Reading over that patent it seems to mainly address the trigger circuit for the SCR.

The real genius in this was realizing a standard coil could be used with a CDI. Was this part of the claim?

Actually, My father designed it to use a standard coil because that was all that was available to him at the time (1960/61 era) . If more efficient coils were available, he probably would have gone that route. However, the reality is that even though there are coils that are more suitable for CDi as far as energy output goes, a standard canister coil is quite sufficient if you throw enough energy at it. The standard coil offers some advantages too. It slows the voltage rise which makes it less likely to damage insulation and also greatly reduces the likelyhood of cross-fire between ignition leads. A lazy coil isn't necessarily a bad thing. Also, with mine, since it has a switch, it allows switching between both systems without changing coils. By the way, that simple trigger circuit can be configured about 3 different ways and work in exactly the same way. Recently, I saw a patent for touch screens that cited my father's trigger circuit because they used the same simple system for cancelling out double taps from jittery fingers. The only drawback to that trigger method is the upper rpm limit which is about 533 sparks/second. Most engines don't rev that high so for most applications it works well provided the right combination of components is used (not necessarily the same values cited in the patent either) Fred

Like so?

http://www.classicretrofit.com/image.../dualspark.bmp

Btw, that's the second spark 'proper' spark coming along 200us later. MOSFET based design (not SCR).

Jonny,
No, Not like that. Yours looks more like the open-circuit waveform other than the first bit. If that's so, your actual duration may be much shorter than what the scope picture seems to depict, but I could be wrong.
With mine, the actual waveform under load becomes more ragged with positive peaks and negative peaks rising much higher than in your scope pic. I interpret this as far as current goes, as several additional short duration sparks within the main spark event. (don't look on the internet for one of my scope pics as they haven't looked like that for the last 2 1/2 years) In your scope pic, it seems as if the capacitor is giving up most of its energy as the spark is first struck, and due to the nature of the nice, clean ringing afterwards, I would conclude that the spark has extinguished. For example, the same sort of thing occurs with mine after the spark series is over. The way I test for duration is to compare waveform shapes from the secondary loaded, to the open-circuit condition. I count the peaks that form when loaded and then subtract the rest at the tail end. It's easier to show than explain, but that's how I do it. It's the same method you can use to prove the duration of the Kettering system, but it takes a little more finesse with the CD because of the alternating nature of the CD waveform. Fred

Well, we don't have a problem with spark duration. Our spark 'interval' is programmable so can be tuned to coil type. Been playing with 120us interval recently, more than 8 times the rate of MSD.

Jonny, It does seem that a spark separation of 200uS or your proposed 120uS is certainly way more useful than the MSD separation of 1mS. That I think is far too long a separation between the onset of each spark event. Programmable is cool. I'm letting it happen naturally and so the effective duration varies with coil. Like I mentioned, there is always more than one way to do something. I prefer the simple analogue approach, but if I had a computer background, it would likely be different. Fred

Assuming the design replaces the SCR with a MOSFET, how do you keep the capacitor from discharging through the body diode of the part? Do you actually wait 200 uSec and trigger another spark?

Somebody help me out here. I'd like to figure this out so I could figure out how that new box works.

If I model the first 12 uSec of the Bosch CDI voltage waveform below as a triangle I get 1/2 * 360 * 12u or 0.00216 Volt-Seconds. Apply that to the coil and at 12 uSec divide that by 8.5 mHenries and get 254 mAmps with 0.3 mJoules stored in the coil (Lm).
That dv/dt also has 40 Amps coming out of the capacitor.
At 12 uSec Vc=0 and Vcoil=0

The energy stored in Lm isn't enough to charge that 1.4 uF cap up to 200 Volts.

Where is all the energy stored? Did I make a mistake in the math? Is the secondary leakage not insignificant (Ls)?

http://i804.photobucket.com/albums/y...psg7pnp7nd.png

https://upload.wikimedia.org/wikiped...39/TREQCCT.jpg

Apparently the assumption the parasitic inductance is trivial is a bad one.
How do you model a spark gap? This can't be it.

http://forums.pelicanparts.com/uploa...1466397032.jpg

Where did you get the values for the standard small Bosch CDI coil used in all the
pre-911 3.2 engines or are you using a different coil?

I Guessed. I was trying to match the voltage waveform of the picture.

I remember the 8,5 mHenries and 2 Ohms from some other thread, The 68 H is the 1:90 turns ratio. The .06 leakage inductance and the 10 Ohm spark gap were guessed at to shape the waveform.

http://forums.pelicanparts.com/uploa...1466436769.jpg

Rick, Have a look here: http://www.worldphaco.net/uploads/CAPACITIVE_DISCHARGE_IGNITION_vs_MAGNETIC_DISCHARG E_IGNITION..pdf

Dr Hugo Holden has done a lot of fine work already that might help you. I don't agree with everything he has written about CD ignition, but you won't find a better place to look if you want to study the subject. Fred

911SC - Primary - .20 mH, Secondary 1.2 H, turns ratio 78
911 3.2 - Primary - 3.6 mH, Secondary 30 H, turns ratio 90

The calculated "reflected" primary inductance based on the measured waveform period is .046 mH.
The actual measured primary inductance with the secondary shorted, e.g. a spark, is .042 mH.