Hello everyone,

This is a writeup on a project I completed this past spring. In short, it's a simple 20 kilovolt power supply in a black box, capable creating 1-inch plasma arcs between the output wires and low energy X-rays in vacuum tubes. I built it using a big flyback transformer from a TV I found on the road and a fluorescent light driver.

Once again, the pictures will be a bit worse than I would like due to the fact that I am away for the break and this device is a couple hundred miles away right now, so I am editing old pictures I have to include.

But first, a very important message: This writeup is for informational purposes only.

EXTREMELY CRITICAL WARNING: DO NOT ATTEMPT TO RECREATE THIS PROJECT!

...unless you have experience with high voltage and 100% know what you are doing. This isn't like my other projects, where the worst thing that could happen is an uncomfortable capacitor shock or a little higher radiation dosage than usual. This device can kill you if improperly handled. From the input line to the driver output to the actual HV output, there are NO LOW VOLTAGE PARTS to this device, and touching ANY part of it or improperly using it puts your life at risk. DO NOT BUILD THIS. DO NOT BUILD THIS AND THEN TOUCH ANYTHING ON IT WITHOUT GLOVES. TAKE HIGH VOLTAGE PRECAUTIONS. YOU HAVE BEEN WARNED. DO NOT BUILD THIS. DO NOT BUILD THIS!

High voltage systems should not be handled without taking proper precautions. This includes but is not limited to wearing insulating gloves/shoes, never touching any part of the device while it is on, using a remote switch to turn the device on and off, and more. If you are not familiar with HV safety, don't trust me and my advice, go read up on it from an expert.

Okay, with that out of the way, let's get to the project:

Overview

This device is a 20,000 volt (measured) output high voltage power supply with approximately 2.5 mA peak current. As such, it isn't as deadly as a microwave transformer, but it is still extremely dangerous and should not be underestimated in terms of physical danger. It runs off of mains 120V AC from the wall. Several of the critical components come from old CRT televisions, which are easy to get for free. Here's a picture of some plasma in a bad vacuum chamber I made using the completed device:

(Image credit: Me) This plasma bends in the presence of magnets

I decided to build this after seeing several videos online of other people building very similar devices. As such, this isn't my design, although as with all DIY stuff like this I had to modify it a bit to get it to work.

The basic circuit works like this: 120V alternating current from a wall outlet comes into a standard fluorescent light tube driver circuit. The driver is meant to, obviously, drive big light tubes, and steps up the 120V 60 Hz current to ~800V high frequency alternating current. This is then fed into a large flyback transformer (extracted from the TV), which steps up the voltage and rectifies it to produce 20,000 volts of direct current at the final output.

The most critical component is the flyback transformer that handles the final voltage step up to 20,000V. Thankfully, these devices are very common, as they drive every single CRT (cathode-ray-tube) TV. Luckily for me, people are still getting rid of their old CRT TV's and replacing them with LCD's. I spent an hour or so driving around my hometown last winter watching the side of the road until I found what I was looking for: An old CRT television set lying out on the road next to the trash. I brought it home and cracked it open.

Taking apart CRT televisions is dangerous in itself. There are several big, ~200V electrolytic capacitors on the main PCB that don't necessarily discharge, which can result in an extremely painful 20,000 volt burst going through the flyback transformer and into you to ground if you try to handle the circuitry without first dealing with the capacitors. The output of the flyback transformer connects with a suction cup to the back of the huge vacuum tube, and is used to accelerate electrons into your screen and produce images (actually, this is essentially the same way the tritium tube from my last post produces light, although in the case of the television, the electrons are electrically accelerated and the phosphor is different). By clipping an insulated-handle screwdriver to the ground mesh surrounding the television tube and inserting the screwdriver tip into the suction cup (while wearing gloves and not touching the conductive parts!), it is possible to somewhat safely discharge the dangerous capacitors and flyback. After a few more taps to all of the suspicious capacitors on the board with my screwdriver, it was time to take it apart.

I took way more parts than necessary, including all of the big capacitors, high-power resistors, and coils, but the only important part is of course the flyback transformer. This is soldered to the main PCB using about 12 pins, so it was somewhat of a pain to get off (the television was manufactured in 2001, so the solder was very old), but eventually I was able to extract the flyback and recycle the remaining television husk. A picture of another flyback I found in storage is shown below (the one in the HV supply is extremely similar):

(Image credit: Me)

Next up was the light driver. I was tempted go to my go-to (the dollar store) and buy some $1 flourescent lights to use the drivers from. I ended up taking some apart, but just because of how incredibly sketchy they were (yes, this entire project wouldn't win any safety awards, but I do have some standards), I ended up buying a cheap light ballast on Amazon. Once it arrived, it was only a matter of putting it all together.

I still utilized my favorite dollar store for the casing, a black toolbox. I cut holes for the output, ground, and input wires using my really crappy, abused soldering iron that fell apart the next week. For the input wire, I used a thick 120V AC plug cut from the same TV I got the flyback from. The HV flyback output wire is obvious (it's an extremely thick red wire with a suction cup on the end), but the negative wire isn't so obvious. You can find it either by measuring relative resistances or checking which pin the flybacks arcs to the most. I temporary fired up the circuit, and checked which pin the output arced to. This is the relative negative for the flyback, so this got the last wire exiting the box.

Several gallons of hot glue later and feet of electrical tape later, it was done! When powered up, there's an audible buzz and the red output wire obtains a potential of +20 kV relative to ground (once again, this is very very very dangerous, do not attempt this unless you very much know what you're doing). 20,000 volts will permanently destroy my innocent $10 multimeter, so I used two Gigaohm resistors (Yup, 10^9 Ohms per, unless I am remembering incorrectly) to drop off most of the voltage, which lets the multimeter measure a much lower voltage that you multiply by a constant based on the resistor values to obtain the true voltage, which I measured as very close to 20,000 volts as expected.

At this point, the device is complete. But you might be asking, what is this dangerous box useful for?

Uses for a 20,000 Volt Power Supply

The first use is to just make arcs. When you bring the +20 kV output wire close to the ground wire, the electric fields increases as the wires get closer together. Eventually, the E-field strength becomes so high that the air in between the leads breaks down into a plasma, which then conducts current across the gap with very little resistance (https://en.wikipedia.org/wiki/Electrical_breakdown). The cool thing is that this plasma, being made up of a bunch of free charges, is affected by magnetic fields as you would expect, so you can do experiments with that. The arc will also split molecules in the air and produce ozone, so I didn't do this without good ventilation, since you don't want ozone building up in your room.

Being high voltage, this device can also accelerate charged particles. In a perfect vacuum, a particle with one electron charge will obtain an energy of 1 electronvolt (1 eV) when it accelerates across a 1 Volt gap. So, with 20 kilovolts, you can theoretically accelerate electrons up to 20 keV, which is well into the ionizing radiation range (you may notice if you read my last post that this is higher than maximum kinetic energy of tritium beta radiation). Of course, this isn't very easy: Homemade vacuum chambers are usually very bad in terms of actually getting all of the air out of the chamber, which will cause accelerated electrons to hit neutral atoms in the "vacuum" chamber and not actually reach their maximum energy. This is why escaping ionizing radiation (other than a tiny bit of UV light) isn't produced when you light a high voltage arc in the air: The electrons and ions simply cannot accelerate to keV energy levels in such a dense gas.

I found an old vacuum tube (a 2X2A rectifier tube) on Ebay and bought it for use with this supply, as I had heard that this tube produces decent X-rays. Theoretically, applying a high voltage across the vacuum gap of the vacuum tube (with no low voltage current on the filament) will cause electrons to tunnel out of the metal and accelerate across the gap. Pre-made vacuum tubes contain very very good vacuums due to the presence of a getter material that mops up left over gas, so the electrons here will easily reach their peak energy across the voltage drop. When they hit the second metal terminal, X-rays are produced in two ways: Either the actual acceleration of the electron hitting charged nuclei in the metal produces braking X-rays, or the high-energy electrons knock inner electrons out of metal atoms, causing other electrons to drop down to the lower energy level and emit X-rays. (You can read more about these two X-ray production methods here: https://en.wikipedia.org/wiki/Bremsstrahlung https://en.wikipedia.org/wiki/Characteristic_X-ray)

(Image source: https://www.radiomuseum.org/tubes/tube_2x2a.html) The 2X2A was the first tube that gave me successful XRay emissions.

As such, I was able to produce X-rays that registered on my Geiger counter. These X-rays will vary in energy and will go up to 20 keV, since 20 kilovolts can accelerate electrons to a maximum energy of 20 keV. Thankfully for me, almost all lower energy X-rays don't escape the glass tube (remember Tritium?), so I could only detect the X-rays very close to 20 keV. Even so, the radiation dosage from leaving such a device on would actually be somewhat concerning, so I only turned it on for a second at a time with a remote switch so that I was several meters away from the source itself and not exposed to any meaningful amount of dose. After a few proof-of-concept runs, I stopped doing this, since there is no point in exposing myself to high X-ray fluxes for no gain. Fun fact: This is how dental X-ray machines work, although they typically use a specially made vacuum tube and a higher voltage source.

I also used this device to create an interesting plasma in my pre-bought vacuum chamber (before I built my own later), using a crappy $40 vacuum pump. Since the vacuum pump and chamber are both quite bad and not meant for high vacuums, the air density in the so-called "vacuum" inside the chamber was way too high for escaping X-rays to be produced, so I was able to experiment around with magnetic fields and somewhat low density plasma without worrying about radiation on top of high voltage risk.

Finally, I was able to produce X-rays from a homemade "tube", but it took a bit of troubleshooting. The vacuum pump I have can draw a better vacuum in smaller containers, because outgassing and leaks are both reduced. In addition, when the high voltage terminals are closer together, the electric field between the terminals increases, which in turn means that the electron acceleration is increased and the neutral blocking atoms aren't as important since there is less space for them to get in the way before the electron reaches its maximum energy. Combining these two principles means that by using two brass ports hooked up the HV supply separated by a few millimeters of tubing with a vacuum in between, I was able to produce X-rays of my own making (picked up on the CDV-700 geiger counter) without resorting to a pre-made vacuum tube! Again, after a proof-of-concept run I stopped doing this, since the whole point was to prove that I could. Below is a picture of an X-ray producing run using my other HV supply (35,000 volts, from a device I got that was about to be thrown away. That's another story).

(Image credit: Me) If you took a longer video of this tube producing XRays, you would actually see the occasional white speck on the camera, indicating an xray interacting with the sensor. The purple region is where electrons are accelerating across the gap and producing detectable XRays on collision.

There's a ton of other stuff you can do with high voltage that I haven't gotten to yet, including Jacob's ladders (where an arc constantly moves up an expanding gap) and literal fusion reactors that fuse deuterium (although this costs thousands of dollars, hundreds of hours, tons of experience, and requires a much more potent HV supply than the one described in this post. They also don't produce anywhere near enough fusion energy to break even so don't expect to power your house (or anything at all) with these. Check out fusor.net to learn more about that)

Here's a terrible picture of the final product, a 20,000 Volt high voltage 50-watt power supply: (Again, my apologies for the bad images, the actual device is back at my apartment quite far away as I write this)

(Image credit: Me)

And once again, do not attempt to recreate this device. It is extremely dangerous if you don't know what you're doing. Please, be safe. I took all possible safety precautions when building and using this device.

(Image credit: http://www.safetysign.com/products/2301/danger-high-voltage-label)

I hope you learned something and enjoyed this project writeup! If you have any questions, let me know in the comments and I'll do my best to respond. Before anyone asks, yes I was using all necessary safety precautions while using this machine.

Thanks for reading!

Tips/Upvotes appreciated!
NAV: NZLuonFWZTRL8mYQ41SwaP7qopSTbgDQ72
BTC: 1MV7pvR9PGzYBMbSpAyRQbMBonJxZyBnrT
VTC: ViuLmSCuyQXLTJKZphKPwmTvFfvjc88Woz