This is a basic Tesla coil schematic (click the image to enlarge). The schematic shows a static spark gap. If you're using a rotary spark gap, simply replace the static gap (the wiring is the same). Older schematics sometimes invert the location of the spark gap and the primary coil. The Tesla coil will work in either configuration, but due to some technical reasons, the configuration shown in the schematic is preferred.

The NST case should be grounded. There is some debate as the proper ground for the NST. Some advocate connecting it to RF ground, others think it should be connected to the mains ground. In the schematic I've shown a switch to indicate the ground could be attached to either.

Neon sign transformers (NSTs) are the preferred power supplies. I'll mention some other types of power supplies later in this section. You should choose a transformer that supplies at least 5kV, otherwise you may have problems with the spark gap not firing.

If a NST has a GFCI (also known as a GFI or ground fault interrupter) circuit, it will "trip" or automatically shut off a NST when it detects an unusual current in the output of the NST. Unfortunately Tesla coils produce current spikes that frequently cause the GFCI circuit to shut off the NST, making NSTs with GFCI circuits unreliable in a Tesla coil. NSTs with a GFCI circuit will usually have a GFCI reset button somewhere on the case or possibly under the top cover. It may be possible to rewire and bypass the GFCI circuit in the NST, although it may be a very difficult process depending on the complexity and location of the GFCI wiring. Adam Johnson has reported that he can use a NST with GFCI when used with a Terry filter. Newer, small NSTs are actually solid state power supplies that are generally unsuitable for Tesla coils. I strongly recommend using an older NST to power your Tesla coil. A good NST should be very heavy and only contain a primary winding, secondary winding and metal core (and probably some potting material). The output frequency should be the same as the input frequency (50 or 60 Hz).

NSTs are usually fairly easy to obtain and are fairly robust when used with the proper protection circuit. Used NSTs are often much cheaper than new ones. They can be found at sign shops and salvage / recycling centers. Typically they either work or they don't. To test a NST, simply connect it to line voltage (wall outlet) and verify it will produce arcs between the output terminals, or each output terminal to the case (assuming the case is grounded).

If a NST dies, the cause of death is sometimes arcing through the internal potting material. Potting is an insulator, usually a hard, tar like substance. The NST can be resurrected by removing the top of the case and heating the NST over a grill to melt the potting material. Baking in an oven is not recommended because of toxic fumes and leaking potting material. Once the potting is melted it can be stirred to remove the short or poured out and replaced with transformer oil. This process is very messy and probably not worth the effort if another NST can be found. The use of solvents to dissolve the potting material may also be an option.

NSTs have shunts or metal plates between the primary and secondary coils which limits the current even when the output is shorted. The current limiting makes NSTs more robust than other transformers. The shunts can be removed to provide a bit more current, but the chances of winding damage increases.

The primary, low voltage side of a NST should be wired through a line filter which is connected to the house or building mains. A PFC cap should be wired across the primary terminals, but the NST can be run without it. Common NST power outputs are 9kV, 12kV or 15kV @ 30mA or 60mA.

NSTs can be wired in parallel to supply additional current to the Tesla coil. Do not try to wire them in series, the extra voltage will short the secondary windings and damage the NST. NSTs with different output currents can be wired in parallel, but if the output voltages are significantly different (more than a few volts), one NST will begin to overheat. Follow this procedure to test NST compatibility...

• Determine the phase of the NST outputs by checking for arcs between the output terminals of the NSTs. Connecting an output terminal of one NST to an output terminal of the second NST (leaving a small spark gap). If you see an arc then the terminals are out of phase.
• Put a mark on the output terminals that are in phase. Also mark the low voltage input terminals since switching one of the input terminals will switch the phase of the output terminals.
• Connect a 1 kilohm 1/4 watt resistor between the output terminals that are in phase.
• Run the NSTs for a few minutes, disconnect power to the NST and feel if the resistor is hot.

If the resistor feels hot then too much current is flowing through the NSTs and they should not be used in parallel.

Other transformers can be used such as oil burner igniter transformers (OBITs), microwave over transformers (MOTs) or distribution transformers used in the power grid, often seen on telephone poles and sometimes referred to as "pole pigs". Pole pigs are sometimes given away by the power companies, but they are extremely dangerous and heavy.

Another power supply option is a bombarding transformer. My information is limited, but they seem to be high power transformers used to make neon signs. They typically operate around 450-800 mA at 22-26 kV. They are apparently very heavy (150-200 lbs), expensive and difficult to find. I'll add more information as I learn more.

The primary capacitor is used with the primary coil to create the primary LC circuit.

The primary capacitor is usually made of several dozen caps wired in a series / parallel configuration called a Multi-Mini Capacitor (MMC). A single pulse type capacitor can be used, but they are harder to find, cannot be adjusted and are more difficult to replace. Also, when a MMC fails it can usually be fixed by replacing an individual capacitor in the array, but if a pulse cap fails it must be replaced.

Other types of capacitors can be made, including salt water beer bottle caps, rolled aluminum foil caps and stacked plate caps. Home made capacitors generally require a lot of work and they often fail. Salt water beer bottle caps are inefficient and it's difficult to know how much capacitance you're working with. Rolling or stacking caps with layers of aluminum foil and plastic insulators have not shown much success. Often the plastic will have microscopic holes or weak spots that quickly short out. Small air pockets between the layers heat up and can explode. Rolled and stacked caps need to be submerged in oil to reduce corona, which can be messy. Despite the higher cost, I recommend using factory produced caps. The primary capacitor is run under extremely demanding conditions. It's exposed to high voltages and very short charge / discharge cycle times. Factory caps can tolerate these conditions much better than anything most of us can make ourselves at home.

Caps usually have a VAC and VDC rating. When using caps as the primary capacitor in a Tesla coil, they will only be charged and discharged for a very short time. Because the caps are "pulsed", we can use the VDC rating when designing the MMC. Although it seems odd, the VAC rating should be ignored.

Normally 1.6kV to 2kV caps are used in the MMC array. Several caps are wired in series to provide adequate voltage rating. It's good practice to construct the MMC to withstand 2 or 3 times the peak voltage rating from the NST. For example, using a 15kV RMS power supply (15000 * 1.414 = 21kV peak) the MMC should have a minimum voltage rating of 40 to 60kV. However, good quality caps can be run closer to their specified rating. Terry Fritz tested three CD942C20P15K capacitors at their rated DC voltage and they lasted for 75 hours before failing. Although 75 hours may not seem like a long life expectancy, most Tesla coils are only run for short intervals. A typical MMC will have about a dozen caps in each series sting. Normally a few series strings will be wired in parallel to provide adequate capacitance. The TeslaMap program has a MMC calculator that makes MMC design fast and easy.

Many people eventually upgrade their Tesla coil by switching to a rotary spark gap or adding additional NSTs. Both of these changes will affect the required MMC capacitance. It's prudent to consider future upgrades when planning and constructing your MMC. The MMC can be constructed with tap points between the capacitors so the capacitance of the array can be easily adjusted. It's also a good idea to consider leaving space to add an additional series string of capacitors in the future. Occasionally a cap in the MMC may fail, so the MMC should be designed to allow replacement of individual caps.

A 1 Mohm to 10 Mohm bleeder resistor should be wired across each capacitor to prevent the caps from holding a dangerous charge. I recommend using high voltage "metal film" or "thick film" resistors. They are commonly available with voltage ratings of 3.5kV and higher. High voltage resistors with power ratings of 1/2 W and 1 W generally have higher voltage ratings than 1/4 W resistors. Some high voltage resistors are physically long, so I recommend that you check the specified length to ensure that the resistors can be easily wired in parallel with your capacitors. The bleeder resistors should not be in direct contact with the case of the capacitor as arcing can occur (refer to the resistor's "dielectric withstanding voltage" specification). It's a good idea to solder the resistors to the opposite side the pref board, or whatever you mount the caps on. When wiring the MMC it's best to twist the capacitor leads together then solder. Don't bother etching copper tracings on the circuit board. The thin copper can't handle the current in the MMC.

I recommend keeping all connections in the MMC as short as possible, especially connections that connect the different series strings. Long or poor connections between the series strings can create an imbalance of current through them. The strings closest (with the least resistance) to the main connection will receive more current.

It's important to use the correct type of caps in the MMC. Most caps will quickly fail when used in a Tesla coil. Look for these qualities in a good MMC cap:

• Polypropylene type caps
• Metal "foil" type, especially foil electrodes
• High dV/dT rating (min of 1000 - 2000 V/uS)
• High RMS current capacity (min of 10 - 15 Amps)
• High peak current capacity (min of several hundred amps)
• Self healing ability

Avoid "metallized" or "metal film" (the metal film is too thin to handle Tesla coil currents). Avoid polyester capacitors.

dV/dT is an important specification in Tesla coil capacitors. It states how fast voltages can change in the capacitor. Tesla coils operate at high voltages and high frequencies so it's important to use caps with high dV/dT ratings. The dV/dT is usually stated as V/uS. dV/dT is calculated as:

dV/dT = 2 x pi x Vpeak x Frequency

For example:
If we have a MMC running at 15kV RMS (15000 * 1.414 = 21kV peak) but we have 10 series caps in our MMC, so each cap has 2.1kV. Assume a 160kHz resonate frequency. The dV/dT is calculated as follows.

dV/dT = 2 x pi x Vpeak x Frequency

dV/dT = 2 x pi x 2121 x 160000

dV/dT = 2132261765 V/S

dV/dT = 2132 V/uS

So under these conditions you should choose caps with a minimum dV/dT of about 2000 V/uS. You can use dV/dT to estimate peak current by using the following calculation:

Ipeak = Capacitance * dV/dT

Using our dV/dT from above with a 0.056uF cap:

Ipeak = 0.000000056 * 2132261765

Ipeak = 119.4 Amps

The following is a good / bad cap list that was created many years ago by several Tesla coil builders. Some of the caps may no longer be available. The VDC rating is used because the caps are pulsed in a Tesla coil.

Recommended MMC Caps

Manufacturer	Part Number	Voltage (VDC)	Value (uF)
Cornell Dubilier	941C series (1) 942C20P15K 942C16P15K 942C20P10K 942C20S47K 942C20S68K 942C series untested* 943C series untested*	600-3000 2000 1600 2000 2000 ? 1600 - 2000 1600 - 2000	0.01-4.7 0.15 0.15 0.1 0.047 0.068
Panasonic	ECW-H16473JV ECW-H16563JV ECW-H series untested*	1600	0.047 0.056 .0047-.056
Seacor	KP25	1600	0.047
SB Electronics	16PSS50	1600	0.05
Arcotronics	RS-114-474 KP1-72	1500	0.047 0.1
Wima	FKP-1 FKP-4 untested*
Phillips	KP/MMKP KP/MMKP-376	2000 1600	0.0047 0.056
Mallory	PVC1611	1600	0.01
Illinois Cap.	683PPB202K 683PPB102K	2000 1000	0.068
BC Components	BC1971-ND	1600	0.01
Evox RIFA (Kerment)	PHE450 (2)	2000	0.15

* Not all of the caps listed in this row have been tested. They should work, but please check the capacitor specifications (dV/dT, RMS current, etc.)
(1) Approved by Dr. Resonance
(2) Tested by Matt

Not Recommended For MMC Caps
Note - Some of these caps can work in a Tesla coil, but they have poor dV/dT specs and will fail sooner than the recommended caps.

Manufacturer	Part Number	Voltage (VDC)	Value (uF)
Cornell Dubilier	940C20S33K avoid the 940 series	2000	0.033
Phillips	MKP336-2
G.E.	42L4102	3000	0.01
G.E.	42L3332	2000	0.33

NSTs do not operate well with resonate capacitance. A resonate sized cap can cause a condition known as resonate rise which causes voltages in the primary circuit to increase far above normal levels. These high voltages can easily damage a NST, so NSTs should only be used with Larger Than Resonate (LTR) primary capacitors. To minimize the risk of a resonate condition in the primary circuit I use a MMC at 1.618 times the resonate size. The ratio of 1:1.618 is known as pi or the golden ratio. Any two numbers in this ratio will have the fewest common multiples which will result in virtually no chance of resonance. Also, LTR caps will transfer the most power through the Tesla coil.

Capacitors can be ordered on the Internet, although they may be difficult to find. Sometimes a fellow coiler will order several hundred and resell them to other coilers. The following links should be a good starting point:
Tesla Stuff
octopart.com

A safety gap should be placed in parallel with the primary capacitance to protect the capacitor from voltage spikes. However, the caps should not be allowed to short directly through the safety gap because the rapid discharge will stress the caps. To prevent the safety gap from shorting out the caps (which is almost as bad a voltage spikes) a high wattage, low value resistor (a few ohms) should be placed in series with the safety gap. The resistor should not be a wire wound type, which may contain inductance and create undesirable effects. No quenching is required in the safety gap.

The spark gap is used as a switch to momentarily connect the primary capacitor to the primary coil. When the gap is shorted the cap is allowed to discharge into the coil.

Many spark gap designs can be used. Spark gaps come in two basic designs: static and rotary. When the gap electrodes are stationary, the gap is referred to as a "static" gap. A rotary gap uses rotating electrodes.

The most simple gap design is a static gap consisting of 2 bolts, wires, drawer knobs, or other conductors that act as the electrodes. The electrodes should be smooth and rounded with no sharp edges that could cause the gap to short erratically. The gap between the electrodes is set to a specific width. The width determines the voltage required for the gap to short. The ideal gap will short just as the primary cap reaches it's peak voltage. The gap should be designed to allow small and easy adjustments to its width. Knobs screwed onto bolts are a good choice. Adjusting the gap width is as simple as turning the knob or the bolt.

Static gaps are simple and easy, but they have some shortcomings. Often the gap will continue to short after the cap voltage has fallen significantly below it's peak - and even below the voltage required to short the gap. This happens because the air between the gap becomes ionized when the gap shorts. The ionized air is more conductive and allows the gap to remain shorted. The performance of a static gap can be improved by blowing air through the gap. This is called "quenching" the gap. The goal of quenching is to blow the ionized air out of the gap. I've used 12 volt computer case fans, others have used vacuum cleaner motors. Generally the more air you can blow through the gap the better.

A Richard Quick (RQ) design uses several copper tubes to divide up the spark gap into multiple smaller gaps. The Richard Quick design usually performs better than a standard static gap with two electrodes.

An improvement to a simple static gap is a rotary gap. A rotary gap uses a motor to rotate the gap electrodes, which can precisely control the gap shorting. Two different types of motors can be used to drive a rotary gap; synchronous and asynchronous (also called "sync" and "async"). Synchronous motors rotate in sync with the power supply frequency (50 or 60 Hz). Sync motors will always run at a multiple of the input frequency. Common speeds are 1200 RPM, 1800 RPM and 3600 RPM for 60Hz input frequencies. Async motors do not rotate in sync with the line frequency.

Rotary gaps come in two basic designs: disk and propeller. The disk design is more common and uses a disk mounted on the motor shaft. The disk has electrodes placed around the edge that rotate and line up with stationary electrodes to create the spark gap. A propeller design looks like an airplane propeller. The electrode is mounted on the motor shaft (but insulated from the shaft) and rotated to line up with stationary electrodes to create the spark gap.

Smaller or weaker sync motors may have trouble turning a disk or propeller. In this case the motor may not start or it may lose sync. When the motor loses sync it will attempt to re-sync. During this time the RPM will vary slightly as the motor "hunts" for the sync RPM. If this is a problem then a lighter propeller gap is a good solution. The rotational power of a sync motor is called torque and is usually measured in in/oz (inch ounces). Torque can be complicated, so I prefer to use watts when dealing with sync motors. For most rotary spark gaps the motor should produce at least 10 to 15 watts. More is always better. I have not had much success with 5 watt sync motors.

Typically the gap is designed to short or "break" 120 times a second (120 BPS) when run from a 60Hz supply. This will correspond to the 60 Hz primary cap charging. It may sound like the spark gap will be firing twice the required rate, but remember that the 60 Hz waveform includes a positive and negative peak, so the gap fires on both peaks.

The number of electrodes will need to be chosen to provide 120 BPS depending on the motor RPM. The following chart lists the required electrodes to produce 120 BPS with different motor RPMs.

RPM = Rotations Per Minute
RPS = Rotations Per Second
BPR = Breaks Per Revolution (required for 120 BPS)

Sync Motor RPM and Required Electrodes

RPM	RPS	BPR	Electrodes
3600	60	2	2
1800	30	4	4
1200	20	6	6
900	15	8	8

I do not recommend a disk diameter smaller than about 5 inches, especially with high RPM motors because they can create a swirling cloud of ionized gas in the gap.

Old sync motors can be found on record players or old computer reel equipment. Some have been found at military surplus stores. New ones can be ordered online. Hurst and Oriental Motor makes good motors.

Terry Blake has a lot of really good information on rotary spark gaps here: http://www.tb3.com/tesla/sparkgaps/index.html

Spark gaps must cope with extremely high currents. Most electrodes will quickly develop burning and pitting on their surfaces. Tungsten is a good choice of spark gap electrodes. It has the highest melting point of any metal so it resists burring and pitting. It can be found as welding rods, in drill bits, etc. Tungsten welding rods come in several different types, each with different properties. The rods have a colored band on the end to identify the type of rod. The color code is:

Color	Additive
Green	Pure
Red	Thoriated
Black, Gold or Blue	Lanthanated
White or Brown	Zirconiated
Orange	Ceriated
Gray	Rare Earth

Note - The color code can very between countries.

Thorium is a radioactive element and can be dangerous to your health. Always use a dusk mask when grinding or cutting thoriated tungsten. Carefully clean any grinding dust and wash your hands. Be careful not to inhale or ingest dust from thoriated tungsten. If your spark gaps use thoriated tungsten, always run them in well ventilated locations.

Before you start to panic, please be aware that thorium is actually quite safe. It's used in very small amounts (2%) in thoriated welding rods. It emits alpha radiation, which is usually not harmful. Alpha radiation is very weak and non-penetrating. The thoriated welding rods are not radioactive because the tungsten blocks any radiation given off by the thorium in the rod. However, thorium can be harmful if you inhale or ingest dust caused by grinding or cutting the rod. But again, it's not a danger if you avoid the dust.

Although tungsten seems perfect for spark gaps, it can be expensive. It's very hard and fairly brittle. I've had some difficulty cutting tungsten welding rods. A hack saw will not work. A Dremel cut off wheel is difficult, but seems to be the easiest cutting method I've found. The welding rods tend to crack easily when stressed. I've been informed that tungsten rods can be easily snapped to size using two pairs of pliers, or pliers and a vise. After being snapped or cut to size, the ends should be ground or sanded to a nice round shape so they arc consistently.

Whatever type of spark gap you choose, it will need to be adjusted for optimum performance. The adjustment procedure is outlined in the spark gap adjustment section.

The primary coil is used with the primary capacitor to create the primary LC tank circuit. The primary coil also couples to the secondary coil to transfer power from the primary to the secondary circuit.

Typically 1/4 inch copper tubing is used to make the primary coil. I've used 6 AWG solid copper successfully, although my hands were sore for a few days after bending the wire. Some people have used flat copper ribbon to save space, but tapping the turns (attaching a wire) can be more difficult. Avoid using other metals like steel due to it's higher resistance at high frequencies. Leave about 1/4 inch spacing between turns. This will prevent arcing and allow space for a tap point. The primary coil can be constructed on just about any non conductive material. The material should be strong enough to support the weight of the copper. You'll need a form with some means to hold the copper turns in place. Plastic wire ties or plastic bars with notches every 1/4 inch are common. If you get copper tubing or wire that is coiled or wound on a spool do not unwind it before making the primary coil. Use the natural shape of the coil to help do the winding. Try not to straighten and bend the tubing or wire too much as this will cause it to harden.

The primary coil is usually flat, called a "pancake" coil. A cone shape, or conical primary is also very common. Some smaller Tesla coils can use a vertical helix shaped primary. Generally, larger Tesla coil use flat primaries and smaller coil can use cone shape primaries. I recommend using a flat coil. Flat coils are easier to build and the conical / vertical helix shapes will raise the top of the primary coil closer to the top load which increases the chances of an arc striking the primary coil. The conical and vertical helix shapes will also increase coupling between the primary and secondary coils. Maximum coupling is usually the goal in most transformers, but Tesla coils need to be loosely coupled. Over coupling (or poor RF grounding) can cause arcing up and down the secondary coil. If you see arcs running up your secondary coil then the primary and secondary coils could be over-coupled and they should be moved further apart. An easy way to do this is to simply raise the secondary coil up a bit. If a conical primary is used the angle should not be greater than 45 degrees.

The primary coil should have a strike ring about 2 inches above the outer most turn. This ring will hopefully stop arcs from the top load from reaching the primary coil. An arc strike to the primary coil can produce a voltage spike large enough to kill the primary caps and / or NSTs. The ring should not be completely closed. One end should attach to the secondary earth ground. Smaller coils that do not produce arcs long enough to reach the primary coil do not require a strike ring, although it never hurts to have one.

Before construction the primary coil you should know how many turns will be required to tune the coil and the length of tubing or wire you'll need. The TeslaMap program can help you easily design your primary coil.

The secondary coil and the top load create the secondary LC tank circuit. The secondary coil also couples to the primary coil and transfers power from the primary circuit to the secondary circuit.

The size of the secondary coil is generally governed by the size of the power supply. For an average sized Tesla coil (about 1kW) you'll want a 4 inch to 6 inch diameter secondary coil. Smaller coils should have about 3 inch to 4 inch diameters, while larger coils should have at least a 6 inch diameter. The height to width ratio (also known as the aspect ratio) is important. If the coil is too short then you'll get a lot of strikes from the top load to the primary coil. The height of the secondary coil should be about 4 or 5 times the diameter in an average sized Tesla coil. For example the secondary coil on a 1kW Tesla coil with a 4 inch diameter should be about 16 to 20 inches high. Remember to cut the secondary form a couple inches longer than the winding height to leave some space on each end! Smaller coils should have a height to width ratio close to 6:1, while larger coils should be closer to 3:1.

The secondary wire is typically thin (22 AWG to 28 AWG) magnet wire. Magnet wire is solid copper wire with a thin coating of varnish as an insulator. It's sold by the pound or the gram. You'll probably need about 2 pounds to wind a typical coil. Double build magnet wire is available with extra insulation, but it's not necessary. Aim for about 1000 turns (+-200) on the secondary coil.

The secondary coil is usually wound on PVC pipe, although cardboard and many other non-conductive materials can be used. White PVC pipe is almost always safe to use. Gray PVC is usually safe, but black PVC may contain larger amounts of carbon which can create problems in some Tesla coils. Some PVC may come with a thin metal strip in it. This is used to help find the pipe after it's buried. Do not use this pipe as the metal strip will quickly short out the coil. In fact you'll want to avoid any metal screws, bolts, plates, etc on the secondary. A non-conductive nylon bolt can be used to attach the top load to the secondary coil.

Before you begin winding your secondary coil, you should calculate how may turns you can wind with the given weight of magnet wire and how long (or tall) the coil will be. This will tell you how much PVC pipe or other form material you'll need. Of course, the TeslaMap program can make all the calculations for you. The PVC pipe should be clean and dry.

Winding the coil will take quite a while. Find a comfortable spot with good lighting and plan to be there for quite a while. A lathe is ideal for holding the PVC pipe while you wind the magnet wire on. Unfortunately, the lathe I used, even on it's slowest speed, was rotating too fast to wind the coil, so I just chucked the pipe in the lathe and rotated the pipe by hand. The spool of magnet wire should be mounted so it will be easy to unwind and untangle during the winding. You may want to wear a thin glove to save the skin on your fingers. Before you start winding the coil, be sure the PVC pipe or other form is clean and dry. Be sure there's no metal shavings stuck on the form. It's probably a good idea to throw a coat of Dolph's AC-43, polyurethane or varnish on the form, inside and out to make sure it stays dry. Start by securing the end of the magnet wire a few inches from the end of the PVC. You can secure the wire with tape or drilling a couple small holes in the PVC and threading the wire through. Be sure to leave about a foot or two of magnet wire unwound on the end. Have some tape handy to easily hold the wire for rest breaks or untangling. Be careful not to leave any space between the windings. Make sure the wire lays flat and straight. Keep some tension on the wire as you wind it. Tape the end of the magnet wire down when finished and leave a couple feet of extra wire on each end. Hopefully if your calculations were correct you have just about a few inches of PVC pipe left on each side. Start coating with Dolph's AC-43, polyurethane or varnish. Remember not to coat the foot of extra wire on each end. I usually coil this extra wire up and let it stick up and out of the way while I varnish around it. Follow the instructions on the Dolph's AC-43, polyurethane or varnish and apply several coats. Keep the pipe rotating as the coating dries. A lathe is ideal, but I've used a hand drill on slow speed to rotate my PVC pipe. You can use other epoxies or sealers, as long as they're non-conductive and won't eat into the magnet wire insulation or PVC pipe.

The top load is acts as a capacitor in the secondary circuit.

The shape of the top load will help determine where the arcs will break out. The doughnut or toroid (also called a torus) is the preferred shape for the top load. As the coil operates a charge will build up around the surface of the top load. A sphere will have an evenly distributed field strength over it's entire surface. By flattening the sphere into a toroid, the field strength will increase around the radius of the toroid. The arcs will break out where the field strength is greatest. The benefit of concentrating the field around the radius is to help direct the arcs outward. Using a sphere will result in more evenly distributed, but smaller arcs.

The size of the top load and the amount of power applied will dictate the size and number of simultaneous arcs that the Tesla coil produces. If the top load is small relative to the input power, then it will produce many simultaneous, shorter arcs. As the size of the top load is increased the number of arcs will be reduced and the arc length will increase. If the toroid is too large the field strength will not be strong enough to allow any arcs to breakout. Placing a sharp pointed object like a thumb tack or a small metal ball (called a break out point) on the toroid will create a disruption in the field and allow the arcs to escape from the break out point.

The most common method of toroid construction is to wrap aluminum dryer duct around an aluminum pie pan. You can also buy a spun aluminum toroid. A top load can be made of practically anything with a smooth shape covered in aluminum foil. Avoid using "metal" paint. Usually there is not enough metal in the paint to create a conductive surface, and even if there is sufficient metal, it's usually quickly burned off.

Generally the diameter of the toroid ring should be about the same as the secondary coil, meaning a secondary coil wound on 4 inch PVC pipe should use 4 inch diameter dryer duct. The overall diameter of the toroid should be about 4 times the ring diameter, so 4 inch diameter dryer duct should be wrapped around an 8 inch pie pan for a total overall diameter of 16 inches.

It's important to physically attach the toroid to the top of the secondary coil. You can get by with just sitting the toroid on top of the secondary coil, but eventually it's going to fall or get bumped off. At best you'll ding up the toroid or your primary coil, at worst there could be a short that blows out your primary caps or something else. A good way to connect the toroid to the secondary coil is to get a PVC end cap for the secondary coil, drill a hole in the middle and insert a nylon bolt sticking up. Drill a hole in the center of the pie pan and slide it onto the nylon bolt. You'll have to use nylon or some other non conductive bolt. A metal bolt will shoot an arc straight up. A wooden mount can be used, but wood should be avoided. Wood always has a bit of moisture and is slightly conductive. It can also swell, shrink, warp and crack.

It's important to have the toroid at the correct height above the secondary windings. If the toroid is too high, you'll see a corona develop near the top of the secondary windings. You may also see some little arcs from the top of the secondary coil. The corona and arcs can degrade the secondary winding insulation. If this is a problem try moving the toroid down. If the toroid is too low you may have frequent arcs striking the primary coil. In this case try to move the toroid up. If you can't find a suitable placement for the toroid you can try adding a smaller toroid just under the main toroid. This can help to prevent corona on the secondary windings and strikes to the primary coil.

Power factor correction (PFC) capacitors are used to correct the power factor of the AC supplied to the NSTs. When a circuit contains a large inductance or capacitance the voltage and current will be shifted out of phase, resulting in reduced efficiency.

The power factor will be degraded due to the large inductance in the NSTs. The capacitance in the PFC cap will realign the voltage and current phases. The amount of capacitance should be matched to the amount of inductance so the capacitance and inductance will cancel each other out. The PFC capacitance does not have to be exactly matched to the transformer. Often the PFC cap is smaller than the recommended size to reduce costs. If you have a suitable capacitor go ahead and use it, even if it's not large enough. Even a small amount of capacitance will help. Multiple small PFC caps can be wired in parallel to increase their capacitance. The PFC caps should be wired across the low voltage inputs of the NST. If you can't get any PFC caps the NSTs can be run without them.

Be sure to use only "run" type capacitors, as opposed to "start" type capacitors. Start capacitors are designed to only be used for short periods of time, to start a motor for example. They will overheat and possibly explode if run continuously. Electrolytic caps should not be used as PFC caps, they'll also heat up and pop.

PFC caps can be found in salvage / recycling centers on AC motors, washing machine motors, refrigerator motors, etc. I believe it's against the law to bury PFC caps because they contain hazardous chemicals, and recycling centers will usually have a pile of them waiting for you. PFC caps can also be ordered on the Internet.

The wire in the NST secondary coil is very, very thin and easily shorted by high voltage spikes generated in the primary circuit. A spark gap and a low pass filter will help protect the NSTs from voltage spikes and premature death.

I've been using a filter known as the "Terry filter" designed by Terry Fritz for several years with great success. Several other people have also had good success with the filter. The filter is a typical RC low pass design that consists of several caps wired in series to shunt high frequency spikes to ground and high power resistors to decouple the NSTs from the primary circuit. My filter has 1000 ohms of resistance and 0.28nF of capacitance resulting in a cutoff frequency of about 570 kHz. A spark gap allows high voltage spikes to pass to ground. The spark gap should be set just wide enough so it does not short when connected directly to the NST output. I omitted the MOVs in my filter. They'll shunt voltage spikes to ground. Each cap has a high resistance (10 Mohm) bleeder resistor across the leads. I recommend using high voltage resistors. The bleeder resistors should not be in direct contact with the capacitor case as arcing can occur. Several caps are wired in series to handle the high voltages from the NST output. The total voltage rating of the series caps should be about 2 to 3 times the peak voltage of the NST output, although good quality caps can be run at their rated voltage. For example, using a 15kV RMS power supply (15000 * 1.414 = 21kV peak).

The type of caps used is not quite as important as cap selection in the MMC. Polypropylene film foil type are preferred. Metalized caps should be avoided.

Line filters are used to prevent high voltage spikes from traveling back into the house or building wiring.

Line filters usually consist of a capacitor to shunt the high frequencies to ground. Most will also use inductors to cut down the high frequency spikes. Some may have MOVs to shunt voltage spikes to ground.

The line filter should be wired in series with the mains power. It should be wired as far from the Tesla coil as possible. If it's wired too close, the wires behind the filter may have induced voltages that bypass the filter. When wiring the filter some people recommend wiring the filter in reverse (the output leads to the house wiring). The logic being that the filters are normally used to protect a device from spikes in the house wiring, but we're using it to protect the house wiring from the device. Other people recommend the standard connection orientation. I think it will work either direction, but I'll let you decide.

Filters can be bought on the Internet or salvaged from equipment. It's possible to design and build your own, but it's usually much easier to buy one. Be sure to use a filter that's rated for the power supplied to it.

All the individual components that make up a Tesla coil (the NST, MMC, spark gaps, etc) should be mounted in some sort of chassis, frame, or enclosure. It's possible to lay all the parts out on the floor (as I use to do) and run the Tesla coil without any chassis, but using a chassis has many advantages. It'll be much easier to move the Tesla coil. Mounting wheels to the bottom of the chassis is a good idea. Mounting the parts on a chassis will prevent them from moving or falling. The parts and will be better organized and the wiring can also be more organized, more permanent and safer.

The most common chassis design is several plastic or wood platforms stacked with enough room between platforms to accommodate the parts. For example the bottom platform will hold the NSTs, and PFC caps. The second platform will hold the NST protection filter and the MMC capacitor array. The next platform will hold the main spark gap. The next platform will support the primary coil and the secondary coil. A box can also used.

The chassis is generally made of wood, plastic or some other non-conductive material. It's needs to be structurally stable to support the weight of the components.

You should have access to a good set of tools and equipment. A good workshop or garage with a nice workbench helps. The tools you'll use can vary depending on your choice of materials and construction technique. You should also have the experience or assistance to safely use the tools. I'll list a few things that you're likely to need.

• Soldering iron and solder
• Multimeter
• Drill
• Saw for cutting plywood
• Hacksaw for cutting wood and PVC
• Wire cutters, pliers
• Measuring tape, calipers, ruler, etc
• Screwdrivers, sockets, wrenches, etc
• Epoxy or glue

I'm sure you'll use many other tools, but this should get you started.

All wiring should be as short an possible. Avoid loops which will create inductance in the wire. Try not to run wires parallel or close to each other which can induce current in adjacent wires.

It's important to use the appropriate type of wiring. All the wiring between the electrical outlet to the lower voltage side of the power supply transformer (NST) should be 120 / 240 volt power wiring of the appropriate gauge. It's usually acceptable to use old power cords, extension cords, etc. House wiring (Romex) can also be used, however I find that it's generally stiff and difficult to work with.

On the high voltage side of the power supply transformer all wires should be high voltage "GTO" or "EHT" wire. Low resistance spark plug wire can also be used. Although I don't recommend it, wire with low voltage insulation can be used, but you'll need to carefully route it away from anything conductive or grounded. The primary circuit will carry very high current, however, the current occurs in relatively short pulses so thick gauge wire usually isn't necessary.

The wire connecting the bottom of the secondary coil to the ground should ideally be braided copper wire. The braided wire will be a better conductor due to the skin effect and high frequency in the secondary coil. However, I've used solid and stranded copper wire with acceptable results.

All connections should be clean. Soldering is the best way to connect wires and leads. When high current flows through a connection, it does not take much resistance to create enough heat to burn the connection. A bad connection will reduce the efficiency of the coil and can possibly start a fire!

Grounding is very important for safety and proper operation of a Tesla coil.

The Tesla coil should have two separate grounds. The first ground is the house or building ground (also known as mains ground). This is the green wire in the electrical outlets. The second ground is RF ground. You'll have to create your own RF ground for the Tesla coil.

Proper grounding of the Tesla coil components has been debated for quite some time. The general consensus is to connect anything you will touch during operation of the Tesla coil to the house or building ground. The secondary coil and anything that may be struck by an arc, or that may experience high voltage transients, should be connected to RF ground. You may refer to the Tesla coil schematic. The general idea is to use the RF ground to complete the secondary LC circuit (the ground plane and the toroid will act as capacitor plates), and to pass all high voltage generated by the Tesla coil to the RF ground. This will hopefully prevent any high voltage spikes making their way into the house or building wiring. The NST seems to be a good boundary between the house wiring and the Tesla coil wiring because the primary and secondary windings are basically isolated from each other. Therefore, anything connected between the house outlet and the NST primary (variac, control panel, line filter) should be grounded to the house ground. The bottom of the secondary coil, the primary strike rail, NST protection gap and filter should be connected to RF ground. Grounding the NST case seems to have caused the most confusion. I recommend connecting it to the RF ground because it's usually more likely to be struck by an arc or experience a voltage spike.

It's important to have a good RF ground. I'll list several ways to create a RF ground, generally in order from most to least preferred. The best RF ground is a metal grounding rod that you pound into the ground. Although there is already a ground rod installed outside homes and buildings, you should not use this rod because it's connected to the house or building ground. You'll have to pound in your own ground rod. The grounding rod should be as close as possible to the Tesla coil, and as far from the house or building ground rod as possible. Generally 6 or 8 foot depth is recommended, but it really depends on soil conditions and other factors. Deeper is always better. Several shorter ground rods can be placed around the Tesla coil if a single rod can't be used. If the ground is very hard or rocky, you may bury your ground rod horizontally at a depth of 1 to 2 foot. If a ground rod is not possible you can create a "counterpoise" ground by placing a large piece of metal plate, chicken wire or mesh under the Tesla coil and use it as your RF grounding. The radius of the plate or mesh should be approximately equal to the height of the secondary coil and top load. If you're on a concrete foundation with rebar (like a garage) you may be able to connect to the rebar in the concrete foundation. This is known as a Ufer ground or a "Concrete Encased Electrode". If you're on a ground floor that's at least semi-conductive you can wet a small area of the floor and put a layer of aluminum foil down connected to your RF ground. This is not recommended for safety reasons and you'll have to use this method at your own risk. As a last resort you can connect to a cold water pipe, but I do not recommend doing so. Putting RF into cold water pipes isn't safe.

Braided copper wire can help the conductance of the RF ground, but regular wire will work just fine. Wetting the ground around the ground rod before running the coil helps conductivity to the earth. Be careful not to damage underground utilities when hammering in a ground rod. Poor RF grounding may not have any apparent effect on the Tesla coil - or it could cause reduced arc length, arcing up the secondary coil, or arcing between the primary and secondary coils.

There is no absolute rule for proper Tesla coil grounding. It's your responsibility to understand the electrical principals of grounding, seek advice and information, take your situation into account and anticipate potential dangers.