A new ignition circuit

[DKGrimm]

I attempted to use an Arduiuo to control the timing advance of my model engine and could never get it working. the Electromagnetic Interference from the ingition system always (eventually) killed the micro controller. I went from a CDI to a TCI (transistor controlled ignition) which was supposed to be quiter from an EMI standpoint. I used resistor spark plugs, I used a ballast resistor, I isolated the microcontroller, I used lots of local decoupling caps and an inductor isolation circuit with an encolsed metal case, I used a seperate battery and opto isolators.... I spent way too much time trying to get it to work until I finally abandoned the project. If you can get the arduino working, you need to let me know what you did. I wrote some really neat code to control the spark advance, but to no avail. Good Luck.

Your experience with Electromagnetic Interference (EMI) struck a resonance with me. Ignition systems generate an unbelievable amount of high-frequency energy that can couple into any circuit they happen to be near. Analyzing coupling between circuits is anything but intuitive. I spent a good part of a 40-year career dealing with EMI in aircraft avionics systems, and it can be done, but it's a humbling experience. H-field (magnetic) coupling is the culprit more often than not. Visualize the path the spark current takes and how it might couple to the control circuitry, either through common impedance (like a common ground path) or by H-field coupling (like between parallel wire runs). Try to assure a single-point ground reference between the spark stuff and the control stuff, and keep the control stuff away from the spark current path.

Experiment. You will probably be amazed at what actually works! Good luck.

Don

 

[Bentwings]

Who cares who it offends! If thats what you want to do then thats what you do. I would love to see it completed. Better idea! Go with dual plugs and 10 coils! Hahahaha!!!!

I agree also don’t forget the spark plug wires automotive wires are 7 or 8 mm although the big top fuel magneto wires can be even bigger they can dwarf the intake manifold tubes on model engines .

 

[dsage]

The Person Steve mentioned in post 255 is Dave Bowes. His engine is called EVIC and has had various iterations. A pretty impressive project.
He has electronically controlled solenoid valves and practically everything else. The control circuitry has SEVERAL PIC processors. He has become the master at eliminating EMI and RFI from his circuitry. So it can be done.
I believe he wrote a couple of articles for Strictly IC, Home Shop Machinist and Model Engine Builder.
Not sure why the link to his web page shows up as "index" but here it is.
http://rbowes1.11net.com/dbowes/index.htm
I see he commented with some suggestions in the next post.

 

[dbowes12]

Some tricks to making electronics immune to spark EMI/RFI:
Put a 10K carbon comp resistor in the HV wire to the spark plug as close to the plug as possible.
Use an EMI filter on the input to the micro from the Hall Effect position sensor. Digi-Key part # M9935CT-ND
Put ferrite sleeves on all the wires to the micro board including power leads, sensor leads and any leads to external devices like LEDs or displays. Digi-Ket part # 240-2112-ND
Twist all wires to and from the micro bord including those to sensors, battery wires, etc.
Not sure the Digi-Key part numbers are still current since I haven't ordered them for a few years.
Hope this helps. If you still have problems contact me directly.
Dave Bowes

 

[dsage]

Thanks Dave.

 

[minh-thanh]

Put a 10K carbon comp resistor in the HV wire to the spark plug as close to the plug as possible.

Dave Bowes

I know some tricks to limit ..
It's a way I don't know
Thanks you !

 

[peterl95124]

this has probably been noted before, but sometimes those of us with flakey memory need reminding...

the improved circuit with time-out requires a possibly non-obvious technique to properly set the timing,

if you turn the engine over by hand in the correct rotation then the spark occurs when the time-out circuit does its thing, long before you've reached the end of the dwell where the spark occurs when running at speed

instead you time the ignition at the point where the LED turns *off* because at speed the spark occurs at the end of the dwell, or

if you turn the engine *backwards* then the LED comes on (and the spark occurs if you're turning slowly enough) at the same point where it will fire in full speed forward rotation because that's where the dwell ends.

 

[dsage]

this has probably been noted before, but sometimes those of us with flakey memory need reminding...

the improved circuit with time-out requires a possibly non-obvious technique to properly set the timing,

if you turn the engine over by hand in the correct rotation then the spark occurs when the time-out circuit does its thing, long before you've reached the end of the dwell where the spark occurs when running at speed

instead you time the ignition at the point where the LED turns *off* because at speed the spark occurs at the end of the dwell, or

if you turn the engine *backwards* then the LED comes on (and the spark occurs if you're turning slowly enough) at the same point where it will fire in full speed forward rotation because that's where the dwell ends.

Unless you have really excessive dwell (points closed - Hall active) then under normal operating conditions - even at cranking speeds - the spark for all intents and purposes happens when the light comes on. Yes it really happens when the light goes out, but as mentioned there shouldn't be much difference. You can do the math. The timeout is about 15ms after the LED comes on. It shouldn't be timing out unless you have excessive dwell.
The best way to time the engine is to turn it over by hand and watch the led come on and set your timing accordingly. Then use a timing light to fine tune it. It won't be far off when the engine is running.
That said - some folks with full sized hit-miss engines with crude contact points or wiper contacts (excessive dwell) will have issues with the timeout. The engine might kick back.
For those cases I suggest you adjust the mechanics to shorten the dwell to a minimum. Again, you can do the math. A coil only takes about 3ms to fully charge. If you have more dwell than that it's just a waste of battery to keep it energized any longer.

EDIT: Some of what I wrote here is jumbled up. See my correction below.

 

[peterl95124]

Unless you have really excessive dwell (points closed - Hall active) then under normal operating conditions - even at cranking speeds - the spark for all intents and purposes happens when the light comes on. Yes it really happens when the light goes out, but as mentioned there shouldn't be much difference. You can do the math. The timeout is about 15ms after the LED comes on. It shouldn't be timing out unless you have excessive dwell.
The best way to time the engine is to turn it over by hand and watch the led come on and set your timing accordingly. Then use a timing light to fine tune it. It won't be far off when the engine is running.
That said - some folks with full sized hit-miss engines with crude contact points or wiper contacts (excessive dwell) will have issues with the timeout. The engine might kick back.
For those cases I suggest you adjust the mechanics to shorten the dwell to a minimum. Again, you can do the math. A coil only takes about 3ms to fully charge. If you have more dwell than that it's just a waste of battery to keep it energized any longer.

Dave, I'm not really sure you've understood the point I am making, we (at least I) "time" our engines by turning them over by hand and watching the LED, not by cranking them at running speed, so the "dwell" when turning over by hand is on the order of 1/8" (magnet width) divided by PI times magnet circle diameter (eg PI x 1" in most of my distributors) which is 0.039 fraction of the circle, or 14 degrees, and 14 degrees is a lot to be off by if you're using the point the LED goes on rather than the point the LED goes off to time the ignition. Does this make sense to you ???

 

[dsage]

Sorry. I made a mistake in my dissertation. And I would go back and edit it but - too much work and too late now..
I have made a note below the last post.
So Yes. Time the engine initially (like you said) by the LED going OFF when turning by hand. Just like you do with points etc on a typical engine. Then use a timing light to fine tune it.
Turning by hand the timeout is, effectively, immediately after the led comes on (only 15ms later). So you will observe a spark when the led appears to come on. Ignore it.
Under normal speed the dwell will be short enough that the time out circuit won't be activated and the on/off of the led is effectively the same thing.
So bottom line, manually time the engine to when the LED goes OFF. Ignore the spark that SEEMS to occur when the led comes ON. And fine tune with a timing light. It shouldn't be too far off.
Does this clarify the situation?

 

[peterl95124]

Does this clarify the situation?

we are in 100% agreement !!!