Paul,

 Is using a DC charged cap, dumped to a coil the most efficient way to
 generate your big multispark?

 You _know_ that a high voltage DC power supply has a fair amount of
 circuitry in it to make the output smooth and ripple free.  This
 circuitry serves no purpose in your system.  And may likely cause you
 operating problems.

 What could work really well is a 12V DC driven buck-boost converter,
 switched by an oscillator that's running at the frequency of your
 spark strike rate.  The oscillator would be gated off when you don't
 want any spark.  The output of this circuit would directly drive the
 coils.

 I haven't run the numbers to see if the AC output from a boost
 converter could be made fast enough properly drive the coils.  Might
 have to play with a SPICE model to see how it would work with real,
 available components.



 Matt-

 Sounds good. Right now I am using an off the shelf,
 300 volt, 30 watt power supply that cost $130. I has a lot
 of built in safety stuff. For example if it senses a short
 it shuts down. It is also regulated to 9 volts which is
 required for starting. My current over all design will work down
 to 9 volts or lower using two 14 pin dip isolated switching supplies
 that put out 15 volts with only 5 volts in. These are used for
 driving the FET's. They too have some built in safety features
 that prevent low voltage damage to the FET's.

 Paul Lamar


Do you have a graphic which shows the balance of time between desired
sparking duration, and waiting for the combustion chamber getting
ready for the next cycle?  It looks like a single CDI spark event
might be 50-80us..


Regards,

Matt-

Is this what you are looking for Matt?

This was the original 4 module plan but I am going to try
and fire one module 220 usec apart because the
time constant of a new .3 ohm FET (old FET was 3 or 4 ohms)
and a 1 mfd cap looks favorable for recharging the capacitor
in just a few usec. If that works it will cut the cost.

This is worst case at 7500 RPM. At 6000 RPM cruise the sparks
are further apart time wise but still 10 degrees apart e-shaft
angle wise.

If a waste spark coil is used for trailing plugs it would
have to fire every 4 ms which is still doable I think.
That would eliminate one module and one coil.

Paul Lamar