4000 Joule Capacitor Bank


The idea behind a capacitor bank is to charge up as much energy as possible to short circuit that energy through small coils, aluminium paper, steel wool, wire and a lot of other things that can conduct a electric current. The short circuit current is enormous for a very short time and that big amount of energy can turn the conducting paths material into vapour.

To get an idea about how much energy 4000 Joule is, here is a couple of examples.

A human heart consumes 1 Joule of energy per heartbeat.

4000 Joule could lit up a 60W light bulb for 66 seconds, using a low energy 11W bulb it could be lit for 6 minutes.

4000 Joule is just enough for cooking 50 gram of water, to bring it from 20 degree Celsius to 100 degree Celsius.


WARNING!: Working with electricity is dangerous, all information found on my site is for educational purpose and I accept no responsibility for others actions using the information found on this site.

Read this document about safety! http://www.pupman.com/safety.htm

DANGER!: High energy discharges can be lethal, the amount of energy released overwhelms what a human limp or life can withstand.


A long time have passed since I built my first capacitor bank, the 333 Joule microwave oven capacitor bank. Since then I wanted to build a larger, but these large capacitors does not turn up often at a reasonable price. Patience was all I needed before picking up a Maxwell energy discharge capacitor from ebay at 40 Euro. It is not a pulse discharge capacitor, but will be well up for the little wear it will see in this use.

When discharging a capacitor into a circuit with a inductance, which could also just be its own equivalent inductance, the voltage will ring between the capacitor and the inductive part of the circuit, resulting in voltage reversal which can be very harmful for the capacitor. There are many ways to counter this and some of them are complex and expensive, for now I will ensure that a part of the circuit will be a thinner wire, that will always explode, and thus cutting the circuit.

Wear and tear on spark gaps is highly dependent on the energy transfer taking place. It is not as much affected by very high peak currents or energy levels, but the charge in Coulombs. Using the capacitor energy calculator on this site shows you both values and you will quickly discover that a high capacity low voltage bank will have a high stored charge versus a low capacity high voltage bank, despite they have the same energy stored, the factor in stored charge is 10 times higher for the high capacity bank.


High voltage supply776 to 1855 VDC from two transformers in step up setup and a voltage doubler.
Full charge voltage1800 VDC.
Stored energy4050 Joule.
Stored charge4.5 Coulombs.
Trigger mechanismSpring loaded spark gap switch.



Having a capacitor bank with a large capacitance, I found it attractive to be able to charge it to different energy levels without being forced to use a variac. Knowing the exact energy stored will also make analysis of the discharges more precise.

By using the multiply input voltage taps on the step down transformer with an input voltage of 230VAC I get a varying output that I put into a step up transformer where I use the 656VAC output tap. Doing this I get a output voltage range from the voltage doubler from 776 volt to 1855 Volt.

Voltage² (Volt) • capacity (Farad) • 0,5 = energy (Joule)

1855² • 0,002423 • 0,5 = 4168 joule

In the following table I have all the charge possibilities listed.

776 Volt729 Joule1 and 7
813 Volt800 Joule2 and 7
854 Volt883 Joule3 and 7
889 Volt957 Joule1 and 6
928 Volt1043 Joule2 and 6
970 Volt1139 Joule3 and 6
1028 Volt1280 Joule1 and 5
1067 Volt1379 Joule2 and 5
1123 Volt1527 Joule3 and 5
1855 Volt4168 Joule2 and 4


I wanted to as many of the parts I have already at hand and focus on using parts that are odd and will have a hard time to find a place in other projects. The first focus of this was on the power supply, looking through my stack of transformers I found two that could be used for a step up transformer arrangement with the possibility of lowering the charge voltage as described above under calculations.

Capacitors and diodes for the voltage doubler is all salvaged from old electronic equipment and is right about on the edge of their ratings. Each string of diodes can withstand 2400 volt and the capacitors can withstand 1800 volt. I hope my decision about the capacitors is good enough, I did it to use a minimum of components and still maintain a capacity large enough to smooth the DC properly.

The spring loaded spark gap switch and charger switch is a copy of the switch I built for the 333 joule capacitor bank. I was a little worried that the rather small construction was not good enough for roughly 12 times the energy. I reinforced the switch with some heavy copper pieces and larger gauge wires and mesh. The biggest advantage of this switch is that I avoid making a protective circuit for the charger as its completely disconnected when the capacitor discharges.

Shot record

In order to keep track of capacitor lifespan, here is a list of the different shots that have been made with it.

100 Joule3x steel wool + 1 Ohm resistor
300 Joule2x steel wool + 1 Ohm resistor
700 Joule2x 10 strands of AWG40
1000 Joule8x steel wool
1x 200cm 0.25mm copper wire
4000 Joule10x steel wool
10x miniature breaker and fuses
25x various electronics like phones

 Crushed cans, sparks and explosions

Steel wool
5th August 2012

The was the first test shot, 1 kJ into a small twist of steel wool.

11th August 2012

The current measurements was done with a Pearson current monitor model 101 connected to a 10x probe and a Rigol DS1052e oscilloscope. Attenuation on the oscilloscope was set to 1x. So numbers should be multiplied by 10.

Current measurements of steel wool, 60 mm length, 10 mm diameter, with 1 kJ energy discharged. Measured peak current 13 kA.

Current measurements of steel wool, 60 mm length, 10 mm diameter, with 4 kJ energy discharged. Measured peak current 29 kA.

A close up of the trigger spark gap doing a shot and the wear caused to it from about 10 high energy shots.


Having only fired the capacitor bank once, at 1 kJ, it is a little early to draw any other conclusions than it works as planned and its an ear deafening loud blast when it fires.

Having now conducted a 4 kJ shot, I can only say that it is a far as one should go with energy discharges in a small room. Feeling the pressure wave from the blast is the point where this continues outside.

Having measured 29 kA through a piece of steel wool makes me very satisfied with this capacitor, it is around 3 times more than I expected from this. Further measurements of short circuits through heavy conductors show level of around 25kA to 30kA.


5th August 2012 – 1 kJ shot into steel wool

19th October 2014 – 4kJ shot into steel wool

19th October 2014 – 1kJ shot into 200 cm 0.25 mm copper wire

19th October 2014 – 4kJ shot into a iPod

19th October 2014 – 4kJ shot through a Mini circuit breaker (SEKO DZ47-63 C10 1P+N)

39 thoughts on “4000 Joule Capacitor Bank”

  1. Pingback: First shot with 4000 joule capacitor bank! | Kaizer Power Electronics

  2. Pingback: 4000 joule capacitor bank fired at maximum energy | Kaizer Power Electronics

  3. How much did your pearson current monitor cost Mads? I had a look on the internet but they appear to be quote only (probably means they don’t sell to average joes like me lol).

  4. Hey Alex

    It requires patience to find a cheap on ebay, I was lucky to find this model 101 that was reasonably priced. Its connector was badly damaged in the shipping so the seller refunded me 33%. Took me about a year to find my first model 110 and then another year to find the 101 🙂
    I did have a spare model 110 but sold it to another 4hv member.

    Kind regards

  5. I had a “play” with my capacitor bank yesterday and noticed something that most people who are into high voltage/high power don’t really mention.

    The two 3ft wires that connect to the discharge area under go quite a bit of magnetic force when the high peak currents flow through them, in my case the slightly messy shape of my grounding stick conductor made it jerk with quite a bit of force. The adjacent HT wire 1ft apart also moved a little bit (its much neater and straighter).

    I just found that quite interesting that the peak currents can be high enough that even a straight piece of wire can make such an effective electro magnet, even with a small 130 joule bank like mine (325v 2480uF).

    Maybe it is a product of capacitance too, since I have seen similar pulling together type of behaviour with homemade HV foil capacitors.


  6. Hey Alex

    It all depends on the construction, if the wiring is made stiff enough or prepared to “bend” away from the short circuit area, you are not likely to see them move because of the magnetic forces which follows the right hand rule in electro magnetism.

    Kizmo had some very big cables fly around with his very big electrolytic capacitor bank, so there are many cases of it, just not many that write about it 🙂

    Kind regards

  7. Dear Mads,

    I have an ignition exciter unit with double 3k volts outputs.
    How i can measure it`s output energy,by simple way,in Jules Unit?

    Best regards,

  8. Hello Hamid

    This sounds like a transformer or switch mode power supply, the current it should be able to deliver must be stated on the unit. Be aware that this could only be rated for pulses since it is a igniting unit.

    One Joule is the definition of one Ampere passing through one Ohm in one second.

    Kind regards

  9. Hi
    Could you please help me out with a circuit diagram to test an electric cable with a ground fault.

    It must be able to work from a 12v car battery and give me 1spark every 3-6 sec. Loud enough to pinpoint the ground fault

  10. Hi Heinrich

    You should look for ignition coil drivers or 2N2222 drivers, those are simple and does not have much energy behind them, so you avoid high voltage damages to the cables you are examining.

    Kind regards

  11. Pingback: Merlin Gerin 1600A vs. 10A Circuit Breaker 4000 FPS Slow Motion

  12. Pingback: Merlin Gerin 1600A vs. 10A Circuit Breaker 4000 FPS Slow Motion - opentesla.org

  13. Pingback: Nokia 6230 Exploding in Slow Motion 1000fps

  14. Pingback: Nokia 6230 Exploding in Slow Motion 1000fps - opentesla.org

  15. Pingback: iPod Nano Explosion in Slow Motion 1000fps

  16. Pingback: iPod Nano Explosion in Slow Motion 1000fps - opentesla.org

  17. Pingback: Samsung Omnia Pro B7330 Explodes in Slow Motion 1000fps

  18. Pingback: Samsung Omnia Pro B7330 Explodes in Slow Motion 1000fps - opentesla.org

  19. Pingback: Nokia Lumia 710 Exploding in Slow Motion 1000fps

  20. Pingback: Nokia Lumia 710 Exploding in Slow Motion 1000fps - opentesla.org

  21. Pingback: Siemens Simatic S5-110 PLC CPU Explodes in Slow Motion 1000fps

  22. Pingback: Siemens Simatic S5-110 PLC CPU Explodes in Slow Motion 1000fps - opentesla.org

  23. Pingback: HTC Snap Exploding In Slow Motion 1000fps

  24. Pingback: HTC Snap Exploding In Slow Motion 1000fps - opentesla.org

  25. Pingback: Sonos Connect Exploding in Slow Motion 1000fps

  26. Pingback: Samsung S4 Mini Explodes in Slow Motion 1000fps

  27. Pingback: Samsung S4 Mini Explodes in Slow Motion 1000fps - opentesla.org

  28. Pingback: Samsung Galaxy Google Nexus Explodes in Slow Motion 1000fps

  29. Pingback: Samsung Galaxy Google Nexus Explodes in Slow Motion 1000fps - opentesla.org

  30. Pingback: TDA7492 Bluetooth Amplifier Exploding in Slow Motion 1000fps

  31. Pingback: iPhone 4S Exploding in Slow Motion 1000fps

  32. Pingback: 10,000 Youtube subscribers & Capacitor Discharge Explosions

  33. Hello. Ive been buildin some capacitor banks and came across the same problem everytime. Discharge breaks some of those electrolytix caps what i used to make the bank. And usually on the anode side of the bank is broken after the discharge. So how to keep them safe from the discharge backfire? Thank you!

  34. Hi STN

    Reverse voltage transients from the short circuit is the killer, opposite potential shatters the internal construction. The only thing to do is prevent reverse voltage from being seen by the capacitor. Very fast disconnect switches or making sure that the short circuit blows up fast enough, to hinder reverse voltage to be too high. Depending on the quality of the pulse capacitor, they are usually rated for something between 30-70% voltage reversal levels.

    I killed this capacitor about a year ago 🙁

    Kind regards

  35. Thanks for your answer Mads! 🙂 this capacitor bank is one of those projects which never get ready xD .. my current setup is a lot of caps because they are around 400V electrolyte caps had to put them also series. basicly what ive sone is that 5 series and 5 rows and then connected the – and plus in parallel to create one cell.. atm build is 3880uF and 2000VDC so if my calculations are right that would be 7760J and stored energy 7,76C.. Still have that many caps that i could build 20 000 J Bank.. I am loading it with microwave transformer through full bridge rectifier.. I was just thinking that cause i ofc dont have variac it always charges the bank full.. was just thinkin that i might not do that much harm to the caps if i loaded it with 1000VDC..? any idea how to get the voltage lower? do i need resistors or something..? And does it affect how much if the microwave mot is still connected to the caps while shorting them? (and no, its not on then) but just connected.. i was thinking that it might affect those current spikes etc.. Here is pic of the current progress, and yes i know its not pretty… Hope you can help. Cheers! -STN

  36. Hi STN

    You would properly need a charging resistor, to limit the current flow a bit, it also suppresses sparks flying and wearing down charge contacts. Even without charge resistors, it would take some 30 seconds to charge 4kJ in my bank, with a MOT. So to get a lower charge, you simply just charge it for a shorter time. Add a analog voltmeter across the bank.

    If you do not disconnect your charging circuit, when firing, you risk blowing up your rectifying diodes and might damage the transformers. Look carefully at my switch and you can see its charging when pulled back, when released its spring loaded to discharge, so it is break-before-make.

    Kind regards

  37. I think your launcher system is awesome! 🙂 There is just more risk to tear down the feed from the mot rectifier thing, but one hand rule and with good 2000V proof pliers it should be ok.. How did you disconnect the caps from the loading power supply?.. mean what is the safest way? I am planning to make my launch switch with 2pcs of steel plates around 2″ or 5cm diameter and was planning to use just original mousetrap (which i gonna mod. a little)[planning to set off the trap by pulling string enough far away from there, and then the circuit would be full] also what ever is gonna be between the shorted poles is gonna be ionized.. Charging resistors are quite uncommon to me.. so where i should connect them? after the full bridge rectifier output or before it goes to the bridge..? or even before that, -> to the mains side? and what kinda valued ones? And what goes to to that sparks flying /i guess you mean between the caps randomly sparking?/, none of that has ever happened or it was totally bad soldering job…

    when i get my mouse trap and nerve to test this thing out i dont want to be anyway near to it, cause that last one what i tested was about that 3450J or (something like that) shorted it outside but ofc didnt have any ear protection xD it was close to my house’s outside wall and that blast hit me like id got a proper slap on my face 😀 (luckily had safety glasses)…. so you suggest that the piece which is about getting shorted could help them caps to survive if i made like aluminium or copperwire coil with multiple turns in it? And is it so that the thinner the wire is the better, chance is there that caps will survive?.. cause then those caps wouldn’t discharge all the way..?

    P.S I don’t have analog high voltage meter, have heard so much bad of those (which come from alibaba, ebay, amazon etc.. but i definetly should buy one to give it a go, maybe there is proper ones too. I only got now my trusted BEHA AMP PROBE its maxed out to 600v, otherwise its my favourite meter. what kinda resistor or what should i use to measure the HV side of the bank with it? without breaking my meter? Sorry if im asking too much but this is not my main region. And i really don’t want to eff things up with my caps bank AGAIN after the first test xD. But I rly understand more about distribution transformers than these. Have been now some years building bigger and bigger ballast to the dist trafo. and i am using it the step up way. Been driving it only with single phase 240V/10A before the ballast. And after the ballast measured Voltage was 380V and peak current were there around 48A on the Low voltage(input) side. HV end was maybe around 30 kV and that info lies only the spark gap what i measured . But thanks for your answer and support Mads my mate! The discharge time is definetly what i should start to look at first. Not sure does the same rule work like 1 0N 1 when the cap bank gets bigger. but if your 4KJ bank took about that 30 sec so nearly 8KJ would be then little less than a minute i assume.. I consider that loading that bank like 15-20 sec would be totally okay for it..? And sorry didn’t had my eyes on your final words on that one comment, R.I.P your caps bank :/ , also i broke this piece…(pic at the end) So I rly know how it feels bro! The best part is that when this cap came to me (thanks to the post ,ups or who ever who effed up the other bushing, it had slight bend to it but otherwise it worked fine.. then i guess i went too far with my experiments and it broke just like 2 weeks after i got it -_- … these are these so called “studying moneys” as we say here in Finland.. But thanks again Mads, learned a lot from you in these couple messages we have written 🙂

    PPS. here is the caps plate as written cause 2 pictures were too much to upload.

    4800V 1phase 95kV BIL (no idea what is BIL
    50KVAR AND OFC 60Hz… realized that just after i ordered it.. 60Hz dosen’t come well with 50 Hz but thats not the reason it broke.. i ran it with DC and that same 2000V.
    it is filled with 1,2 gallon combustible liquid
    But R.I.P also to this..


  38. Hi STN

    See here: https://kaizerpowerelectronics.dk/wp-content/gallery/2012_06_01_-_4kj_capacitor_bank/IMG_7707_close.jpg

    When pulled back, by the white plastic string, it hits the charging position, when released, the spring forces it to hit the discharge copper terminal and just after it breaks and spring goes back to upright position, no touching charging or discharging terminal. Quite clever design I made there, if I have to say it myself 🙂

    You can use analog or digital volt meter, most have a input impedance of 1M Ohm, then you just add more resistance to change the ratio. Add 9M Ohm for a 600V meter to become 6000V. You loose precision and response time, but for a capacitor charge indication, its good enough.

    BIL is short for Basic Insulation Level and is a high voltage standards measure for the highest possible voltage a piece of equipment can withstand without failure. So its not a rating you can use in operation.

    Kind regards

Leave a Reply

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

The maximum upload file size: 10 MB. You can upload: image, document, spreadsheet, text, archive, other. Drop file here

Scroll to Top