Kevin Alderman,

If you prefer there is a mechanical way around the dwell time
requirements for smart
coils like the LS1 and RX8 coils. These coils have a built in power
transistor for switching the
coil on and off and only require a 5 volt pulse. 5V on (charge) 0V off
and fire.
No computer and no software.

Use a 4 ms 555 one shot trigger by the CAS mounted on a movable plate.
Something that can be made on a lathe. The 4 ms is a minimum
   charge time
for the coils.

You can mark the lever for start, cruise and full power.
Just like the old Ford Mode T :)

Hall effect sensors are the hot tip for the relatively clean aircraft
environment.
No VR board required.

They are low cost and reliable.

Paul Lamar

   This might be usful:

http://www.elektor.com/news/__crazy-summer-deal-elektor-s-__embedded-linux-board.2538177.__lynkx

<http://www.elektor.com/news/crazy-summer-deal-elektor-s-embedded-linux-board.2538177.lynkx>

   Murry I. Rozansky


   You can get a full blown PC mother for that much money
   and it will run Linux and boot off one of 4 or 5 USB ports.


   Better yet boot real time  DOS and write the software in 3rd grade
   QB :) All free.

   Paul Lamar




1. I already have a board that works with the code, it was around
$100.00 Yes I think the arduino mega or due would work- the others wont.
The short answer to why is that the others- like the UnoR3, etc. do not
have enough analog inputs for the sensors. The mega has a better
processor and 16 analog inputs for ADC. The mega android also has a chip
specifically for output to an android OS- not just a display but control
as well.

2. I have an ARM board that operates faster than the arduino and it
boots in linux, but I decided against this approach for a simple reason-
this board uses an SD card as a boot medium, and it takes over a minute
to completely boot up. So, either leave the master on and discharge the
battery to keep the computer on, or wait a minute after the master is on
to start the plane.

3. An arduino that will operate the coil ignition would not be that
difficult- using both top CAS pickups, and the bottom. Read the top at
202 degrees btc, delay it based on RPM, and, charge it for 4ms, then
fire. In theory this could be run around a bit farther for high rpm lean
mixtures. Not sure at the moment how to incorporate the additional
advance automatically, unless it is tied to the manual leaning. Have to
think about that. I guess the timing could be advanced manually with a
pot, but then you would have to have a piezo sensor incorporated in to
make sure you didn't go too far. More thinking-

4. I like the board that I have with the x gate- the x gate allows the
normal repetitive functions to be independent of the main processor- and
leaves much of the processor power for calculations. I also have a
Renesas board that has standard CAN communication protocol that lets it
communicate with other boards- It has way more power than is needed
also. For now, I will use the xgate board for the EFI.

The efi code as written (and modified by me so far) needs quite a few
sensors- intake temp, ambient air temp, MAP and ambient pressure, water
and oil pressure and temp, fuel pressure, CAS and the lower 12 tooth
gear as well. Will also need to monitor TIT and EGT and battery voltage,
all used to control the engine and let you know if something is wrong.
It should be noted that the FreeEMS guys as a group do not use the hall
effect sensors, they prefer the stock CAS. The consensus is that the CAS
is more robust than the hall effect sensor. That being said, the program
has several alternatives for using toothed wheels, or an opto isolator
wheel.  The system can use either.

The code is in C, and shouldnt be too difficult to adapt the code to
take the xgate portion and put it back into the main or at least a
header file so that it will run on the arduino. Its on the list of
things to do- just need more time.

Paul- I have a Garrett T04 turbo- talking to my mazda racing buddies in
Miami, 8 psi should get me over 300 hp, and 12 maybe 325+. I will be
setting my rev limiter for 8500, and really hoping I will have enough
power to need it :). If I can get 300hp at 15000 ft, I will be tickled
to death. If you remember, I want an almost automatic system- so
eventually the limiter will be down to 7500, cruise at 6000, etc. Push
it to max- 7500 for T/O, first detent 6000 cruise- then descent, flight
idle, ground idle, etc. With manual override as well, of course.

KA

As far as the computer is concerned I am not only looking for
engine controls but a hi rez cockpit display as well.
It is best to do the engine control and the cockpit display
with the same program on one computer. It is all the engine
parameters such as EGT's, fluid temps and intake manifold pressure that you need on a cockpit display. That is the reason I like a stock mini form factor PC mother board. They all have super VGA built in.
Tremendous market demand is forcing an ever  higher level of integration with an ever lowering cost at ever higher speed. They are manufactured in the millions.

IMHO the days of the high pin count single chip computer are numbered.

It is simple to get 16 channels of A to D with one 8 bit port.

With a 2.85:1 the prop tips run in to the mach 1 barrier
at 7500 RPM.

You can go to a much smaller diameter prop but the efficiency of the
prop will drop at least 10% and maybe more. That is a big drop
in thrust. Thrust is the bottom line.

More HP will just burn more gas without going any faster at 7500 RPM.

Best to up the boost at 6000 RPM cruise and use wider blades or
three blades.

Make sure you put a turbo tach on the turbo charger.
IMHO it's life will be numbered in less than several hours if you get
it much over 80,000 RPM. This because the strength of the
turbine blade material is vastly reduce by the high EGT's of the rotary.

Paul Lamar