Skip to content

Half bridge ignition coil driver

Project description

Ignition coil drivers that utilize an oscillator chip like NE555 and a single switching device are a very common beginner HV project. In my own version, I pointed out some of most important problems with this kind of approach. Firstly, I wanted more power, which apart from producing good looking arcs may make ignition coils more useful HV sources that may be used to power even a medium sized spark gap Tesla coil. Obviously, this implied producing more current at lower voltage than usual in order to obey ignition coil’s insulation limits. Instead of storing energy in the coil and releasing it in pulses, I wanted to use the coil as a real transformer, transferring the energy in both half cycles for maximum efficiency. In order to acheive this, a half bridge topology was chosen, which is very common in power electronics. The ignition coil makes a rather poor transformer, considering it has a non-closed iron core which is simply a single rod of iron laminations. This provides plenty of leakage inductance, which is a good thing because it helps to limit the current while drawing arcs or charging capacitors. A bad side-effect is presence of relatively large reactive magnetizing current in the primary. This current seems to be easily handled by the 14A mosfet switches, though it contributes significantly to heating of the coil.

The half bridge is powered from rectified and filtered mains voltage, which is around 325V DC. This results in primary voltage amplitude of around 160V, which is a very large value compared to usual 12V in the ignition coil’s original application! Still though, the coil handles this well because it’s insulation is designed to withstand flyback pulses upwards 200 volts – the main limiting factor is now overheating instead of over-voltage.

A half bridge topology is commonly used by HV enthusiasts to drive TV flyback transformers – but I’ve never yet seen anyone use it for an ignition coil, and was really curious how would it turn out. A great advantage of this topology is the fact that it is no longer possible for the ignition coil to saturate under arc loading, since the half bridge enforces no DC component regardless of the load!


The drive signals for the MOSFET’s are provided by a TL494 PWM chip and amplified by a pair of PNP-NPN transistor ‘totem poles’. The tricky part was making an electrically isolated connection of these drive signals to MOSFET gates, as it is required by the half bridge topology – for high frequency bridges operating in range of tens of kilohertz, the common choice is a gate drive transformer consisting of several windings of PVC insulated wire on a small toroidal ferrite core. This application, though, required drive signal frequencies as low as few hundred hertz, which would require way too many turns on the ferrite core and be impractical to construct. On the other hand, switching losses are quite negligible at the frequencies used, allowing a decrease in gate drive speed. Having this in consideration, I extracted some isolation transformers from old dial-up modems and tested them as gate drive transformers. They had to be carefully selected due to variation in their turn ratio – but in the end they performed very well as gate drive transformers!

The circuit was built on a piece of prototyping PCB, separately from MOSFET’s and gate drive transformers. The switching devices chosen were again IRFP450’s, and they were mounted onto a CPU heatsink for cooling. Mains voltage is rectified and filtered using two series connected 200V, 470uF capacitors which act as voltage divider for the half bridge.

Frequency and duty cycle of the TL494 are independently adjustable with onboard trimmer resistors. The frequency was chosen by slowly increasing the supply voltage and frequency while monitoring the supply current – the DC current into the half bridge should stay in range of few hundred milliamperes – much more than that indicates saturation and required an increase in frequency. Ultimately, a frequency around 1500Hz was chosen as this seemed to be the lower limit of saturation with full mains voltage input, with the no-load current of around 300mA. It’s strongly recommended to use a fuse on the mains input, as well as NTC inrush current limiting resistor, which can be found in most switching power supplies. It is also recommended to ground the lov voltage section GND to mains ground, to reduce the risk of electric shock in case of insulation failure in the gate drive transformers.

Circuit schematic. All resistors are 0.25W and capacitors 50V, unless otherwise noted


Finished PCB


A picture of a full bridge power section, an attempt that resulted in a dead ignition coil. A half bridge is enough for 230V mains! Four isolation transformers extracted from modems are visible.


The following video shows an example of an arc drawn from the setup with full mains voltage input. The arc is thick and bright, obviously carrying a lot of current, unlike the typical ignition coil outputs. The arc starts at less than 2cm but stretches up to almost 10cm, demonstrating that long arcs may be produced with relatively low voltages as long as there is enough current to heat the arc and keep the plasma conductive. This idea has been taken to extreme with my ATX transformer HV source.


This kind of ignition coil driver is an interesting innovation that allows lots of power throughput, and has potential to be used as a power supply for spark gap Tesla coils, charge large capacitors, or simply as a big arc generator. The tradeoff here is similar to one seen in Tesla coils: transformer quality is sacrificed in favor of enhanced HV insulation. A winding around a rod core is easier to insulate, at cost of increased reactive power that needs to be handled by inverter (actually not much of a problem for modern switching devices). An approach like this is known to be used in some X-ray machine designs, where the HV transformer is also rod-cored to enhance insulation.

Links and references


[1] TL494 datasheet

[2] 4hv forum thread

One Comment
  1. ALEJANDRO permalink

    Hi powerfull tesla and very high current output, the ATX transformer HV source is used in this case It was connect to base of secondary coil and then tuned to resonance frequency or have other system for produced this SSTC.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: