HappyTime 2 Cycle Ignition Circuit Explained with Diagram

[VTBikeman]

I don't think there is an embedded processor as it wouldn't like the Hv environment. It triggers on the rising side of the wave form. When I drive it with an amp I'll know if there is any advance at all. The spark occurs about 10% of the way into the rising waveform and it looks like there is plenty of room for the speed to double or triple before it runs out of room. That is faster than the motor can probably go. 60 hz= 3600 rpm.
The circuit charges on the negative half of the waveform so technically I guess that is the start of the cycle which puts the spark at about 60 to 70% of the time in.

When I post the photos you can take over figuring it out. My head hurts.

Again, timing is independent of voltage in.

 

[arceeguy]

That makes sense.
So I guess you have a dual trace scope and took photos of the input and output of the CDI?

I have a variac, and an isolation transformer so I could duplicate your test. I also have a signal generator and an amplifier that could feed a 12V transformer as a "step up" to drive the CDI. The big question is if the CDI likes rock, country or rap music. (or perhaps even disco?)

 

[VTBikeman]

I took photos of both the input voltage and input current across a 10 ohm resistor in the ground leg of the transformer, not the box. I left the box on the bike, ground connected as usual, put a screw in the spark lead and held it close to the motor to see the spark. The spark trigger is obvious in the scope trace with sharp spike and some HF ringing on the input voltage wave form. I used a 12 watt 1:1 transformer which limits current. The motor alt winding is 320 ohms or so so it also limits current.

I think the diode may short out the negative input pulse due to the heavy loading on the voltage wave form and the current waveform shows current draw. This is consistent with a steady diode drop of about .595 volts. The heavy loading on the negative side shows up as severe clipping with some tilt so some cap is getting charged somewhere. It may a lower voltage CDI running on 120 volt peak, not rms.

Timing is not voltage dependent so it isn't a zener or other voltage dependent device tripping it.

I also don't see any hash that might indicate an inverter.

Maybe a 555 timer chip with a film cap to determine delay timing in the cycle??? This is all relatively slow so the delay may not require super accuracy. Engines run even if the spark isn't exactly where it needs to be.

Will post photos later.

 

[VTBikeman]

These are shots at 60 hz (=3600 rpm).

I ran it as high as 500 Hz (=30,000 rpm) and it still fired but things looked worse. Of course that is faster than any of these engines will ever go.

The timing from A to B seems to be constant even when I drive it from an amp into a step up xfmer and vary the frequency.

The - input voltage peaks under load at about -20V and the negative current is fairly sinusoidal, with less distortion than my off - the - line ac wave. There is some kind of load that behaves resistively after the diode. If it was a capacitor it should show a tilt. It could be an inverter but it is awfully quiet.

I ran neg DC into the box from a current limited supply. I could dump 275 ma at 15 volts into it continuously. That doesn't indicate to me an inverter whose input current should go down as a cap charges. It looks like a fairly low ohm resistor.

At higher frequencies the negative current was greater, off the scale.

At B you have the spark evidenced by ringing and overshoot.

Note the small positive current starting at A and note that it doesn't start until the input voltage in the reference frame hits 50 volts. It also peaks and goes to 0 current at spark time. That indicates to me that there is a cap or a circuit that is 50 volts above ground.

There is no neg current drawn until the input wave falls to 50 Volts. Note the heavy negative current and note the assymetrical wave.

It is your turn to figure this out. I have to give the designer credit, it is quite elegant.

This is like doing the NYT Sunday xword puzzle. Quite a brain teaser.

I'd like to kill my first post with the photo as it is obviously wrong and the post on the other list which is obviously wrong. It was a good guess but no cigar.

 

[VTBikeman]

Photo at 500 hz, (30,000) Rpm.

0 point of the lower trace is 1 division higher and shows -600 ma peaks.
As you can see it is running out of time on the positive peak to fire.

Varying the frequency without changing the sweep rate showed a constant time from A to B as in the other photos.

I attribute some of the differences to the driving amplifier/transformer characteristics. The amp didn't like being driven too hard at 60Hz but was OK at 500.

 

[arceeguy]

Great information, thanks!

We may not know exactly what is inside one of these two wire ignition boxes, but thanks to your super sleuthing, we know that it is a "true" CDI and it does not have any sort of rev limiter.

I wonder why these units are subject to SIDS. "Sudden Ignition Death Syndrome" Could it be transient spikes from the magneto instantly destroying the circuitry? If so, maybe a couple of back to back Zener diodes or even a small capacitor across the magneto could prevent them from dying? Hmmmmmmmmm.......

 

[VTBikeman]

SIDS? Not so far. May be just that it isn't made to mil spec or component selection isn't high quality. I doubt that a zener or a cap will save any given all the odd waveforms. It may even delay the spark a degree or 2 if the cap is too big or it may load the alt.

I'd bet a bad ground could kill it. The ground and hot connectors are cheesy and could easily open from vibration.

Who TF knows? Anyone with an xray machine out there?

I've spent way too much time on this nonsense. I get on the bike and it goes forward which is all I care about.

I have some other ideas to think about.

 

[VTBikeman]

Sorry, I didn't mean to sound so mean about it. It was great fun but at the expense some other things that I SHOULD have been doing.
You know . . . Would-a . . . Should-a . . . Could-a . . .

 

[arceeguy]

Maybe the CDI is more reliable than I thought. I guess we only hear the bad things about the engines, while the vast majority are chugging along just fine. I dunked my CDI in a glass of water while the engine was running and it didn't miss a beat submerged for 10 minutes - but then again, I haven't started it today.

 

[emdude]

for yor considereation...

...regarding the ignition system:

Based on the thread, the scope shots and speculation as to how the CDI box works I have the following comments. It is in fact a CDI unit, that is a "capacitor discharge ignition" system. I looked at the magneto coil in my engine (haven't even installed it in a bike yet) in order to get some idea of how the thing must work and what if any timing advance there may be. Noting that the engine runs clockwise when viewed from the left side and noting TDC of the piston, it becomes clear that the magnet rotor passes the (approximate) center of the coil a few degrees before TDC. Ignition ALWAYS occurs when the positive going voltage begins to decline. Or so I surmise based on the research done and explained in the thread. There will be a very slightly advanced spark at higher revs because the "pulse" as it were is of shorter duration and the declining voltage does so a bit sooner. However I expect this to be nothing more than a side effect of the design and not any intended advance.

As to how it works:

It's simple...and yet maybe not. The high voltage from the magneto coil to the CDI unit...or what ever you really want call it clearly indicates a capacitor discharge system. The diode and cap shown in earlier drawings are correct. The cap is charged on the rising voltage. The "primary" of the ignition coil is most likely in series with an SCR to ground. The SCR's gate is held low until the voltage begins to decline. A few volts or so after peak, it is triggered by some circuit which of course fires the coil by discharging the cap through it. Up to that point it's simple. A circuit to do this could be made from what most have in their junk boxes.

A problem may be that since once the cap is discharged the process could start all over even on the now only declining voltage. So there may be something in the unit that only allows the above triggering to occur once for every positive going pulse. But this could be worked out as well.....actually on second thought, that may not be that difficult either.

I ponder all this because I wanted to know if the timing on these engines is static or has some type of dynamic advance or retard. It seems two strokes do better with more advance at low rpm's and less advance at rpm's. If this engine is set up for optimum advance at lower to midrange rpm's, gains could be had by creating a new CDI unit that functions as the above but has electronic delay for higher rpms. i.e. retarded timing.