Index the 3kW Tesla Coil

Introduction: Introduction of this project and overview.
Construction: Day by day construction details.
Testing: Testing, debugging, improvements and experiments.
Pictures and Videos: Multimedia of the coil in action!
Others: Miscellaneous

Current Project Status:
  Project Completed, Tesla Coil Stable and working!
       - 4 MOT Stack completed and plugged in!
          > Including photos and a video! (27th Jan 05)
  Tesla Coil Completed! (31st October 2004)
  Construction Period (25th July 2004)
  Just Started (18th July 2004)

After the success of my 180W 40mm Tesla Coil, I decided to build a much more powerful Spark Gap Tesla Coil (SGTC), but not being too big to be difficult to transport. This page catalogs the construction and testing of the new Tesla Coil as well as performance pictures and videos. 

My previous Tesla Coil had a 40mm diameter secondary, and ran at some 180W from a 6kV 30mA Neon Sign Transformer (NST) (Previously a 750W 30mA unit, but that one died so I had to change a transformer). My previous experience told me that NSTs were really difficult to find cheaply... at least in this area... and thus I opted for a Microwave Oven Transformer Stack to power the new coil. The secondary diameter was decided to be of a medium 4" diameter and an Asynchronous Rotary Spark Gap (ARSG) would be built for the new coil.

The 180W Tesla Coil was built in November 2003, during the school holidays. I decided instead of concentrating my efforts all into the tesla coil at one, I would build this slow and steadily. Project construction will begin on 18th July 2004. There isn't much time during school term so I'll be working on the tesla coil only when time permits, a bit at a time.

This page will document my progress and I will document the whole construction and debugging process. I'm thankful for the many people and websites which have helped me in the construction of my first tesla coil and thus I would like to contribute too. If this page can help or inspire a single person to build another tesla coil, it is all worth my effort.

Day by Day Construction Index

Today marks the start of the construction of the new Tesla Coil! One of the most important components would be the Power Supply Unit (PSU). This calls for a relatively high voltage power supply to charge up the tank capacitor. As such, a PSU is very important as a Tesla coil could not run without it. Furthermore, other components directly, or indirectly are affected by the PSU. This is why I am building the PSU first.

Power Supply Unit

The Neon Sign Transformer is the ideal choice for a power supply. They are of relatively high voltage (up to 15kV) and the largest I know of is a 15kV 120mA model (These are extremely rare! I've only seen it in a catalog, but have never seen it anywhere in real life or known of anyone who has. More common  ones are 6kV to 15kV at around 30mA). Furthermore, they also have attractive features such as internal current limiting, effortless paralleling (for increased current), and comes in many different voltages and currents. However, for some people (especially for those on a tight budget like me), NSTs might be very difficult to obtain, or find cheaply. As such, an alternative power supply has to be used. Enter the Microwave Oven Transformer (MOT). These are big transformers found in Microwave Ovens. These are much easier to find (in old Microwave Ovens) and are also usually available for free, or at a very low cost. They also provide lots of power. However, unlike the NST, they are usually of a low voltage (typically 2000VAC+) but at a huge current (300-1000mA). 2kV is too low to reliably fire a spark gap, and the huge current will overheat most spark gaps. Furthermore, MOTs are not as well current limited as NSTs and thus may seem undesirable for Tesla Coil use. But these problems can be solved.

A few months ago I came across Greg's Tesla coil page, and in it was described a dual MOT power supply. I found it to be quite suitable forthe coil I would be building and thus I have adapted it for use in my new coil. This circuit uses voltage doublers to increase the  voltage from two MOTs. I modified it slightly to fit the 240V mains here. Below is a schematic I've drawn for the new tesla coil dual MOT power supply.

Here's the plan. For the power supply, I have 2 similar MOTs in parallel from a 240V 15A 50hz outlet. The secondaries are in series, and both cases are grounded. The operation is rather simple. On the first half of the AC cycle, Diodes D1 are forward biased and the doubler capacitors (C2) are charged up. On the next half of the cycle, the current reverses, and the diodes are reverse biased, and behave like an open-circuit. The secondaries of the MOTs are now in series with the charged capacitors (C2). Thus, the sum of all four potentials is now across the diodes, making a 12kVDC (roughly, depending on MOTs) pulse up to several hundred milliamps to the tesla coil tank circuit through the chokes L1. This is a classic voltage doubler circuit. The chokes L1 and bypass capacitors (C3) form an RC low-pass filter to prevent any RF feedback from the tesla coil. The final output is in the form of 12kVDC pulses at 50Hz. Greg has used this circuit without any filter and smoothing, but with excellent results. All parts can be found from two old microwave ovens and are are easy to fit an assemble together. More capacitors (C2) can be added in series to further limit the current draw. Estimated output would be around 3kW, 12kV at 250mA - more than enough power!

Construction on this power supply began today. I mounted the two MOTs side by side on a scrap aluminium sheet I had lying around. The MOTs were then screwed on with self tapping screws. (see above right) The aluminium base also helps for easy ground connection as both MOT cores are connected together. Primary side connections are done. I'll try to fix up the doubler circuit next week, hopefully if school is not too busy. As of now, the dual MOT stack weighs 9.5kg and the cost is $10 (All costs here are in Singapore Dollars; 1.7SGD = 1USD)... bought 2 old microwave ovens from a junk shop. The aluminium plate is free.

This afternoon after school, I went to get some components for the power supply.

On the left you can see all the parts.

1 100m reel of cheap PVC insulated wire (for winding the chokes), $7; 4 5mm brass screws with 2 nuts and 2 washers each, for connections ($1); 20mm PVC pipe with end caps (for diode stick) and 2 3" PVC end plugs ($3 total); 6 3kV 0.01uF Ceramic Capacitors (spare), 30 1N4007 Silicon Diodes (15c each, I had some already), and some spade connectors. Total amount spent today was $12.50, which makes the total cost $22.50 so far..

I constructed the diode stick with the bypass filtering ceramics too. The 6 ceramic capacitors were soldered in series (giving a 18kV 1.67nF cap, the values are not calculated, they just seemed like a good idea..), and were inserted into the PVC tube. 24 1N4007 Silicon Diodes were soldered in series in one long string, and coiled around the PVC. Electrically, the diodes and caps are in parallel. The only purpose of the ceramic disk caps is to allow any RF leakage from the Tesla coil to bypass the diodes, possibly sparing them from RF destruction. I will begin building the pancake type chokes soon.

I had more time to work on the power supply unit today, and I completed the pancake chokes. These chokes are used for RF filtering. Instead of buying power wire-wound resistors, output chokes were wound instead. I used a similar idea as Greg's Dual MOT power supply, and wound my own 'Pancake chokes'. I used 3" PVC end-caps, which are used to cap off 3" PVC pipes. I drilled holes though the centres and put a brass machine screw through it. I then cut out a round plastic piece from 3mm clear acrylic, so that it just fits inside the end-cap, then drilled a hole through the centre. I soldered the 5 stranded PVC insulated wire to a brass nut, and screwed it inside the end-cap. I chucked the screw into a drill, and spun the wire on, and put everything together. The wires comes out from a small hole at the side of the PVC end-cap. It's difficult to really explain the construction so below is a diagram of how everything goes together. This results in a nice little package, and the whole thing weighs slightly more then 200grams. The wire was just scramble wound onto the brass screw.

The first photo is the completed pancake choke. The clear acrylic allows see-through, and looks nice, however any insulating material like other plastics would work well too. I happened to use acrylic since I had some spares left lying around.

The second photo is the components of the power supply. There are 2 MOTs, 2 Microwave oven Capacitors (2100VAC, 1uF each), the diode stick, and the two pancake chokes. The only thing left is to make the safety gap, and the stick everything down on that 1/2" thick medium density board, although I might change to a wooden board because its cheaper :).

The power supply is finally complete! (photo below)

[On the left is a top view of the setup.] Yesterday, I sawed out a 30.5 x 33 x 0.9 cm plywood board, and sprayed it bright red (it's fluorescent too), so it was time to finish up the whole power supply. I bought some #4 1/2" tapping screws (costing 50cents, making the whole cost to be $23). 7 holes (3mm) were drilled into the aluminum plate (at the sides and corners) and the dual MOT pack was screwed on to the wooden board. It's strong enough to hold the 10kgs of MOTs even upside down. A power cord was then connected to the MOT pack.

The capacitors are free standing, but held down by a nylon fishing line (connected to two screws by the side, like a tent) and the blue piece of plastic is to separate the capacitors and keep the fishing line under tension to hold the caps down. This worked rather well.

The diode stick is held above the pancake chokes by two L-shaped brass supports I cut out from a spare brass sheet. This saves space too so the footprint of the power supply will not be that big. See the first picture on the right.

The spark gap (right-most picture) is made from two L-shaped copper supports cut out from a spare copper sheet, and two machine screws act as adjustable electrodes. They are currently galvanized steel, but I may change them to brass once I buy more of them. The pancake chokes are not exactly held Down, but are held In Place by 4 screws screwed down beside them to keep them from shifting around. (More visible in the top picture of the power supply unit). The whole setup is quite heavy and weighs slightly more than 10 kg.

I plugged the power supply into a 15A 240V outlet.

The voltage multiplier certainly works. The spark gap is set at almost 10mm, yet it arcs across. Since the output is pulsed DC, you can distinctively hear the 50hz pulse from the output arc. The arcs are similar to that from arcs drawn from a single un-ballasted MOT, but this time it is much hotter and brighter, although not as smooth because it is pulsed DC.

After the completion of the power supply, my next plan is to build the ARSG, though I need to find a suitable material for the rotor...

I bought more parts today.

I got a slightly used (in fact, it looks rather new!) 500W 11,000rpm angle grinder for only $20. (I have to thank the shopkeeper for selling it to me so cheap). This would be used in my Asynchronous Rotary Spark Gap (ARSG). The high speed allows for very high break rates, but the speed can be controlled with a variac.

I bought other brass screws, wood screws and also a U bracket and a bolt for the rest of the project. Since I had some time, I decided to start working on the rotary gap...

Several tesla coilers have successfully built the ARSG using an angle grinder. However, I had difficulties obtaining a suitable rotor material. Good materials would be FR-4 Fiberglass (G-10 Garolite) (unclad printed circuit board), Tufnol, phenolic... which have high tensile strength, and work well under temperature. However, I couldn't obtain any, so I'm using chopping board material. I'm not sure what exactly it is, but it's probably Ultra-High Molecular Weight Polyethylene (UHMW PE). I cut a 13cm dia rotor (small, so there will be less forces), with 4 electrodes (basically 5mm brass bolts and nuts) 2cm from the edge of the rotor. The center hole is 10mm to fit the arbor. UHMW PE is easy to work with and I built it in a short time. However, I ran out of brass bolts so I have to get more. I do not have much complicated equipment to fabricate the rotor... I cut out the rotor blank with a jigsaw, marked the centre holes and electrode holes with a sharp point, and drilled the holes. I attached the cut out rotor blank and pressed it against a metal file as it spun. There is no noticeable vibration at that speed, so I guess it worked. I tested by screwing two brass bolts (only had 2 left) to test if the UHMW PE would hold, and it did very well.

However, there are problems. UHMW PE's working temperature is less than 220F (100C), which is very bad. I am worried that the brass electrodes would get too hot, soften the plastic and send the electrodes flying out at high velocities... I'll continue looking for better materials.

All parts for the ARSG have been obtained.

Many brass bolts, screws etc have been obtained, costing $10 total. Brass is not cheap... the M8 x 4" bolts are very expensive at $2 each. A 8mm drill bit was also bought. I was very lucky and found a suitable material for the rotor today ($5), making the total cost $60.50.

This material is bakelite (7mm thick). Bakelite, the brand name of a versatile, heat-resistant resin called polyoxybenzylmethylenglycolanhydride, and is produced by combining carbolic acid (or coal tar or phenol) and formaldehyde. It has excellent insulating and heat-resistant properties. Bakelite is also known as phenolic, and I have seen many rotary spark gaps made from this same material.

A 15.5cm diameter disc was cut out using a jigsaw and a 10mm hole was drilled in the centre (to be attached to the angle grinder arbor). 4 5mm holes were drilled 2cm from the edges, and brass screws were used as the electrodes. For the fixed electrodes, 2 copper L-shaped brackets were made to hold the adjustable electrodes (in this case, M8 4" brass bolts). They are so long so I can feed them as they erode away.

In case you were wondering, I balanced the rotor but attaching it to the grinder and spinning it against a coarse grinding stone. There is no noticeable vibration when this thing is running at full power so I guess this worked. The stand is made from plywood (spray painted bright orange to look better). I cut a slot in the wood to accommodate the rotor, and the angle grinder is fastened by a U-bolt and another bolt screwed into the threaded socket on the side of the grinder head intended for an extra grip handle. This results in a good 3 point mount. More work will continue on the stand tomorrow.

I decided to complete the whole ARSG today.

I sawed out the supports (from plywood) and tap screwed them together. The electrodes are brass and are screwed on to copper brackets I made from some copper sheet. Multi-stranded copper wire connects another copper bracket to the main electrode bracket, whereby the output wires can be screwed onto. The power supply and ARSG are now completed - 1 step closer to completion!

Due to the National Day celebrations in school today, we were dismissed at 10am. I proceeded to acquire more components, mainly for construction of the secondary coil.

I bought a 26" long PVC pipe, with an inner diameter of around 4", and an outer diameter of 110mm, as well as two end caps. ($6.80). I plan to construct a grounding strip on the secondary, and the stainless steel ring ($2) would clamp it down. Other parts include a scrap piece of brass, nylon bolts (there should not be any conductive parts inside the secondary coil) and other brass nuts and washers for affixing the secondary to the platform and the toroid ($2.90), and 1 liter of clear gloss polyurethane varnish ($11). $22.7 spent. This brings the total amount to $83.20...

Today I gave the secondary a coat of varnish. As you can see, I connected a drill to the coil form, and spun the pipe as I applied the varnish. It's a crude setup, but just works. The speed of the drill is varied by a variac. I need to make a better setup however.

For the wire, I plan to use 0.5mm wire (AWG 24), wound for 55cm, making 1100 turns. However, I have yet to find a source of r the wire... Anyway, the polyurethane takes 24 hours to dry completely.

Last Saturday (7th Aug), I located a good source for the magnet wire. Look at the photo on the left. Notice a small reel of wire, a large reel of wire and a Microwave Oven Transformer. For my Tesla Coil 1, I used 0.2mm wire, bought from RS for $31.10 (about 18USD)  per 500g (about 1600m) - quite expensive. I decided to try my luck on yellow pages and search for motor rewind companies. I found a small shop in some industrial area, and they indeed had cheap magnet wire.

 I bought a half used reel of 0.5mm wire (smallest reel they had.. new reels are 10kg). This reel contained 3.8kg of 0.5mm wire! Since my secondary is 110mm dia, I would need around 350m (around 1000 turns) of wire, which means I only need around 600+g of wire. 3.8kg is enough to wind 6 coils. However, I did get the wire cheaply, at only $10 (around $5.80USD) per kg. I paid $38 for the whole reel. Should I have bought from RS, things would be different. RS sells the wire at $25.90 for 250m, which means I would have spent $36.26 for the wire I need. Saved $30! I will definitely go back to motor rewind shops to buy wire. About $6 of wire was used (to be counted in the cost for this project).

Lets move to the secondary now. I made a mistake in putting too much varnish over it. Even after 1 week, it had to dried fully and there were many blobs all around. This made the whole pipe useless. I went out the the nearby shop and got another 27" of 4" pipe. After sanding to remove all marks, washing it, and drying it under the hot sun, I started winding the coil.

I used the same winding technique as for my previous coil which used 0.2mm wire. This is a 'poor man's winding jig'! The wire reel and secondary is placed between two chairs, and everything is slowly wound entirely by hand. This coil too around 1 and a half hours of boring winding to complete. Notice how nice and tight the turns are - it looks like a copper tube! I wound for 55cm length. There should be around 1000+ turns of 0.5 wire. I hooked it up to my drill with a variac, and spun it slowly as I applied a THIN coat (learnt from previous mistake) of clear gloss polyurethane varnish, and let it spin for another 45 mins until the varnish was tacky and could not drip. Upon close inspection, I realized that a lot of bristles from the varnish brush had came off and were stuck on the secondary. Using a thin wire, I painstakingly picked out the bristles, wasting another half an hour of my time. Next time I'll probably use a small paint roller instead... also, I need to redesign the spinning mechanism. At such slow speeds, the drill is very unstable, and stalls easily. Spin too fast and all the varnish files off. Anyway, the coil is looking great and shiny :-).

On the left is a simple to-scale drawing of the primary supports, the secondary and the top load.

I completed the secondary section today. Since the windings and the coats of polyurethane were done, I proceeded to make the secondary ground. A rectangular section of an end cap was marked out and scored with a penknife. A grove was cut using a Dremel rotary tool and a cutting disc for the wire. A rectangular sheet of copper was then cut out. The base wire of the secondary coil was then hammered flat using a hammer (copper is soft; insulation scrapped off first), and it was soldered to the copper sheet. The ground sheet was then epoxied in place. This grounding terminal is very effective. A grounding strap can be strapped to this copper terminal, and is easy to attach or remove. 2 long M10 plastic bolts were screwed in the end caps, and both end caps were epoxied on the pipe. The large plastics bolts are much better as they are much larger. Both of them cost $2 together. To see how this will be attached, look at the drawing above. It's difficult to explain how it will all fit together. As the project goes along, you'll see how everything will be fixed together. The primary is meant to be adjustable too so I can change the coupling.

Total cost of the project is now S$91.20.

Due to school being irritatingly busy and stressful, it was difficult to find time to work on the tesla coil, until now. It's a 1 week school September holidays. I decided to start making the adjustable primary supports. A 2 foot diameter circle was cut from 1/2" plywood using a jigsaw. (I got the wood free from some construction site nearby which junked lots of wooden boards) A hole was cut in the middle of the circle (where the secondary will be) and 6 slots were cut for easy tapping of the primary from below. 1 litre of Nippon Bodelac Wood/Metal paint and a brush was bought ($14). I choose Blue Marines because it seemed to be the best colour. The primary deck was given two coats of paint. The result is rather nice. I'll be buying 50' of copper tubing tomorrow (hopefully) and then construction of the primary coil can commence!

Total cost of the project is now $105.20

Several more parts were bought today.

50 Feet of 1/4" Flexible Copper Tubing for the primary coil ($20), 3 metres worth of 3/4" PVC pipe for the primary supports (3 metres was the shortest length they have, but it's cheap for only $1), and many bolts and accessories - 6 big M10 Stainless Steel bolts for the adjustable primary, along with many nuts, washers and lock washers; 20 smaller stainless steel bolts for wire connections and M10 brass nuts and washers for the secondary top attachment to the toriod. I probably bought too many, but these can be used next time for other projects. The stainless steel stuff are not cheap, total cost being $16. Since the primary deck paint was still wet at the edges (due to an unskillful application of too thick a layer of paint...), construction is delayed. I cut a grove at the top of the secondary for the wire to wind up to the topload. Everything is looking okay, but progress is painfully slow.

Total cost of the project is now S$142.20

Today I completed the primary coil. It has slightly less than 10.5 turns, and is made of 1/4 inch diameter soft copper tubing wound in a flat spiral, with 1/2 inch separating each turn. The insulating supports are 23.5cm pieces of 3/4" PVC pipes. They are tied down at the ends by cable ties. (photo 1). In photo 2, you can see the lovely 50 foot roll of copper tubing. Around 120+ 5mm dia holes were then drilled beside the PVC supports (not very well done by it works..). The cable ties go through the holes, over the copper tubing (which rests on the PVC supports) and back down through the other hole and firmly clasps the tubing to the primary deck. (Photo 4 and 5). The coil is held in place by over 60 cable ties. I might even use 1 more per support to make it even stronger. Compare the size of it with my mini 40mm 180W tesla coil (photo 6). The finished coil is kept in a plastic bag to protect the copper from oxidising.

This is an extremely simple method for winding a flat primary, and yields a very sturdy finished product. I thought it might look quite messy, but it's actually quite neat. It requires no complicated machining, notching or precision work. Only common hand tools were used to make it... a drill, a ruler and a marker. The main drawback is that it is rather time consuming. I took around 2 - 3 hours to complete the primary. Although easy, it was very tedious lacing and clinching the many cable ties. Many people report all sorts of nightmarish tales of primary construction, including turns that come loose and don't stay, plastics that are impossible to glue, difficult to cut precisely etc.. This is very different with this method. Easy, but tedious and rather boring. Here's how I marked out the pipes for the copper tubing.

Although the top looks relatively nice, the bottom is chaos. There are many things that can be learnt. Firstly, cheap plywood is lousy to work with. Everyone knows that when you drill plywood fast, the bottom ply tends to crack off. This is exactly what happened in this and the whole bottom looks like cracked wood. I'm not so concerned about it because it will be covered by another deck which this primary deck is resting on, but it still looks ugly. Secondly, don't apply too thick a coat of paint and make sure it doesn't form drips or blobs because it takes an exceedingly long time to dry. Next time, if I make something like this, I'll probably be using PVC sheets all the way. For now, I'll still stick to wood and paint. In any case, I still got the wood free, and I don't want to waste the paint.

This primary coil design is by Greg Hunter, who uses it on his 4" Junk box coil and 6" coil. http://www.hot-streamer.com/greg/. Credit must go to him for such a innovative and good design.

School has been *extremely* busy, so I had no time to work on this coil at all. However, the end-of-year examinations are just over, and the holidays are coming, so I will be able to resume work soon, and hopefully complete the coil in early November.

I went out the the shop and got some 1.5" Nylon wheels and some stainless steel Pan Self Tapping Screws and washers to attach the wheels (base will be wood). Total cost is S$12.80 bringing the project cost to S$155.00. I was planning to cut the wood to make the box to house everything, but to my horror, the plywood I had was either too small or was disintegrating. This made me unable to carry on building. I shall buy some new wooden boards sometime this week.

The coil is actually mostly complete. I need to build the toroid (5x20" out of PVC ducting covered in aluminium tape), the capacitor bank (will be using a MMC, though the caps won't be cheap), the frame to hold everything together, the wiring, safety circuit and finally to paint it. I expect to complete it in two weeks if everything goes well. I shall order the MMC capacitors soon.

Yesterday was the last day of school. It's the November/December Holidays. I went out and got everything else I needed to get for my tesla coil.

Refer to the above photos. The two 3' x 2' x 11mm Plywood boards ($6 each) for the box to hold everything together. Supports will be made out of the 40mm PVC pipe ($2.50) and held together by the threaded rods ($2 for 2m, will be cut). The Toroid will be made out of the 5" Inner Diameter PVC flexible ducting ($24 for 1.5m). The 2' x 1' x 4mm Clear acrylic (covered with brown paper to protect) ($10) is for the capacitor bank. For the primary wiring, I will be using AWG8 Ed Acoustic Super Oxygen Free Copper Interconnect  High Current Cable ($7.50 for 3m), with the suitable cable lugs ($4 for 16) and clear heat-shrink ($2) to keep it nice. For the primary tapping, I will be using either the large crocodile clip (90c) or the fuse holders (60c). For connections, I will be using either the wire lugs or the large industrial sized (these are huge compared to the 15A ones for normal use!) connectors ($0.70). The grounding rod will be the copper pipe ($2) and will be connected to the Tesla Coil via a 5m long wire (the green and yellow one) ($4.80 with lugs). The copper sheet is for other misc connections. Finally the frame will be held together with the M6 Nuts, washers and lock washers and others ($1.50).

The capacitors and resistors have also arrived!

The capacitors and the yellow one. I bought 65 of them These are 0.1uF 1500V Axial Polypropylene capacitors from RS Singapore. I quote from their catalog, "Film-foil polypropylene capacitors protected by polyester wrap and epoxy end seals. With very low loss dielectric suitable for continuous use a high ac voltages. It will withstand fast rise time pulses and has an excellent high frequency performance." Each capacitor costs $3.42, stock no. 114-480. I will be wiring 5 strings of 12 capacitors yielding a capacitor bank of 41.7nF rated at 18kV. Total cost is $222.30

For the resistors, I bought 60 of them, each costing $0.30, costing $18 in total. The resistors are 0.5W 10MegaOhm each. I quote from the catalog "The VR series of resistors comprise of a metal glazed film deposited on a high grade ceramic former with end caps and welded tinned electrolytic copper termination wires. The body is protected with a light blue insulating lacquer. These resistors are for applications in which high resistance, high stability and reliability are required at high voltages. The resistors meet the safety requirements of IEC65." Although the total cost is $240.30, I managed to get a discount and got the resistors and caps for $214.99 instead.

Total amount spent today: $70.50 + $214.99 = $285.49, making the total project cost (as of now) to S$440.49! An expensive day.

I decided to build the toroid today. A nice toroid is important, but a professional spun aluminium toroid would be very expensive (few hundred). As such, I decided to make my own 5" x 20" (turned out slightly bigger) toroid. Building my own toroid cost around $30 and about 3+ hours of work. The method I used was from Easternvoltageresearch's page, which can be found here.

First, a 10 inch dia centre disc was cut out from 9mm plywood using a jigsaw. (picture 1) I drilled holes around the edge at roughly 1 inch intervals. I drilled more holes where the PVC ducting connects to have a stronger joint, as it will need more support. Notice the centre hole. This is drilled to accept a M10 plastic bolt from the top of the secondary coil. I sprayed the disc with 3M Super 77 Spray Adhesive and stuck aluminium foil on it (photos 2, 3), and cut out the excess foil with a penknife. The PVC ducting was then attached using wire. The wire goes through the hole in the plywood, around the ducting, and is twisted with pliers (photo 4). Twisting will fasten the ducting securely to the centre disc. Bend the twisted wire into the groove of the ducting. This part was the most difficult especially at the part where the PVC ducting comes together. More wire does it. Photo 5 shows the completed toroid without its aluminium covering. After that, I used 2" wide aluminum tape and taped up the whole toroid. The last photo shows the result compared with my 40mm mini Tesla coil.

Let me tell you a story about the aluminum tape I used. I first bought this tape before building the toroid of my 1st tesla coil. I didn't know where to get tape like this back then. One day, when I was at a shopping centre with my mum, I saw this tape at a home-fix shop, and promptly bought it. It was $15, but I couldn't wait. Anyway, few days after that, I went looking for more parts at the industrial area, and found shops selling the tape for $6 only - wasted $9. What is the moral of the story? The moral of the story is that we should never buy stuff like this from the shopping area as we will most definitely get ripped-off. Anyway, I was surprised that the tape managed to last for so long - two smaller toroids and this big toroid. Anyway, the toroid is looking good. To be fair, I will add $10 to the project cost as I estimate I used about 2/3 of the $15 tape.

Project cost is now: S$450.49! I will continue work tomorrow.

I did lots of machine work today.

I cut out two 2' squares from the 11mm thick plywood I bought yesterday using a jigsaw (photo1) and drilled appropriate holes for the supports (threaded rod). Here's how the primary, secondary and top board looks like now (photo 2). The primary deck will be adjustable. I subsequently drilled 6 larger holes on the top deck for primary tapping (photo 3). I finished the boards with a coat of Nippon Bodelac Wood/Metal paint, and left it to dry overnight. I will paint the other side tomorrow.

I also started work on the capacitor bank. The capacitor bank is an essential part of a Tesla coil. It had to be able to withstand repeated charging and discharging and the oscillations of the tank circuit. Furthermore, the radio frequency places enormous stress on the capacitors. Professional pulse capacitors needed for Tesla Coil like the one I'm building will easily cost several hundreds or even thousands of dollars. I have opted to use a MMC (Multi-mini capacitor) bank. This is basically several small professional pulse capacitors wired up in series/parallel. The capacitors arrived yesterday.

I marked out the acrylic board (photo1), drilled 120 holes and cut it using a jigsaw (photo 2). The capacitors are slotted in and are wired in series (twisting the leads) (photo3). The resistors are added for charge equalization and for safety (photo 4), and everything is soldered together (photo 5). The is one string of 12 capacitors. I have 4 more to, but I'll do them tomorrow. It's getting late.

I will complete the painting tomorrow morning, finish up the structure in the afternoon, complete the capacitor bank and wire everything up by night. I should be able to get it fired up on Sunday!

I built most of the frame today. The paint had dried already, so I screwed the four nylon wheels on the bottom frame with 16 stainless steel pan screws. (photo 1). I sat on them (around 55kg) and it rolled easily, so it can definitely support the whole Tesla Coil. I also cut out 4 30cm 40mm PVC pipes (photo 2) as supports, and sprayed them florescent orange (photo 3). The two boards are screwed together by 4 long threaded rods (photo 4). The complete frame can be seen in photo 5. The photo on the left shows the ground connection under the top deck. If you look closely at photo 5, you will be able to see a L-shaped copper bracket which is connected to the bottom of the secondary coil. The ground wire is screwed here. Everything is looking good. The photo above right shows a mock setup with the adjustable primary deck, secondary, toroid, spark gap and power supply. However, due to the busy schedule today, I didn't manage to do much. The coil is almost complete, more work will continue tomorrow and I hope to fire it up tomorrow night!

I did the final bits of work today.

Take a look at the 60 capacitor MMC bank! (photo 1). 5 strings of 12 capacitors, making a 41.7nF 18kV capacitor bank. Everything is mounted on 4mm clear acrylic. Primary connections are 3 thick copper wires with output lugs. I think it looks rather good :-). The second photo shows the copper ground rod which will be pounded in the earth. It has a 5m wire attached to it. photo 3 shows the primary tap I made out of 4 fuse holders. There is a lot of surface area, and is easy to tap. All primary wiring is done by 8AWG Oxygen Free Copper high current audio cable. Luckily I had foresight and bought a large 60W soldering iron, otherwise it would be almost impossible to solder the thick wires of the capacitor bank and the fuse holders etc. I completed everything, and the last photo shows the coil in my backyard. I have to run this thing off 2 mains outlets. One 15A 240V one for the dual MOT power supply, and a 13A 240V outlet for the 500W angle grinder (spark gap).

Construction is finally complete!

--- 31st October 2004, Construction is Complete! ---

Testing, Experiments, Etc.

*First Light*

I completed the coil today, so I plugged it in and fired it up. The secondary ground is connected to a 50cm long copper rod pounded in my garden (the yellow/green wire). I used two mains outlets to power it: a 13A 240V outlet for the 500W ARSG, and a 15A 240V outlet for the main power supply. I'm still not sure how much power it is drawing, but I will do measurements soon. It was getting late so I lashed it up quick, placed a metal rod on the toroid, shifted the clothes hanger nearby, and tapped the primary at turn 7. The ARSG was set at full power.

I turned on the power and the whole coil roared and erupted with streamers! I turned off the power, set up my camera, and proceeded to take the photo you see on the left. However, the fun ended when my dad came and told me to continue tomorrow... it was so noisy. (I was wearing ear muffs so it wasn't that loud to me :P ). Arcs are really thick and scary. I didn't measure the arc length but I estimate the arc length in the photo to be roughly 90cm long. Notice how bright the main gap is. You can also see the safety gap firing. I will be tuning it tomorrow night and I hope to get better performance. Everything is looking very good. I can't wait to fire it up again tomorrow!

The second run! Alas a few seconds after turning on the power, the coil suddenly stopped. I did a quick check and everything seemed to be okay. I disconnected the power supply from the primary circuit and turned it on. The MOTs buzzed and nothing happened, no arcing at the electrodes. As I went to check the power supply, my hand brushed against the 24 1N4007 string. It was then that I realised that the diodes were *very* hot. Since everything else seemed ok, I got two Microwave Oven Diodes in series and replaced the diode string... and it worked. However, the big diodes are still getting hot, limiting the runs to only a few seconds at a time. Anyway, it was getting rather late again and my dad told me to keep the stuff and do it again tomorrow (earlier), so I haven't really done much at all. I did try the primary taps though. At turn 8, there is hardly any spark output. Turn 6 is similar to turn 7 but I'm not very sure yet. I took 3 photos of the coil in action. Meanwhile, I will try to find out the problem of the diodes. The MOT caps remain cool.. only the diodes get very hot.

I think that the filter chokes are actually doing more harm than good. Some detrimental oscillations could have been created which could have blown the diodes. Today I removed the chokes and tested the coil again. The coil worked for around 2 seconds, and stopped. The microwave oven diodes are hot and they probably died too. I'm not sure if they died because of the removal of the chokes. However, the MO diodes were running with the chokes last night so maybe that might have destroyed the diodes already. I will be getting some resistors (around 50W 100ohm resistors), and get more 1N4007 diodes and see how they perform. I can't do anything with the power supply blown.

I bought some new stuff for around $10 (left). I replaced the diodes with 24 new 1N4007s and the chokes with 100ohm 30W ceramic resistors. I ran the coil for a few seconds, and everything was fine. The resistors got slightly warm and the diodes were not hot at all.

I ran it again, but suddenly, it stopped. This time the diodes were hot. I guess the diodes died again.

I can't get the coil running without the power supply...
However, I have some ideas. Firstly, I realised my safety gap is too small. I set it small enough for it to arc across when power turned on so it shorts the supply. Also, there is no 3rd electrode to ground in the safety spark gap which might also be a problem. 100 Ohms might be also a bit too little.

Anyway I will be going for a holiday (school trip) so I'll fix this when I return.

I bought new components (diodes and resistors for $15.50) and fixed up the power supply today. Click on the thumbnails on the left for a larger photo with description. The left most picture shows the fixed power supply (11kg) and other things. I got a working 15kV 30mA old Neon Sign Transformer (12kg) free today from an Neon shop. Need to find a use for it. The thing in the background is the bottom of my 180W Tesla Coil 1.

Anyway, here's what I fixed.... I replaced the burnt diodes with another new set of 24 1N4007s; removed the previous 30W 100 Ohm resistors and replacing them with 60W 220 Ohm ceramic wire wound resistors; Added a new electrode from a squashed copper pipe which will be connected to RF Ground; shortened the diode PVC tube; and replaced the older ceramics with 2 new 10kV 1000pF (1nF) ceramics in series. (Yellow ceramics in the photo on the right).

I tested the power supply (not plugged into the Tesla Coil yet) and it works.. arcing between the gaps. however, I noticed something.. even after a few seconds run, the resistors get blistering hot! Just for fun, I took a small piece of tissue, wet it and placed on top of the resistors for cooling testing.. after less than 20 seconds the water in the tissue started boiling... but anyway these ceramics are made to withstand around 400C so it should be okay. Furthermore, by making the gap small and arcing it, the power supply is loaded uselessly and will also cause the doubler caps to be discharged at high rates which will heat the resistor up significantly, probably more than the normal load. The arcing will also cause HV spikes bad for the diodes.

For the next test I will open up the safety gap, and hope everything works all right! I'm not sure if I need more resistance.. I can always add the 100 ohm resistors in series if its inadequate. I hope I don't blow anymore diodes.

No updates for a long time. The new school term has been *very very* busy. Today is Hari Raya Haji and we get a 1 day holiday. So I decided not to put off the tesla coil for too long and fix the power supply. As you might have guessed, the diodes blew again. In a fit of frustration, I decided that there will be NO MORE SILICON in the power supply! As such the idea of the 4 MOT Stack  was born. (MOT = Microwave Oven Transformer).

Having accumulated several MOTs in the past few months, 7 in total, it was time to put 4 of them to good use. I was deciding on whether to  have a 5 MOT stack of 4 MOT stack, but I decided 4, because it's lighter and will provide enough power. After checking the multiplying voltage of each MOT, all four, when connected in series will give around 8600V. After looking around, I found Greg's Page: http://www.hot-streamer.com/greg/4pack.htm, which had excellent information on his own 4 Stack MOT. After modifying his circuit a little, I get this:

I have 4 unmodified MOTs. The first two have their core connected to ground, and the primary windings in parallel. The two series ends are connected to 4 Microwave Oven caps in series each, which is connected to the cores of the other two MOTs, which are separated from each other. But the primaries of the last 2 MOTs are also in parallel with the primaries of the first 2 MOTs. the Secondaries of the last 2 MOTs then have a voltage of around 8600VAC. I measured the multiplying ratios of all 4 MOTs and the voltage will turn out to be 9 x 240.. .which is 8600VAC. So this stack will be capacitively ballasted with the microwave oven caps (C2) and will churn out ~3kW if all is well.

As you can see in the schematic, I have 7 caps, not 8. Not sure if 7 caps will work. Doing some calculations, 8 caps will yeild around 2.8kW while 7 will yield 3.2kW. I shall test and see what happens.

So today I went to NTUC and bought a large Polypropylene Container to put the MOTs in. ($6.90). The MOTs will be covered in oil (probably motor oil) to insulate it better, and to provide some form of cooling.

There is cardboard around the MOTs to protect the plastic box from scratches, and to provide insulation between the second 2 MOTs. The cardboard will soak up the oil and everything should work out well. I don't have enough caps so I'll be getting a few more MO caps from a shop tomorrow, as well as a 15A relay otherwise the 4 MOT primaries in parallel will kill the breaker when I turn them on. Hopefully this will turn into a good, well-behaved and reliable powerful power supply for my Tesla coil. If all works out well, it can even power a very big jacob's ladder! I really hope this one turns out all right. Wasted too much time and money on the Dual MOT setup. The only drawback in this setup is that the whole thing is *Very* heavy... currently weighing 20kg without oil. I hope the box can stand the weight.

Very busy day today. I went to the electronics shop and spent $28 buying stuff... 4 0.83uF 2300V Microwave Oven caps for ballasting, a 15A relay in case I need it, 6m of thick wire (good for 15A) and many spade connectors. I went back home and did all the wiring and arranging in the rest of the afternoon.

As you can see on the left, it's very messy! I also went to the provision shop and bought 3m of clear silicone tubing for $1.80 for extra insulation. The high voltage wires are sleeved with this tubing. After fixing together the wires, I plugged in my variac at 10VAC to test... adding up the transformer multiplying ratios, which is 9.25 + 8.7 + 9.1 + 8.9, I should get 359.5V out... after fiddling around with the phasing for 10 mins, I finally got it right, with an output voltage of 356V, close to the predicted value! So since everything was working fine, I went to the petrol station and got 4l of synthetic motor oil.. it's Shell X100, and cost $20.90. It's some thick gooey stuff which looks greenish yellow...

I poured it into the container until it covered the windings... which took almost entirely 4l.. which made the box VERY heavy. I let it sit for a while, and then put it into a plastic bag, and connected it to a vacuum cleaner. This should suck quite a bit of air bubbles out of the thing. I don't have a professional vacuum chamber but this should be ok for now.

So I set 2 electrodes close together, connected it to the HV output (to see if there were any arcs), and plugged it in with a 22Ohm 50W resistor I had lying around at the primary... it worked! I tried again, and the resistor promptly blew apart (with quite an explosive force!). Okay not very smart but at least I know my power supply works. Anyway I got the resistor free from my friend.

So with no more primary resistance I just turned it on.. the breaker held for about a tenth of a second before tripping. Not good. I got MANY extension cords and wrapped them over pieces of iron, as an attempt to increase inductance... well it worked to a small extent. Not too many circuit trips. I tried to draw some arcs.. very nice. Take a look at the video frame capture above right. Looks just like a single MOT arc (which is expected.. around same power level). Anyway I got some arcing inside the box (I think it's happening on the surface of the oil) so I added more oil (all 4 liters) and some cardboard spacing (the cardboard will soak up the oil). Didn't have time to test anymore so I'll do so tomorrow. I hope everything is okay and the arcing inside will stop. I need a better box but I guess this will hold for a long while. I'll get a bigger box to put this box into incase it fails and all the oil spills out. Finally, I need sometime to prevent current inrush when I first turn that thing on.

My friend, Raptor, suggested to me a method of 'soft starting', i.e. limiting the current when I first turn that beast on. The idea is simple, basically, the power to the MOT stack is first connected through a resistor, which is connected to a relay. When the power is turned on, the current will first flow through the resistor, preventing a surge.

Then the current will flow through the relay coil, which will cause the relay to turn on, shorting the resistor. Even though this happens quite quickly, the relay delay time will be enough for a soft start. The first few cycles of the AC will pass through the resistor first. I tried it with a 12V relay, and it works! Next will be to try it with a 240V coil relay so I don't need a 12V adaptor. Less wires, and more elegant. I tried to measure the current draw, but it's very erratic, bouncing from 6A to 16A. Anyway suppose I don't get 3kW, I can always connect less capacitors. The capacitor leads are not under oil so I won't get my hands oily. Also, arcing inside has stopped! The box is holding up well and I'm confident it can hold up quite well as long as I don't stress it too much. Check out the power arcs this beast can put out, without getting warm at all. (well at least not in this short test). I took a video and you can download it here, or click the photo of the arc on the top right.

The next thing to do will be to make proper connectors for the box so I don't need to leave a gap for the wires to go through. I guess I'll be using brass bolt as connectors; and of course use the more elegant mains powered relay. Once done I can put it on the tesla coil and fire it up! I am pleased with the performance. Arcs are similar to a single unballasted MOT, but start further apart (4 times the voltage) and I can draw them slightly longer.

On Tuesday afternoon, I went to the electronics store and got lots of stuff. I got a 4 x 5A mains triggered relay, some 15A connectors, a resistor and some RCA audio jacks, for a total of $7.80. Got to work once I got home.

So I wired up the soft start circuit in a nice box with some connectors at the top. The photos show the final result. The relay is a 4 contact 5A relay so it should be good for up to 20A. The resistor is a 25W 47 ohm aluminium resistor. It was a rather tight fit but I got it all in. It will be connected to the box by nylon cable ties. Tested the relay and it seems to work. However it was getting late so I continued the next day.

So today (Wednesday), I installed the soft-start box on top of the MOT container. I used the RCA audio jacks for the HV connectors. It looks nice and is easy to use. I forgot to take photos but I'll do so tomorrow.

First test was promising and everything seemed to be okay. However, I tried a few more times and the breaker still tripped occasionally! Apparently the relay is closing too fast. However, I know what the problem is. I've wired the relay coil before the resistor, so I should have wired it after the resistor, which will give more delay time. Not a big problem, I just need to switch some wires. It's getting late so I shall continue tomorrow. Hopefully I can wire everything up right tomorrow and plug it into the tesla coil! I should get the coil running earliest tomorrow, but definitely before the end of this week. I will try and post a circuit diagram once I have time.

On the left is a photo of the high voltage feedthroughs using RCA audio jacks.

Notice the use of clear silicone wire tubing around the HV wires. This should be sufficient to insulate other wires from it. Besides, it's around 8.5kV, significantly less than the 15kV of some Neon Sign Transformers. Anyway, I fixed the relay coil, and placed it after the resistor. Tested it a few times and everything seemed good! I powered a nice big jacob's ladder with it. Much better than MOTs as the voltage is high enough for it to start itself at the bottom of the ladder. However I didn't take any photos of that in action so.. well I guess another time. Impatient to get the coil working, I plugged the power supply to the Rotary Spark Gap (without the tesla coil caps first)... as you can see it's quite bright already. (Taking a photo of the spark gap while the coil is running would be mad.. the spark gap will be *extremely* bright. So since everything was working good, I plugged it into the coil.

The results are Impressive!

Check out the 4 photos (Thumb-nailed ones) above. You can click them to enlarge. The first one shows arcing to ground to a clothes rack which was placed quite near. This was at primary turn 7. Notice how thick and hot the ground strikes are! Bright white!. Then I moved the clothes rack away and the breakout point. (photo 2). Streamers from the toroid and some primary strikes. Didn't harm to coil one bit but it's not so good. I guess I need to work with a breakout point. Still very impressive.

I did a turning check and moved the primary tap to turn 6. The clothes rack was placed just over a meter away. Results are much more impressive! Turn 6 is better than turn 7! Arcs are over a meter long! The 4th photo shows several ground arcs. The coil is totally electrifying. The noise generated is insane (i need to wear ear muffs) and the arcs are just totally scary in real life. Success at last! And best of all, no more fried diodes, no more tripped breakers, and excellent performance!

Now all I need to do is to tune it properly! I believe I can achieve arc lengths up to 150cm. As for now, I'm very pleased with the coil. I shall insulate the wiring in the tesla coil properly with the excess silicone tubing. Also, I hope the 25kg MOT stack is not too heavy for the wooden base of the coil.

Today is a good day!

Here is a video of the coil. My camera doesn't do well in low light situations so everything is rather dark, but it'll do for now.
Click here to download. [610kb, Windows Media Video format, right click and save target as].

I did some more testing today, and managed to get good results at turn 5.5 on the primary, and I took many photos!

In the first photo, the coil is breaking out to air. Notice how bright the spark gap is! The streamers are longer than they look in the photos because my camera can't capture them very well. The second photo shows the tesla coil destroying a ladder. I'm not sure why but many coilers like to have ladder strikes so I might as well have one too! Third photo shows arcing to a grounded rod. There's a big light bulb on top of the coil, note the interesting patterns. The fourth and fifth photo shows some longish arcs and steamers. I especially like how bright and thick the ground strikes are.

Today's arc length record is 110+cm point to point. Not too bad I say! Maybe I'll remove 1 MOT cap and see what happens with the increased power! Or maybe I can add another capacitor. As for now, I'm very pleased with the coil. Now to push it to it's limits and see what this thing can really do.

I got to run my Tesla Coil again, and took a few more photos. It seemed to perform rather well! This time it didn't even trip the breaker once and I ran it for over 5 times for quite long. I'm pleased that it's still holding up very well. The spark gap gets very hot, but not too hot to cause me to worry too much; even the primary wires get warm showing that significant current is flowing through. Performance is good! Enjoy the photos