ZVS CW Tesla Coil - Complete Guide

A while ago I had this idea — take an induction heater coil, wind a secondary next to it, and use the high-frequency field to light up noble gases in glass tubes. The glowing plasma effect I had in mind looked amazing in my head. Then it hit me: what I was describing was literally a Tesla coil. A continuous wave solid-state Tesla coil, to be exact.

So I ordered a ZVS induction heater module online, rewound the primary, added a secondary coil, recalculated the capacitor values, and after tuning the resonance — it actually worked. This is my CW (Continuous Wave) Tesla coil, built around a modified ZVS driver.

The coil is not finished yet — I'm still improving it — but it already produces real arcs, and I want to share how to build one yourself.

What is a CW Tesla Coil?

A ZVS (Zero Voltage Switching) driver is a self-oscillating half-bridge circuit. Originally designed for induction heaters, it generates a high-frequency sine wave through a resonant LC tank — your primary coil and capacitor bank. The secondary coil sits nearby and picks up the oscillating magnetic field, stepping the voltage up to tens of kilovolts through resonant transformer action. That high voltage ionizes the air at the top of the secondary and produces plasma arcs.

The key word is resonance. The primary LC tank must oscillate at the same frequency as the secondary coil's natural resonant frequency. When they match, energy transfers efficiently and voltage builds up. When they don't match, almost nothing happens.

Optimal frequency:

The sweet spot for a ZVS driver is around 200-330 kHz. At this range the MOSFETs switch efficiently, losses are low, and streamer formation is good. My coil runs at around 500 kHz which is on the high end for a ZVS — it works, but efficiency is lower than it could be. If you are designing your coil from scratch, aim for 300 kHz. You will get better efficiency, cooler transistors, and stronger arcs at the same input power.

Core components:

1. ZVS induction heater module (standard board from AliExpress, 1000W rated)
2. 2× IRFP260N MOSFETs (stock on the board) — or upgrade to IRFP4768 for higher voltage operation
3. Capacitor bank — any polypropylene film capacitors rated for your target voltage. I kept the original capacitors from the induction heater board and recalculated how many to put in series to hit the right capacitance for my frequency
4. Secondary coil former — I used a PVC pipe, 75mm diameter, 200mm winding length. Any diameter works, just recalculate in JavaTC
5. Magnet wire for secondary — I used 0.38mm. Any gauge works, it just affects turn count and resonant frequency
6. Primary wire — I used 2.5mm copper wire. Any thick insulated wire works
7. Aluminium foil for ground connection
8. Topload or breakout point (a simple bolt works)

Power supply:

I powered the coil from a lab power supply feeding a boost converter, which let me adjust voltage from 20V all the way up to 60V and beyond. This combination is ideal for tuning because you can start low and raise voltage gradually while monitoring current and arc behaviour. Any adjustable supply works — the boost converter just extends the range if your base supply voltage is limited.

Tools:

1. Oscilloscope (essential for tuning)
2. Adjustable lab power supply with ammeter, OR fixed supply + inline ammeter
3. Boost converter (optional but very useful)
4. JavaTC calculator (free, online at http://javatc.teslacoil.co.nz)
5. Multimeter

Step 2: Build the Secondary Coil

Wind your magnet wire onto the PVC pipe former. I wound 1000 turns of 0.38mm wire onto a 75mm diameter pipe, giving a winding height of around 20cm. Keep the winding tight and even. Seal the finished winding with varnish or epoxy to prevent arcing between turns.

The exact dimensions are up to you — any coil geometry works. Just enter your dimensions into JavaTC before winding to find out the resonant frequency. That frequency is what you will need to match with your primary and capacitor bank.

If you are designing from scratch, choose your dimensions so that JavaTC shows a resonant frequency close to 300 kHz for best results with a ZVS driver.

The bottom terminal of the secondary coil needs a ground connection. I used a piece of aluminium foil laid flat on the table under the coil, connected to the negative terminal of the power supply with a wire.

That is all that is needed. The foil acts as a ground plane and gives the RF current a return path. No earth ground or special grounding equipment is required for a small coil like this.

The primary coil and capacitor bank together form the resonant tank circuit. Their combined resonant frequency must match the secondary.

Calculate your capacitor value:

Where f is your secondary resonant frequency and L is your primary inductance. You can calculate L using JavaTC. Any polypropylene film capacitor works — just make sure the voltage rating is at least 3× your supply voltage to survive resonant spikes. I used the original capacitors from the induction heater board, connected in series to hit the calculated capacitance.

Wind the primary:

I used 2.5mm copper wire wound into 3-4 turns with a small gap between each turn to spread the winding height. Any thick wire works. The exact number of turns depends on your target inductance — calculate it in JavaTC.

Initial primary placement:

Start with the primary sitting at the base of the secondary coil form. You will fine-tune the position in the next step.

The stock capacitors on induction heater boards are typically rated for around 50 kHz. At 300-500 kHz they will not perform correctly. Desolder the original capacitors from the board and connect your calculated capacitor bank in their place.

If your target frequency happens to match the original board design, you can keep the stock capacitors and simply add or remove capacitors in series to hit the right value. Calculate first, then decide.

This is the most important step. If the primary and secondary are not at the same resonant frequency, almost nothing works — you will see very low current draw and no arcs regardless of how much voltage you apply.

Method 1 — Oscilloscope:

Connect your oscilloscope probe to the primary. At resonance you will see a clean sine wave with maximum amplitude. Adjust primary position, turn count, or capacitor value until the waveform is cleanest and amplitude is highest.

Method 2 — Ammeter:

Connect an ammeter in series with your power supply. Power up at low voltage (20-30V). Move the primary up and down along the secondary and adjust the gap between them. Watch the current reading — at resonance current draw increases noticeably. That position is your tuned point.

If you don't have a lab power supply, use any fixed supply with an inline ammeter — the principle is identical.

Tuning adjustments:

1. Move the primary up or down along the secondary — in my build the best position was around 4cm above the base
2. Adjust the gap between primary and secondary — my optimal gap was around 1.5cm. Too close causes overcoupling and arcs disappear even though current is high. Too far and energy transfer drops
3. Add or remove half a turn from the primary
4. Adjust spacing between primary turns to slightly change inductance

How to measure frequency safely(Oscilloscope):

DO NOT connect your oscilloscope probe directly to the MOSFETs or any part of the live circuit. The high voltage spikes will destroy your oscilloscope instantly.

Instead, simply hold the probe tip and ground clip near the Tesla coil — without touching anything. The oscilloscope will pick up the electromagnetic field radiating from the coil and show you the frequency on screen. This is completely safe for your equipment and gives you an accurate frequency reading. The stronger the signal on screen, the closer you are to resonance.

Start at low voltage — around 20-30V — and raise gradually while watching the ammeter and checking transistor temperature by hand every 30 seconds.

Voltage guidelines:

1. 20-30V: weak arcs, good for initial tuning
2. 50-60V: best results with stock IRFP260N transistors, strong self-sustaining arcs
3. 70V and above: replace stock IRFP260N with IRFP4768 MOSFETs

Why upgrade transistors for higher voltage:

Resonant voltage spikes on the drain can reach 2-3× the supply voltage. At 80V the spike can hit 240V. That is safely within the IRFP4768's 500V rating but dangerously close to the IRFP260N's 200V limit. The IRFP4768 also handles more current and runs cooler.

Mount transistors on proper heatsinks — minimum 8×8cm — with thermal paste and a fan. Running without adequate cooling will destroy the transistors quickly.

My power setup:

I used a lab power supply feeding a boost converter. This is highly recommended — it lets you raise voltage in small steps and see exactly how the coil responds at each level rather than hitting it with full power immediately. (https://www.ebay.com/itm/)

I also recommend to use big capacitor to cover pulses when turning it on otherwise lab power supply can go to current limit and IGBTs will fry out.

With 50-60V supply the coil produces self-sustaining arcs. Bringing a grounded object close pulls longer arcs. Without a toroid the bare wire end of the secondary acts as a natural breakout point.

The toroid I originally planned (25cm × 5cm) dropped the frequency too much and caused the coil to lose resonance completely. For now I am running without it. Adding a toroid and retuning the primary to match the new lower frequency is on the list of upcoming improvements.

Before spending money on expensive components, check industrial scrapyards and electronics recyclers. Welding inverters, industrial UPS systems, and variable frequency drives contain IGBTs rated 600V and 40A and above — exactly what high-power Tesla coils need.

I will be visiting local scrapyards and showing you what is inside these machines and which transistors are worth pulling out. A single scrapped welding inverter can contain everything you need for a much more powerful build, often for almost nothing.

in case you don't want buy induction heater module, you can just build it by yourself using info from this link: https://www.open-electronics.org/how-to-build-a-1000w-zvs-induction-heater-using-a-resonant-rlc-circuit/

This is just the beginning. The coil works and produces real arcs, but there is a lot of room to improve.

Upcoming improvements to this coil:

1. Better resonance tuning for maximum efficiency
2. Higher supply voltage with IRFP4768 transistors for stronger arcs
3. Proper toroid topload with breakout point, retuned to the new resonant frequency
4. Scrapyard IGBT upgrade from salvaged welding inverters

Future projects:

After finishing this ZVS build I will be moving to the Labcoatz DRSSTC — a proper double-resonant solid-state Tesla coil with feedback control and much longer arcs. I will build it, modify it, and push it further than the stock design. After that, the goal is a QCW DRSSTC — the type that produces those long, slowly growing sword-like streamers you have seen in high-voltage videos.

Each project builds on the last. Start here. Build this. Then we go bigger.

Subscribe and follow along — every future build will be documented here with full details, schematics, and lessons learned.