TESLA COIL CONSTRUCTION DIARY

I first found out about tesla coils when I was about twelve or thirteen. I was reading a book about some guys who built the early mainframe computers back in the 60s. Of one of the characters, the book said, he came from an eccentric family, the kind who would more probably be "gathered around a Tesla coil than a television" of an evening. I didn’t know what a Tesla coil was, but the fact that it was more fun than TV stuck in my mind. Eventually I went to university where I got unlimited internet access. One day, must’ve been about four years ago, I got a notion to type "Tesla coil" into a search engine. What I found was incredible. It wasn’t so much the big guns like Greg Leyh’s Electrum, or the Tesla Technology 13M, though they were certainly impressive. Rather, it was the sheer enthusiasm and have-a-go spirit of the average Tesla coiler. The pictures told the story: In living rooms and garages across the world, bolts of blue lightning were shooting forth from inflatable pool toys, pieces of PVC drain pipe, timber, toilet ballcocks, microwave oven parts, and beer bottles immersed in buckets of brine. White-hot discharges seared through apartment ceilings and set fire to furniture on the floor above. In a back lot somewhere in Texas, a man cowered under a ladder wrapped in chicken-wire while his "friend" unleashed a 13-kilowatt electrical storm on him. And my personal favourite: With its bank of industrial triodes glowing furiously, a large vacuum-tube coil filled its creator’s hallway to the ceiling with a sheet of sparkling blue-golden flame. Yep! Tesla coils were definitely more fun than TV.

About a year after that, I was visiting a friend who was into electronics and I mentioned that I wanted to build a TC. He went into his garage and came back with a big metal box with a lightning flash on the lid. "You can borrow my neon sign transformer… Don’t burn it out…" Yikes! When I got home, I had to "test" that transformer. I took it down to the basement, where I rigged up a Jacob’s Ladder from two bits of metal rod gripped between bits of wood in a vice. I attached the H.T. wires and threw the switch. Bzzzpppp… bzzzppppp… OK it appears to be functional. Whoa, the wood is on fire! Satisfied, I put the transformer on a shelf where it lay for ages. In fact I moved into a tiny apartment and lost the workshop facilities.

It was now one year since I got my transformer, and still no sign of a tesla coil. I had forgotten about the whole business. Then, in the university library I found a biography of Tesla. Reading that got me determined again. I decided I would tackle the most difficult part first: the secondary coil. I checked out other people’s coils on the internet, and subscribed to the tesla coil mailing list, to pick up some ideas. It seemed the important thing was to get the secondary the proper size, determined by the power input. Too small, and it would break down and burn up. Too big, and the sparks would look wussy. I had a 10kV/25mA neon sign transformer, which I reckoned would be good for about 200-250 watts. Most folk were using a secondary coil about 15"-18" tall at that power level. The recommended aspect ratio was about 3 or 4:1. So I was looking for a tube 18" long, about 5" diameter, and made of some RF-friendly insulating material. I managed to "find" a giant plexiglass tube that was pretty much ideal for the job. If anything it was a bit too tall and thin (20" x 3.5") but hey, room for expansion!

Now I had the tube, there was just the matter of getting thousands of turns of really thin wire onto it. 800 to 1200 was the recommended number. I calculated the right thickness of wire to give about 1000 turns, and bought a reel of that and some "Insulating Varnish". I made some wooden end cheeks to hold the tube, and put a length of threaded rod through the whole lot to hold it together. One end of the rod I chucked into a small cordless drill, the other I put through a hole in an angle iron bolted to the bench. Here I made my first mistake. I thought I would be clever and coat the tube in varnish first, then wind the wire on while the varnish was still tacky. That way, I figured, the wire would be totally encased in varnish for best insulation. The spinning tube splattered varnish everywhere, and once the wire was wound on, it stuck and couldn’t be undone if I made a mistake. By the time I was finished I was covered in varnish and the secondary looked like it had been wound by the banjo-playing kid from Deliverance. To add the finishing touch, I dropped the tube and cracked it.

The coil and transformer lay around for another year gathering dust, while I worked on other projects, like uni stuff, and trying to find a drummer for my band. So it stayed until last month, when the band split, and the lab work at uni finished. While writing up my thesis I researched the beginnings of electrical power transmission. The first ever AC power line was between Niagara Falls and Buffalo, and it was designed by… Nikola Tesla! Him again! In the recesses of my brain, something began to stir. I remembered that somewhere in my parents’ house was a metal salad bowl, of a smooth, rounded, oval shape. I asked my mother: "Oh that horrible thing, you can have it." Another part of the tesla coil jigsaw had fallen into place.

I cleared the floor and put my R.F. signal generator in the middle of the room. On top I balanced the secondary, crowned by its salad bowl topload. I plugged the generator into the bottom of the secondary and swept it, monitoring the output on a scope. I found the notch where it was resonating, and fine tuned it, lying on the floor to minimise capacitance. Then I plugged in a frequency counter: 268 kHz. All I had to do now was find the tank capacitance, and I could design the primary coil.

I dug out the NST, and did some tests on it to find its equivalent circuit. I wanted to know the leakage inductance so I could resonate it with the tank capacitor for maximum power. I measured the short-circuit current (25mA) and the open-circuit voltage. I didn’t have a meter that could withstand 10kV, so I plugged the H.T. winding into 240 volts and measured the voltage on the L.T. side. I calculated that the open-circuit voltage would be 10kV. D’oh, now I know why they called it a 10/25 transformer. I also measured the secondary DC resistance (30kOhm) From these measurements I calculated that the leakage inductance was around 1,000 Henries. The capacitor to resonate that at 50 Hertz was 9.4nF.

I heard some bad things about resonant operation, so I put together a Pspice model. I found that the voltage would rise to 20kV on the first half cycle, 30kV on the second, and so on. So if the spark gap doesn’t fire, the voltage will rise until something blows. The transformer was designed for 10kV RMS which is 14kV peak. If I set my spark gap at a bit less than 20kV then there should be nothing to worry about. With 100 bps, and a 9.4nF cap charged to 20kV, the power input to the primary is 200 watts. I checked out the Farnell catalogue for caps. They were selling yellow cylinder-shaped ones, 47nF @ 1.5kV, pulse rated 1000V/us. I looked on the net and found that some coilers in the UK had used these before with good results. 15 in series would give me 3.1333…nF @ 22.5kV, and three of that in parallel would give 9.4nF, using 45 caps. I ordered a pile of caps and 10-meg high-voltage bleeder resistors.

Back down in the basement, I cut out two plexiglass disks and epoxied them into the ends of my secondary, to seal it. Then I prepared two X-shaped bits of M.D.F. and epoxied them into each end of the secondary. Now I could screw the coil to a baseboard, and screw the salad bowl to the top. Where would it end? Would I be able to re-animate dead tissue soon? Not without a spark gap. I thought about the RQ design made of ½" copper pipe but I reckoned it was overkill for my 200 watt coil. Rotary gaps were out, because I didn’t have a synchronous motor. Well strictly speaking I did, but it would mean trashing my reel-to-reel tape deck. What could I make a spark gap from? I got some pennies and drilled holes in the middle. No, I wasn’t trying to forge Chinese currency. British pennies are made of solid copper, and I reckoned they would make a fine mini-RQ gap. I mounted them in a row with .025" gaps, and zapped them with the NST. At 10kV RMS I was just lighting 5 gaps in series. That makes 2.8kV (peak) per gap, so 7 gaps would fire at 20kV. In theory… Better make it 6 for safety.

Last weekend, the weather was nice and sunny (if somewhat cold) so I sat out in my folks’ garden assembling the MMC capacitor. I got a bit of fibreboard that was perforated on a .25" grid and messed around until I found a high-voltage friendly way of putting 45 caps onto it. I finally succeeded by piling them up two rows deep. The 45 bleeder resistors fitted easily underneath. I made busbars out of copper clad PCB and to finish off I fitted binding post terminals.

Next job was a frame to hold all the parts, and the primary coil. I made the frame out of three bits of white particle board held together with woodscrews. I put plastic threaded standoffs on it (by drilling holes and epoxying them in) to mount the capacitor and spark gap. I fitted the gap with four mini cooling fans while I was at it. Then I made four formers for the primary, out of plastic chopping boards bought from Woolworths. I chose chopping boards because they are usually made from polythene which is a great R.F. insulator. Unfortunately mine were cheaply made and the plastic tended to split apart as it was drilled and cut. I fixed the formers to the stand, then started to make the primary coil.

For the primary I used some copper pipe I got from the university scrap bin. I think it was hydraulic tubing of some kind, quite small at only 4.7mm O.D. That would let me wind the primary on a closer pitch. This was important because I was going to end up needing a lot of primary turns. I designed it this way on purpose because I heard that coils with more primary turns have less spark gap losses. I used a spreadsheet to calculate the archimedean spiral inductance formula I got from Matt Behrend’s website. After a bit of messing around I got a design that had a large enough I.D. to clear the secondary, and was not too big. I ended up using 13 turns on a 10mm (0.4") pitch. I gave it a slight conical angle and lined up the inside turn with the bottom of the secondary. I started by soldering a wire into the inside end of the pipe. Nothing but the best, I used some Shark speaker cable that I had lying around. Then I carefully bent each turn to shape and clicked it into the formers. All went well until turn 10, when I ran out of pipe. Luckily I managed to get more of the same and join it by soldering. I added a strike rail made of some larger copper tube, maybe a bit optimistic considering the power level I would be running at, but it looked neat.

The beast was nearly complete. I just had the wiring to do. I used more speaker cable to hook up the rest of the primary, and a fuse holder for the primary tap. I decided to use an odd number of pennies in my gap and ground the middle one to the mains earth, in the hope of protecting the NST. I also made ground leads to connect secondary base and strike rail together, and a flying lead with a big alligator clip to hook it all to R.F. ground. (I don’t like hooking the secondary to the primary: I think it would send RF back into the NST and cook it) Hooking this all up took a whole evening, but finally it stood ready on the workbench, topload just clearing the low ceiling. (OK so I made it a bit too tall.) I let out a Frankenstein-like evil genius cackle: The very next day I would harness the forces of lightning! If it worked…

I rose early the next morning armed with R.F. generator and scope. I got the coil and sat it on some bits of wire mesh to act as a groundplane. I clipped the ground lead onto these. Then I hooked the generator up to the secondary and tuned it till I found resonance. Next, with the secondary disconnected, the gap shorted and the generator connected across the tank cap, I tried to get the primary to resonate at the same frequency. I couldn’t detect any obvious peak but the signal seemed to get slightly larger in the neighbourhood of turn 12, so I left the tap there. I was going to hook up to a variac and power it on gradually, but my parents stopped by to see some Tesla action before they left for work. It was full power or nothing…

I hooked up my cooling fans to a cordless drill battery. They spun up with a little whining sound. I plugged the tesla coil into one end of a long extension cord, and retired to the other end of the basement. I was getting pretty nervous. What if something went wrong? The capacitors could explode, the spark gap could fall apart and blow the NST. And even supposing it worked, what would it be like? I had never seen a tesla coil working before, only pictures. Everybody stand clear! I inserted the plug and threw the switch.

BBBBRRAAATTTTTTTT…. Holy twisted metal Batman… It worked! Dim streamers of purple light danced around the salad bowl! The air was rent with an evil tearing racket! There were shouts of astonishment from my folks! Suddenly the capacitor bank started arcing, white-hot power arcs right into the particle board stand. But nothing blew up or burned and the coil carried on firing. I cut the power.

The whole place was reeking of ozone as I went over the circuits for a damage report. Everything was A-1 except for a tiny hole punched in the particle board. The cap bank had arced over to one of the screws that held the frame together, and from there to the NST case, about 0.5" in total. That was a lot of voltage and I was worried about the transformer. But I was also desperate to see my creation in action again. I moved the tap to turn 13, turned the lights out, and hit it again.

This time the noise was even louder. The basement lit up with a dull blue flickering light. The streamers twisted and forked through the air, reaching out further. I think they made it to almost 18"-19" but I certainly wasn’t going near it with a measuring tape. As for touching the streamers, forget it! I shut down for a while to let the air de-stink.

The next time I fired up, I had installed safety gaps on the NST, set to .3" per side. I took a photo of the coil working, and I also tried lighting some neon tubes. I put the tube (an old part of a neon beer sign) on top of the topload and zapped it. Sparks came out both ends of the tube and it lit up dimly. I also tried putting it on a support about 8" away with the end of the tube facing the topload. The coil generated a single big blue-white streamer that hit the tube and lit it. Bizarrely, after that I couldn’t get it to light again. The ozone was getting too stinky by this point and I had to pack it in.

So there you have it. The culmination of three years of daydreaming, looking at Tesla websites when I should have been working, and idle boasts to friends and colleagues. In the end, it only took about three days’ work to put it together.