High voltage triac switch

[yougarage]

I want to build a simple "switch" to interrupt the supply to a moped ignition coil (Capacitive discharge ignition), it is fed with AC voltage, varying from 200 to about 500 volt depending on the RPM of the engine.
I need to drive this switch with a microcontroller, so i thought to use a triac and an optocoupler, but due to this wide voltage range, how do I select the gate resistor?
If I choose a low OHM value, the gate will melt at higher RPM, and a high value will not provide enough current to the gate during the engine start.
Is there a way to regulate the gate current dinamically?

 

[alec_t]

AFAIK all cap discharge ignition systems discharge a DC voltage (built up on a cap by rectifying an AC signal). The switching/control would normally be in a DC domain. Can you post the schematic of your system?

 

[yougarage]

Hi Thanks for your reply

Here is the schematic, I also posted some pictures of measurements made with my cheap scope
The voltage scale is on the bottom left side
The high voltage wire is the green one, the signal wire is the red one

The "cranking disconnected" picture shows the measurement across the green wire, disconnected from the coil during the engine start.

The "high voltage connected" shows the same wire, but with the engine running and the coil connected.
The voltage scale for the above is 80v for each row, the 0 is located at the central row.


The "pickup connected" and "pickup disconnected "show the measurements across the red wire, with engine running and engine starting respectively, here the voltage scale is 1 v for each row.
I need to find a way to stop the engine for a while for a rev limiter, and I'd like to find a solid state solution. Maybe it's better to work on the low voltage side? (the red wire)
I tested some commercial rev limiter, but they short the high voltage wire to ground to stop the engine, and sometimes burn the coil inside the generator.

 

[alec_t]

Maybe it's better to work on the low voltage side? (the red wire)

That would be my choice. Can't make out from the scope traces for the red wire where 0V is. Is 0V at the bottom of the screen? Or does the waveform have + and - parts?
Anyway, you could try this:

 

[yougarage]

Hi the ov is located on the center row, so the waveform have + and - parts.
Anyway only the + side of the waveform triggers the scr, I tried to interpose a 1n4007 between the red wire and the ignition coil, and works the same.
The negative "noise" you see in the waveform is generated by other magnets on the rotor that serve lights and horn.
Ok I will try the transistor asap
Thanks!

 

[alec_t]

I tried to interpose a 1n4007 between the red wire and the ignition coil, and works the same.

That's encouraging, because the transistor drops only a tiny bit more voltage then a 1N4007.
BTW, I forgot to say that the mechanical switch shown at the transistor base can be replaced with a second npn transistor controlled by your micro.

 

[yougarage]

Hi that worked pretty good!
You suggested to use a second npn transistor, can I use an optocoupler instead, or I have to use the optocoupler to drive the second transistor?
I have some 4n25, could they be fine?
Thanks!

 

[alec_t]

Hi that worked pretty good!

It's always good to find practice agrees with theory!

can I use an optocoupler instead

A 4N25 should work ok instead of the second transistor. When the opto-diode is ON the ignition will be OFF.

 

[yougarage]

perfect, the 4n25 did the job!
Now to learn some theory, I was reading the datasheet of the 2n2222,(I'm a beginner in electronics)and I was wondering if is there a limit in negative voltage applied to the base of the transistor.
In this case we're applying to the base an ac voltage that swing about from -2v to +2v, what would happen if the voltage was swinging from say -20v to +30v? would the transistor burn?
And also, is there a reference for the voltage drop on the datasheet?

 

[cachehiker]

Yes. As I understand it, the emitter base junction of a 2N2222 and most NPN transistors will behave somewhat like a 10V zener diode when reverse biased. It's probably best to clamp or otherwise limit the reverse bias to about half of that. Too much leakage will overheat the junction.

I assume you're asking about the collector emitter drop when the transistor is driven on. That would be the saturation voltage on the datasheet. Somewhere in the vicinity of 200mV IIRC if Ic << Ib*hfe.

 

[alec_t]

perfect, the 4n25 did the job!

Hurrah!

is there a limit in negative voltage applied to the base of the transistor

Beyond about 5V negative the base junction can break down. Shouldn't be a problem with the waveforms shown, but if there's a risk of -5V or more negative then a reverse-biased diode can be connected between the transistor base and emitter.