This looks like a real simple way of using megasquirts on rotaries and you
have
redundancy. The engine will run on one rotor. It solves a lot of problems
using just
one Megasquirt on a rotary.
I has the advantage of using a notebook computer with a USB connector to
tune and
save injector tables to share with other people merely by emailing them.
Megatune is very sophisticated and easy to use tuning software that runs
on a notebook.
IMHO this is a great idea.
Paul Lamar
-----------------------------------------------------------
Using MegaSquirt ‘out of the .box’
www.not2fast.com/megasquirt/doc/rotary_injection.doc
MegaSquirt introduction.
The MegaSquirt was design by Bruce Bowling and Al Grippo to be a simple
and cheap
batch fire fuel injection system. Been around for years.
There are 2 injector driver outputs on the MegaSquirt, running off the PWM
channels
of the MCU. This allows both saturated and peak and hold injectors to be
used on the
same hardware: very clever. This allows 2 different injection events to
occur in
parallel, which suits our purposes very nicely.
Specific needs of Rotary engines.
The rotary engine, due to the rotor layout has a few key differences with
a
reciprocating engine. Many of these work in our favor. And remember, the
otto cycle
is the otto cycle:
There are 2 rotors, spaced 180° apart. EFI.
Build 2 MegaSquirts.
Fit a pair of hall effect sensors one at 22 degrees BTDC and one 180°
later (top and
bottom) Hook one MegaSquirt to the top sensor and one to the bottom.
Burn in injector staging code if required.
Ignition.
The ignition pulse on 1 rotor corresponds to 22° BTDC on the inlet cycle
of the other
rotor, give or take the amount of advance being used. Very handy.
Use 2 MSD capacitor discharge units per rotor. One per plug. This is what
the racers
do and is it very simple and straight forward. Each Hall Effect sensor
fires two MSD
units 22 degree BTDC.
Injection.
Around 220-280° is available for injection. In a reciprocating engine, you
can ‘puddle’ fuel on the inlet port to increase the injection time. In a
rotary there
is little time for this before the next inlet cycle starts. This forces
relatively
large injectors to be used, with staged injection being a
useful method of getting high max flow rates and a good dynamic range. For
example,
at 8000 RPM 1 revolution is 7.5ms, giving around 6ms for injection.
In terms of injector strategies there are a couple in general use
The best for our use here is 2 injectors in each runner.
The code mods are very simple, if needed at all
You get the full 6 ms for injection at high RPM
2 MegaSquirts are needed so adding a full $100 to the costs
Tuning will be a little more tricky as you have to set each parameter
twice
----------------------------------------------------------------
I had an idea of using 3 megasquirts that fed a logic box that compares
the outputs
and forwards the one that at least 2 ms boxes agree on. Not sure if it's
really
feasible since that would induce some lag in the outputs.
Joshua
---------------------------------------------------------------------
SymTech Labs' coolant temperature sensor is a standard GM CLT sensor
capable of
interfacing with all MegaSquirt ECU's. It features 3/8" NPT threads and
can also be
used as an intake air temperature sensor (IAT/MAT). Each unit is shipped
with a 6"
pigtail for easy installation.
Intake Air Temp/Manifold Air Temp
The big question that I have is there seems to be conflicting info. Often
they talk
about a MAP sensor but this block diagram does not show it. A Manifold Air
Pressure
sensor (MAP) can be used to control the amount of fuel injected just like
an O2
sensor. That is what Tracy does.
Any type of feed back from an O2 sensor will not work as 100LL is death to
O2 sensors.
The MS flow chart shows a read of the MAP but the schematic shows no
connection to a
MAP sensor.
Can anybody help me out on this? Is the sensor on the mother board? Where
is it
connected?
BTW The megasquirt has the built in capability to send engine parameters
back to the
notebook in real time for use as an EFIS. The notebook software is already
available
to display these engine parms in real time. This is a real cheap EFIS.
Paul Lamar
Separate Megasquirts for each rotor is a great idea as long as there is no
danger of fire or explosion if the ignition for one rotor fails but the
fuel in-
jector continues, and unburned fuel air mixture continues out into the hot
rotary exhaust environment. Perhaps an electronic switch can be in-
corporated into the Megasquirt circuitry that allows the fuel injector to
fire
only after a recent ignition signal--"recent" as established from voltage
remaining in a capacitor which is otherwise charged from ignition
activity.
I am very interested in this one-megasq-per-rotor approach and would
appreciate if anyone has any ideas. Thanks.
Gene Kahn
Gene,
Had a computer re-installed in my 86 RX7 and an injector pin got crossed
over into a ground. The injector went to a steady on state. Was totally
undriveable, smoked like all get out but didn't seem to pose any fire risk.
John
Finally found the MAP sensor. It is mounted on the back side of the board.
Tracy uses the same sensor.
There are dozens of files in MS on the internet. I don't know how anybody is
expected
to sort all this stuff out. If I did not have extensive experience in this biz I
would be lost.
http://www.rotaryeng.net/IEEE-road-test-sys.pdf
Paul Lamar
Paul,
I am currently building up the fuel system for the P Port. Oil and cooling come
next. Iwas planning on using Megasquirt and am interested in how 2 units will work
together. Could you post how the ignition will work? Hoe do the other aftermarket
CD systems get wired into the system?Also one of the vendors is offering plug and
play as well as kits for all of the MS family of boards. It would be interesting to
see what version you pick for aircraft use as some of the latest versions have many
options that we do not need.
Joe Berki
Limo EZ
I am going to order one next week and test it. I'll let you know what I find out.
I'll test it on the bench with a pulse generator to simulate the CAS or Hall Effect
sensor, an O-scope to check the injector pulse width and hand vac pump to simulate
the MAP. I will buy the early/cheap version with asm language as they are easier for
me to modify the code to suit the rotary. I have a set of boards that I will sell
cheap if any one wants to build a kit. Cash only.
The CD ignition is real easy as you have no dwell to worry about. A CD system fires
the plug the instant the pules comes in from a CAS or Hall Effect sensor.
Just use 2 sensor 180 degrees apart on the flywheel. One for the front rotor and one
for the rear rotor. That eliminates all the kludges trying to get the Megasqurt
ignition to work with a rotary. Use the Megasquirt for fuel injection only.
However using one MS per rotor may simplify the ignition situation and I might be
able to get it to fire an RX8 or LS1 coil. I don't know yet.
Paul Lamar
Look at the jimstim and if you have a ms3x the jimstimx. See
DIYAUTOTUNE.COM for a source.
chris smale
I don't think I need it Chris. I have and complete selection of pules gens and
O-scopes. I do have the stim PC board unpopulated so far.
BTW do you still have your airplane? I could drive up in the RX8 and take a look.
Paul Lamar
--
The Rotary Engine News Letter. Powered by Linux.
ACRE NL web site.
http://www.rotaryeng.net
You Tube key word PaulLamar2
Copyright 1998-2012 All world wide rights reserved.
