Paul,

  My situation [flex plate failure] was exacerbated by a lengthy period of
 tuning
 issues. Turns out it needed diodes on each of the fuel injectors.  There
 was
  also a problem with the Renesis style CAS, which I later had converted
  back to the stock CAS.  I'm sure the rough running engine didn't help
  the gearbox any.

  Mark

  The RX8 style trigger wheel with one CAS is difficult to handle in
  software over a wide RPM range. It is easier to use two CAS's 180 degrees
  out. The RX8 software must discriminate between 3 different wave shapes.
 It
  takes a very fast processor and maybe an EFI specialised computer chip to
  do that. Or missing pulse hardware.

  Which CAS system did you revert too? The 2nd gen or 3rd gen?

  Tell me more about the diodes on the fuel injectors.
  FET's usually have them built in.

  Paul Lamar

  Paul,

 I reverted back to the 2nd gen CAS... the one that came with the engine.
  My EC-2 was initially set up for the 2nd gen CAS, but I decided to change
 to the newer Renesis "crank mount" style.  Tracy had to "convert" my EC-2
 to work with that system.  After over a year of trying to tune it, I
 finally had it changed back.  It was an improvement, but there were still
 tuning problems.  It wasn't until after Tracy told me about the 3-rotor
 applications needing diodes in the injector circuits (very near the
 injectors) was the problem fixed.  I don't think the diodes are needed on
 the 2-rotor engines, or with the new EC-3 controllers.

 Mark

 That version has two mag sensors. One for the RPM and one for 22 degrees
 BTC. That is
 much easier to deal with in software. I have an early EFI and I will take a
 look at
 the FET's and see if they are the kind with the built in diodes.

 Paul Lamar

 Timers are used in EFI CPU's to interpret the data coming in from the CAS.

 A little background on CPU timers.
 Obviously one of the primary uses of timers is to measure time. Duh :)
 When a timer is used to measure time it is also called an "interval timer"
 since it
 is measuring the time of the interval between two events.
 It will increment by 1 every machine cycle or 921,583 times per second in a
 CPU.
 8000 RPM is only 133 revolutions per second.

 In a 2nd gen CAS there are 24 teeth on the lower wheel but it rotates at
 half engine
 speed. That means there are 12 pulses coming out of the lower mag pickup in
 one rev
 of the engine.

 Times 133 and we get a frequency of 1596 PPS.  That means the timer in the
 CPU can
 easily determine the length of time for one of those teeth to go by the
 lower mag
 pickup.  The resolution of the measurement is 921,583/1596 or one part in
 577.
 More than enough resolution for RPM and we are checking the RPM 12 times per
 engine
 revolution.

 The drawback to this method is the processor is being interrupted 1596 times
 a
 second. That is once every  625 usecs. It takes about 2 usecs to complete a
 asm
 instruction so we must get everything that needs to be done in about 300
 instructions. Does not leave a lot of time to send engine parameters to the
 instrument panel display tablet or cell phone.

 Now having said all that I am not sure any given EFI program works in that
 fashion.
 There are other ways of doing it.

 Paul Lamar



Paul,

Thanks for the simple explanation of how all that counting happens.
It has always amazed me at how the ECU can determine the exact time to
fire each plug with the engine running at 7k and above.  As you have
described, this is a piece of cake for those little microprocessors.

I've been lurking and reading about your EFI project with the intent
of possibly using it for a 20b test stand project.  Eventually, it may
find its way into something with wheels.

Mark

There is another point about that 2nd gen CAS method and those 12 pulses can tell the
computer where the eshaft is to the nearest 28.8 degrees. Not enough to make fine
adjustments in the ignition timing with out resorting to delays. Delays are a major
waist of precious time in the microprocessor world. At 8000 RPM there are 8,000
microseconds per rev (8 Milliseconds)  or 22 microseconds per degree if my arithmetic
is right.

One would like to advance the ignition at low manifold pressures at high altitude up
to 50 degrees BTDC because the BSFC gets better.

Here is another interesting tid bit. Since the dwell time required for the coil on
plug type ignition is 4 msec. A signal to charge the coils from the computer must go
out 200 degrees BTDC so the coils will fire at 20 BTDC!!!

Paul Lamar