Featured on Hack a Day:

The Capacitor discharge microspot welder has been featured on HaD, see it here . Thank you guys!

The circuit diagram and complete source code is available at the bottom of this article. This project is based on Jiri Pittner's welder for the Atmega16, and so the source code (Below) is released under GPL.
welder sparks 7 welder sparks 6 welder sparks 5 welder sparks 3
This project took a lot of time to complete. This article is structured on several parts, the bottom containing the final results. The microcontroller source code for the Dual Pulse Welder/Cutter is included at the bottom part of this article.

Progress summary:

2012-02-08, First tests
2012-02-09, Choosing the LCD
2012-02-22, 1,000,000uF Capacitor update
2012-03-01, Project box
2012-03-03, Power supply and some innovative electrodes
2012-03-18, Tungsten electrodes
2012-07-15, Source code and circuit diagram

Part 1: February 8th, 2012

I've just completed the first tests of a new challenging project, a pulsed discharge micro spot welder and cutter tool. It stores energy into a huge capacitor bank, and discharges it via two electrodes in the given target, regularly metal foils/sheets. The logic and the precise timing (of the order of micro-seconds) is controlled by an AtMEGA16-16PU microcontroller running at 16MHz. It can be used for spot welding and for plasma cutting.

The power switch is handled by 5 power mosfets, IRFP2907, set in parallel and driven by a darlington array IC with further transistor amplification. A single mosfet is rated for 75V 209Amps and an extremely low Rds. Do the math for the 5 working together!

Here is one of the construction tests, where I cut a beer can:

The molten material sticks to the discharge electrodes, making the operation difficult. I plan to try manufacturing them our of Graphite or Tungsten. Let's see which one works best!

This project is work in progress. My capacitors bank is aprox. 60000uFarad, but I need to get close to 1 Farad to have enough energy for correct operation. I will probably use a Car Audio capacitor.

Schematics, Microcontroller software, PCB layout to be added. Video showing spot welding will follow soon.

Some construction pictures:

The tool currently operates in two modes
1) Micro spot soldering
A predefined (uC+LCD+controls) high current pulse is delivered to the two electrodes, when the trigger button is pressed. Metal wires or sheets pressed between the electrodes will be soldered.

2) Cutting
A burst of short pulses are sent to the electrodes to generate plasma and cut thin metal foils.

The pulse duration, repetition rates are configurable. I got best results with pairs of pulses. Duration in the range of tens - a few hundreds of microseconds. For welding, the first pulse melts the material, and the second consolidates the soldering. For cutting the first pulse does the damage, and the second pulse helps to clean the electrode off the molten micro spot.

More operating modes can be added, so please give me ideas!

- I need to find better electrodes, that do not stick to the material , I'm considering tungsten or graphite. Suggestions?
- bigger capacitor bank: I need more power, and will probably go for a 1Farad Car Audio capacitor . Those Sprague I'm currently using are not enough for ticker sheets.
- finish and publish the software
- publish schematics and PCB layout.

Part 2: February 9th, 2012

I need to decide what type of LCD to use: the 2 lines HD44780 LCD, or the bigger one, with 4 lines. The code I wrote works with both. The larger one would have the advantage of allowing more text to be displayed. The smaller one is on the other hand more compact, better suited for my small case.
2012-02-09 14.02.32 2012-02-09 17.06.06

Part 3: February 22nd, 2012

I purchased a suitable capacitor. Finally! It's a Pyramid Royal Blue, CAP160DBL 1.6 Farads . The first thing I did was to dismantle the useless electronic part and the display. By doing so, the new size fits perfectly to my project box. For those complaining that this type of capacitor, when used for car audio systems, doesn't switch off and depletes the car battery, simply unscrew the connectors to take off the display part. You won't be needing that anyway:
pyramid 16farad 1 pyramid 16farad 5 pyramid 16farad 6
The price for this one was aprox. 55USD from a local shop and this makes it the most expensive component of this entire project, as expected.

Part 4: March 1st, 2012

The project box, in the final shape is ready. I had to cut some more wood, but that's an easy task if using the proper tools (I got a nice pendular saw as a gift, last Christmas).
For a quick test I tried using a halogen 12V electronic transformer. That was a waste of time, as those things are unstable, and not working properly (they are barely oscillating):
capacitor dual pulse welder 1 capacitor dual pulse welder 2 capacitor dual pulse welder 3
The box exposes the heatsink, a standard PC power connector, and a fuse holder. The big capacitor is mounted on the top, as it looks nice and is too big to fit inside anyway.

Part 5: March 3rd, 2012

I resisted the temptation to throw that halogen electronic transformer to the trash. It was such a waste of time! Instead I found a charger, from an old laptop. It is a 90W, 19.5V SMPS supply, with the perfect size for my box. Initial tests shown it is capable of charging the capacitor (with a short initial blackout), and even hold the intense Arc cutting mode power consumption. As this was another Chinese cheapo, some optional components such as filter, capacitors, etc where missing from the PCB, and I added those first. Next I did a heatsink upgrade, to make sure it will run continuously, for as long as I need to use the tool:
smps 2 smps 3 smps 4
So far this proved to be an excellent choice! My device can now be powered from a 110V-220V outlet. Alternatively, you can choose one of the SMPS power supplies presented here.

For the control panel, I needed a few potentiometers and some push buttons. I used a few NOS Russian 47K pots, I got from Gintaras Ebay show, KWTUBES. You can get some high quality components there at fair prices.
kwtubes 1 kwtubes 3 kwtubes 4
Some pins are used to interface with the floating cables coming from the PCB modules. By doing so, I can easily dismantle and re-mantle everything together.

Here's how the setup looks like now:
setup 1 setup 2 setup 3

For the working electrodes I ended up using copper. But not just copper, also a new innovative design: the thick wires are connected to the electrodes using a screw, a screw nut and a washer. The electrode itself is a 10cm metal hollow tube. Inside, copper wire can be inserted, and while working, it can be pulled outside when needed. Exactly like a rotring pencil : the cooper electrodes is eaten while in use, but conveniently can be extracted from within the electrode, or replaced.
electrodes 2 electrodes 3 electrodes 4
Here's a video showing the working principle:

The positive electrode also holds the discharge command push button that can be pressed with one of the fingers.

Finally, here's the completed project:
welder_ready 1 welder_ready 2 welder_ready 4

A few sparks, as recorded in some interesting photos, timing was crucial to get them, but I managed to do that eventually:

The power components are interconnected using thick copper wires. The wires to the electrodes are Copper Litz wire with section are of 16mm^2.
The 1.6F Capacitor that I used, has an ESR of 0.0016 Ohm.

Part 6: March 18th, 2012

I recently received a few Tungsten Carbide rods. This excellent material has a high melting point (2,870 °C), it is extremely hard (8.5–9.0 Mohs scale) with low electrical resistivity (~2×10−7 Ohm·m) making it an good electrode material for the Capacitor Welder. I hope to see it solves the molten sticky metal issue I had with copper electrodes.
Results show excellent performance:

Here is another video:

After a few minutes of use, some signs of damage are observed on the electrode, but they are minor at least when compared to copper:
tungsten_carbide-2 tungsten_carbide-1

Components used:

There were many modules involved in this project, but I will try to mention the most important:
1] 5x IRFP 2907 21USD . Source . These were used in parallel, on the large heatsink, to switch the capacitor discharge on and off
2] 2x IRF2807 3USD . Source . One handles the charging and the other the discharging of the capacitor, to keep it at the predefined voltage level.
3] 1x ULN2803 5USD (10 pieces pack). Source . TRANSISTOR ARRAY-8 NPN DARLINGTONS used in the transistor driver board.
4] 1x Atmel ATMega 16-16PU 12USD. Source .
5] 1x 16x2 HD44780 Character LCD 3USD . Source .
6] 1x Power Supply 19.5V 4.5A 13USD . Source . I had a few of these laptop adapters and it seems they are well fitted for my Capacitor Pulse discharge tool, for both welding and cutting.
7] 1x 1.6F 20V Capacitor 50USD . Source: local car audio shop.
8] 4x 47K Potentiometers 9USD (50 pieces pack). Source .
9] Other 20USD (16MHz Crystal, connectors, switches, wires, etc)

Part 7: July 15th, 2012


The eagle PCH can be downloaded here.
PCB Layout:

If you need to print the PCB for your own board, please use this PDF file.
Software, released under GPL License:
Here is my own software implementation, featuring the C++ LCB library that I described in a previous article:
Hex file and complete source code

Other resources:
You might also want to see Jiri Pittner's welder, presented here.

Looking forward to see your own variants of this cool project!

Here are some variants created by my readers

David's capacitor discharge spot welder
David did an excellent job at building this tool, with a very nice case and a pedal as a trigger
david_capacitor_discharge_welder_1 david_capacitor_discharge_welder_2 david_capacitor_discharge_welder_3 david_capacitor_discharge_welder_4
david_capacitor_discharge_welder_5 david_capacitor_discharge_welder_6 david_capacitor_discharge_welder_7 david_capacitor_discharge_welder_8

(Visited 2,839 times, 55 visits today)
Tagged on:                                                     

116 thoughts on “Capacitor Discharge Microspot Welder / Cutter

  • January 9, 2014 at 5:38 am

    Hello Dear Radu…
    It’s about IRFP2907

    The DRAIN is connected to POSITIVE capasitor terminal..
    The SOURCE is directly for WELDING Stick, Is it right dear Radu????

    Thanks so much Mr. Radu..

  • January 9, 2014 at 11:58 am

    @ddeeen: 1. yes 2. yes , but I mostly use avrdude – as a software developer I’m used to command prompt apps

  • January 9, 2014 at 3:22 pm

    It’s about IRFP2907

    The DRAIN is connected to POSITIVE capasitor terminal.
    The SOURCE is directly for WELDING Stick, Is it right dear Radu????


  • April 14, 2014 at 12:20 pm

    All parts that needed to built this project is complete right now..
    I’ve bought some parts from Ebay.
    I hope making this welder become success…!!

  • April 16, 2014 at 6:11 pm

    Dear Radu,

    some times ago i found your project and this of Jiri Pittner.
    Then i started to build my own one.
    I use 10xIRFP2907 as main switch, mounted on some 150mm² and 400mm² copper bars.
    To switch them fast as possible, i changed the mosfet drivers.
    Capacitors are two 1Farad in parallel, but one of them is almost scrap. It was marked as 1.2F and under 1.5mOhm ESR, measured it is 0.8F and 2.6mOhm, and the capacity goes down if i charge faster (?).
    The other one is ok, real 1F and 1mOhm.
    I also have added a big freewheeling diode, because the fast switch off of the mosfets causes high induktive voltage spikes.
    These spikes were only limited by reaching breakdown voltage of mosfets.
    I have measured spikes of 85V, 50µs long. With the diode i measure 40V/5µs. I think my Mosfets are lucky about this.
    The software is a mix of yours and Jiris.
    At the moment the whole thing works fine, and the casing is nearly done.
    Thank you very much for the inspirations.
    I have some photos, if you are interrested.
    Best regards, Frank

  • April 17, 2014 at 6:37 am

    Dear Frank,

    I’d love to see your work and maybe post some of your pics here. If you want feel free to send them over, you can find my email under the about section.

  • April 30, 2014 at 10:06 am

    Sir, do you have skype??
    Please let me know it..
    my skype is= ddeee-duos

  • May 19, 2014 at 4:59 pm

    Hi Mr Radu, thank you for this great work.
    This tool would be very useful for me , so I have decided to try
    to build it.
    I’ve been reading your explanations for days. It is hard for me ,
    because I’m not really good in electronics, nor in English. I think I have understood the most important and I have got most of the parts.

    But I have a problem with the module 1 power supply => I see 2 additionnal capacitors on your picture ‘capacitor_discharge_welder_06.jpg’ than has been drawn on the schematics,
    => and I suppose, but I am not sure, that what is shown on ‘capacitor_discharge_welder_02.jpg’ are diodes.
    Could you please explain this to me.

    Best Regards,
    David Another’One

  • November 4, 2014 at 7:54 pm

    just wanted to suggest using a tungsten electrode used for tig welding sharpened to a point with your cutter,here’s a link explaining a bit on the theory of it) https://www.arc-zone.com/pdf/GrindingTungsten.pdf ( doing so will concentrate the current flow to release from a single point, giving you a stabilised arc that you can control. try attaching something (non conductive) to use as a guide to hold the electrode at a set distance off the material. you should be able to produce a krisp clean cut as thin as paper. the closer to the material the more concentrated the arc and as you move away the arc starts to wander resulting in a wider, rougher cut.
    hope this helps a bit.
    btw, cool project!

  • November 4, 2014 at 8:29 pm

    @john, thanks for the interesting suggestions!

  • November 26, 2014 at 6:35 pm

    Salut Radu! Frumos proiect! Vreau sa il construiesc si eu. Doar o mica corectie… Ai 2 tipuri de mosfeti desenati de tine in schema electronica, dar nu se pupa cu ce zice producatorul despre ei…:

    IRF2807PBF, N-channel MOSFET Transistor, 82 A 80 V, 3-Pin TO-220AB
    IRFP2907PBF, N-channel MOSFET Transistor 209 A 75 V, 3-Pin TO-247AC

    Amandoi sunt N-channel, tu ai desenat un N si un P.

    Stima si respect!

  • November 27, 2014 at 2:54 am

    Multumesc Dorcky. Am sa corectez schema cand o sa am ocazia.

    Iti recomand sa folosesti un uC mai mic si mai ieftin, de ex. un atmega8 . Nu e nevoie de atmega16

    Si apoi IRFP2907 sunt suficienti 2-3.

  • December 9, 2014 at 12:04 pm

    what is the current rating of capacitor which is used here

Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>