Ignition coil driver

Hi everyone!

In my starting sequencer manager based around an Arduino, I need to trigger (switch?) @ 20Hz a high voltage ignition coil. It is an automotive type, plain and simple.

Using an arduino output, I made this simple schematic and would like your feedback on it.
I know there are plenty of circuits out there but I can’t seem to find an explanation / a universal solution and would like to understand why this or that would work or not.

So I based it around a 4N35 optocoupler and a IRF640 MOSFET.
Obviously I’m still a rookie in electronics but I’m trying :slight_smile:

Here are two of my interrogations :

  • some people say the MOSFET with it’s flyback diode will impede the coil to spark as the current will go through the diode and not be shut completely, thus not creating the spark.
  • Can’t I use a TRIAC? They seem to be cheaper and have higher voltage ratings for the same price
  • Some of the 4N35 I’ve seen have 6 pins and one connected to base of the phototransistor, I thought this wasn’t connected?

Thanks for your help!

Marc

One problem with using an opto isolator to drive the mosfet is that ordinary opto isolators turn off rather slowly. This might affect the strength of the spark.

I believe that the usual way to drive an ignition coil these days is with an IGBT. See for example Automotive - Infineon Technologies. You cannot use a triac in this circuit, because you can't turn a triac off except by reducing the current through it to near zero.. Capacitor-discharge ignition systems used to (and perhaps still do) use an SCR.

The intrinsic body diode in a MOSFET will indeed soak up the power that would have gone into a spark on the secondary.

The secondary of the ignition coil (remember its a transformer or autotransformer) will
arc before the primary exceeds a particular voltage. Just ensure the Vds of the MOSFET
or IGBT is greater than this and it'll work.

Use a MOSFET driver chip like the MIC4422, not an opto isolator?
You then get a really good hard turn-off - the MIC4422 can sink many
amps from the gate-drain capacitance as its output resistance is a couple of
ohms.

Hi everyone,

Thanks for the answers.

So I can or cannot use a MOSFET for ignition coil driving? :~
Seems opinions differ on this one...

I will look into the mosfet driver possibility but it doesn't answer the possibility of using a MOSFET question? :smiley:

dc42, I actually looked into IGBT's but they're that much harder to source and the dedicated, ignition coil use I can find are quite expensive @ 6-7$ for one!
And what about the TRIAC, couldn't I reduce this current to turn it off? Using a diac maybe?
I know CDI use SCR, I recently built one with a BT151. it's not working anymore though... XD And they do work the same way as TRIAC, don't they?

Thanks for the help!

Marc

Triacs cannot possibly work, you need to rapidly turn off the current, triacs do not turn off when current is flowing.

Any device that can interrupt a current will work, be it BJT, mechanical switch, MOSFET or IGBT. IGBTs are favoured for high voltage as they are more robust with high drain voltages
and high dV/dt values.

But you need to know the primary voltage before selecting a device. For instance a
coil that generates 30kV with a 1:100 primary to secondary ratio will exhibit ~300V
spikes on the primary and needs a device rated at 600V or thereabouts. If the ratio is
1:500 then the primary voltage spikes are about 60V.

Another consideration is that if you run the system with the spark plug leads disconnected, the voltage appearing across the mosfet or IGBT will be higher than usual. The mosfets I have come across that were designed for this application had built-in protection diodes so that the voltage spike wasn’t all dissipated as avalanche energy.

I suggest that you either use an IGBT designed for this application, or you use a capacitor discharge circuit and SCR.

Ideally you also want a capacitor across the switch device so that the coil primary "rings" when the supply is triggered; otherwise you'll just get a single transient spike in the secondary. In the old contact breaker type of ignition systems, used in petrol engines, the contact breakers had a capacitor across the contacts to provide the "C" component of the "LC" circuit. Without this capacitor the ignition system simply did not work. Since you want a 20Hz trigger I suggest you measure the inductance of the primary and calculate an initial value of C. Any loading on the secondary side will affect the resonant frequency of the circuit so the value of C may need modifying to suit.

Ok ,will consider the IGBT solution then.

Jackrae, just one thing : I thought the capacitor across contacts in old systems was to avoid too much of an arcing and rapid deterioration of the so called "points".

New systems, as far as I know, do not include a capacitor in the design!

Without the capacitor there is indeed a fat spark across the contact but when it's included in an inductive circuit it also serves as an essential component of an oscillator.

If you look at a CDI circuit you'll see that they all contain a capacitor in the primary circuit. Although not immediately evident, the capacitor and ignition coil forms an oscillator circuit by virtue of the supply battery being part of the LC series circuit.

However a fair description is given here :-

In CDI the capacitor is high voltage, connected to the coil suddenly to induce high voltage
in the secondary, no oscillation or LC action (a transformer is not the same as an inductor,
people get confused by this).

In standard old style points-based ignition the capacitor was to limit the
risetime of the primary voltage to allow the spark to quench on the contact
points (overwise it would both destroy the contacts and remove all power
from the spark plugs). Because it increased the risetime the sparks at the
HT side were less powerful or reliable than with CDI, hence the move over.
There is no deliberate resonance in the system to my knowledge, the spark plug
should be cutting short any resonant behaviour anyhow as all the energy
should end up there.

CDI ignitions caused fatalities with some mechanics who were used to touching
spark plug leads to check for HT (CDIs put a lot more energy into each spark,
again increasing engine reliability especially with open throttle when cylinder
pressure is maximum.

MarkT:
In CDI the capacitor is high voltage, connected to the coil suddenly to induce high voltage
in the secondary, no oscillation or LC action (a transformer is not the same as an inductor,
people get confused by this).

In standard old style points-based ignition the capacitor was to limit the
risetime of the primary voltage to allow the spark to quench on the contact
points (overwise it would both destroy the contacts and remove all power
from the spark plugs). Because it increased the risetime the sparks at the
HT side were less powerful or reliable than with CDI, hence the move over.
There is no deliberate resonance in the system to my knowledge, the spark plug
should be cutting short any resonant behaviour anyhow as all the energy
should end up there.

CDI ignitions caused fatalities with some mechanics who were used to touching
spark plug leads to check for HT (CDIs put a lot more energy into each spark,
again increasing engine reliability especially with open throttle when cylinder
pressure is maximum.

I would have assumed cylinder combustion chamber pressure is dependent mostly on cylinder compression ratio, not throttle position or rpm.

I would have assumed cylinder combustion chamber pressure is dependent mostly on cylinder compression ratio, not throttle position or rpm.

Well no, it depends on the charge (closely linked to the throttle angle aka alpha) and also RPM's since volumetric efficiency isn't the same at all rev's.

So the LC oscillator circuit seems to be a point of discussion... Interesting...

@MarkT : Did it really cause fatalities or is it yet another myth? Even though it might pack a bit more energy (to me it's more that it is more compact / draws less current --> used in motorcycles and NOT AT ALL in cars), I cannot see a couple of high V's & extremely low amps killing anyone.