Isolated Ignition Coil Driver

If you have an old ignition coil lying around, it's easy to make a 30kV continuous spark generator. Perfect for your Frankenstein laboratory in the basement.

An ignition coil has two windings. The primary 12V winding is between terminals + and - , the high voltage output winding is connected to - and the HT connector in the middle. The circuit switches 12V through the primary coil. The 3.3V digital control signal has an on-period of 1ms and an off-period of 2ms, so the frequency is about 333Hz. This switches a 1200V MOSFET. The switching signal can be supplied by a microcontroller or a signal generator, and is passed through a PC817C (or similar) optocoupler to provide full isolation from the high voltages and noise. The primary inductor has a 100nF foil capacitor (1kV or greater) wired in parallel. This creates a resonant L/C tank circuit, with a resonant frequency of about 5kHz. When 12V is turned on, the inductor "charges up". When it's turned off, the tank circuit resonates (see trace below), which induces a constant A/C voltage in the HT secondary to create a continuous spark.

The 1N5408 (1kV 3A) protects the MOSFET from the huge back-emf, and the UF4007 (1kV 1A) is a high voltage fast-switching diode to provide additional protection.

I used an old SCT2540KE MOSFET, but any >=1200V MOSFET which handles at least 3A will do. I also used a heat sink, see image below, but it doesn't get very warm so you probably don't need it (depending on the MOSFET and the primary coil resistance).

DO NOT CONNECT YOUR OSCILLOSCOPE TO THIS CIRCUIT - IT WILL PROBABLY GO BANG!
AND KEEP THE SPARK WELL AWAY FROM YOUR DIGITAL CIRCUITRY AND USB CABLES!!!

Matt's Tip #327: Never use your tongue as a voltmeter.

Below, the blue trace is the 3.3V digital control signal (x1000). The green trace is the HT output. You can see that the tank circuit oscillates when the control signal is off, f = 1 / (2π × √(L × C)) = 5.0329kHz. These traces are from the LTspice simulation.

I used a 12V 8A power supply, but 3A would have been enough. I made the mistake of trying to make a spark between the HT and the mains earth wire. This instantly blew the TL431 regulator chip in my switch-mode power supply. Luckily, I had some of these lying around, so I was able to fix it very quickly. It is safest to keep the HT circuit isolated and floating (not grounded), discharging the spark directly to the -ve terminal of the coil, NOT to ground.

My coil produces a spark that jumps a 12mm gap. Folklore has it that this is 30kV (30kV per cm in air), higher than the voltage from the simulation.
Here's the video...

Optically Isolated Ignition Coil Spark Generator (30kV)