It is pretty easy to devise a new ignition system.

You might have to cut a new trigger plate and mount it on
the e-shaft. It should look like this. Just about any MSD
ignition system should work with this. Slot the mounting holes
so you can adjust the fixed spark timing to about 20 degrees BTDC.

The trigger plate can be cut out of 1/4 inch thick mild steel
with a computer controlled plasma or laser cutter.
...
Paul Lamar ...No rotor no motor.

<trigger-wheel.jpg>

I thought for sure someone would call out Paul on the obvious error in
the "Midwest Simple Ignition" diagram.

He depicts TWO nodes on the trigger wheel.  That would fire the both
rotor ignitions every 180 degrees.

I included another diagram depicting a correct implementation of this
idea (using separate sensors for each rotor coil set).

While I was at it, I threw in sensors for other engine events as well
(using injectors as an example).  Just taking the idea to it's logical
conclusion...

(By the way, as shown, injecting at 90 degrees BTDC is WAY too late.
Just use this diagram to get the IDEA.)

-Vince Orton


Yes you are right Vince. Thanks for catching that.
Here is the original diagram again with one too many teeth.

However as I recall at the higher RPM's the injector can be
open for an entire revolution (8 ms) or more so 90 degrees before ignition
may not be optimum. As more fuel is needed the pulse gets
wider at the higher RPM's and WOT. As I recall the duty cycle on the
injector pulse width can vary from 10% to 80% and is limited to some
pulse rate which I forgot what that pulse rate is.

"Injector pulsewidth typically ranges from 4 ms/engine-cycle
at idle, to 35 ms per engine-cycle at wide-open throttle."
According to Wikpedia. Obviously at high RPM injector timing
is irrelevant. That also implies a max pulse rate of about
25 pulses per second if my arithmetic is right.

I'll have to go back and research this as we explored
this several years ago as I recall. I think the injector timing
at low RPM can be done in software using the ignition trigger to initiate the
process. That is what I originally intended. Saves two reluctance pickups.

Assume for the moment the back ground loop is incrementing a memory
location starting from zero. The computer receives a hardware interrupt from
the ignition trigger. When the ignition pulses arrives the count is noted
and set to zero again. This gives a count that is an inverse of the RPM.
Large number low RPM. Small number high RPM. Shifting this number
right one bit divides by two giving a count to half way around one rev
or 180 degrees out from the 20 degrees BTDC. When the count reaches
this number on the next revolution the injector can be fired at low
RPM's. You software types can check me on this.

BTW rotaries, unlike piston engines, are on the intake stroke almost
100% of the time. In other words the intake port is always open
except for the brief period of time the apex seal crosses the intake
port. In the PicAXE the pulse width generator is running 100% of the
time and only the pulse width is normally changed by the software
depending on the MAP or mass airflow reading.

Paul Lamar ...No rotor no motor.

I assume the "ignition device" and the discussion about injector timing
are related by using the pulse from the ignition device to time both
ignition and injector timing. Am I right?

What bothers me is that there seems to be no provision for the
trailing plug ignition. I am using 15 degree difference. Is this correct?

The way I handled it in software was to compute the time it takes to move
15 degrees and then fire the trailing plug. I am not happy with this as it
is
expensive in cpu cycles. On the other hand it provides great flexibility
which can be used to implement automatic spark advance/delay,
among other things.

What do you think about using the first pulse <from the 24 lobe sensor in
the stock CAS> after the ignition pulse? (360 degrees / 24 = 15 degrees).

I am also considering a simple hardware only solution. I am not happy
with it either because of the lack of flexibility. It does have the
advantage
of being independent of the CPU, so if the CPU goes belly up, you
would still have spark. But what good is spark without fuel?

Doc Custer
Retired Software Engineer and sometime Audiologist


The conclusion is ejector timing is not needed or even
possible at high RPM. The rotary intake port is always
open. The rotary is almost like a turbojet when
it comes to continuous combustion. It does in one rev what
a single cylinder piston engine takes two to do.
In effect the one rotor rotary does all four strokes in one rev.
The PicAXE pulse width generator runs at a constant pulse rate
not synced with engine rev. One just needs to change the pulse width
from time to time.

Tracy has been firing the leading and trailing plugs at
the same time for years with no apparent ill effects.
One advantage is: if one fowls the engine will run on the
other.

However if you insist use my background loop count and you
will always know how far back or forward to go :) Low overhead :)
Cryptic statement: slow software math can always be done away with using
frequency generators, counters and hardware interrupts :)

The techniques are outlined in this paper. Someday I will have
a pdf.

Microprocessor-Based System for Roll-Down and Acceleration Tests
by D.K. Lynn, C.R. Derouin, and Paul Lamar
Los Alamos Scientific Laboratory
Published in IEEE Proceedings 29th Vehicular Technology Conference
(The technical details of the Lamar Instruments road test system.)
Arlington, Illinois, March 28-30, 1979

Paul Lamar ...No rotor no motor.

The Rotary Engine NewsLetter. Powered by Linux.
ACRE NL web site. http://www.rotaryeng.net
Copyright 1998-2007 All world wide rights reserved.