PowerLabs Twin Tesla Coils!



This is my 3rd attempt at a Tesla Coil, and this time the main idea was to build a coil as good looking and efficient as possible.
Of course, living in an apartment, size and power level are limited, so, instead of making a large, poorly built coil and than pumping ridiculous amounts of power into it until sparks are obtained, as so many of the self-acclaimed “coilers” out there do, this project was focused around a coil that would be small and low power (this brings the added advantage of safety) but yet would still produce nice sparks due to its high efficiency. As for the size, I reckon one foot (30cm) is about as large as I can build and still call it “tabletop”, so that was the chosen size. Taking that and the power level (270Watts) into account, the rest of the coil was designed to match those two parameters. The final design calls for a twin (half wave) system, which will completely eliminate the need for a ground, and will therefore make it completely portable, with zero setting up time. The twin configuration also produces twice as much output voltage for any given power, with far lower ground losses (as there is no ground); this makes it more efficient than the quarter wave system.
In fact, this system has been so successful, that I’ve had 7 beginners copy my design with similarly impressive results. I am always willing to help someone with their projects (this is what this page is for in the first place), so long of course as I get credit for it.
Anyways, here is the system:


Here is a picture of both secondary coil forms. These were precision cut from a 3mm thick, 4.2cmCoilforms diameter translucent acrylic tube using a table saw, and caps were machined for them out of 3mm thick acrylic using a Dremel Moto tool with a 30000RPM angle grinder and sanding wheel attachment. The caps are glued onto the pipe with cyanocrylate superglue (nothing holds acrylic better) and filed to perfectly match the pipe’s diameter. Each pipe was toughly cleaned and dried prior to winding, but the coil form itself was not varnished inside and out -as is a custom with PVC- due to acrylic’s low water absorption coefficient. Acrylic was chosen because it has a lower R.F. dissipation factor and water absorption coefficient than PVC, the next choice.

Winding secondary coil by handWinding each one of the coils by hand took me about 4 hours. Here you can see how I did it: I used magnifying glasses to better position the wire, paper sheets under the coil for it to turn smoother, and rubber gloves to get a grip on it and to prevent the sweat in my hands from all those hours of work from getting into the windings, as the mineral salts and oils contained in sweat may cause all sorts of insulation problems for a finished coil.

Wound coilAnd here you can see the (just) wound coil… Notice how tight those turns are: The coil shines like a copper tube! The wire used here was 180oC C-class insulation modified polyurethane wire, normally used for making high temperature transformer and motor windings. Besides being slightly darker than normal enamel wire, it also has a somewhat better voltage standoff resistance, which is advisable for high performance (spark length vs coil winding height) Tesla Coils such as this one.

Completed coilsAnd here is the finished product… Both coils wound and varnished, with their copper ground straps on place (not really visible due to poor lighting on the picture) and their space-wound top inch, which will be used for further distancing the toroid from the strike rail. The coils received 4 layers of high gloss Polyurethane varnish each, with fine sanding between the 3rd and the 4th coat, in order to make them extra smooth.

 Now the Secondary Coil specs: (note: both coils are identical).

 Diameter of secondary coil: 42.00mm (1,65″).
Winding length of secondary coil: 254.00mm (10″) (note: Coils are 30cm (1′) tall).
Aspect ratio (D/H): 1/6.
Wire diameter of secondary coil: 0.287mm (29AWG).
Spacing between windings: Approx. 0.00mm.
Secondary turns (assuming 98% winding space is filled up): 885.00.
Secondary wire length: 116.77m.
Secondary inductance: 4.98mH.
Self capacitance: 3,79pF
Approximate resonant frequency: 1.150MHz
Secondary quarter wavelength resonant frequency with 8.49pF top load: 642.27KHz


Primary Coilform Segment marked for cutting

 Here you can see one of the primary coil form segments (number 8, more precisely) marked and ready for cutting. All segments are made from 4mm thick PP sheet, and were cut and drilled by hand…

Primary coilTo your right is the finished Primary. Two identical units were built. The silver plated Litz wire coming out from its right is connected to the first turn, and a fuse holder is used to tap the required turn (7th with 1 toroid, 8th with two using 6.8nF capacitance). The first 5 turns are never used and so were varnished with 3 coats of PU varnish so as to decrease the possibility of primary-secondary sparks. All turns are held together with .6mm nylon rope (invisible).

Primary tapHere is a close up picture of my Primary tap. It is made from a fuse holder and the cable that connects it is 5mm dia. multi stranded copper cable, insulated with a 1mm thick PVC jacket.

 The Primary Coil Specifications are the following:

 Primary Coil inner diameter: 6.6cm (2,6″)
Outer diameter: 24.5cm (9.6″)
Height: 4.5cm (1.77″)
Conductor diameter: 5mm (0.197″)
Length of conductor: 4.21m (13.82′)
Inter turn spacing: 5mm (0.2″)
Total Inductance: 0,0123mH (0.007mH when tapped at turn 7)
Geometry: Inverse conical with 15degrees slope.
Best coupling when the first turn of the secondary coil is level with the first primary turn. Attempts to increase coupling resulted in racing secondary sparks. With the current system it tunes in at turn 7.

Top Load:

Toroid materials 4 Toroids were built: The two used for the twin coils are 5.5cm (2,15″) in diameter, and have a 20cm (8,2″) cross section.  To your left you can see the materials used to make those: A 30cm long, 5.5cm diameter flexible aluminum ducting, a 30 meter roll of heavy-duty aluminum tape, and a 10cm diameter, 5mm thick polypropylene disk which stays in the center of the toroid and gives it shape.
A second set of toroids was made later, which employed a 21cm diameter plywood disk as their center, and two tubes as the actual toroid. These have 26cm cross section and 5.5cm tube diameter.
Finished toroids Here you can see the two finished toroids, covered in aluminum foil and smoothened with the back side of a spoon… Notice how the middle is also covered in foil: This makes connections easier, and adds to the capacitance a bit. In order for single breakout to occur (desirable as it allows the maximum possible power to be transmitted to the sparks), the toroid has to be as smooth as possible, otherwise power is wasted in multiple streamers. Each toroid has a capacitance of about 8,85pF

Big toroid...As I pushed up the power on a single coil I noticed that the voltage had become so high that streamers were beginning to strike the primary coil. Therefore a 2 new toroids were built, each employing two tubes together around a 21cm diameter plywood disk. This new toroid sits on top of the original one (picture near bottom of the page) and provides the streamers with more than twice the energy, while at the same time keeping them away from the primary. It was using this that I obtained my current spark record for one coil (50cm using the same setup!), with the added benefit of eliminating primary strikes. The sparks now are also brighter and hot enough to burn wood! Each one of these has 13pF total capacitance.

Spark Gap:

Spark GapInspired on a linear adaptation of the classic RQ series static gap, this gap, which was first designed, I believe, by Terry Fritz, is responsible in no small part for this coil’s excellent performance. It consists off a 10 section series static gap, air cooled (vacuum style) Each segment is a 10X 2cm copper pipe, held by four washers (two on each side, one behind and one after the pipe). These pipes are held by round head screws on top of a polypropylene box and two fans (one on each end of the box) suck air through the gaps between the tubes. Connections are made through 1cm wide, 1mm thick multi stranded copper wire and the fans are powered by a 12V 500mA wall adaptor. It has been run for over 30 minutes at full (360W) power non stop without overheating or showing any signs of performance deterioration. It is also incredibly quiet, its sound being completely downed out by the coil even at the lowest power levels.
For the final version of the twin coil a synchronous rotary spark gap is being assembled, which will allow the tank capacitance to be increased and will provide the maximum possible efficiency.

Tank Capacitor:

Copper bussbars, capacitors, nuts, bolts and washersTwo 50kV 6.8nF capacitor banks with integrated adjustable safety gapsThe single coil and the first Twin Coils used the same bank of 4 parallel TDK UHV�12A 173K Strontium Titanate doorknob-type pulse capacitors, from a nitrogen Laser, as is used on the other coil. Each capacitor is rated at 1.7nF, 50kV and 12A RMS current, giving the bank an overall capacitance of 0.0068uF, which is less than optimum for mains resonance on the 9kV transformer the single coil uses (I should be using at least 0.0088uF for it), and just slightly larger than resonant on the 12kV transformer used on the twins (0,0066uF required for mains resonance).
Good part of this coil’s excellent performance is owned to these exceptionally good capacitors, which are designed to produce high peak currents at high frequencies with minimal losses. Even after running the coil continuously at maximum power for half an hour they remain at room temperature. The low capacitance on the 9kV design gives the coil a very fast break rate, which is interesting to observe (the sparks can be seen to grow and move around fast, sometimes even rotating around the toroid). The capacitors are inter connected by 3mm thick, 15mm wide, 23cm long pure copper buss bars, and each bank has its own adjustable safety gap at the end of it.

On the final version of the Twin Tesla Coil, both of the capacitor banks pictured above (to the right) are used in equi-drive configuration, for 0,0136 uF @ 50kV. The larger tank capacitance allows the use of larger toroids on both coils on the twin system, and more than doubles individual spark energy, making the overall performance better and far more impressive. I hope to implement the equi-drive system shortly.
 Update! I decided to sell the coils before I moved to college in the US, but I did not want to get rid of the capacitors, so I built an alternative, cheaper capacitor bank. The bank is made up of 28 individual capacitors each one rated for 47nF at 1kV. The capacitors are arranged as 2 parallel units with  14 caps each, and is therefore rated at 14kV, 0.0067uF. One advantage of these capacitors is that they have a self healing dielectric, so they can run closer to their maximum design voltage and suffer voltage surges without any permanent damage; if a dielectric punch trough occurs the metalized layer on the polypropylene dielectric vaporizes around the arc and the capacitor continues to work with only a slight decrease in its capacitance…

Line Filter:

Two line filtersUsually not necessary on small coils, as they don’t have that much power to go into interference in the first place… When I first ran this coil I had my computer and stereo system both on and they were unaffected except for some crackling sounds on the computer’s speakers (but not on the stereo) as the main gap fired. Still, I had those two integrated line filters left over from a microwave oven transformer that I took apart and decided to put them in the coil anyways, as any extra protection wouldn’t hurt performance… The result is that I can now run the coil 1 meter from the computer and video it at the same time 🙂

Power Supply:

Variac, mains filter, power factor correction capacitor, and high voltage transformer A 9kV, 30mA (270W) Neon Sign Transformer (back of the picture), controlled by a 1A variac with 15uF 60Hz power factor correction (last of the cans taped to the variac). You can see that the variac has the PFC and the line filters taped to it… It makes the setup a lot tidier… And helps decrease setup time.
For the Twin system a 12kV 30mA neon was later used, as the second coil allows the system to handle more power.

Completed Coil:

Tesla Coil ready to be runAnd here is the completed coil (primary, secondary and the two toroids). In this picture the ground strap (the wire coming out of the front) has been connected to the first turn of the primary, completely eliminating the need for a ground and the possibility of a primary/secondary strike. This is similar to the Oudin resonator, and there seems to be no performance loss from doing it on small coils like this one. Still, the Oudin configuration is only used when a proper ground cannot be obtained as it injects dangerous AC currents on the output streamers. This configuration achieved a maximum spark length of 50cm at 270W (9kV 30mA NST), limited by sparks striking ground. Interestingly enough, the primary coil has never been struck with the large toroid in place. The sparks go over it and hit the ground up to 50 cm from where they originated… The excellent e-field (Electrostatic field) control achieved by the two toroid configuration combined with just the right coupling eliminates the need for a strike ring, which would limit performance by actually drawing the sparks towards it.Tesla Coil Setup And to your right is the completed system, ready to be run. From left to right: 9kV 30mA (270W) Neon Sign transformer, 10 sections series static gap, 12V battery to run gap fans (later replaced by a wall adapter), tank capacitor, Tesla Coil.Twin system test setup. Gap between coils is 63cm and represents the maximum output obtained as of today.And here the two identical coils are seen next to one another in twin configuration. Since there is now twice as much inductance in the circuit, the tank supply was upped to 360Watts (12kV 30mA), and the tank capacitor was increased to 0.078uF so as to make tuning with the extra capacitance possible. The basic setup is still the same, with all the components laid out as seen above, but now two coils are wired in anti-phase so that when one coil is positive, the other is automatically  negative. Because there are two coils, twice as much voltage is built up (in this coil this means 450Kilovolts), resulting in slightly larger sparks and better performance than a single coil of the same size and input power. Performance here is limited by the fact that I could not tune in larger top loads into the system (not enough primary turns). With the new equi-drive system and two larger toroids this coil is expected to achieve as much as 70cm (29″) spark length. As of today the record is 60cm (2′) with, and 55cm (23″) without breakout points. In the background a fluorescent lamp can be seen.


This coil has surpassed all expectations and proven to be one of the most efficient Tesla Coils ever built for its size. At full power (270W), a single coil produces 4 or more simultaneous sparks that could reach out as far as *50cm* (on a coil that has only 25cm winding height!!!). There is enough energy in the system to make the sparks white even when they are not striking ground, and ground strikes are very loud and hot enough to set wood on fire. The twin configuration produced some even more spectacular sparks that could occasionally connect at 60cm (2′) distance! Overall, I am very satisfied with this coil and its performance, and I have been contacted by 4 beginners wanting to get into the hobby by constructing a coil like this one. This kind of interest assures me that I have really build something that is impressive and outstanding. The next step from now will be to double tank capacitance, construct a synchronous rotary spark gap, and put it all on its power box with a separate variac. Until than, you may enjoy the pictures and videos of the coil’s current performance (a special thanks goes to my father for videoing these while I operated the coil).


 Below you see a small exposition of Tesla Coil spark pictures. Run your mouse over them for a brief explanation.

Branched sparks first strike a grounded pole 45cm away from the coil, than move away to stretch out 50cm and hit a grounding plate beneath it

 It is interesting to note that the fluorescent light, one meter away from the coil and therefore well outside its reach, lights up to maybe 1/3rd its natural brightness by the electrostatic field alone.


A loooong ground strike. Click on picture to watch video.Here the power level on the (single) Tesla Coil is slowly stepped up from 0 to 100%, and the sparks can be seen to grow gradually until large ground strikes such as this one are produced. In the middle of the video arcs can be seen being drawn to a screwdriver which I am holding in my hand (no, I don’t feel any shock from doing this). 1.43MB…

The 1 foot long arc curves upwards due to its heat. Click on picture to watch video.Here the Twin Coil system is seen arcing to 1′ (30cm) length with the lights on. 452KB.
Here is an 292KB version of the same video but with the lights off, and the arcing length increased to 40cm (16.3″). In the middle of the video the safety gap on the capacitors can be heard firing (a loud snap). If you have a slow connection or don’t have the time to be downloading all videos, than definitely download this one!
Here is a 708KB version of this video with the lights off and the arcing length increased to its maximum, 60cm (2′)

A power arc between the two towers of the Twin Tesla Coil system. The light on the background is lit by the electric field alone. Click on picture to watch video.Two coils in the dark producing (without any breakout point) arcs 55cm (23″) in length to one another. 874KB.

 If for some reason you can’t see the videos, leave a note on my guest book so I can look into it. The videos were compressed with Intel Indeo Codec 5.04 to 85% fidelity, at 15 frames per second, 176 X 144 resolution, 24 bit color  and 44100Hz 16 bit mono sound, and than re-compressed into .mpg format with XING Mpeg encoder, to its original format. They are meant to be played full screen.


 A power arc forming2,27MB video showing the twin coils operating at full power (360W) on a tabletop. Arcing distance here is a little over half a meter (20″ or so), limited by the table edge clearance. Notice how thick the arcs look!

Power arc in the dark4,16MB video showing the coils operating over the tabletop at full power for a longer period of time. As with the other video, it shows the coils running both under lighting (180W incandescent lamps) and at semi-dark, so that a good perspective of how bright the sparks are can be obtained.

Sparks trail over the glass surface of this bottle into my bare hand!Funny video showing sparks striking a glass bottle I am holding in my hand. Shows both lights on and lights off. Power level here is sub 80W and repetition rate is 5 sparks per second. 758KB

Check out the PRESENTATION given to a physics class with this coil.