One of  the vexing problems we have had for years was the difficulty  of
  installing off
  the shelf capacitor discharge ignition systems on the rotary.  The very
  common 2nd gen
  crank angle sensor put out only one pulse per engine rev while two were
  required.  One
  180 degrees out of phase with the other.  One pulse for each rotor.

  A four cylinder piston  engine puts out four pulses per rev of the 1/2
  speed cam shaft
  and the distributor sends the spark to the right cylinder as needed and
  one CD system is
  used.  If we could somehow get two pulses 180 degrees out of phase we
  could use two CD
  systems with the benefit of multiple sparks at a
  much higher RPM than would be possible with a 4 cylinder engine and a
  single CD system.



  Yes we could use two trigger wheels on the e-shaft  and two sensors
  mounted 180 degrees
  out.  The spark wheels would have to be slotted so the spark timing could
  be adjusted.
  The 2nd gen CAS has two sensors and it is already slotted for timing
  adjustments.  They
  are a dime a dozen on e-bay and are the most common way of triggering
  rotary engines.

  The top sensor puts out one spark per E-shaft rev and  the lower sensor
  puts out 12
  sparks per E-shaft rev.  The lower sensor is used to send RPM data to the
  computer.  Why
  does the top sensor have two tits on the trigger wheel?  Simple!  The CAS
  runs at half
  E-shaft speed left over from the days when the rotary had a distributor.

  What I have come up with is a simple and cheap two chip counter circuit
  that counts six
  pulses on the lower sensor and puts out one pulse for the second rotor.
  Ideally I'll
  sell an assembled PC board unless someone else wants to do it.  We don't
  need variable
  advance as the engine runs in a narrow RPM range of 5000 to 7500 RPM and
  real aircraft
  engines don't have variable advance.  If you want variable advance get a
  computer and
  program it yourself :)

  Here is the timing diagram. The counter chip I will try  to use is a CMOS
  4017 decade
  counter.  A hex inverter chip will provide both positive and negative
  going triggers for
  different CD systems depending on their needs.

  http://www.kpsec.freeuk.com/components/cmos.htm#4002 The count advances
  as the clock
  input becomes high (on the rising-edge).  Each output Q0-Q9 goes high in
  turn as counting
  advances.  For some functions (such as flash sequences) outputs may be
  combined using
  diodes.

  The reset input should be low (0V) for normal operation  (counting 0-9).
  When high it
  resets the count to zero (Q0 high).  This can be done by the top trigger
  pulse.

  The disable input  should be low (0V) for normal operation. When high it
  disables
  counting so that clock pulses are ignored and the count is kept constant.


  -------------------------------------------------------------
  With  the right drill motor you can pull the roll pin and gear and chuck
  the CAS for
  testing the circuit.  The shaft diameter is just under 1/2 inch.

  BTW you can use two CD systems on the leading plugs and two  Corvette
  style
  LS1 coil on
  plug systems for the trailing plugs I think.  I'll have to test that as I
  think the coil
  on plug systems require a certain limited dwell time.  I suspect if the
  dwell time is too
  long the GM coil on plug systems over heat.  In that case perhaps a couple
  of 555s will
  put out a fixed pulse width to the coil-on-plug systems.  One can always
  use 4 CD
  systems.  Racing Beat uses 4 CD systems on their racing engines.  Six on a
  3 rotor.

  Also BTW I have a report on the GM LS1 coil on plug  systems: "Turns out
  that LS1
  igniters always draw power whenever available.  Not much -- only about .02
  amp each.  But
  that will (and did) drain the battery over time for sure.  So I rewired to
  put a switch
  inline to disable them the same way the injectors need a similar switch to
  not keep
  sucking juice.  Ammeter and voltage level testing now confirm batteries
  don't get run
  down anymore."

  Paul Lamar
  ----------------------------------------------------------------------

  Or, grind off all but 2 of the 12 gear teeth for the bottom pickup coil.
  Run the MSDs from each of the coils.
  The top pick up for the front housing and the lower pick up for the rear
  housing. Have the MSD drive a double ended MSD (or other suitable) coil to
  fire  both plugs at once.

  No advance curve is required. The engine will start and run fine with
  anything from 20 to 27 degrees of advance. Mark out two sets of timing
  marks in
  the spaces between the teeth on the flex plate. One set for the front
  housing  and one set for the rear housing. Use epoxy paint. Very easy to
  see.
  Make a  fixed pointer on the PSRU mounting plate.

  Keep the plug wires short and use spiral core wire.

  Lynn E. Hanover


  That will work no doubt but it makes a mess if you do it while the wheels
  are still in there. Easy to have the die grinder slip. If you take the
  shaft
  out it is a fair amount of work to take it out and put it back in.

  I like my electronic way as it gives more flexibility on the type of
  output
  trigger. Up or down and width. Here is the counter out put and the final
  output trigger for the lower pickup or rotor 2.

  Take your pick. Both ways will work.

  Paul Lamar

  Paul,

  I'm guessing this approach wouldn't work on a 3-rotor?

  When I first received my 3-rotor I wanted to hear it run, but didn't have
  the engine computer.  Not wanting to invest a lot of money into a
 temporary
  ignition system, I modified the CAS to accept 3 separate pickups, one for
  each rotor, triggering off the dual-pointer rotor.  I fed the signals into
  three late 70's GM electronic ignition modules which I used to trigger
  three
  GM dual tower coils, each coil firing the leading and trailing plugs on
 the
  same rotor.  And low and behold...it worked.

  I'm thinking that I could use this basic design to trigger three MSD CDI
  units instead of the three GM ignition modules/GM dual coils. With each
  MSD-CDI firing only one set of plugs, it should be able to provide
 multiple
  sparks up to 7500 rpm.

  How much (if any) improvement in bsfc do you think this would give?


  Mark S.

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

  I would like to see some pictures of that 2nd gen CAS you
  modified for 3 rotor use. That is real interesting.

  I have not figured out how to do that yet so I am reluctant
  to say it won't work :) Some bright electronics type might
  come up with a solution without using a computer.  It can
  always be done with a computer. Electronically dividing by
  4 instead of six might work.  I think one may need 2 4017
  decade counter chips instead of just one. Perhaps it could
  be done with a 4518 dual decade counter.

  In your case you could always use 3 pickups
  and one tooth on an e-shaft trigger wheel which
  would constitute the sledge hammer approach :)

  I am hoping for a lot of BSFC improvement but I don't
  really now if one can measure it short of a dyno run.
  Perhaps getting one of those HyTek Automation data ack
  systems and recording the injector pulse width, the intake
  manifold pressure and the RPM and you have what you need
  for a pretty good way of measuring the relative BSFC as you
  will know the relative HP and the absolute fuel burn.  You
  have 8 channels of 12 bit A to D to work with.  That is
  pretty high resolution in anybodies book.  That is one part
  4096 parts.  The other five channels can record outside air
  temp, air speed and altitude among other things.

  To re-word what you said MSD claims there is not enough time
  to fire the plugs more than once over 3000 RPM.  I am
  hoping that translates to 6000 RPM since we would be using
  one unit per rotor.  A one rotor engine is equal to two
  cylinders while a 2 rotor engine is equal to four cylinders
  as far ignition requirements are concerned.  MSD is talking
  about a four cylinder engine in this pdf I think.

  There is little doubt in my
  mind that multiple sparks at high RPM would be beneficial
  as our combustion chamber moves past the spark plugs unlike
  a piston engine combustion chamber where the plugs are on
  one end.


  Paul Lamar

 Paul,

 I'll look around and see if I can find it and send you a picture.

 Mark S.

 You are up early :)

 I just had an idea. A way of testing a highly multi spark
 system in a rotary with off the shelf hardware.

 Since we have a moving
 combustion chamber that stays in contact with the leading spark
 plug for over 270 E-shaft degrees we need an entirely different
 type of ignition. If we fire 20 degrees BTDC the combustion
 chamber is still in contact with the leading spark plug for 220
 degrees  of e-shaft rotation.

 In fact it is more than that. The center
 of the fire stays in contact with the spark plugs much
 longer than a piston engine due to the geometry and
 centrifugal force on the mixture molecules.

 Here is the idea. Use 3 MSD units per rotor.
 MSD claims the spark duration is 20 degrees.
 One fires at 20 degrees BTDC. The second fires at
 TDC and the third 20 degrees ATDC for a total spark
 duration of 60 degrees. Three times longer than what
 MSD claims.

 One tooth on the trigger wheel and 3 pickups for
 each group of 3 MSD units. The groups of three are
 180 degrees out for a 2 rotor. For a 3 rotor of course
 you would need nine and 120 degrees out.

 What do you think?

 If that does not improve the BSFC I don't know what
 will.

 This is for testing purposes only of course.
 It is far too expensive for production. A custom
 designed ignition system for the rotary is what we
 really need.

 Paul Lamar


Paul,

Remember, there's a 2 hour time difference.
You're the one that's up early.
;-)

What ever happened to your idea for a
120v multi-spark CD ignition system?
Sounded good to me.

Mark S.

The design is not complete. I am looking
for something we can test right now with
off the shelf hardware to see if it is
worth pursuing.

Here is a set of graphics on the idea.
When you look at this you wonder why
Mazda did not try it years ago.

Paul Lamar