Nikola Tesla's famous air core transformer--aka the Tesla Coil--was developed in the 1890s as a source of high voltage, high frequency electricity. In those days it was a valuable component in the emerging field of wireless telegraphy, and Tesla had ambitions to charge the earth and upper atmosphere in order to supply electrical power throughout the world without costly and cumbersome power lines. His concept did not work out for various reasons, and the Tesla coil survives today as a project for hobbyists. The design below is based on popular designs published for amateur builders circa 1910 or so. I have modernized the frame by using PVC, but the horizontal coil was very popular with early radio experimenters in that long ago time.

Step 1: The Frame

The frame of the tabletop Tesla is made entirely of half-inch (12.7 mm) PVC pipe. There's no point giving you exact dimensions of the stringers and risers because they depend on how big your secondary coil is. You can scale the frame up or down as you wish. I used a 12 inch long cardboard tube for mine, making the footprint of my coil 14 inches by 11.25 inches.

Study the pictures and you will see how the frame is made. The configuration as made requires the following joints:

(4) 90 degree elbows;
(8) "T" joints
(2) end caps

All the rest is straight half-inch tubing, cut to length. NO CEMENT WAS USED TO JOIN THE PARTS and none should be used. The friction fit of the tubing is reasonably strong, and leaving the joints unglued allows you to take the frame apart to work on the coil, make adjustments or repairs, etc.

The center uprights consist of three T joints each, stacked vertically. Short lengths of tubing connect these. If you make a bigger coil these length will have to be adjusted accordingly.

The cross piece that runs underneath and parallel to the secondary coil has to be drilled for the primary supports. Find the center point of the cross piece and drill two holes so that the primary form in centered on the secondary. Again, I can't tell you exactly where, because it depends on what you use for your coil forms. But center it and it will be fine. See the page on the Primary for more details of the mounting method.

The secondary is supported by plumbing caps and tubing adapters inserted into the cardboard tube. The tube I used is 1.75 inches in diameter (it's a thick-walled cardboard paper towel tube). I had to experiment at the home center to find off the shelf PVC plumbing pieces that would fit, but I found ones that slip in closely. Again, no glue was used. You want to be able to remove the secondary for maintenance or replacement.

In the two topmost T joints insert plugs to support the brass terminal posts. These can be anything non-conductive--cork, rubber, etc. I found wine corks fit nicely. Push them down equally on each side until they are level with the top edge of the secondary form. Above them fit a length of PVC tubing. Exact height is not too important; they should be tall enough to keep the terminals away from the active coils to avoid arcing. Mine are four inches tall each.

Drill 3/16ths holes in the center of two PVC pipe caps for the terminals. Drill small holes--about 1/8th inch--in the support tubes opposite the ends of the secondary coil to feed the secondary's wires through. See the step about the Terminals for final connections.

Step 2: The Secondary Coil

The secondary coil in a Tesla coil is generally a long coil of fine wire. It is not connected to the primary coil, but picks up current from the primary by induction.

For this tabletop model I used a heavy duty cardboard cylinder formerly used in a roll of commercial paper towels. It's about 3/16ths of an inch thick, 1.75 inches in diameter, and 12 inches long. You can also use PVC tubing, or any other type of insulating plastic. Avoid colored tubing. Sometimes the dye is conductive, which will cause your project to fail. Avoid black tubing. It often has carbon in it, and carbon conducts electricity.

If you use cardboard as I did, coat it well inside and out with several coats of polyurethane varnish or shellac. These help stabilize the cardboard and greatly improve its insulating properties.

Once your secondary form is hard and dry, apply two strips of double-sided carpet tape to the outside, 180 degrees apart. This is very sticky tape used to hold carpets down. On a 12 inch tube, I used two strips of tape, 11 inches long each. This will help hold your secondary wire in place as you wind it.

The wire I used is 24 gauge magnet wire. It is solid copper with a coat of enamel insulation. Start winding half an inch from the end of the tube by taping down the wire (temporarily) and winding slowly and carefully, keeping the wire smooth and snug. Do not cross the wire at any time, or your work will come to naught. It's boring, but I wound the whole secondary by hand in about 80 minutes while watching TV. I wouldn't bother with winding jigs unless you plan to make several coils.

You can use other fine gauges of wire--22, 26, 28, 30, etc.--and the wire you choose will affect the resonance of your coil. I had 24 gauge on hand, so that's what I used. There are 515.9 turns on this coil, for a total of 249.7 feet of wire.

Once your secondary is wound, paint it several times with clear polyurethane or shellac. Let it dry thoroughly between coats. Some coil makers put on many, many coats of varnish. This both insulates and glues the wire to the form. I put on four coats of shellac. That seemed like enough.

Carefully remove the anchoring strips of tape from each end. I used a bead of hot glue to fix the wire to the form. Leave about eight inches of wire free beyond the glue. These will connect to your terminals.

(The secondary form is supported by PVC plumbing caps inserted inside. I spent quite some time at the home center trying out different bits of plumbing connectors and caps to find ones that fit the cardboard tube nicely, but I eventually found perfect pieces. )

Feed the wires through the holes bored in the upright supports. Scrape off the enamel insulation (I used fine sandpaper for this) from the last three inches of the wire. Push the tubing and PVC cap up, and attach the bare wire ends to each terminal. Solder if you want. I used aluminum HVAC tape, which is strong and conductive. Push the supports back into top T joints. The ends of the terminals will be supported by the cork or other stopper previously inserted into the frame.

Before you push the PVC caps in place, you may want some friction pads to keep the terminals from swinging around two easily. I cut a rubber stopper in half lengthwise and put one piece in each upright. With the caps in place, the terminal rods are jammed against the rubber, which allows fairly good adjustment of the terminals.

See the Terminals page for details of making them.

Step 3: The Primary Coil

The primary form is a so-called 2 inch PVC pipe coupling. It actually measures 2.75 inches in diameter. The first thing to do is mark and drill holes for mounting it on the frame. Drilling in-line holes in a round object can be tricky without a jig or clamp to hold the work in place. I used the edge of an open drawer (a natural 90 degree jig) and marked two spots about 3/8ths of an inch from each end of the coupling. Use a 1/4 bit and drill carefully. Next. line up the cross piece that runs beneath the secondary coil. Using the primary coupling as a guide, center the two tubes and use the holes you already drilled to mark the cross piece. After spoiling several piece of PVC I learned it was better to drill the four holes one at a time; that is, drill the top two holes, then align your bit by looking down the inside of the tube to drill the holes through the other side.

You will need 2 1/4-20 nylon bolts (2 inches long in my model) and six nylon hex nuts. Insert the bolts through the holes in the primary form. Put one nut on each and tighten. Put a second nut on each bolt but don't tighten yet.

Now let's make the coil. Stick a 2 inch wide strip of double-sided carpet tape around the middle of the PVC coupling. On my model I used 12 gauge stranded wire with a heavy nylon insulation. This is sold on bulk reels in most hardware stores. At first I put on 9 turns. Later, on the advice of TeslaMap, I reduced this to 5. I would start out with 5 or 6 turns and work on tuning the coil later. The carpet tape will help hold the 12 gauge wire in place, but the insulation and thickness of the wire resists conforming, so once you have the primary wrapped, anchor the wire in place with a couple of layers of electrical tape.

Slip the primary around the secondary. Let it hang loose for a while. Push the plugs into the secondary to support it and snug up the various PVC joints. Make sure the cross piece is in place underneath the secondary and loose primary. Align the holes in the cross piece with the bolts. Attach the last pair of nylon nuts to the underside of the cross piece and tighten them just until they are flush with ends of the bolts. Now tighten the middle pair of nuts down against the cross piece. The idea is to support the primary from below at just the right height to center it around the secondary. Screw the bottom pair of nuts on each bolt up or down as needed to align the coil. If needed, you can also twist the cross piece forward or back to help center the primary. Study the photo; it should make things clearer.

Once you have the primary centered, use nylon zip ties to tie the primary wire to the cross piece. This will support the wire against pulling and help hold the coiled wire in place.

Step 4: The Terminals

The twin terminals are made from 3/16ths inch brass rods. Cut two equal lengths--mine are 18 inches long--and bend them at the same point about 2/3rds of the way. To make sure they bend at the same place I taped the rods together securely and bent them at the same time over a length of steel pipe. Don't use clamps or pliers or you may mar the brass, Gouges and burrs are not good for electrical terminals. You will have excessive corona leakage at such spots.

Insert the terminal ends through the holes in the PVC caps. Wind a generous amount of stripped secondary lead around each rod and secure them with aluminum HVAC tape.

The electrodes at the tips of the rods are solid brass balls sold in home centers and lamp shops as pull chain knobs. They are 0.7 inches in diameter (18 mm) and bored through for the pull chain. With a 3/16 bit, open the holes enough to accept the brass rods. Wrap 1/2 inch wide strips of aluminum HVAC tape around the rods, about 1/2 an inch back from the end until you have a good friction fit for the brass balls. If you're good with a torch and solder, you can probably sweat the electrodes on, of course.

The rod ends rest on the corks or plugs you put in the frame earlier. To create friction dampers for the rods so they don't swing around too freely, cut a rubber stopper or cork in half lengthwise and press these pieces (one half per post) into the PVC terminal posts. With the caps in place, there should be enough pressure on the rod and stopper pieces to control the motion of the terminals.

Step 5: The Spark Gap

The spark gap is made entirely of brass hardware and HDPE (polyethylene) plastic. The base is a HDPE cutting board 8 by 6 inches. My first spark gap was a simple two electrode model, but it quickly overheated in use, so I put together the four post model you see here. You will need 4 90 degree brass brackets( 2 x 2 inches), 2 0.75 inch 8-32 brass bolts, 2 2.5 inch 8-32 brass bolts, two 8-32 brass nuts, 4 8-32 brass thumb nuts, and four 0.4 inch (10 mm) solid brass balls, drilled and tapped for 8-32 threads. These balls are sold as lamp repair parts in many hardware stores and home centers. You will also need 6 short machine screws.

The two center posts are mounted on the HDPE base with a single machine screw in the rearmost bracket hole. The screws should be short enough they do not pass all the way through the plastic base. If they do you will have arcing through them when the coil is energized--not a good idea. Using one screw in each bracket allows you to pivot them slightly to adjust the spark gap.

A 0.75 inch long bolt is put through the top hole of each of the center brackets and a nut is turned on tight. Add a brass ball to each post.

The other two posts are adjustable as to length and spacing. Place one at right angles to the center posts and screw it down with two machine screws. Put on thumb nut on one of the long brass bolts. A high voltage wire from the power supply will attach here, and the wire going on to the capacitors. Put the second thumb nut on after the bolt has passed through the bracket. Tighten, then add the ball electrode. Repeat this assembly for the last post, which is mounted at a 45 degree angle to the rest. (This angle is strictly a matter of convenience due to the size of the base. If you have a larger base, you can set the last post perpendicular to the center posts). The other lead from the power supply goes here, and the HV line going on to the primary coil.

Adjust the four electrodes so that they line up and have about 1/16th inch gap between them to start with--about 1mm or slightly less. Adjust as necessary to get good output.

In operation the multi gap still gets warm. When the air is ionized, it adversely affects the rate at which the spark gaps fires. The result is fluctuating power levels and poor output. The gap can be cooled by a stream of air from a vacuum cleaner with its hose reversed, a hair dryer set on unheated air, or even a converted computer cooling fan. Quenching the gap will greatly improve the performance of your coil.

Do not look at the spark gap when it is operating. It gives off intense UV light, like a welder, so avoid looking at it or else shield it in some way. When the coil is operating there is lethal voltage going through the spark gap--do not touch it!

Update 6-25-09: Spark Gap Mark III. I rebuilt the four post spark gap with much larger components. (See pictures). Instead of the little 10 mm brass balls, I installed four brass knobs designed as drawer or cabinet pulls. The two egg shaped knobs are lacquered brass; the spherical knobs in the center are satin nickel plated brass. The brackets and supporting bolts are the same as before

The larger knobs work very well and handle the heat of spark gap operation better. They are superior to the smaller electrodes, and and being threaded for the same 8-32 bolts, they are easily exchanged.

Update 8-3-10: For even better performance, try making a pressurized spark gap, as described here: http://www.stephenhobley.com/blog/2008/09/10/petes-quick-hyperbaric-spark-gap-tm/. I used smaller components than shown--half-inch copper pipe and a half-inch sized PVC junction box--and it works very well when connected up to a small vacuum cleaner. See this video for example.

Step 6: The Capacitors

Two things matter most in choosing capacitors for a Tesla coil--capacitance and voltage rating. I'm not going into the theoretical requirements here because there are many authoritative texts who do that much better than I ever could. Suffice it say you need at least double the DC voltage rating of a capacitor for it to survive life in a Tesla circuit, and more is even better. With a 7.5 kV power source, I would want at least a 15kV rating, and more is better.

To figure capacitance required by the coil, a design program like TeslaMap is invaluable. For various practical reasons I ended up with two 4nF, 20kV ham radio capacitors I got off eBay. I linked them in series, doubling the voltage rating to 40kV. I measured the capacitance of the two caps in series and got 2.75 nF. I tried using one of the caps alone and it burned out. Perhaps it was defective, or perhaps the voltage rating of 20kV just wasn't enough.

(Since starting I have burned out two more of the ham radio polystyrene capacitors. They apparently don't stand up to the severe stress of Tesla circuits very long. I don't recommend them.)

I next tried my homemade Leyden jar array.

Update 6/16/09: I reworked my soda can Leyden jar battery (12 jars linked in parallel) by joining them with strips of HVAC tape and binding the whole array tightly together to insure contact between the individual jars. This upped the capacitance to 5.76 nF, which works very well with the bipolar coil as built. The Leyden battery takes the power from the spark gap and NST with no signs of over-voltage, heating, etc. Homemade proves better than factory made, in this case!

Update 9/14/09: New photos added of my homemade Leyden jar array. This battery has 18 jars, for a total capacitance of 8.05 nF. It works very well with all of my tabletop coils, upright and bipolar.

Update 10/7/09: I have added a complete Instructable on how to make the Leyden jar battery for a small Tesla coil.

Step 7: The Power Supply

For power I use an Allanson 7,500 volt neon sign transformer. It is current limited. As Mitch Tilbury writes in The Ultimate Tesla Coil Design and Construction Guide:

"Small Tesla coils that typically use NSTs will not require circuit protection as the NST is current limited and will draw less line current than the 15-A [15 amps] rating that is typical for building electrical outlets." (page 275)

I installed two RF chokes on the lines running to the coil. These are snap-on types sold at Radio Shack. They are there to block or reduce RF interference.

The Allanson unit is handy in that it has a pull-chain switch with a long string attached. You can easily stand two or three feet away and turn the apparatus on and off from a safe position. I mounted the NST on a pine plaque, well-shellacked, and put four hard rubber feet on the plaque to raise it and give it a more finished appearance.

Step 8: How It Works

I was fortunate the coil worked the very first time I tried it. I was frankly startled at how easily and well it performed. After a few tests, I burned out a capacitor or two and learned my first model simple spark gap wasn't adequate. I have since run the coil with a bank of 12 or 14 handmade Leyden jar capacitors with varying success. I have also run it successfully with ceramic doorknob caps made in the former Soviet Union (check eBay!).

At its best, the bipolar coil makes loud, thick arcs when the electrodes are closest together (a gap of 2.5 inches). With an insulated wand I have pushed the electrodes as far apart as six inches and had a continuous arc. Beyond six inches I get lively brush discharges--purplish coronas around both brass electrodes. An ordinary clear incandescent light bulb, mounted on a insulated rod, is filled with violent, fiery streamers when placed between the terminals. Fluorescent tubes glow several inches away, though oddly enough they glow brightest at about four o' clock low instead of up close to the terminals.

Cooling the spark gap with a stream of air intensifies the arcing, making it louder and brighter. Merely switching to the multi-point spark gap array had much the same affect.

So that's my bipolar Tesla coil. I will continue to experiment with capacitors and spark gaps and report my results in updates to this page.