I'm looking to build a small spark ignition system for a butane appliance. (It's not a lighter, but that's an easy way to think of it.) It needs to be triggered by a microcontroller, and small size and low-cost are my main considerations. I don't need many sparks per second, I don't need a ton of energy in the spark, and I don't even need the spark to happen immediately after triggering. Just one little spark, every once in a while.

I have a working circuit with a push-button control from a commonly-available cigarette lighter. It basically charges a 330uF electrolytic to 12v, and then uses a snap-action microswitch to dump the cap's charge into the primary of a small autotransformer. I'm sure it's not great for the switch contacts, but you can't argue with the simplicity.

To me it looks a lot like a baby CDI system. I tried using a small-signal BJT to replace the snap switch (since I had one handy), but although it did successfully discharge the cap through the primary, it didn't spark. I'm guessing that's either because the voltage drop across the BJT was too high or peak current too low, etc.

Question 1: I know SCRs are generally used for CDI systems. Are they inherently better than a transistor for this low-energy application?

Question 2: If I need to use an SCR, how do I go about choosing one? The transformer's primary is low-resistance so my peak current needs to be high-ish, but my continuous current is basically zero. Continuous voltage is 12v but transients could be large maybe? I need a physically-small device (TO-92 or smaller), but I obviously don't want it to blow up.

Question 3: Given that I don't need latching, is there some other device that would do this switching just as well or better? I would like to trigger it from a microcontroller's GIO pin so low drive current/voltage would be nice.

Thank you!

 

You could use a 555 set up as a monostable to charge the primary of a small dual-winding choke ("flyback transformer"). At the end of the charge period, the energy will discharge into the secondary at high enough voltage to arc across an air gap. You'll need an NPN on the primary with sufficient collector-to-base breakdown voltage, the magnitude of which will depend upon the gap size and turns ratio.

 

You could use a 555 set up as a monostable to charge the primary of a small dual-winding choke ("flyback transformer"). At the end of the charge period, the energy will discharge into the secondary at high enough voltage to arc across an air gap. You'll need an NPN on the primary with sufficient collector-to-base breakdown voltage, the magnitude of which will depend upon the gap size and turns ratio.

Thanks for your reply. If I understand correctly, this would be storing the energy in the choke versus in a capacitor, and it's the interruption of the primary current which would cause the spark by collapsing the field. Like a points/condenser automotive ignition? What appealed to me about the CDI approach was that I could charge the cap slowly (i.e. over several seconds) with small currents from a charge pump. I don't know a ton about transformers and chokes, but would I need to move a fair amount of current in the choke to "charge" it?

The other thing that appeals to me -- and I admit this is just me being lazy -- is that the lighter from whence this circuit came costs about $20, so I know that this tiny, nicely-potted autotransformer must be available in quantity for low $, and all the engineering has already been done. If I can replace the snap switch with a suitable device, I'm done.

 

The question is: how much energy do you need to ignite the flame? From that, everything else is elementary math. A larger inductance will charge more slowly and store more energy at a given current -- but will be physically larger. A smaller inductance can be quite a bit smaller, but require a larger charging current (but! energy goes up with the square of current, and only linearly with inductance -- E = 1/2LI^2).

A cap is going to draw most current when it is discharged, and inductor when it is charged. In either case, you will need to supply enough energy to ignite the flame and compensate for losses.

 

I could charge the cap slowly (i.e. over several seconds) with small currents from a charge pump.

Alarm! I have repaired furnaces that took, "several seconds" to ignite. The best part was that nobody got hit by the metal cover.

 

You could also slowly charge a cap then quickly discharge it into the inductor. *Just like* automotive ignition.

 

Alarm! I have repaired furnaces that took, "several seconds" to ignite. The best part was that nobody got hit by the metal cover.

Not to worry...first, I have control over the gas so I could charge the cap, then turn on the gas and then fire the spark. Also, I'm dealing with tiny gas flow rates here. A delayed spark would create a small "pop" at worst.

 

You could also slowly charge a cap then quickly discharge it into the inductor. *Just like* automotive ignition.

Right, so that was what I was trying to ask in my first post -- how to do that fast discharge. A regular BJT doesn't seem to do it so that's why I'm asking about SCRs. Any thoughts about that?

 

Yes, an SCR is an excellent choice for dumping a cap into an inductor.

But I don't think it's necessary.

 

Yes, an SCR is an excellent choice for dumping a cap into an inductor.

But I don't think it's necessary.

So this is where I'm stuck. An SCR would be great -- but how do I spec it? Meanwhile, an SCR isn't necessary as you say, but on the other hand I can't use any old BJT either. Hence my original question: Assuming I want to dump a given cap into a given inductor, what type of device should I use and how do I spec that device?

 

Ok. You have a circuit that already works with a switch (i.e. 12VDC, 33o uF, and an autotransformer), and you would like to replace the switch with an SCR.

Figure the peak discharge current into the cap through the transformer primary. This'll be a function of C, primary inductance, and power supply voltage - "on-state" voltage.

Choose an SCR that'll handle the peak current. Then try it.

For some hints, do a search for Xenon Flasher schematics, and concentrate on the trigger voltage generator portion of the circuits. This is similar to what you are trying to do.

 

Sounds good, that's what I'll do. Thanks for your help.