1
\$\begingroup\$

I'm currently working on a project designing an experimental engine control unit, which ignites a flammable mixture with a standard spark plug and standard ignition coil (the old style https://en.wikipedia.org/wiki/Ignition_coil). The primary is first charged to a nominal voltage (12V I think) then, to generate the spark, the primary voltage is quickly dropped to zero and the collapsing magnetic field in the secondary inductor causes a voltage spike across the spark plug. I'm controlling the primary voltage with a microprocessor and MOSFET and charge the primary when the engine is "armed" and turn off the coil when "arm" is switched off.

I'd like a way of making the spark coil fail-safe, i.e. if there is a sudden power cut, no spark will be created due to the sudden voltage drop, but be able to generate a spark by intentionally dropping the voltage whenever needed. This needs to be 100% reliable, as it could be safety-critical if there is a build up of explosive gasses and an emergency shutdown is needed. I understand usually you'd connect a flyback diode (plus Zener/resistor) in reverse bias across the coil to stop the voltage spike, but the voltage spike is needed to generate a spark...

I looked at make-before-break relays to connect the flyback diode before the primary current is switched off, but couldn't really find any available and I'm not sure how they could be connected to stop the spark in the event of a power cut or when the "arm" switch is turned off. Likewise with using solid-state switches to include a diode in the loop, I'm not sure if it could be connected quickly enough to beat the voltage spike.

Thanks in advance for any input, this has been puzzling me for months now. I'm a mechanical engineer at heart, so please be patient if I've misunderstood anything in the above!

\$\endgroup\$
1
  • 1
    \$\begingroup\$ What’s wrong with the standard method of pulse resonance with a switch cap of suitable value. Conventional autotransformer coils have a voltage gain of over 1000 in linear mode so if the primary switches to 200V the secondary is well over 20kV with a primary safety switch. SRF’s I have measured are above 20kHz \$\endgroup\$ Jan 19 at 0:57
0
\$\begingroup\$

Instead of charging up the coil in advance, can you just apply a pulse to your FET when you want to create a spark?
Or
Is there some scenario where a 100ms pulse would need to be prevented from creating the spark after the rising edge has already happened?

EDIT: Thinking more about this, if it is safety critical, then you should have a shorting relay on the output. When the emergency button is pressed or power is cut, then the output of the coil is shorted. Obviously you'd have to find a relay that can stand off the sparking voltage.

\$\endgroup\$
1
  • \$\begingroup\$ Thanks for this, yes I really don't know why I didn't think of that before... A case of the classic "The old system worked like that so I didn't question it". I'll try and find out or measure the inductance (or even just the time taken to reach supply voltage) of the ignition coil so as to be able to apply the minimum length pulse. I can't see there ever being a situation when a spark a few tens of ms after shut-off would cause a problem. \$\endgroup\$
    – AndrewG
    Jan 19 at 18:49
0
\$\begingroup\$

In the abstract, you could put a normally-closed switch in series with the reverse diode, just as the question suggests. By default, the closed switch allows the diode to act, and no spark is generated. To allow spark to be generated, open the switch. It might be necessary to add protections to the switch, especially if it's solid state. It might also be necessary to add circuitry to delay the current turn off long enough for the switch to close first, but this is not long.

\$\endgroup\$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.