Paul. I have looked into the Ford EDIS ignition system after reading about
it on the newsletter. I bought the EIDS-4  which is "electronic distributer
for four cylinder"   It reads a 36 tooth minus 1 tooth trigger wheel to do
the timing which is 10 degrees BTDC from the installed position of the
trigger wheel with the missing tooth aligned with TDC. Mounting it 6 degrees
advanced will give a 16 degree BTDC spark event. I also bought the 4 coil
pack which is for lost spark use. The lost spark wires the two plugs in
series connected by the engine block giving one spark electrode to shell and
the other shell to electrode. The majority of the energy goes to the plug
under compression with little expended on the exhaust stroke plug.

I also
bought an assortment of EEC-IV Ford engine computers. They cost from $1 to
$44 dollars on E-Bay. These were chosen by pre coded ignition key(1996) and
after 1988 to be mass airflow with 1993 thru 1995 my preference as these
also work sequential injection and multi spark for the lower rpm range. They
run off the stored tables at wide open throttle and below normal operating
temp. In betwen they use algorithms to determine lamda=1 from a wide band
oxygen sensor.

Any oxygen sensor voltage between .33 and .80 doesn't change
mixture, a pretty leniant range. The ignition timing comes from the variable
reluctance sensor as it falls thru 0 volts to the EDIS module which
conditions the signal and sends it to the computer. The computer (EEC-IV)
sets a pulse width signal output back to the EDIS every 5 spark events which
advances the timing to a maximum of 59 degrees. This range is a result of
the binary count, nothing to do with wanting that much lead. The MAF sensor
and injectors have to be a matched set such as an 80 MM meter and 19 thru 24
pound injectors for the table values to be correct. The Map sensor is open
ended and used for barometric corrections and in some installations reads
the manifold pressure and sets the barometric presure at key on and wide
open throttle. If the throttle position sensor is not used only the look up
tables are available for mixrure control making the oxygen sensor also
redundant.

The tables devide the airflow into 30 divisions with multiple
tables for all the perameters sort of like a multiple axis matrix nor just a
single table. The computers have a 60 pin connector and if from an economy
box type car about 40? of the pins are unwired. There are two types of
injector firing for V6 and V8 engines, both sequential and firing every
injector on a cylinder bank called bank fire. IT appears either system works
the same on a rotary since we need only two injector events. The processor
generates a result aparently in 1 microcecond when using the tables ans 3
microseconds using the algorithms. The avarage injector duty is 30% and
maximum of 50% and mid RPM pulse width is 3 milliseconds.

The EEC-IV can not
be reflashed internally hoever a second connector called J3 on the other end
allows using aftermarket chips th vary injection from 0 to 720 crank degrees
and ignition from plus 59 to minus 20 degrees from TDC.


The next
generation  beginning 1996 can be reflashed but has a 104 pin connector
nearly all of which connect to something like transmission hydraulic
pressure and shifting, turning off the air cond  at WOT or extended idle,
cooling fans.......and the list goes on and on.

When you get your PC based injection running it would be nice if
you could post a table of optimum air fuel ratios for a P-port on your dyno
from 4000 to 7000 RPM. I am convinced the MAF sensor is right for me and it
seems you are going with the MAP sensor now that you have the signal
generation worked out from the oil ports.Also I want the ignition timing for
the lead plugs to be computer controlled from load and MAF on a separate
EDIS-4.


With the 4 rotor I have 8 plugs and will need 2 of the EDIS
controllers to run them. Having the second plug at a fixed timing and
independantly controlled seems like the safest use of the parts. I am also
considering putting two of the composite MAF sensors in series and using two
EEC-IV controllers for redundancy on the injectors with only one powered up
at a time. I am looking at the "tweecer" brand of recalibration chip for the
J3 connector and having the optimum air fuel ratios would be very helpfull.


 I started back on the CAD drawing last week. A student version of
Autocad 2010 opened  most ofthe Acad 14 drawings but would not render the
solids, just a black screen. The loft command balked and displayed invalid
sections  even trying to loft some circles. Either the student version
doesn't loft or my install is not right. The screen is mostly occupied by
"ribbons" with only half the screen available for drawing due to tripple
redundant places to start command menues. It ran well of the command line
but without the loft it is useless so I threw it away and put in an old hard
drive with Rhino. The loft worked on the firt try and the used interface was
very easy to use. It is not photoreal but does a colored shaded loft with
one keystroke.

I finalized my ports for the end housings but Rhino won't
open 3D autocad files of the housings to apply the ports to them.

I am in
the process of separating all the 2D sections out of the autocad drawings to
import into Rhino to recreate the solids in Rhino. It's the old two steps
forward and 1-1/2 steps back.

On a positive note the Meshcam3 program easily
created the G-code files from the stl files except two parts which  have an
error. I can run the ports as a separate Meshcam3  program to make the
casting patterns but would rather shift everything into Rhino and have all
the engine stuff in one place.

The airframe work was finished in Autocad so
no translation is needed just printing the drawings to scale on my wide
format roll printer.I was hoping to have a core engine ready for sun-n-fun
but the redrawing will be too much  delay for this years event.

george grimes

Stock systems are not a bad way to go but in the air leaning is the
problem. Perry has been flying a stock RX7 computer for years. O2
sensors are fouled by 100LL but Perry tells me the system defaults
to open loop. Open loop is too rich so the fuel burn suffers.

Paul Lamar