my logo

DMX 512

switch pack
strobo mod

high voltage

tesla coil



...the best of Henne's Sites in english...

tesla coils

The tesla coil was developed by Dr. Nikola Tesla - the inventor of the technical alternating current - around the next to last turn of the century. He even called it his most important invention, because it put the foundation-stone for his "free energy concept". In this concept, Tesla described a wireless power supply, which he wanted to manage with the help of the "magnifying transmitter". The magnifying transmitter - his advancement of the tesla coil - was a very high performance ELF-transmitter, which should transmit on a harmonic of the Schumann frequency. He claimed to have determined this frequency during a thunderstorm from a resonance behavior of the earth. There was also a press demonstration of this installation, during which he wirelessly brought 200 bulbs to shine.

While concrete construction plans for the magnifier are no longer present, an easy reproduction of his tesla coil is possible. High frequency output voltages of several megavolts (1.000.000V) are possible. This high output tension might not be lethal, since the HF current cannot penetrate into the body due to the skin effect. Therefore a lightning strike frequently results "only" in serious burns. (this does not apply to the primary circuit of a tesla coil: There the voltage is definitely lethal due to the 60 Hz!)

typical schematic of a tesla coil

The function of a tesla coil is quite simple: A high voltage transformer with an output voltage of at least 7500V and 20mA  charges a high voltage capacitor. If this reaches a certain voltage, a spark gap fires and the cap discharges through the primary coil. This consists of few turns of thick wire. Primary and capacitor form the primary resonant circuit. This resonant circuit induces its energy into the secondary resonant circuit, which consists of the long secondary coil and a head capacity. The  frequencies of both resonant circuits must be identical for an effective coupling. A high output voltage is reached, because the inductance of the secondary is enormous in relation to the primary coil inductance.

Now to the individual parts of a tc:

high voltage transformer

As a source of high voltage mostly old ignition transformers from oil burners (OBITs) are used. They supply 10kV at 20mA. Alternatively you can use xfmrs for neon signs (NSTs). They have an output voltage of approx. 7,5kV at 60mA. You can feel lucky, if you own a small distribution transformer (pole pig) or a potential transformer, because these xfmrs are rated at 10-15kV and a power of more than 5kW! You can increase the spark length of your coil by switching the outputs of several xfmrs in parallel, so increasing the current. Since it is quite difficult to get these powerful transformers in Germany, many coilers work also with two transformers from microwave ovens (MOTs) producing approx. 12kV (switched in series and with a level shifter). I have appropriate schematics posed under "MOT Supply" on my site...


Capacitors which stand several kilovolts at high frequency usually are too expensive. Therefore most coilers connect many MKP caps in series to increase the dielectric strength. Such a cap is called "MMC" (multi mini cap). Because of the HF current you should oversize the MMC about 20% of the dc rating. And even then reliable continuous operation cannot be guaranteed... The advantages are  a small space requirement and no losses by heating up or corona.

If you want to build a cap that is more simple and cheaper, I recommend saltwater caps - also called Leyden jars: To build one, you have to wrap a glass bottle (1l) in alu foil up to the neck and to fill it with saturated saltwater.  Then you place these bottles into a suitable box, the bottom of which is covered with aluminum foil or a metal plate. So the bottom is used as electrical connection to the wrapped bottles. Long nails in the bottles serve as the second electrode. The glass is the dielectric of the capacitor. If the corona effects at the edges of the bottles disturb, you might give a bit of oil as an isolation layer into the saltwater.

A1l bottle has a capacity of approximately 1nF and a rating of > 30kV. An alternative are 1l PET-Cola bottles with a dc rating of 20kV and 2,2nF. Here the oil is important, because the corona destroys the plastic.

Although Tesla accomplished all of his experiments with Leyden jars, glass should be a worse dielectric than plastic or ceramics.

For maximum spark length the impedance of the capacitor should correspond to the impedance of the xfmr :

X=  XC

XT = U   &  XC =        1
                     I                    (2π·f·C)     


U = output voltage of the xfmrs; I = output current of the xfmrs; f = frequency; C = capacity

MMC Cap 4*3,5nF/16kV

spark gap

The spark gap is the only active part of a tesla coil. Therefore innumerable building suggestions exist.

Serial spark gaps - many small spark gaps one behind the other - and rotary spark gaps are used most frequently. With latter variant the spark jumps from a static electrode to a rotating wheel with many electrodes and from there to another static electrode.  This expenditure is done to quench the spark as fast as possible, as soon as the energy from the primary circuit was transferred into the secondary circuit. If the secondary reinduces the primary and this again the secondary (etc.), one speaks of a "quench failure". The optimal spark gap quenches after the first transfer of energy to the secondary circuit (1st notch quench). The following example shows, how important a good spark gap is: With a simple single spark gap - buildt with two screws - I got only a streamer of 8cm length. With a series spark gap made of four single spark gaps I obtained already light blue discharges of 35cm length. (6"-tc with an OBIT)

static series spark gap (4fach-Sucker-Gap)

2nd notch quench


There are three different kinds of primaries:

  • the normal Helix (just like the secondary): This form is rather unpopular among experienced coilers, since sparks can easily flash from the secondary into the primary, which leads to the destruction of the capacitor. In addition to this it tends to overcouple with the secondary...
  • the flat spiral: With this coil there should be no flashing overs. It is used very frequently even in larger tesla coils. Its disadvantage is the very strong em. field in its center, which decreases outward very fast. That entails a very strong demand for the isolation layer of the secondary.
  • the best choice is an "inverse conical saucer": this coil is a spiral essentially again, whose turns rise with 15-30°. This design ensures an even field and prevents flashing overs. Sketch:

o               o

o        o

o  o

As material either thick wire or thin copper tube is suitable. The smaller the resistance, the better is the efficiency. The conductor may be hollow due to the skin effect.

Along with the power of the xfmrs, there also was an increase of the capacity of the capacitors. In order to maintain the same frequency, the primaries became smaller. This led to a tremendous loss of energy in many of today's coils, since a tiny coil cannot produce a strong and even magnetic field. This phenomenon is called "undercoupling". On the other hand an "overcoupling" may be recognized by "racing sparks" on the secondary which destroy their isolation and by a visible ionization of air in the proximity of the tesla coil. In this case the secondary should be removed from the primary, until the critical point (no more racing sparks) is achieved: now the tc is optimally coupled!! So better choose the inductance a little bit too high than too low... (12-15 turns are a good appoximate value.)

The inductance can be calculated using the following formula:

L = µ r ·µ o ·n 2 ·A


L = inductance; µ r = 1; µ 0 = magn. field constant; n = number of turns; A = cross-section area of the coil; l = length of the coil

primary (test version) 10 turns


The inductance of the secondary can also be estimated with the formula mentioned above. The secondary consists of thin enameled copper wire and has 1000 -1600 turns. The aspect height : diameter should amount to approx. 4:1, in order to achieve the largest inductance with the smallest length of wire. This is necessary, in order to keep the resistance of the coil as small as possible to ensure a small oscillation damping of the secondary.

Please note that the voltage between two neighbouring turns may amount to max. 1kV. A coil that produces sparks with a length of 1m (approx. 1 MV) therefore needs at least - already due to smallest irregularities in the winding - 1300 turns!!

Under ideal conditions a standing wave spreads in the secondary. So the length of the conductor should be a quarter of the wavelength (dominant mode):

l D = C

        4 f

l D =  length of wire; f = frequency of the primary circuit; C = speed of light

This formula assumes the wave spreads with speed of light in the coil. The real value however deviates a little from this, so for quarter wave (Marconi-) antennas an empirical dilution factor is used. I have not found an appropriate dilution factor for tc's yet (for information I would be grateful!). For antennas a higher bandwidth can be achieved by mismatching with smaller efficiency. For tc's this means that you don't need to tune the resonance too exactly, but a quantity of power is lost; with optimal adjustment the secondary has only a  small "sweet spot" but a much higher efficiency.

The suitable wire size can be calculated with:

dD = 200

 dD = wire size (in mm); f = frequency of the tesla coil

PP is the best secondary coil form material because of its small dielectric losses. (PVC pipes are worse but still OK.) After winding the turns are fixed with several layers of insulating  spray  or epoxy resin.

Do not coat the coil form before winding!! The performance with a coating is not much better but if the coating is dissolved by insulating spray after the coil is wound, you get a problem...

4"-secondary (approx. 1400turns / aspect 4: 1)

head capacity

The natural frequency of the secondary circuit is much higher than that of the primary circuit due to the small self-capacity. To lower the frequency you need a head capacity. For this all sorts of metal spheres are suitable. Otherwise there is also the torus design, made of a ring of "Aluflex". The capacity of a spere can be calculated with:

C = 4πe0 · R

C = capacity; R = radius of the sphere; eo = el. field constant

Torus capacities are mostly used, because their electric field shape prevents the last turns of the secondary from a spark breakout. Since the secondary can oscillate freely due to its loose coupling, you should build a large head capacity. Richard Hull - the technical designer of the famous Nemesis coil - claimed: the larger, the better!

The frequency of the two resonant circuits can be calculated by Thomson's oscillation formula:

f =       1
             2π √L·C

f = frequency; L = inductance; C = capacity

In practice it worked satisfying to determine the actual point of resonance by tapping different turns of the primary coil. If the coil works, you found the point of resonance.

If possible avoid wood and metal plates in the coil proximity, otherwise energy is wasted because of eddy currents.

The only coil fulfilling - according to my opinion - all conditions is the "Nemesis" made by Richard Hull, which broke all records in efficiency...

our first version of the tesla coil (max. 10cm output...)

Any adjustments of the coil should only be made with pulled mains plug.

When the coil is off, capacitors should be short circuited. Otherwise they may get a lethal charge during the dwell time (dielectric memory effect).

The reproduction of my tesla coil as well as your own experiments on the basis of this site are performed on your own responsability and risk.

Tesla discharges are impressing, but very dangerous!!!

so the result should look then

WebRing Logo

Next Prev Skip Next Skip Prev
Random Next 5 List Sites Join Ring