Van de Graaff Generator

About a year ago I attempted to build a Van de Graaff generator. Typically back then if I wanted to build something i would just gather what I could find for materials and then just wing it until I got something I could live with. for a lot of the things I made this ended up working okay, but it did not at all work for the VDG. There were numerous problems with belt tension, roller alignment, and motor power. After a while I simply gave up on the idea and moved on to other things.

Recently partially because of my work on my science fair project, I decided I was going to make a Van De Graaff Generator, but the right way this time. In the past any of my builds involving any kind of moving part hasn’t ended well, so I wasn’t too confident in my ability to successfully build anything mechanical going into this project. I had much more experience with designing, using tools, and putting things together in general this time, so as predicted the build went much more smoothly and was a pretty good success.


I designed the VDG to be built entirely from off the shelf parts since I have no way of machining anything myself. Luckily the VDG is very simple, so this task was pretty easy.

A VDG works by utilizing a difference in triboelectric properties between the two rollers, to get a charge on a rubber belt which carries charge to and from a metal sphere at the top where the charge can build to enormous amounts. The charge is transferred to and from the belt by two metal combs or brushes at the bottom of the and top of the belt.

The first thing I made for the VDG was the rollers. I started out using some PVC end caps, but later changed the design to use bed rollers instead, because they were the perfect size and shape. To allow the rollers to spin in the VDG I attached skate bearings using a threaded rod and some nuts. I attached a pulley in a similar way to the bottom roller so It could be coupled to the motor later on.

The rollers have to be triboelectrically different, so I covered the bottom roller with Teflon tape, and the top roller with aluminium tape. Teflon and aluminum are at opposite ends of the triboelectric series, so this worked very well. The orientation of the rollers determines the polarity of the output, and with teflon on the bottom and aluminum on the top this VDG is negative.


The next parts I made were the brushes. I made them by soldering straight pins on some sheet copper at regular 1/8 inch intervals. these brushes worked very well.


Next I had to make the body of the generator. For this I used a piece of three inch diameter PVC pipe. I needed the rollers to be separated by about 16 inches so the belt would be stretched to about 18 inches when installed. I settled on making three holes on the bottom, two for the roller, and one to help with the installation of the brushes. For the top I decided I would drill two holes for the roller, but then cut the PVC about halfway up these holes which would leave two arcs in the PVC for the roller to sit in, but would make the belt much easier to install. To hold the sphere on the top of the body I bent a coat hanger into a V shape and glued the two legs into two small holes I drilled in the top of the PVC. The sphere would get held up by the top of the wire V.



Next was the belt. I would have tried to make the belts myself if I could find any fitness bands that weren’t non-latex. In the end I just ordered a  latex rubber VDG belt from Edmund Scientifics. Installing the belt was pretty straight forward. I had to do some disassembling of the bottom roller to get the bed roller part inside the PVC and the belt around it, but once I had that done I just had to pull it through the tube and put it around the top roller. The belt tension albeit not much, was enough to hold the top roller in place quite nicely.



To position the brushes, I nailed one brush to a small platform of scrap would and screwed into place under the bottom roller. the top roller was attached to the wire V I used to hold up the sphere.

The last thing I had to do was assemble the drive system. I was originally going to use an old drill to power the generator, but decided to use a surplus AC motor instead. The AC motor worked very well. To step up the speed a little I decided to attach a large pulley  to the motor and a smaller one to the bottom roller and use a belt to attach the two. \The ratio between the circumferences of the pulleys ended up being about 2:1 so the roller was spinning about twice as fast as the motor. This setup worked quite well.


To make the Sphere I used two stainless steel mixing bowls. I cut out the bottom of one with tinsnips, and taped the two together with electrical tape. I also Added electrical tape to all the sharp edges to reduce corona losses.Image

As of now The generator is putting out very strong arcs in the 4 inch to 6 inch range, but it can put out some weaker arcs that are a little over a foot long!

For more detailed instructions and a video visit where I posted this project. You can search 350kv Van de Graaff Generator or copy this address:


ZVS Driver

Last summer I encountered some videos of the ZVS driver online. I had derped around with single transistor flyback drivers, but it wasn’t anything too serious. The ZVS on the other hand was promised to be the most efficient driver around, and the videos of it only supported this. I had kind of forgotten about this driver until I was browsing Mouser one day looking for some Mosfets, and thought I recognized them from somewhere. After a few google searches I found that I had indeed seen them before, on the Mazzilli ZVS driver schematic. I quickly looked around to find the other necessary components  and found that I could get enough components to make two of the drivers for under $30! After knowing that I couldn’t possibly keep myself from this wonder of electrical engineering any longer. I ordered the parts and got building. Before long I had my own ZVS driver.

001 002 004

I made sure to use thick wire for all the connections because up to 10A could be going through the circuit. I wound the 5+5 primary with 18awg enameled wire and it has worked fine. I also put a rubber pad between the wood and the bored as well as adding a power switch and some 12 awg wire for connection to the batteries on the power input terminals. I also replaced the inductor, but I have no idea of the value of either inductor that I used. Everything else has stayed pretty much the same and matches the specs recommended on the schematic.Mazzilli ZVS Flyback Driver

The ZVS is a push pull oscillator. In its purist sense it would consist of nothing more than two mosfets a capacitor and an inductor. The driver relies on the fact that no two components are exactly alike and uses small differences in the mosfets’ internal resistances and other properties to start the oscillation. The capacitor prevents the current at the primary from rising to the point where the core saturates, and the inductor prevents AC from getting back the the batteries. Other components are used as you can see though. The fast diodes prevent the fets from turning on when they shouldn’t and going suicidal. The rest of the additional components act as gate protection to make sure excessive voltage does not reach the gate of the mosfets and kill them.

In operation the capacitor and primaries form an LC circuit, so you can easily change the frequency that the driver opperates at by changing the value of the capacitor and/or the number of turns on the primary.

To power the thing, I tried to use 6v lantern batteries at first, but these are not at all designed to handle a 10A current draw. I found that with 6 of them in series the voltage would drop from about 36v to about 8v! This fact made them absolutely useless for this driver.

what you really need is a few sealed lead acid batteries. Although big and relatively expensive they can pump out some serious amps without any trouble, and are absolutely perfect for the ZVS. I started out by ordering two 12v SLABs from ebay, though I plan to get one or two more when I have some more money.004

Despite running at only 24 volts, this driver was putting out some very impressive results. I estimate the voltage is 20kv-25kv, which is decent, but the current had to have been at least 100ma! I was able to draw the arcs which started at a little over an inch to 5-6inches!008 009 010 011

The ZVS is a truly marvelous device, and I can already think of a lot of projects I would like to include it in. The only problem I have had with it so far is that it has a tendency to destroy every capacitor I use with it after a while. I originally had an MKP cap on it, but it inexplicably failed on my second day of using the driver. I bought a mica cap with the necessary ratings as a back up, but it over heated and eventually failed as well. I found another random film cap in an ATX power supply a while ago with pretty good specs so I tried it. It worked surprisingly well for about a week before it too over heated and failed. As it stands I am still looking for a suitable capacitor, but I’ll update you when I find one.

If you attempt to build the ZVS driver, always keep it on an insulating surface, and use a chicken stick! a 2 foot long piece of white (black colored pipe can indicate carbon which can become conductive) PVC should work well. Never touch a live wire of the ZVS while it is on even if it has insulation! Also make sure you always disconnect the driver from the batteries if you are going to touch anything on the driver. Power switches are convenient, but you shouldn’t entrust your life with them as it is easy for them to get turned on by mistake. Also it is just a good idea in general to have multiple checks to make sure a high voltage device is off so you don’t loose your life.

Wine bottle CRT

So some time ago I was browsing youtube when I cam across the most wonderful video I had seen up to that point. Someone had constructed an old school cold cathode electron accelerator (called CRTs or discharge tubes) out of a wine bottle and some stuff from a local hardware store. I made a mental note that I wanted to try it, but when I saw the cost of the necessary vacuum pump I kind of lost interest. a few months later or so I was looking at instructables, when I saw that Daniel Kramnik had posted instructions on how to make it. This rekindled my interest, but I still didn’t have the money to buy a pump. That summer I saved a little money and at last I had one. Also that summer I had gone to a local neon sign shop and picked up a small neon sign transformer for $20! For the vacuum port on the bottle (also acts as the cathode) both of my sources for making this recommended using a concave surface which helps focus the beam. The original video said that they had used a drawer handle and the instructable said to use a mini doorknob. I couldn’t find either at my local hardware store, so I made my own design for the vacuum port. I had previously tried to setup some kind of system for connecting my vacuum pump that could act as the standard for all of my future projects. When I bought the pump it was claimed to have both a 1/4″ and 1/2″ port on the pump. When I got it, it looked more like a 1/4″ and 1/8″ ports. This was not much of a setback however. More frustrating was that I could not find a single flared hose barb fitting at any of my local hardware stores. I have found some that will work on, but they don’t come cheap. It’s about $10 dollars for one of them. After a while I resolved to just tighten the hose I was going to use directly onto one of the ports threads. I had a length of vinyl tubing sitting in my garage that was the right diameter, but its walls were too thin and it collapsed under vacuum. To combat this I found some 1/4″ fuel line that was strong enough to withstand the vacuum, and used a small piece of the vinyl tubing to connect the vacuum pumps port to the fuel line. Hose clamps were used on both connections and lots of teflon tape on the pumps port. For most of my experiments I simply connected the bottle to that one length of tube directly to the pump and it worked surprisingly well. I am currently working on trying to connect a thermocouple gauge tube to get vacuum readings, but am having some problems with leaks  on new connections. For most purposes the single hose connection will be fine. Image This shows how I made the connection with the hose. A flared hose barb will be used in the future. Based on my system for connecting to the pump (using the 1/4″ fuel line) I decided to just buy a whole bunch of hose barbs that fit the fuel line to connect the chamber to the hose. The hose barb size ended up being 1/8″ pipe thread to 1/4″ barb I think. To make the vacuum port on the wine bottle I used one of the hose barbs described above, a 1/8″ pipe nipple, two washers with holes that roughly fit the nipple, and some epoxy (I have tried both five minute epoxy and JB weld successfully). I started by gluing together the two washers. Once that dried I tightly threaded the hose barb onto the nipple (teflon tape is not needed because epoxy will seal it). I then put the nipple through the washers, and glued the bottom of the hose barb down taking care to leave no gaps in the bead of epoxy. I then aded more epoxy around the edges of the hose barb to ensure a good seal. once the assembly was complete I used a rather large amount of epoxy to secure the washers to the top of the bottle once again put more epoxy on the sides to fill in any gaps and get a good seal.ImageImage The vacuum port also acts as the cathode so we will not have to worry about making at least one of the electrodes. The anode however does need to be added. I found a 3/16 (I think) glass/tile spade bit at my local hardware store which was perfect for drilling a hole in the side of the bottle to allow for a wire to be put through to act as the anode. Drilling the hole is pretty strait forward. You just put the bit in your drill apply a little pressure and start drilling slowly and carefully. I found the hole can be a little difficult to start, but putting a piece of masking tape on the spot you want to drill helps keep the bit from wondering away from where you want to drill. You should probably do this before you attach the cathode/ vacuum port, because if the bottle cracks when you are drilling it is trashed and you can’t use it. Once the hole is drilled You just need to slip the wire for your anode in the hole (the shape doesn’t really matter. I’ve put it in a loop and just bent it downwards and it doesn’t really affect performance), and epoxy it in place. I used a thick layer of epoxy on top of the initial seal to help keep the anode wire from breaking from being bent around a lot (my original CRT had to be decommissioned because the anode wire broke off). Image It’s really messy, but it provides a good seal and good strain relief for the wire. The next step is to make a power supply. As a rule of thumb you want a DC power supply that can provide around 10kv. You can use higher but beyond 30kv x-ray production starts to become a concern, so you don’t want to exceed it. You can also use a lower voltage (as I did). I’m not sure what the minimum voltage required to get it to run is, but I would bet you need at least 2kv. I made my power supply by using a 6kv 30ma neon sign transformer with a microwave oven diode to rectify the AC. To find the DC voltage that is output by rectified AC you use the folling formula: Vin (1.414) = Vout. To wire up the CRT I connected one HV terminal of the NST to the cathode of the CRT, and connected the other terminal to the anode of the microwave oven diode. The cathode end of the diode connects the anode of the CRT. The low voltage terminals of the NST get connected to the hot and neutral or mains. Pay attention to the stripe on the diode which indicates the diode’s cathode when wiring up your CRT and always use wire rated for the voltages you are working with. ImageImage All that’s left to do now is fire up your pump, turn on the high voltage and watch the plasma!ImageImageImageImage Nothing happens at atmosphereic pressures because the free mean path is too small, but once your pump has been running for a while plasma begins to form as the accelerated electrons excite the gas (nitrogen) inside the tube. You will notice several features that occur with in the tube which deserve explaining. One of the first features of the plama you will experience is what’s known as a Faraday’s space, and it is the gap in plasma that starts to form at relatively low vacuum, where the collum of plama begin to separate from the cathode. This occurs because there is a higher concentration of electrons near the anode (which electrons are attracted to) than the cathode resulting in more plasma. After pump runs a little longer you will notice the plasma forming stripes. These are called striations and they occur because there are going to differences in the glass (and any other material in the tube) from one part of the tube to another resulting in a higher concentration of electrons, and therefore plasma, in some parts of the tube. These descriptions are probably over simplified, but they get the idea across. If your pump is good enough and your bottle sealed well enough, then after the striations form, the faraday’s space and striations disappear and the bottle gets filled with a uniform light blue plasma. After this happens you may notice another very interesting feature. Starting from the cathode the plasma will disappear, and be replaced with emptiness within the tube. The glass around the emptiness however will fluoresce under the bombardment of electrons, and as the vacuum increases so does this empty space until the amount of gas inside the tube is too low to facilitate cold cathode emission any longer and the tube goes dark (it is unlikely you will reach this low of vacuum with your pump). This is called the Crookes dark space and to me it is one of the most fascinating features that occurs in discharge tubes. This project was a lot of fun, and now I can tell my friends that I have built a particle accelerator! I am far from finished with CRT projects though. In the future I intend to make a CRT oscilloscope from scratch using an electron gun assembly. I also want to try an experiment that uses an old 2 inch CRT along with some other equipment to determine the charge mass ratio of the electron, just like Thomson’s third major CRT experiment. Also for those who are interested I am currently working with some professors from DU to make a more serious and more powerful particle accelerator (atom smasher) to perform experiments I hope to enter in the ISEF. Well that just about does it for this project. I encourage you to try this and have fun with it, but be very careful when dealing with high voltage and vacuum, and make sure you are not producing x-rays! Also You should probably limit the amount of time you leave this running to a minute or so to decrease the chances of something bad happening.

Seeing alpha particles

A few months ago I deccided it was about time for me to actually do an experiment, which I hadn’t done in a while. There was one that I had heard about from many different sources and wanted to try.

Invent Geek as well as numerous videos on Youtube claimed that with just a smoke detector and a webcam you could indirectly see alpha particles. All you had to do was extract the americium 241 from the smoke detector and place it close to the sensor of the camera and you had yourself a device which could detect alpha particles.

Aside from sounding too good to be true, Americium is technically not an exempt source since its proton count exceeds that of uranium. However after seeing the price of other exempt sources online, and seeing how many people bend the rules, I decided it would not be a big deal.

Extraction of the americium from the smoke detector is relatively simple. There were no screws holding the plastic case together so I just pryed it apart with a couple of flat head screw drivers. From there I removed the PCB. The americium is housed an a bulky aluminum case which is held to the board with two solder joints. After desoldering those contacts the case was removed, exposing a plastic base with metal plate in the middle connecting to one pin of an IC. After desoldering the IC pin I broke of the tabs holding the plastic to the board, and carefully pushed the metal button out of the middle of the plastic base. This button has a thin foil of americium 241 on one side of it and that button is what I was after.013 021

Next was the webcam portion. Alpha particles (or a high energy helium nucleus) are very positively charged, and consequently have a habit of ionizing every atom they come across which takes a lot of energy. This means that not only do they have a range of only a few cm in Earth’s atmosphere, but they are also stopped by just about anything. Because of that, they are not able to go through the lens of the camera.

The lens on the camera i used just screwed right off (I used a $5 PC cam I found on amazon), and allowed me easy access to the cmos sensor to place the americium in front of. In fact, the back of the lens acted as a perfect place to mount the button. By simply gluing the americium to the back of the lens, I killed two birds with one stone. I got the sample close to the cam’s sensor with no barriers, and I blocked all the light that would have otherwise come through the lens.

016 017 019

To put the finishing touches on the device I painted an old plastic container black and added that classic identifying symbol for all things radioactive.025

And now of course the most important part, the results. I plugged in the webcam to my computer and watched the screen. I was amazed and excited to be greeted by small flashes of light which are a result of the alpha particles hitting the sensor. I thought for a while the flashes may be the sensor detecting x-rays resulting from Bremsstrahlung, but more likely the particles are pulling electrons from the sensor resulting in an electron hole which causes a flash to occur on the screen. Either way it was a fun and pretty simple project.

I will try to post a video soon, but until then you can see more detailed instructions and a video on an instructable I made about this (I got first prize in their Hack it contest with this). Just go to and search “how to see alpha particles.”


     For as long as I can remember, science has been my biggest fascination. As my fascination has become an endeavor to understand the nature of the world around me firsthand, I have found myself diving deep into a world of DIY devices and jury rigged experiments. Whether I am making Tesla coils, homemade CRTs, or burning lasers, I always find great pleasure in what I do and hope to inspire others to follow suite. At any rate, as my number of completed projects begins to grow I feel more obligated to share what I’ve done so it can be useful for others.Image