Introduction: 50,000V High Voltage Power Supply

Picture of 50,000V High Voltage Power Supply


  • This is an extremely dangerous project, and it should only be attempted by people with experience in electronics, and specifically, High Voltage. If it's your first time seek help for your own safety.
  • Homemade High Voltage supplies are unlikely to meet any international standards, the safety and correct operation is NOT guaranteed at all, and will depend on the skill level of the builder, the effort put into it and most importantly, the common sense.
  • This is not intended to be a tutorial on how to build a high voltage PSU, I only intend to show how I built it. This PSU has been built according to my criteria, and it can be useful if you're looking for inspiration, but I discourage anybody who wants to replicate it blindly without applying his/her own common sense. I'm not an expert on the subject.
  • I'm not responsible for any damage or injuries caused by the use of this information.
  • Please, be extremely careful with High Voltage.

Step 1: Intro:

Picture of Intro:

This high voltage power supply has been designed to output a fixed voltage of around 50kV, it could easily be converted to an adjustable supply by connecting a variac in case of using transformers or by adding some extra circuitry to regulate the power going in. I initially thought about a high frequency PWM to regulate the power going into the capacitors, but I abandoned the idea. I found that adjusting the frequency is enough to make the voltage vary by a significant amount, allowing some control over it, this happens because the flyback must operate at a certain frequency in order to maximize the output.

The total cost of building it is around 10 to 15€ since most of the parts (transformer, bridge rectifier, heatsink, flyback, switch, connector, cables...) have been salvaged, the only parts that I bought are the components of the 555 driver, the connectors and the capacitors. This shows the importance of having a big pile of electronic junk, specially old stuff with chunky electric and electronic components, it doesn't matters if you have to pick it from the dumpster, it can save you tons of money on the long run and by repurposing these devices you're being Eco-friendly. A good practice it to save the tin when desoldering and avoid throwing it into the trashcan and when you're done with the board or there are no more valuable components you can take it to a place where it can be recycled properly.

Caution has been taken in order to isolate the high voltage output from the user and the internal circuitry.

Step 2: Materials:

Power input:

  • Transformer + bridge rectifier + capacitors


  • Switching PSU

Both must be rated rated 5 amps and 20 volts at least for higher voltages.

  • Separate* input for the driver
  • Switch of choice and connector
  • Shrink tube

Driver circuit:

  • Perforated board of PCB to be etched
  • 555 timer
  • 8pin socket
  • 7812 (if the power intput to the 555 is > than 14.5 or lower than 35)
  • Small heatsink for the 7812 or other regulator (if needed)
  • 2*100nF
  • 1*1uF
  • 1*10nF
  • 1*68uF (or 100uF if you wish)
  • 2*4148 diodes
  • 3*10k
  • (1 per MOSFET) 10R
  • 1*680R
  • 1*470R
  • 1*10k pot
  • 1*100k pot
  • 2* pot knobs
  • 1*2N2222 and 2N2907 (or other NPN - PNP pair)
  • 1*Infrared sensor
  • 1*Infrared LED
  • 1*BC547 (or similar e.g: 2N2222 or 2N3904)
  • 2*Banana female connector (or isolated high voltage connector)
  • MOSFETS (I used 3* IRF540N but I recommend 1* IRFP260)
  • Heatsink for the mosfets (and fan if needed)
  • Pushbutton

High voltage:

  • Flyback transformer from an old TV or computer monitor
  • Thick copper cable (around 1 meter)
  • Epoxy

*You can try using the same supply used to power the flyback to power the driver, but I'm not sure about if this would cause the driver to operate under rough conditions and lose overall efficiency. I've used a 12V 1A PSU from an old router. Remember to connect both negative terminals of the power supplies together, otherwise it won't work.

Step 3: Calculations

Picture of Calculations

The only calculation we'll need to perform is to obtain the value of the capacitors (in case you're using a transformer and not a switching PSU to power the flyback) we use the formula provided by the picture.

In my case I used 20000uF, this causes quite a lot of ripple, but it isn't really important since the flyback won't be picky about it, I'll maybe add 10000uF or 20000uF more just to see the effect on the output. I discourage wiring the driver to the same source than the flyback, the ripple created by the changing currents could alter the correct working of the driver, resulting into lower efficiency and smaller arcs.

Step 4: Building the Case

Picture of Building the Case

Every power supply needs a case to hide the components from the user, and this is a high voltage power supply, so extra care must be taken. The obvious choices are wood and plastic or variants, although plastic is preferable. Metal is not a good option for a beginner, specially without further isolation and using proper rated connectors, cables... even one of the sides being metallic adds a tremendous amount of danger for you and the circuitry.

Even though the use of plastic is encouraged, I chose wood because I couldn't find a plastic case big enough, this has allowed to further isolate the high voltage parts avoiding corona discharges and other potential dangers.

After applying my basic woodworking skills I came up with this, the steps are detailed in the pictures.

Note: If you plan to paint the case, check the paint isn't conductive at high voltages when dry, oil based paint is recommended if it's too dense mix with a thinner (e.g: acetone). Water based paint is obviously forbidden.

WARNING: Although wood is a very good isolator it can't be compared to plastic due to it's ability to retain moisture. Using wood is discouraged if it's going to be in a very humid environment, and extra care must be taken in order to keep the wood away from water. It's a good advice to let the wood dry in an oven before applying paint or treating it.

Step 5: The Driver Circuit:

Picture of The Driver Circuit:

I've made this small driver circuit it revolves around a 555 timer with adjustable frequency and duty cycle (from 5-50kHz and 5-50% duty cycle), it has it's own 12v input independent from the transformer, which has an 8 volt output, unfortunately this one doesn't delivers enough current to charge the gates fast enough, leading to efficiency problems which translate into lower output voltage at the flyback.

The MOSFETS are three IRF540N in parallel, I used these because I had them lying around and they're relatively cheap, you can also use the popular IRFP260N. In this configuration they barely get warm even under full load.

I was going to add an optocoupler to the circuit but I decided not to, Instead I just wired a button with a 1k resistor where the IR sensor should be placed. You can edit the circuit and remove the transistor, leaving the button connected to pin 4 with a 10k resistor going from it to ground.

Note: To make the duty cycle go from 5 to 50% (and not from ~5% to 100%) you have to place a 10k resistor at the potentiometer as shown in the picture, this resistor must be placed in series with the diode facing the capacitor. If you connect it in series with the other diode you'll end with an adjustable duty cycle from 50 to 100%.

You can download the Eagle schematic and the board files below if you plan to etch a PCB.

Step 6: Wiring Things Up:

Picture of Wiring Things Up:

After making sure the circuit works correctly I wire the MOSFETS in parallel, to do this I join all the drains and sources with high amperage cables, add a 10Ω at each gate and joint them together.

Note: If you parallel MOSFETS add a 10Ω resistor at each gate and bridge the 10Ω resistor on the PCB.

I also attach a mains connector and a switch to the case and wire them, it is very important to use heatshrink tube or other kind of protection when making connections with the mains, you don't want exposed connections around. I screwed the ground connection to the case just in case I want to wire something afterwards and so it doesn't moves.

If you want to make an optocoupler to further improve the safety of the supply you can make one with an IR LED and sensor and a piece of heatshrink tube (white is better since it reflects light better than black), wire the IR LED to a batery pack in series with a resistor and a button and you're good to go. Just make sure this circuit isn't in close proximity to the other one or you'll defeat the purpose of the optocoupler, which is preventing you from getting shocked. Without an optocoupler current can travel through your body in case you touch or get near the high voltage output, reach the button and enter the circuit again.

WARNING:A pushbutton is extremely preferable over a switch, in case of accident the button will spring back and shut the circuit down. NEVER use a switch to turn on the high voltage output unless everything is always under special conditions that make it impossible for an accident to occur.

Once the driver has been assembled I can start testing without having assembled the PSU yet.

Step 7: Wiring the Power Supply:

Picture of Wiring the Power Supply:

After I found a suitable power source I attach it to the rest of the circuit. I connect the 12V PSU to the grid along with the transformer with a single pole switch wired as shown in the pictures. I connect the transformer to the bridge rectifier and then to the capacitors using plenty of heatshrink tube.

Wiring the power supply to the flyback and MOSFETs is shown in the next step.

Step 8: Preparing, Wiring and Isolating the Flyback

Picture of Preparing, Wiring and Isolating the Flyback

To prepare the flyback I just need to find the negative pin, this is usually done by powering the flyback up, then swing the high voltage cable around the pins until it arcs to the right pin. I solder some thick wire and cut and cover all the pins with epoxy, this is a good practice, since it avoids corona discharges and other nasty things, hot glue works too, but epoxy is better since it soaks better.

Then I wind around 10 turns of thick wire around the core of the primary. The positive output of the PSU is connected to one of the ends of this wire, the other end gets connected to the drain of the MOSFET, finally, the source is connected to ground. You can use a terminal block to swap the connections around (see picture), it's quite convenient.

One very important thing to notice is flyback transformers have polarity, they have a built in diode which only allows current to pass in one direction. To test which way you should connect it just power the flyback (lowering the power just in case) and place the positive and negative wires at a certain distance, then reverse the connection, compare both spark lengths, if the spark or corona discharge is bigger one way or the other you'll know you've connected it right, I usually mark the negative side with a permanent black marker in order to avoid confusing them.

To isolate the flyback I pass the high voltage wire to the next compartment though a small hole, roughly the diameter of the wire, I solder the wire to the banana connector and then stuff the cavity with plastic film, this will ensure there's no arcover or corona discharges, i cover the plastic film with some plastic foam. I cover the flyback compartment with the same kind of foam.

You've probably noticed that small cup around the connector, I drilled a small hole and passed the connector though, then I bolted it to the case, it might not be as good (and expensive) as a standard high voltage connector, but this cup is a good protection against accidental arcovers.

EDIT: I added a 22uF 250V capacitor across (or parallel to) the primary in order to achieve some resonance, this has improved the current output and I would say the voltage is a bit higher now, I might be wrong, but the transformer and bridge rectifier don't seem to get as warm as before, or at least I haven't felt any bad side effects of using the capacitor. I recommend using a non polarized high voltage rated capacitor since the voltage spikes at the primary can be nasty.

Step 9: Test Run:

Picture of Test Run:

After everything is connected you can test if everything is working fine. Be extremely careful when doing your first test, it's extremely advisable to set a spark gap and a wood stick to push the button just in case. If you're using two separate power sources for the driver and the flyback, remember to join the negative terminals, otherwise it won't work.

After completing the first test without problems I would test it again in total darkness to see if there are corona discharges where there shouldn't be and fix this with epoxy, corona dope or some other kind of insulating material.

To optimize the flyback's output power we'll need to adjust the frequency and duty cycle, to do this you can set a spark gap and adjust the knobs until the spark length is maximum. You can also measure the current flowing through the secondary in and adjust the potentiometers to obtain the maximum amount.

Step 10: What Can I Use It For?

Picture of What Can I Use It For?
  1. X-ray tube PSU
  2. Gas ionization
  3. Ozone generator
  4. Nitric Acid generator (Birkeland-Eyde process)
  5. Metal vaporization and deposition
  6. Smoke precipitator
  7. Jacob's ladder
  8. Franklin's bell
  9. Ionocraft
  10. Electrostatic motor
  11. SG tesla coil PSU
  12. More random experiments (and burning things up)

And the list goes on...

You could possibly convert it into a plasma speaker by connecting an audio input to pin 5 of the 555, you can add a small amplifier if the sound isn't loud enough.

Step 11: Some Safety Advices:

Picture of Some Safety Advices:

I found this instructable which explains in detail the procedures and guidelines when dealing with high voltage as well as identifying potential electrical dangers.

I also found this pdf I already posted in other high voltage related project.

It's extremely advisable to read and comprehend this information, you shall not proceed with any high voltage project without knowing the basic safety precautions and procedures and without previous experience.

Special equipment like properly rated rubber gloves should be worn.

Also, when high voltage is mixed with vacuum there's a chance to produce X-rays, vacuum tubes and some low pressure lightbulbs are known to produce X-rays when a high voltage is applied. Be careful an follow general radiation guidelines to avoid accidental exposure to ionizing radiation.

Step 12:

Picture of

I hope this has been useful to obtain some inspiration and gain experience about experimenting with flyback transformers and high voltages.

Thanks for watching and please be safe.


LauraMoon (author)2017-10-23


In shematic is two transistors T2 - 2N2222 and T3 - 2N3904 connecting Emiter to Emiter, but on the plate they are connected Emiter to Colector. Should not be turned to the other side T3 ?

Victor805 (author)LauraMoon2017-10-25

Well spotted, the reason the transistor is flipped is because I used P2N2222's, which have reversed pinout compared to the normal 2N2222.

Dragonite123 (author)2017-10-14

Looks dangerous.

Victor805 (author)Dragonite1232017-10-16

It quite probably is :P

kishorb8 (author)2017-04-14

can i use 3524 push pull circuit to drive to drive mosfet.

Victor805 (author)kishorb82017-05-07

I guess, MOSFET drivers are probably a better option here.

namindu (author)2015-09-02

Can I use arduino board as 555 timer

Victor805 (author)namindu2015-09-03

Of course, I used Arduino before to drive a flyback with nice results. The only problem I found was to generate a square wave, since the program I wrote was too bulky to run at a higher frequency than 17kHz.

could you provide us with the code maybe make it adjustable?

Arduino has it's limitations, so try to optimize the code as much as possible if you want to go above 20kHz

I don't have the time, but you can try and expand this basic piece of code:

f = 17000; //Frequency
d = 45; //Duty Cycle (restricted from 10 to 60%, otherwise damage may occur, 45% set as the standard duty cycle)
t = 0;
t = 1000000/ f; //period in microseconds
x = (d/100)*t;
y = (100-d)*t;
//Put this in a separate loop:





rah187 (author)2016-08-14

Cool! As a high voltage hobbyist, this is valuable information. I have no idea if I will be doing this anytime soon due to time limitations, but it's good to know that I can build a high voltage power supply with relative ease.

zappenfusen (author)2016-07-08

Late comment but having once worked in a Sugar refinery powered with 575 volts having no earth grounding and hence no relation to ground was a learning experience all should have. The building of indicator devices giving visual warnings a short was existent was only one of many changes to understanding of distribution systems in large industry utilizing HV systems. One tester lead on a phase, one on known ground would give indication of 0 volts. Very weird to apprentice in training. Lord forbid hang yourself between phases though! Coincidentally that refinery went up in an conflagration leading to numerous OSHA studies on Dust Control and cleanliness in Dust intensive environments. These studies resulted in exoneration of all executives responsible for lost and ruined lives of tenured, loyal employees. Of course. Money talks!


Amir HosseinS2 (author)2016-04-30

Hello guys .

to tell you the
truth i couldn't follow the structure ,i found it very confusing,
specially i don't know anything about transformer which was used in this
project. i wanna to have a clear map for creating this . may somebody
give me a clear map ?

i really approciate it

Victor805 (author)Amir HosseinS22016-05-01

Sorry mate, but all has been written with enough detail so it can be
easily understood by anyone with at least some experience. With all my
respect, if you don't understand it, this project isn't for you.

Try other projects and gain some experience before trying to do something as dangerous as this.

HaroldV10 (author)2016-04-15

blue button used for what ?

I didn't see any button in your circuit

Victor805 (author)HaroldV102016-04-16

Sorry, I did not mention in detail that the button should be connected to the diode in the optoucoupler or directly to pin 4 of the 555 with a resistor to ground.

HaroldV10 (author)Victor8052016-04-21

so add a diode and resistor to the schematic ?

Victor805 (author)HaroldV102016-04-21

The IR diode emitter and resistor are out of the circuit, it's an optocoupler, the diode should be powered by a couple of AA batteries and a resistor of about 100 ohms, it is very simple, it's all explained in step 6 with a picture of the optocoupler itself.

zappenfusen (author)2015-09-01

Being an electrician by trade, retired and alive, I've experience many self prescribed treatments for stupidity. Only once did my life flash before my eye's. All warnings should be taken seriously and avoid grounding yourself at all cost. Neat project explaining the xfmr's.


nwonharp (author)zappenfusen2016-01-19

Yes ,

I have worked with many things that were deadly in my life , and want to caution everyone to be careful .!!!

Cheers !!

namindu made it! (author)2015-09-04

Thanks! It works. But, I have got a small spark about 0.5 cm. I wind 15 terns to my primary coil. and I used 555 with 2N3055. So I want to know why the spark got small

Victor805 (author)namindu2015-09-05

A spark so small could be indicative something is wrong with the driver, here are some things you could try:

-3055 power transistors take a lot of current through the base in order to go into saturation, the driver circuit might not be powerful enough to switch it properly and fast enough, try changing it for a MOSFET like the IRFP260N. This circuit has been designed to work with MOSFETs not BJTs

-Check the power supply can deliver enough power.

-Check the 555 is working and outputting a proper signal, an oscilloscope would be nice for this task, but a multimeter with a frequency function could be helpful too.

-Adjust the frequency until you get the biggest arc possible.

-Check the polarity of the flyback.

donfrench (author)2015-09-03

Would there be an application for this in very high speed flash photography?

Victor805 (author)donfrench2015-09-03

I don't really know since I don't know much about that field, it could be used as a power supply for the flash unit, but I guess those units work with lower voltages.

donfrench (author)Victor8052015-09-03

The very high speed flash device I was thinking of uses a 35,000 volt capacitor and a 45,000 volt trigger transformer. But apparently you don't need a high voltage power supply to charge the capacitor.

MrCheatak (author)2015-09-02

Max amperage?

Victor805 (author)MrCheatak2015-09-03

Let me measure for you:

30mA output amperage, the input amperage must be around 5 Amps

Omnivent (author)2015-08-29


50 Killer-Volts and a wooden box - In what world is that safe?

I'd recommend anybody wanting to play with high voltages, to stay away from wood for box material, even if it's "just" containing mains voltages (or at least impregnate the wood with a generous amount of epoxy or similar) - wood as is, is only an isolator when completely dry and unfortunately, even moist from the air is enough to make it a less effective isolator (which I'd have thought that anyone tinkering with HT would know?!).

A proper grounded/earth connected metal box would be a lot safer, but a good quality plastic box would be my preference.

cmarcus1 (author)Omnivent2015-09-02

Actually it's not clear that using a grounded box would be a lot safer - it depends on the type of fault. If it's a momentary spark from the output to the box shell, the inductive impedance of your ground cable may be too high, causing the box to momentarily hit several kV. So I'd exercise careful judgement on this matter...

Omnivent (author)cmarcus12015-09-03

One should always exercise careful judgement when dealing with potentially dangerous stuff - that goes when posting stuff that children can read as well.

I'm sorry, but your argument about a high impedance ground doesn't play, as part of the definition of a ground is that it is low impedance and if you have it otherwise, you're having a defect ground and as such should repair it, rather than using it as an invalid argument - A car with e.g. a broken brake line is in the same category, but does that cause you to do a full 14 hour checkup of any car you're about to drive?

So yes, a (properly) grounded box is a lot safer!

cmarcus1 (author)Omnivent2015-09-03

I'm not arguing against a properly grounded box - of course that is safer.

I'm just pointing out that most grounds (like wall plug grounds) are actually quite high impedance at high frequencies. They won't necessarily be able to protect you from a high speed fault. In that case, a metal box is actually less safe.

Normally this wouldn't be an issue, but with 50kV it certainly is.

Victor805 (author)Omnivent2015-08-30

I've had no problems so far or felt anything weird, but I agree plastic would be preferable. In my case I let the wood dry before applying the oil based paint. As you said, dry wood has a high resistance, even higher than than glass or rubber, but spilling water over it or being in extremely humid conditions could render it useless. After informing myself I thought it was safe to use wood, it's quite dry where I live, so I thought it wouldn't be a problem, but I'm not sure if I would recommend using wood for someone living in England for example, although I don't think this is a potential danger, considering that the distance (from the contact point to the the high voltage output in "low resistance" humid conditions) necessary to feel a shock would probably be enough for an arc to jump from the output to your hand.

In any case, a wood case would require special care to avoid water getting near it.

I'll add a warning about this, thank you.

treesquatch (author)2015-08-28

What kind of devices/machines would this be used to power?

guds777 (author)treesquatch 2015-09-02

Coil gun, rail gun all the way.

Victor805 (author)treesquatch 2015-08-28

As exposed in step number 10 there are many devices/machines that rely on high voltages, CRTs, X-ray scanners, night vision googles, mass spectrometers.... It is also used in the industry to manipulate small particles or particles in the air thanks to electrostatic forces.

nwonharp (author)2015-09-01

Kind of a neat project , but someone could get " fried " if they didn't pay close attention to the safety aspects . I agree with Omnivent that a properly grounded/earthed metal box would be the safest . If you choose to use wood , you could put the wooden box in your kitchen oven and bake it at say , 225deg F or 110deg C for about an hour ( you may need to experiment ) to evaporate all of it's moisture , then seal it with a suitable sealant . Make sure that your wires on the high voltage side are rated for the task . I have an old 25,000 volt transformer that was used to power the neon lights on a motel . It and the remaining neon tubes were given to me when the motel was closed down . It makes an impressive " Jacobs ladder " with just a couple of coat-hanger wires at a 15-20 deg angle for the uprights . The spark will start at the bottom , climb to the top , then " snap " and start at the bottom again , due to the ionization of the air , and produces ozone . Only do this in a well ventilated area Anyway , be very careful around high voltages !

Cheers , take care , and have a good day !......73

nate71173 (author)2015-08-29

Very cool

About This Instructable




Bio: I'm an electronic engineering student. I don't usually have much spare time but I like to work on random projects to keep myself ... More »
More by Victor805:Quick and Easy Office Chair FixMake Golden Coins (really Easy)The Joule Blinker
Add instructable to: