While only a few tens of milliwatts will kill you, you need the voltage to drive that current. 10v (in most cases) is not enough potential to drive a lethal current through your skin. Human skin has a resistance of 100s of kOmhs when dry. Skin soaked in salt water will still have a resistance of a few hundred Omhs.

        The back of capacitors used in this hack may have the capacity to deliver 1000s of amps, but it is only charged to 10v. The most current that will be produced is I = V/R. If ‘R’ is a human body, ‘I’ won’t be very big.

    3. I worked in telecom with a guy who died from a 10v header at 2 amps for just a few seconds. Forehead-to-hand. We’d thought at first it was the 48v dc bus, but power monitoring audits and the red mark on his head showed otherwise.

      I treat everything as dangerous, even my EL wire drivers. Supercaps are a wily bunch, getting better as time goes on. They act just like batteries with near zero ESR, hence the high current output.

  3. [P]oints have to be subtracted for the sheer kludge of banging to copper blocks together.

    Fixed it for you.

    Don’t hack together something quickly just to make it work and then ALSO hack together a shitty mechanism to activate it. At least put time into properly designing and implementing the parts that can kill you or cause you serious harm. So tired of seeing this.

  4. I’ve used a spot welder many times before when I worked in sheet metal. I also have two very thin traces of stainless steel in my hand when a part I was welding apparently had water or oil between the two pieces and blew a small hole in my work, sending the stainless flash into my hand where I can still see it today.

    Copper electrodes are a must. I’ve never seen a spot welder without them. Also, there needs to be some way to adjust the voltage as the thickness of the metal being welded I doubt would be consistent. Some way to clamp the electrodes in place and a foot pedal switch to fee up your hands to hold the pieces being welded is also necessary. Depending on the voltage and current being used, an NO SCR triggered by the foot switch would be ideal, Depending on the current, though, a two pole Mercury relay would be a safe bet. Just be sure to make a snubber if you’re using any microprocessor circuitry to control the voltage/current.

    Banging together two copper blocks is not only bush league but dangerous.

    1. I believe he’s using graphite electrodes because he’s welding copper. Wouldn’t make much sense to me to use copper electrodes to weld that. That said, copper works great on steel.

  5. Instead of making 2 carbon electrodes he could have just went to a welding store and bought a pack of carbon arc gauging electrodes which are the exact same same thing he built, but they’re made to a higher standard.

  6. Why would you wire capacitors in series. that will only lower capacitance. {1/C+/C(n+1)}^-1 this is dump. make a real arc wielder. use a car battery charger or something and some graphite.

    1. It lowers the capacitance as seen from the ends of the series, but he’s charging the capacitors in parallel.

      Plus, the energy of the discharge doesn’t change, because at the same time the energy stored in the bank increases at the square of the voltage.

    2. Four 2.5V @ 2,600 Farad:
      Since the equation for ENERGY storage is 1/2CV^2…you get more bang out of a potential voltage increase. A good analogy is momentum where capacitance is mass and voltage is velocity (1/2mv^2).

      Series => 10V @ 2,600F => 130,000J
      Parallel => 2.5V @ 10,400F => 32,500J

      With that said, I have always connected them in parallel because in this application, lower voltage is advantageous.

      1. @poe
        They store the same amount of energy if they are hooked up in series or parallel, you forgot to decrease the capacitance in your formula for when they are connected in series. They are connected in series so that it can deliver more current faster; like 600F of capacitance is probably overkill for this project.

  7. I have worked on several spot-stud welders that are capacitive discharge. These are the welders that weld the little threaded studs or standoff that you see in metal cases and electrical enclosures.

    They used two large electrolytic caps, probably about 400v, they were unmarked so I don’t know the capacitance. An adjustable voltage regulated doubler circuit controlled the charge voltage to the caps setting the weld power. I think maximum was around 250v. A large hockey puck style SCR was used to discharge the caps into the work. Very simple and reliable.

  8. Hi everyone! hope you like (or at least are entertained by) my stupid idea. I got some answers for some recurring/related questions you guys have asked.

    I went with ultra caps because they were actually cheaper (about $10-15 ea. surplus), smaller, and I thought more durable. I understand they were pulled from electric bus systems, so I figured they were designed for running at high power. Plus they’re rated for millions of charge cycles, unlike batteries.

    The banging blocks technique is obviously a bit, um, unorthodox, but it’s extremely robust. As Haku mentions, MOSFETs and the like can fail catastrophically; I have a hard time imagining my copper brick doing the same. Ultimately, a proper MOSFET switch with a 555 pulse timing circuit is probably the way to go, but requires a hefty bit of planning and thought.

    The motivation for this project was simple- a group I work with needed a way of connecting copper tabs. When it seemed that all commercial systems cost $10k+, I scrounged $50 and built this thing in a couple hours as a proof of concept, nothing more.

    In the end, I think I’ve stumbled across a quite useful method for welding copper- there are obvious flaws in the execution, but the central premise works rather well.

    1. The banging together is actually not bad. Nothing says that we have to use microcontrollers or transistors to control a load. The issue I see is a way to do it reliably and safely. Heavily insulated holders for both ends? :D

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.