Chapter 8

The Tesla Coil — A Resonant RF Machine

Strip away the mythology. This is what it actually is, and why it works.

What a Tesla Coil Actually Is

A Tesla coil is a resonant air-core transformer optimized to produce high-frequency, high-voltage output. That sentence contains three operative words: resonant, air-core, and high-frequency. Each distinguishes it from a conventional power transformer.

A conventional power transformer (Chapter 6) uses a laminated ferromagnetic core to tightly couple primary and secondary windings. The core concentrates flux, maximizes the coupling coefficient k (approaching 1.0), and operates at line frequency — 50 or 60 Hz. The output voltage is determined almost entirely by the turns ratio N₂/N₁. There is nothing resonant about it.

A Tesla coil discards the iron core entirely. Primary and secondary are wound concentrically with an air gap between them. The coupling coefficient k is deliberately set low, typically 0.10–0.20. The circuit is then operated at the resonant frequency of both LC loops simultaneously. This design choice produces voltages wildly beyond what the turns ratio alone would predict.

The Six Components Power supply. Any source capable of charging the primary capacitor to several kilovolts. A neon sign transformer (NST, typically 9–15 kV at 30–60 mA) is the standard for amateur builders because NSTs have inherent current limiting. A microwave oven transformer (MOT) is more powerful but is not current-limited and is significantly more dangerous.

Primary capacitor (C₁). Stores the energy that drives each oscillation. Polypropylene film capacitors are standard — they tolerate the high peak currents that surge through the primary circuit on each spark-gap firing. Electrolytic capacitors will fail immediately.

Spark gap. Acts as a fast, high-power switch. When the voltage across C₁ exceeds the gap's breakdown voltage, the gap ionizes, its impedance drops from megohms to a fraction of an ohm in nanoseconds, and C₁ discharges into L₁. The gap must quench — go non-conductive — within one to two half-cycles of the oscillation, or energy flows back from the secondary and is dissipated.

Primary coil (L₁). Few turns (4–12), wound from heavy copper tubing or heavy-gauge wire (8–10 AWG). Peak currents at resonance can reach hundreds to thousands of amperes for modest power levels. Wound as a flat Archimedean spiral or conical helix. Always has a tap point to adjust turns and tune f₀.

Secondary coil (L₂). Many turns (800–1500 typical), wound from fine wire (22–28 AWG) on PVC or acrylic. Sealed with polyurethane varnish to prevent inter-turn corona.

Top load (C₂). A conductive toroid at the secondary's top terminal. Its geometry determines its capacitance to ground (typically 10–30 pF). Combined with the secondary inductance, it sets the secondary resonant frequency — which must match the primary.

The circuit topology is two coupled resonant LC loops: primary (L₁, C₁) and secondary (L₂, C₂), both tuned to the same frequency f₀. When the spark gap fires, C₁ discharges into L₁, creating a decaying oscillation at f₀. The oscillating current in L₁ radiates a time-varying magnetic field linking with L₂. Because the secondary is tuned to the same f₀, it responds strongly — energy transfers over multiple cycles, the secondary voltage builds, and continues until the voltage exceeds the dielectric strength of surrounding air or all energy has transferred. For a well-tuned coil with a rotary gap, energy transfer efficiency approaches 90% [Corum & Corum, 1987].

LC Oscillation and Resonance — The Physics

An inductor of inductance L stores energy in its magnetic field:

EL = ½ L I²

A capacitor of capacitance C stores energy in its electric field:

EC = ½ C V²

In an ideal LC circuit, energy oscillates between these two storage elements without loss. The oscillation frequency is:

f₀ = 1 / (2π √(LC))

This follows from the equation of motion for charge q on the capacitor:

L d²q/dt² + q/C = 0

This is the harmonic oscillator equation. The solution is q(t) = Q₀ cos(ω₀t + φ), with ω₀ = 1/√(LC), giving f₀ = ω₀/(2π). This is not an approximation — it is exact for an ideal LC circuit. For a real circuit with resistance R, the resonant frequency shifts slightly downward to ωd = √(ω₀² − (R/2L)²), and the oscillation decays. For Q >> 1, the shift is negligible [Terman, 1943, p. 135].

At resonance, inductive reactance XL = ωL and capacitive reactance XC = 1/(ωC) are equal and opposite. The circuit presents purely resistive impedance — ideally zero if resistance is zero. In a Tesla coil, the secondary is a series resonant circuit: at f₀, its impedance is minimum and its response to the driving field is maximum.

Q Factor The Q factor is the dimensionless ratio of energy stored to energy dissipated per radian of oscillation: Q = ω₀L / R = (1/R) √(L/C) = f₀ / BW where BW is the −3 dB bandwidth of the resonance peak. A Q of 100 means the circuit loses 1% of its stored energy per radian; the resonance peak is sharp. For Tesla coil secondaries, Q values of 50–300 are typical [Corum & Corum, 1987]. The limiting resistances are: DC wire resistance, skin-effect resistance at the operating frequency, and radiation resistance.

Skin effect in 28 AWG at 200 kHz: skin depth δ = √(2ρ/ωμ) ≈ 0.148 mm. The wire radius of 28 AWG is 0.160 mm — slightly larger than the skin depth. Skin effect raises AC resistance modestly (factor ~1.1) for this wire gauge at this frequency. For the heavy primary copper tubing, skin effect is more severe per unit volume, but the primary Q (10–50) matters far less than secondary Q.

Resonance as coupled pendulums. Two pendulums on the same string, tuned to the same period: swing one, and energy flows to the other, which swings higher and higher while the first goes quiet, then reverses. This is exactly the Tesla coil's energy behavior. The coupling is the mutual inductance M = k√(L₁L₂), where k is the coupling coefficient (0.10–0.20). The number of primary oscillation cycles needed for complete energy transfer is approximately 1/(2k). For k = 0.15, this is about 3.3 half-cycles [Corum & Corum, 1987, p. 14].

Voltage Multiplication — The Full Picture

The turns ratio N₂/N₁ is not the dominant factor in Tesla coil output voltage. Students trained on transformer theory expect V₂/V₁ = N₂/N₁. The Tesla coil operates as two resonantly coupled oscillators, and voltage multiplication must be calculated from energy conservation.

Energy conservation approach. When all primary energy has transferred to the secondary:

½ C₁ V₁² = ½ C₂ V₂²

Solving for the secondary voltage:

V₂ = V₁ √(C₁ / C₂)

This is the theoretical maximum. With C₁ = 10 nF, V₁ = 15,000 V (peak from a 15 kV NST), and C₂ = 10 pF (top-load capacitance):

V₂ = 15,000 × √(10,000 pF / 10 pF) = 15,000 × 31.6 ≈ 474,000 V

Nearly half a megavolt from a neon sign transformer. This matches empirical observation: medium NST-driven coils with 6-inch secondaries routinely produce 300–500 kV and streamers 24–36 inches [Freau & Lowndes, 1968].

Two modifying factors:

Coupling efficiency. Spark gap losses and primary circuit resistance reduce V₂ by a factor of √(efficiency). Practical coupling efficiency for a well-designed rotary gap coil is 80–90% [Corum & Corum, 1987] — about an 8–10% reduction in V₂, not dominant.

Quench timing. If the gap does not quench at the first current zero, secondary energy flows back into the primary and is dissipated. Rotary gaps quench reliably. The first notch in the primary current waveform is the ideal quench point: all primary energy has moved to the secondary at that instant.

Typical Secondary Performance

Secondary DiameterApprox. TurnsTypical Peak VoltageStreamer Length
3 inch900100–200 kV6–12 inches
6 inch1,200300–500 kV24–36 inches
12 inch1,5001–2 MV6–10 feet
Colorado Springs (51-ft primary)—~12 MV (Tesla's estimate)30–40 feet

Colorado Springs voltage from Tesla's notebook [Tesla, 1978, p. 61]; supported in order of magnitude by Corum & Corum [1987].

Streamers — What They Actually Are

A streamer is a propagating plasma channel in air. It is not a flame, not a spark in the technical sense, and not a continuous arc — though it can become one if the energy source sustains it.

Initiation. Air at atmospheric pressure breaks down at approximately 3 MV/m for uniform fields (Paschen's law, tabulated in IEC 60060-1 [IEC, 2010]). Near a toroid edge or sharp point, the field concentrates and the local threshold is lower. A free electron — always present from cosmic-ray ionization (~1 ion pair/cm³/s at sea level) — is accelerated by the field. If it gains sufficient energy between collisions to ionize the next molecule it strikes, an electron avalanche begins. The avalanche produces a conducting channel where the streamer tip acts as an extension of the electrode. The streamer propagates outward at 10⁵–10⁶ m/s — not because electrons travel that fast, but because it is the wavefront of enhanced field and ionization that propagates [Raizer, 1991, p. 327].

Branching. Ionization is a stochastic threshold phenomenon. Small inhomogeneities in air density and humidity create multiple competing tips. When one branch grows faster, it screens the others by reducing the local field nearby. The result is a fractal branching pattern whose spatial statistics follows a Laplacian growth model [Niemeyer, Pietronero, Wiesmann, 1984].

Color. The blue-white luminosity comes from emission of excited molecules in the discharge channel:

The visual blue-white is the blend of N₂ violet bands with weaker oxygen lines [Bibinov, Fateev, Wiesmann, 2001]. The UV component (337 nm) is significant, invisible to the eye, and dangerous to unprotected skin and eyes with prolonged exposure.

Sound. Each streamer event heats a thin column of air to several thousand Kelvin in microseconds. The rapid pressure expansion is a miniature thunder clap. Repeated at spark rates of 60–500 per second, these merge into the characteristic crackling hiss. The sound spectrum is broadband, peaked in the 1–10 kHz range.

The Faraday Cage — the skin effect as a party trick. Science museums worldwide demonstrate Tesla coil (or Van de Graaff generator) discharges using a person inside a metal cage. The discharges crawl spectacularly across the cage exterior while the person inside sits unaffected. This is the skin effect and the Faraday cage principle working together: at high frequencies, all current flows on the outer surface of any conductor. A metal cage is a conductor. The electrical energy stays on its surface — it has no path to the interior, and no reason to take one. The Museum of Science in Boston's Theater of Electricity has done this demonstration since 1956 using the Van de Graaff generator built by Robert Van de Graaff at MIT in 1931. Faraday himself demonstrated the shielding principle in 1836. The cage bears his name. The visual drama is Tesla coil–era physics made visible at human scale.

Why a fluorescent tube lights at a distance. The tube contains low-pressure mercury vapor (ionization potential 10.4 eV vs. 15.6 eV for N₂). The high-frequency electric field from the Tesla coil secondary penetrates the glass envelope and excites the mercury discharge without any physical contact. This also demonstrates that the high-frequency E-field extends meters from the top load — the hazard region is larger than the visible streamers suggest.

Safety Primary circuit (C₁, gap, L₁): 9–15 kV from an NST; stored energy ½C₁V² = ½ × 10 nF × (15,000)² = 1.125 J. Sufficient for cardiac fibrillation. Capacitor retains charge after power removal — always discharge with a shorting stick before approaching. This hazard is identical to any other high-voltage capacitor bank.

Secondary output: High-frequency (100 kHz–1 MHz) does not eliminate biological hazard. RF burns from streamer contact are second- and third-degree injuries at the contact point. At high power levels, RF current magnitude can disrupt cardiac rhythm [IEC 60479-2, 2007]. The skin-effect "safety" argument (Tesla made it himself in his 1891 AIEE lecture) is not a reason to contact the output.

Electromagnetic interference: The spark gap produces broadband RF across the full spectrum. Electronic equipment within meters — GPS, pacemakers, hearing aids, oscilloscopes — may be damaged or disrupted at tens of meters.

Operating procedure: Use a safety interlock (dead-man switch) that physically disconnects primary power. Never operate alone. Discharge C₁ before approaching any part of the circuit.

Tesla's Use of the Coil

Tesla did not build the resonant transformer as a demonstration device. In 1891, no practical RF source existed above 100 kHz. The spark gap Tesla coil was the first one.

His first patent on the resonant transformer dates to 1891 (US Patent 454,622). His lecture before the American Institute of Electrical Engineers that same year — published in the AIEE Transactions — is the technical introduction of the device to the engineering world [Tesla, 1891]. He demonstrated high-frequency discharges, illuminated phosphorescent bulbs wirelessly, and passed high-frequency current through his own body to light a lamp. (The skin-effect argument for safety, made publicly by Tesla in that demonstration, does not justify contact with high-power secondaries — see Safety note above.)

At Colorado Springs (1899–1900), Tesla built the magnifying transmitter on a scale large enough to test wireless power transmission. The laboratory housed a primary coil 51 feet in diameter and a 200-foot mast with a copper ball top load. His notebook — 400 pages of calculations and observations in his own hand, published by Nolit in 1978 [Tesla, 1978] — records operating voltages, frequencies, capacitor values, and experimental results in engineering detail that rewards careful reading. Entry for July 3, 1899:

Primary Source — Colorado Springs Notebook, p. 61 "Determined to-day that a potential of probably 12,000,000 volts is obtainable with the magnifying transmitter as it has been set up... the discharges extended from 30 to 40 feet in length and the streamers were of extraordinary breadth and luminosity." — N. Tesla

This is consistent with the energy-conservation voltage calculation applied to his known capacitor values and the physical evidence of the streamer lengths.

The Solid-State Tesla Coil — Modern Implementation

The spark gap erodes, produces broadband RF noise, is inefficient (gap losses are significant), and cannot be precisely controlled. By the 1990s, high-speed IGBTs and MOSFETs capable of switching kilowatts at hundreds of kilohertz made solid-state replacement practical.

SSTC (Solid-State Tesla Coil). The spark gap is replaced by an H-bridge or half-bridge inverter using IGBTs. The bridge drives the primary at the secondary resonant frequency, with frequency locked by a feedback oscillator sensing the secondary's own resonance (via a current transformer on the secondary base or antenna pickup). Self-resonance tracking is necessary because the secondary resonant frequency shifts when streamers form — the extending plasma column changes the effective C₂. The SSTC allows modulation: varying the duty cycle controls spark rate and therefore the audio output of "singing" Tesla coils.

DRSSTC (Dual Resonant Solid-State Tesla Coil). Adds a resonant primary capacitor C₁ matching the primary to a resonant condition, allowing primary current to ring up over several switching cycles before discharging into the secondary. The DRSSTC is the current state of the art for amateur high-performance builds and large public demonstrations.

ZVS (Zero-Voltage Switching) driver. For the experimenter's first Tesla coil: two N-channel MOSFETs in push-pull configuration with a resonant tank circuit that causes each MOSFET to turn on only when its drain-to-source voltage passes through zero, minimizing switching losses. Complete ZVS boards are available for $8–$15 online (search "ZVS driver 12V") and operate from 12–48 V DC. This is the starting point for Experiment 8.1.

The Magnifying Transmitter — Tesla's Advanced Concept

Tesla carefully distinguished between the "Tesla coil" (two-resonator system) and the "magnifying transmitter" (his primary invention — a three-resonator system). The three components of the magnifying transmitter:

  1. Driver transformer and primary capacitor — provides power at resonant frequency
  2. Primary coil — the driven resonant circuit
  3. Extra coil — a high-Q resonator, loosely coupled to the primary coil, that magnifies voltage

The extra coil is physically separated from the primary coil and is self-resonant at the same frequency. It is driven by conductive connection at its base to the primary circuit, which acts as a current driver. The extra coil's top voltage rises in proportion to its Q:

Vtop ≈ Q × Vbase

Tesla's theoretical claim was that the magnifying transmitter, operated at the resonant frequency of the Earth-ionosphere cavity, would couple energy into a global standing wave. The resonant frequency of the Earth-ionosphere cavity was not derived until 1952, when W.O. Schumann calculated f₁ ≈ 7.83 Hz [Schumann, 1952] — now called the Schumann resonance. Tesla's estimate of ~6–8 Hz for Earth resonance [Tesla, 1978, p. 142] was surprisingly close.

Whether the Colorado Springs machine actually excited the Schumann resonance is not established. Its operating frequency was 50–150 kHz [Corum & Corum, 1987], not 8 Hz. Corum and Corum argue the physics is consistent with Tesla's observed results if interpreted as coupling into the Earth's surface as a transmission line rather than the Schumann cavity [Corum & Corum, 1986]. This remains technically contested but peer-reviewed.

The practical result: the three-coil design produces higher output voltage for the same primary energy than an equivalent two-coil design. This is well established empirically and is used in modern high-performance amateur builds [Golka, 1989].

Experiment 8.1 — Build a ZVS Mini Tesla Coil

Materials

Procedure

  1. Wind the secondary. A power drill in a vise spinning the PVC form makes this manageable. 28 AWG is fragile — keep tension to a few ounces. Apply two coats of polyurethane varnish between layers and a final coat over the complete winding. Cure 24 hours.
  2. Connect the secondary's bottom terminal to ground. The top terminal connects to the toroid.
  3. Wind the primary around the lower 15–20% of the secondary body with an air gap (do not wind on the secondary form). The secondary winding should begin above the primary turns.
  4. Connect ZVS driver output terminals to the two ends of the primary coil. Do not apply power yet.
  5. Measure secondary resonant frequency (see Experiment 8.2). Adjust primary tap (number of turns in circuit) to match.
  6. Apply power. Streamers should appear from the toroid. If none appear after 30 seconds, adjust primary coupling (height) and retry. Do not run continuously for more than 60 seconds until you've verified MOSFET temperatures.

Expected Results

Safety

ZVS primary is only 12–24 V. Secondary can produce 50–100 kV. Do not contact the top load or streamers while powered. Use a fluorescent tube as a non-contact indicator. Keep operating time under three minutes per session until you understand the thermal limits of your specific MOSFET package.

Experiment 8.2 — Measure the Resonant Frequency of an LC Circuit

Purpose

Verify the formula f₀ = 1/(2π√(LC)) with a direct measurement. Develop intuition for how small changes in L and C shift resonance — the same adjustment you make when tapping the Tesla coil primary.

Materials

Circuit

Series connection: function generator output → 1 kΩ resistor → inductor L → capacitor C → return to generator ground. Monitor voltage across C on the oscilloscope.

Procedure

  1. Calculate theoretical f₀. Example: L = 200 μH, C = 47 nF → f₀ = 1/(2π × √(200×10⁻⁶ × 47×10⁻⁹)) ≈ 51.8 kHz.
  2. Set function generator to sine wave. Sweep from 0.5×f₀ to 2×f₀.
  3. Watch the oscilloscope. At resonance, voltage across C peaks sharply. Measure the frequency at the peak. Should match calculation within 5% (dominant error is capacitor tolerance, typically ±5–10%).
  4. Slide a ferrite core partly into the coil. Inductance increases; f₀ shifts downward. Observe the peak shift in real time. This is exactly how the tappable primary coil tunes a Tesla coil.

Measuring Q

At the resonant peak, note Vpeak. Find frequencies f₁ and f₂ where voltage = Vpeak/√2 (the −3 dB points). Then Q = f₀ / (f₂ − f₁). For a well-wound air-core coil with polypropylene capacitor, Q will be 30–100 at these frequencies.

References

  1. Tesla, N. Colorado Springs Notes 1899–1900. Nolit, Belgrade, 1978. Primary source — Tesla's own laboratory notebook.
  2. Tesla, N. "Experiments with Alternate Currents of Very High Frequency and Their Application to Methods of Artificial Illumination." AIEE Transactions, 1891. Reprinted in Nikola Tesla: Lectures, Patents, Articles (Nikola Tesla Museum, Belgrade, 1956).
  3. Tesla, N. US Patent 454,622. "System of Electric Lighting." Filed 25 Apr 1891.
  4. Tesla, N. US Patent 645,576. "System of Transmission of Electrical Energy." Filed 2 Sep 1897, granted 20 Mar 1900.
  5. Freau, R. and Lowndes, J. "Resonance Characteristics of the Tesla Coil." IEEE Transactions on Industry and General Applications, IGA-4(3), 276–279, 1968.
  6. Terman, F.E. Radio Engineers' Handbook. McGraw-Hill, 1943. Chapter 3: resonant circuits, Q factor.
  7. Corum, J.F. and Corum, K.L. "Vacuum Tube Tesla Coils." Corum & Associates technical report, 1987. The most rigorous engineering treatment of the magnifying transmitter available.
  8. Corum, J.F. and Corum, K.L. "A Physical Interpretation of the Colorado Springs Data." Proceedings of the International Tesla Symposium, 1986.
  9. Schumann, W.O. "Über die strahlungslosen Eigenschwingungen einer leitenden Kugel, die von einer Luftschicht und einer Ionosphärenhülle umgeben ist." Zeitschrift für Naturforschung A, 7(2), 149–154, 1952.
  10. Raizer, Y.P. Gas Discharge Physics. Springer, 1991. Authoritative reference on streamer physics, Chapter 12.
  11. Niemeyer, L., Pietronero, L., and Wiesmann, H.J. "Fractal Dimension of Dielectric Breakdown." Physical Review Letters, 52(12), 1033–1036, 1984.
  12. Bibinov, N.K., Fateev, A.A., and Wiesmann, K. "On the Influence of Metastable Reactions on Rotational Temperature in the Positive Column of Nitrogen and N₂/He Glow Discharges." Plasma Sources Science and Technology, 10(4), 579–588, 2001.
  13. Golka, R. "Long Arc Simulated Natural Lightning." IEEE International Symposium on EMC, 1989.
  14. IEC 60060-1:2010. High-Voltage Test Techniques — Part 1: General Definitions and Test Requirements. International Electrotechnical Commission.
  15. IEC 60479-2:2007. Effects of Current on Human Beings and Livestock — Part 2: Special Aspects. International Electrotechnical Commission.
  16. ARRL Handbook for Radio Communications, current edition. American Radio Relay League. Chapter 9: RF fundamentals, skin effect.
  17. Cheney, M. and Uth, R. Tesla: Master of Lightning. Barnes & Noble Books, 2001.