Archive for the ‘ High voltage ’ Category

Class E Solid State Tesla Coil


As an introduction to the world of solid state coiling, I have chosen to replicate Steve Ward’s class E tesla coil. The design was very simple, yet he seemed to have achieved good results, so I thought it would be a good first solid state coil.



I have always been intrigued by Tesla coils. Last summer I constructed a spark gap Tesla coil from some things I could find around my garage, and a neon sign transformer I got for about $20 at a local shop. While the it worked, the coil was much less than ideal.

The way SGTCs are built just seems very limiting, which made me want to try going solid state. Solid state Tesla coils are much more numerous in design, and the results you can get are just as numerous. There are huge diversities of spark length you can get, and plenty of designs to choose from. For a first attempt at solid state coiling I wanted something a little simpler than a half bridge. I felt Steve Ward’s class e coil was simple enough for my first attempt, so decided to build it.

I should probably explain a little bit about how Class E Tesla coils work. As you have probably heard before, Tesla coils are air cored resonant transformers. Which means when the primary circuit is oscillating at the resonant frequency of the secondary, voltages go through the roof. With spark gap Tesla coils, an LC circuit is setup with the primary to get it to oscillate at the resonant frequency of the secondary. With solid state coils, this oscillation is accomplished using electronics. The trouble is, the smaller the secondary is the higher the resonant frequency is, and with smaller coils, it can be rather difficult to use conventional driver topology to achieve the higher frequencies. To combat this, topology commonly used for radio applications is often utilized. This allows smaller coils to be driven effectively.



Above is the schematic of the coil, but you can also find it on Steve Ward’s website.

I guess the basic design already existed, but I decided I wanted to etch a board. I have never etched before and I thought it was about time I learned to. This circuit was simple enough that I felt comfortable trying it for my first etched board.

I drew up a schematic in eagle relatively easily. I did spend a while trying to untangle air wires on the board before I finally settled on a layout I could live with. I went with a double sided board (although in the future I think I will try to get my boards to be one sided)007

The above picture shows the basic layout. I did find one mistake on the board, and I think it is a problem with eagle’s autorouter. Pin 1 of the FET driver chip should have been grounded (and I did it in the schematic) but it wasn’t on the board. That is an easily correctable mistake though. If you want the Eagle file you can email me at


Here is the printed layout. One of the layers is actually a mirror image so that it will be correctly lined up when placed face down on the board.

I plan to use toner transfer to make this board. The only problem is I don’t actually have a laser printer, So I will need to take the printed layout and get it copied onto photo paper somewhere. I’ll hopefully have a board etched by next week (It’s 7/14/13 now).

I should note that the diodes recomended in the schematic to clamp the antenna got lost (I forgot to order), so I used some point contact germanium diodes I had around. Although a 1N4148 would have worked fine, these are better suited for higher frequency applications, but mostly I think they look cool (the glass case makes them look like little vacuum tubes).


July 19, 2013

A lot has happened. I should start from the beginning.

I had the board design as you can see above, but I do not have a laser printer. Since I planned on using toner transfer as the etching method, this was a problem. Luckily most office stores use laser printers in their copying machines so I just got the design copied onto photo paper with one.

My first attempt to transfer the toner ink failed. Luckily I had two copies made. I was much more successful the second time. After I removed the paper however, I found there were still several spots where there was quite a bit of ink missing. I tried to combat the issue by filling in the areas with sharpie. Just an FYI to anyone who may try this, it is not as reliable as you might think.

After etching the board (with ferric chloride of course) I have to say I am not terribly impressed with toner transfer. Admittedly some of the problems that came about are probably due to my lack of experience, but I still think I am going to try other methods in the future. Even so, I absolutely love etching. It for sure beats the hours spent putting something together on perfboard, and the additional hours afterwards spent finding and fixing mistakes.


Below is the board once populated.


To connect some of the off board parts/larger components, I put some pads on the board to connect to wires.  The three in the top left are for the transformer (although only two are used) , the three in the bottom left are for connections to the primary, and the three next to the bottom left go to the mosfet.

Up to this point in this project, I had kind of forgotten the reason coilers are called coilers. I had up to this point put all my thought and effort into the electronics, and kind of forgot about the coil winding part. But of course it was inevitable that I would would wrap wire around a form for a couple hours. After all, it is a Tesla coil. 

The coil was made from roughly 500 turns of 28 awg wire (8 inches tall) wrapped around a piece of 3 inch PVC (3.5 OD).


I threw together a little jig  made from some books to assist in wrapping the coil. While it wasn’t the sturdiest, it served my purposes fine.


I was pretty happy with the end result, but it wasn’t quite finished. The coil had to be coated in several layers of polyurethane to prevent further turmoil down the road when temperature and humidity begin to change (winter is coming…). If this step is skipped the wire can expand and contract causing it to cross over itself and render the coil useless.021

After varnishing, the coil doesn’t really look all that different, but it definitely is not coming unraveled any time soon.

the primary coil is still in the works. I got a 4 inch PVC coupler as the former for it and plan to wrap a few turns of 12 awg solid core wire around it. How many turns exactly is yet to be determined.

Now comes the fun part. Wiring everything together and seeing if it works. At first, it didn’t. One of my absolute least favorite things is working for a long time on a project and then turning it on for the first time only to have nothing happen.

Trouble shooting was made much easier because of the etched board, but it still took a while to figure out exactly what was going on. And to be honest I still am not quite sure of the cause of those countless anti-climactic moments proceeding the press of the power switch, but alas, I got it working.

What I do know is that there was one bad trace I didn’t catch before, and one of the capacitors had a too low voltage rating to perform its duty (it paid the ultimate price for this flaw). Also, stranded wire seems to do some weird things to the electronics, but that is mostly straightened out now.


While I did get first light, there is a lot of improving to do. First, you’ll notice there is a lot of breakout/corona discharge occurring along the wire to the main breakout point. I plan to shorten the wire and give a more permanent breakout point to help prevent losses in spark length. Second, it would be nice if this didn’t stay sprawled out over my table for much longer than it needs to be, so I will be making a frame to hold everything and make it more aesthetically pleasing, portable, and in general less likely to fall apart when touched.

July 23, 2013

I have been experimenting with how many turns to include on the primary. Steve Ward said a 1.2 turn primary worked best for him, but I found a little more, closer to 1.5 turns, worked best for me. I also played around with the coupling a little bit. It worked best about 1/3 up the secondary, but the primary former I bought wasn’t that tall, so I just moved the primary as high as I could without getting arcs to the secondary.

In the schematic you’ll notice there are two potentiometers (actually trimmers on the board) that control the on and off time of the signal from the 555 chip. Since 555’s output is connected to the enable pin on the gate driver chip, the its signal actually affects the characteristics and size of the arcs. Also, you can hear the frequency from the 555 as the coil operates  since it is effectively turning the coil on and off. You could probably even get the coil to play music by tying the output of a microcontroller to the enable pin of the gate driver, but I didn’t really etch the board to allow for that addtion. Either way, I adjusted the 555 to get the longest arcs. this is a good way to adjust arc size and characteristic without actually tuning the coil.

Arcs to the secondary actually became a bigger problem than I thought. It happened several times when the primary was a little closer to the secondary or didn’t have the former between it and the secondary. Luckily none of the arcs burnt all the way through the insulation on the secondary, so it still works.


To make things look better, and to make the coil take up less space and be more portable I made a little frame to hold it. it is an 8″x8″x4″ box with open sides. the electronics are mounted to the top of the bottom face, and the primary and secondary were mounted to the top of the top face.

I also added a nail to act as a break out point. With just the wire sticking up, there were a lot of other breakout points that took away from the length of the main arc.

015 016

the two faces were made from 1/4″ plywood cut with a chop saw (I had to cut it from both sides), and the legs were made from four 4″ pieces of 1/2″x1/2″ wood (also cut with a chop saw). I mounted the board close to the back to with the trimmers on the edge to allow for adjustment if needed, and put the transformer in front of that. I made sure to put a fan on the front of the board to get circulation across the whole board which had a few other components (rectifiers, regulator and one of the capacitors) that were getting rather warm and either didn’t have, or had very small heat sinks. I added a danger sign just to add to the look, and then I was done.


With optimum tuning and coupling, I managed to get arcs up to 4 inches (to a grounded piece of metal), which is pretty close to the 5 inches Steve Ward managed to get.005 006

I also got a CFL bulb to light up from 1-2 feet away.

Aside from that, I managed to make a small coil that will be easy to take places for demonstrations if needed, and will be very quick (almost plug and play) to setup if I ever want to show anybody which is pretty nice.



I managed to get good results from this coil despite it being my first solid state coiling project. It worked very reliably and is easy to adjust. (I also learned to etch circuit boards)

I am very happy with the outcome of this project, and I would say it would make an excellent first SS coiling project, or even first Tesla coil in general.

Ignition Coil Driver


Ignition coils can produce up to 60 kv AC. As an amateur scientist, I have endless uses for such a power supply. This is my design for a driver of such a coil, to be used in numerous future projects.











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.