Repairing CDI ignition unit

On a sunny day (Sun, 05 Apr 2009 20:45:51 +0200) it happened Helge wrote in :

I am not sure there should not be a cap on the input between L2 and D3, if so then it is a voltage doubler, else it just rectified the peak. So, if you have a scope, look for the peak to peak voltage on L2, that will be what C2 maximum is charged to. The positve impulse from the pickup coil L1 triggers the thyristor, causing C2 to scharge into the igniton coil primary. R2 limits the gate current, so if you know the peak voltage on L2, and max gate current allowed for the thyristor, then you can calculate R2.

R2 indeeds depends on the voltage L1 gives. Looking at the elco of

L1 gives a voltage that give a peak current throug R2 and C1 into the gate. And then C1 slowly discharges over R3. I expect R2 to be a factor x smaller than R3.

If something fails in this circuit, I expect it to be the thyristor. Of course any component can fail, it is all pulsed and used in higher temperatures.

P.

I agree Jan ! Its a voltage doubler. The capacitor will be in the region of 500 - 550 volts wg. I would guess about 470 Ohm for R2.

The circuit is not too dissimilar to the units I used to make 20-25 odd years ago. In those I used a pair of power transistors in a self oscillating inverter producing about 300v across the cap. The caps used to fail quite often. Probably couldn't handle the current through it.

R2 has to be lower than R3, R3 discharges the cap after the thyristor has completed a pulse. This normally only takes place after a firing trigger from the fly wheel as gone by.. You have more time in discharging the cap while it's waiting for the flywheel to present another trigger.. The trigger pulse of course must present a much more announced current pulse in the gate.

The spec's on the thyristor indicates 200 ua on the average to trigger how ever, a practical design does not operate components at their limits. The max is .2 amps. the reverse gate voltage is no more than 6 volts so, the minimum turn on is ~ .8 volts. This leads me to think that maybe you have ~ 2 volts to work with. This leads down to a rather smaller resistor. I would assume a 5 ma drive pulse would be in order to trigger the gate. And picking a standard out of my head 470 Ohm R should do it.

On the cap, if the body is large, .75 " or more, I was say a 600V cap.

"Helge"

Oil filled paper /foil types are reckoned to be the best ( longest lasting) here.

..... Phil

I just checked. There is definitely no cap there.

I (obviously) can't start the engine, so I have to measure while pulling the starter cord. That gives a good 150V peak on L2, and just under 10V peak on L1. This is measured with the coils connected directly to the scope, no other load. The voltages will probably be higher when the engine starts, since it will be spinning much faster. Here are a couple of screen shots from the scope:

I'm not sure I understand the circuit completely.

The data sheet says gate trigger current max 200uA, and gate trigger voltage max 0.8V. That should give R2 a theoretical maximum value around 10k? The max gate current is 200mA, so around 50 ohm theoretical minimum for R2 at this speed?

The motor will be spinning at least three times as fast when running. Will the voltage be three times as high? If so, R2 would need to be at least 150ohm plus safety margin.

OK. I can't use a polarized cap here, can I? I'd guess the inductance of the ignition could would cause the cap to be charged a little "backwards"?

Thanks. I was just about to order a polyester one.

I'll see what I can find.

On a sunny day (Tue, 07 Apr 2009 19:20:40 +0200) it happened Helge wrote in :

Ok, well then I do not quite understand why they short the negative pulse to ground.

Cool!

Think of the primary ignition coil as a short for a moment for DC. Now the right side of C2 is grounded, and C2 is charged to +150V by the impulse from L2. A little later the magnet passes by L1, L1 sets the ignition timing, its position is probably adjustable. When the L1 pulse comes, the thyristor is switched on, and grounds the left side of C2. Now a current flows through the primary ignition coil and thyristor from C2, C2 discharges, until the thyristor current drops below its hold current, and the thyristor stops conducting (voltage may reverse too, same effect).

With 10 V positive pulse on L1, .7 V drop for D1, and .8 V drop for the thyristor, you have 8.5V left. So 100 Ohm makes 85 mA....

No, the voltage will not be much higher I'd think, but the frequency will be higher.

PS Actually maybe the voltage will be higher, but I dunno how much. Better try a higher value say 330 Ohm for R2, measure the voltage then with full speed. Same for the capacitor, try a 400V one, just measure it. If not enough, then use a higher voltage one. Next one up maybe 600 or 1000 V. Those caps need to be able to do the current too.

ground.

To build up a large current in the coil, which will give a bigger pulse when the magnet moves across it and the induced polarity flips over to the other side.

Jon

Polyester (or any plastic film) caps may not be able to stand the heat in an engine application.

I think an oil-paper designed for high temp operation is going to have to be the only choice here.

Jon

ground.

To find out just how much bigger, I just did a test with the diode (D3) connected. It made quite a difference. This is the voltage across L2 and D3:

The picture is actually lying a little. The probe is about 110X, while the scope thinks it's 100X, which means the voltage peak is now a full

This test is done without any cap to load the coil. I expect the cap to dampen the voltage peak a bit, but I'm still beginning to doubt whether a 630V cap will be enough?

On a sunny day (Thu, 09 Apr 2009 21:15:48 +0200) it happened Helge wrote in :

to ground.

That is very interesting. I'd never expect a voltage dobler to work that way:-) Something to remember for the bag of tricks.

e pulse to ground.

e pulse to ground.

Here is something which might help.

1. Try the R2 of 39 or 47 ohms.
2. The capacitor C2 is Styroflex, or Mylar, as shown in the picture. This variety is non polarised. They come in ratings above 100 Volts, generally. For this case, use any Voltage ratings above 600 Volts. Will be interested in knowing the outcome!

I couldn't find an identical thyristor, so I just bought a cheap one that seemed to handle the voltage and current, without looking carefully enough at the other specs. It turns out this new thyristor needs a lot more current and voltage to trigger. As a result, our calculations on R2 are not directly applicable.

I tried 1K first, but the thyristor did not seem to trigger, so I tried 100ohm. That gave a nice spark, so I kept that one.

I got a 630V polypropylene one. The dealer's web shop says it's suitable for high current applications. It will have to do for now. It is, unfortunately, much larger in size than the original one, so it won't fit into the original casing. I'll just have to get a little creative, I guess. There's plenty of room, so it should not be a problem.

The outcome could not be better: The engine starts easily and runs smoothly. I just ran it for 10 minutes without any problems.

The bad news is, my cheap Chinese scope totally freaked out once the engine started, so I was not able to get any measurements on the running circuit. The scope behaved as if the buttons were wildly pressed at random. I'll see if I can figure out what happened.

Thank you very much to everyone who helped out.

Since you don't know the gate specs of the device you have, you should use the highest value that gives consistent triggering.

The polypropylene one should be fine ! I think those are supposed to be self healing. I found that the cap was the component most likely to fail SC. They tend to get quite warm in use.

Sounds good !

There are some very high voltages and fast transients generated when the engine is running. I used to be able to get inch long sparks out of mine open circuit. About 16-17Kv on the plug.