Hundred Dollar Capacitive Discharge Welder

[Robert] needed to weld metal tabs on a few batteries. In a proper manufacturing situation, this is usually done with outrageously expensive welders. Not wanting to spend thousands of dollars to attach bits of metal together, [Robert] built his own capacitive discharge welder for only $100.

Instead of the giant transformers you’d find in a spot welder, a capacitive discharge welder uses a huge bank of capacitors – greater than 1 Farad – to weld pieces of metal together. Huge caps like these are commonly used for ridiculous car stereo setups, so with the addition of a car battery charger purchased from Walmart, [Robert] had most of a welder on his workbench.

To control the mass of power coming from his huge cap, [Robert] used a 13o amp Silicon controlled rectifier to improve the control of his welder. With the battery charger, cap, and SCR, [Robert] only needed a few bits of heavy gauge wire to tie the entire build together.

[Robert]’s build welds metal tabs on battery terminals beautifully, but the possibilities don’t end there. This welder could easily be repurposed to build the skeleton of outrageously intricate dead bug circuits, or maybe even keeping that thing you made with your Erector set in one piece permanently.

40 thoughts on “Hundred Dollar Capacitive Discharge Welder

    1. I don’t believe it. They are way, way too small. Like four orders of magnitude too small.

      Some breakthrough in capacitance happened when nobody was looking so the inventor starts selling them on ebay for the same price as a 1F capacitor? I think not.

    2. I wouldn’t even want to think about what would happen if you picked up one of these the wrong way.

      When I was in Jr High, we used to leave charged electrolytics on our benches at the end of class to discourage the kids in the basic electricity class from messing with our projects.

    3. I have 650 Farad Boostcap in front of me now, which I’ve been experimenting with for a couple of years, including making a small spark welder for replacement battery packs: I had some *very* old laptops that I was going to use for portable controllers and data collection, and my cordless tool batteries that were hitting End of Life.

      They won’t work in a pure capacitive welder. They just don’t discharge fast enough, and the low voltage is a significant hindrance. I did manage to get it to work in a primarily inductive spot welder, but the welds really didn’t hold up that well over time in my tools.

      The spec sheets I read offered real promise (ESR ~50 milliohms = a theoretical 54 A spike), but whether there were factors I overlooked, or if they degraded in their prior usage (most of these are from gigantic capacitive bank from utilities like Ontario Hydro, which are retired on a fixed schedule, and sold dead cheap to US bulk buyers vs. paying significant hazardous waste fees.

      You can do a lot with these, but I’d say their performance is more like a supercap than a true ultracap, at least in the condition usually found on eBay (at the time, I couldn’t even get the distributors to return my calls for unit 1 pricing on a NEW BoostCap)

      A tip: a timed dual (or more) zap worked much better than a single pulse. The initial spark seemed to clean the arc surface, so the more powerful main spark could weld better. I set pulse duration/timing with a crude analog circuit, because the slow initial discharge made the power of the inductive kickback spark very nonlinear. There was no point in trying to set values except by trial and error. If you use a better performing cap, you may consider using a microcontroller like the pro battery welders do.

      1. I forgot to mention — the low voltage was also a big issues. I’d previously tried 1-10 Farad 5.5V supercaps, and the 650F “ultracap” took just as much coaxing to draw a spark. Aside from discharge rate (a huge factor that can’t be overlooked) the power stored in a capacitor is CV²/2, so a 5.5V supercap stores four times as much energy per Farad as a 2.7V ultracap

        (But I suppose that’s more relevant to other ultracap projects and experiments I’ve done, such as harvesting of ambient environmental energy, back before Linear Technology, Texas Instruments, etc. came out with chips that made that easy.)

      2. i bought some maxwell bcap3000 capacitors used and i love this b****es. yes, they are used. but even with my thing wire connections my meters have shown 1000 amp peaks and over 1000 amp pulses (pegged my ammeter. so far i have 1 15 volt bank of 433 farads that’ll start a diesel engine in the cold on the first try, im slowly buying these bcap3000 models so i can do so extreme pulse power experiments. they also come in super handy as fast charge power supplies. nothing like charging them at 150 amps for a few seconds and having a nice work-light in the field for a couple hours. me end goal is to have a 1000 volt bcap3000 array. at that level i can do some serious research into welding, plasma cutting, extremely pure power source, and maybe even harvesting energy from solar/wind.
        when i was a child about 15 years ago, i read about ultracapacitors in a book saying how they were still theoretical devices and only in labs. the fact that im finally at an age where i can afford them just makes so happy that im finally living one of my early life dreams. my child self would be so proud if he could see where i am right now <3

    4. thank you for laying this out so clearly. im trying to rebuild an x26c taser battery. they are 50 bucks a pop, which is ridiculous. The taser safety often comes off (its still safe enough) which activates the light which drains the battery. So I do through a lot of these. Honestly this 100 project will still save me money.

  1. Then again, there’s the approach my dad used when I wanted a better connection for the batteries in one of my electronics projects. He soldered the leads to the battery. Only, he didn’t have a soldering gun, so he used the kitchen stove. 80

  2. Good project i have needed it overtime for battery cells.

    I ended up doing it with the soldering iron with satisfactory results.

    The trick is to carefully sand/grind the surface of the plate to a rough look and quickly spit on the cell a hot ball of lead and join either a wire or a strip treated the same way.

    However capacitive discharge is way better as you can throw the cells out of the window and still have them attached.

    A cheaper and better solution is to join them mechanically.

  3. Definitely had good luck with truebassb’s approach for rebuilding old calculator battery packs. More so with wire than with metal strips: with the strips, there is weakness in the joint on both sides, rather than just on the battery, though there are spatial advantages to the strips if you are in a tight squeeze.

    If you can get a small ‘pool’ of solder to form on the battery, rather than a blob stuck on, you’re in good shape- just apply a little flux and tack on a tinned wire. The mechanical strength is not so great though, so it’s best to keep the wires as short as possible without straining them and get the whole battery complex reinforced, i.e. by wrapping in electrical tape and carefully stuffing it back in the pack it came from. Since the wires add a mm or so the fit will usually be tight, ideally applying pressure in the right direction to mechanically support the connection.

    Of course, this is not a great idea for if you want something super-reliable or for more than one-offs. Robert’s system looks great for small- to medium-batch production, plus it’s super cool; or again for new designs just use a battery box as suggested above.

    BTW the ‘low discharge NiMH’ batteries (eneloops or whatever) are great stuff. Power density is not Li Ion, but they will charge with simple circuits, are safe, and hold a charge for a long time. Good for dropping into older equipment or for projects where maximal power density/size is not a requirement.

  4. I wish hackaday would remove this article. This has been up on the net for a long time. It’s been proven to be FALSE. The setup he used can’t do much of anything and there is a lot of misleading info.

    1. I have personally done this with capacitors from a satellite aiming controller (8 foot commercial dish). We used it to attach the old tabs and wires from a laptop bios battery to new CR2032 batteries. We had nowhere near 1F but at 12v it still worked.

    2. You mean, for example, how he talks about inductance and resistance but neither a resistor nor inductor(coil) can be seen in his setup?

      Yeah, I wouldn’t trust this thing to do anything but overload the power supply.

      1. Ummm, at low currents you can usually ignore inductance and resistance in straight wires.

        But when you pulse a few tens or hundreds of amps though a wire, you definitely have to take inductance and resistance into account.

    3. Indeed, Bradley, I do wish you would show me where it has been proven false.

      I’ll have to let the battery packs that I have built with it over the years know that they don’t work and never did.. If you can prove that it doesn’t work, then by golly, I’ll take the entire thing off the website.

      Yes. I’m the one who developed it and did the writeup, although Phil Pemberton did his first.

  5. ws = (F/2) * V^2 ??

    Umm… this equation is essentially useless. If you’re going to use Watt’s Law, use it right. Watt’s/Ohm’s Laws only deal with 4 quantities: power, resistance( or impedance for AC circuits), current and voltage.

    Farads is a measure of capacitance, which must be converted into reactance( the AC form of resistance) before those laws can be used. And let me not get into the fact that the frequency of the source is part of that conversion. No frequency, no conversion.

  6. Just a note, you can buy spotwelders on ebay for 200-300$ that are usable for these things from china. Of course building them is more fun, but you can have them far cheaper then “thousands of dollars”.

    1. At the time I built this project… about 2007 … Those cheap chinese welders did not exist, and the cheapest commercial model available had gone up to $3000 for the single pulse model about a month after I wrote up the original pages for the website. and the dual pulse model was about $4200

  7. If you are going to involve a car battery you might as well just use that to do the welding, no capacitors required. It is done all the time in remote areas to spot weld metal. You need to be careful though as the instant current a car battery can deliver is enough to melt tabs like those on small c, d, etc cell batteries into a molten mass.

  8. I think I’d have to add some circuitry inside the foot switch to be able to shape the pulse going to the SCR. Of course, that would complicate the project (and I do like the simplicity), but I’d want to play with pule amplitude and timing to see what worked best.

  9. I found a place that sells battery holders at a really good price.

    I bought some 4 AA packs with leads for 0.17 each.
    Shipping is reasonable as you can see below… I got 50 of them.

    4 x AA Battery Holder x50 $8.50
    Standard 7 – 16 Days Delivery) $5.99
    Grand Total $14.49

    Total delivered price for each holder was a bit less than 30 cents.

  10. Electrodes sticking is usually caused by either too little pressure on the electrodes or too little hold time.

    The excessive erosion of the electrodes is caused by misalignment and too little pressure on the electrodes.

    A decent fixture to hold the electrodes would probably fix both of his problems. Might as well go with a proper parallel electrode setup at the same time.

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.