Boost retard

[David Doddek]

alain wrote:

>
>I would like to build a simple ign retard box similar to the ones by Jacobs
>and many others, i have a 
>84 RX-7 GSL-SE this is the first year they put a 13B with fuel injection., i
>will be adding a Vortech supercharger soon and i was thinking a using a MSD
>with boost retard but hey this list is up to this task.
>The rotary engine has a similar firing as a 4Cyl engine so don't be afraid
>of it ! 
> 
>I would prefer to use a 68HC11, i have a proto board for this.
>
>I plan on locking the mechanical and vacuum advance and us a lookup table.
> 
>The ign pulse would come from the magnetic pickup process it, then send it
>to the amplifier and to the coil.
>
>Ultimately there would be a knock sensor just for safety if ever i go over
>board with the timing or get a bad batch of gasoline, but the main input for
>controlling the amount of retard would be vacuum/boost and  RPM.
>
>If this is to complex maybe just have everything left alone and retard the
>timing proportional to boost.
>
>I could build the board program EPROM's but where i lack the most is in the
>software skills kneaded for this type of project any one interested give me
>a buzz.

The software is really not all that hard.

Your mag pickup will have some number of pulses per revolution.  Since you
said it is like a 4 banger, it would be 2.  This gives you an input of
information about crank position every 180 degrees, which is where you need
to fire.  Now if you set the distributor to give these pulses 90 degrees
before TDC or what it would be in the case of a rotary, you have a starting
point for timing.

Using an input capture, measure the time between two successive input pulse
sets (actually this will be three pulses with one being at 0 time).  Take
the newest or latest time and subtract the ondest or first time.  This will
give an indication of the acceleration of deceleration rate of the
crankshaft.  Now take the newest time and add/subtract this delta time to
predict what the next time period will be.  You now have the time for the
engine to turn 180 degrees.  If you divide this time by 180, then you have
the amount of time per degree.  Follow closely here.  Take 90 degrees and
subtract off the desired timing ie 15 degrees.  This will leave the amount
of time to delay before turning the coil off.  Multiply the time for 1
degree times the amount of degrees to delay and add this time to the time
that the last distributor pulse was seen.  Send this time to an output
capture set to turn off.  I will leave it up to you how to calculate the
time to turn the coil on for dwell.  See, timing is not all that hard.

Hope I helped.

David J. Doddek                                          |pantera at pobox.com
Owner SGD Electronics & Development Engr for Caterpillar |h 309 685-7965
Formula SAE Team Sidewinder 94-95                        |w 309 578-2931
89 T-bird SC,  69 Fairlane w/SGD EFI                     |fx 217 428-4686
74 Pantera w/Electromotive Tec-II Twin turbos and Nitros |
Hey, If you are going to go fast, go REEEAAL FAST.       |