I found this circuit on the web.
http://www.mcmanis.com/chuck/robotics/tutorial/h-bridge/bjt-circuit.html

Seems to work great on 12 volts or so. I used a 1.5 ohm
resistor to simulate the coil.  Works to 125 PPS
or 7500 RPM.

The trick is to use opto isolators to trigger the FET's.

I'll know  more when I get the 300 volt power supplies.


Next step is build another one and hook them both
to the same resistor.

Basically what this does is substitute for the half bridge
driver chip.

-------------------------------------------

Here is what two CD ignition systems looks like.
Eventually it will be just one but it will fire 10 degrees
apart four times from idle to 8000 RPM in the interest
of a better BSFC. So far rotaries have used piston
engine ignition systems. IMHO not the best as the
combustion chamber is moving past the spark plugs.

I built this dual prototype low voltage version  to explore
any possible issues of having multiple 1 uF capacitors
ultimately charged to 300 volts connected to the same
coil. In this case a 2.5 ohm resistor. Better to burn
up parts at this voltage rather than risk $100 high
tech 300 volt power supplies. Also on the way.

That Internet article mentioned above ironically discussed
controlling this with a micro computer which is
what I will do to simulate the final stand alone
system. Each FET switch will be turned on and off
at appropriate times with individual ATMel 1280 bits.

Saves making a Hall Effect trigger PC board which
the final design will use. It also prevents both
FET switches being on at the same time in a totem pole
arrangement like this. That would be an instant
disaster.

I may or may not use the opto's in the final design.
Depends on how easy it is to use the Aussie type
half bridge driver IR2155 chip which is on its way from
God knows where :)

Stay tuned.

Paul Lamar

Paul,

Be aware that a resistor is a very poor stand-in for an inductive
load, such as a spark coil.  You would be better off to use
any type of spark coil laying about and add a spark plug to
the coil output to limit the voltage.  .030 inch gap or so.

CD systems can be quite tricky because of the energy stored
in the coil inductance after the 300V driver pulse.  The
optimum pulse width can be difficult to determine.

Good luck!

Craig Taylor

True but the recharge time is much slower than the cap
dump times so you will see the big differences in the
dump portion with perhaps some oscillation. I am going
slow here to resolves a few unknowns before I go
whole hog. I want to fully understand this system as I go
along.

I am in the process of getting some replacement parts
for the Aussie systems so I can get that working again.
I also have a new 12 volt supply that is  more like
a 12 volt battery. I'll use those things to investigate
the exact timing needed.

I had no idea a low voltage on the power supply could
kill it. The original designer perhaps did not know that
either and perhaps that is why several sets of FETs
were obviously blown.

Paul Lamar


From memory, the driver has two power mosfets in
a "totem pole" configuration.  The circuit that
drives the gates of this stacked pair need to
have enough supply voltage available to accurately
control the timing of the gate turn ON and turn OFF. If the gate voltage amplitude and/or timing are incorrect,
one fet can be ON when the second turns ON.  That allows
large currents to flow from the power supply directly
to the supply return thru both low impedance fets. Doesn't take much of that to cook the mosfets.

Craig Taylor


Yes I agree. That is probably what happened.
The new power supplies cuts off when the drive
or input voltage is inadequate. The also cut off
with short circuit protection I won't have to
worry about that anymore.

Here is the pdf on the one I am using.

I think I can eliminate the upper FET's
for the CD part. If I cut off the supply
during the time the caps are discharging.
I only loose a small amount of charging time.
One high voltage switch is all I need.
Saves 3 FET's.
Paul Lamar