Subject: Sorting the trees from the forest in megasquirt.
From: rotaryeng
Date: 7/22/2012, 10:49 AM
To: AAAA Put this in the To box


 There is bewildering amount of info and options in Megasquirt.
 The relative low cost ($250) and very simple early board known as the MS I V2.2 is
 still available and will work on a rotary. It does not have injector dampers but you
 don't need injector dampers with Mazda high impedance injectors. Also it cannot stage
 the injectors.

 Both sets of two injectors fire at the same time on each rotor. No staging of the two
 injectors on any one rotor. That means you will get the engine to run well at medium
 and high RPM with a prop load but it may idle very poorly as the dynamic range of
 most injectors is limited. For example the RX8 engine has 3 staged injectors on each
 rotor and Tracy uses 2 staged injectors on each rotor. The injection is not timed
 simply because a rotary rotates too fast to time the injectors.

 Also the software does not support timed injection. It is referred to as "batch fire"
 and that is probably what Tracy is doing as well. Timed injection does not make much
 sense in a p-port rotary because the intake port never closes. That is unique to an
 intermittent class combustion engine. Yes the rotary is an intermittent class
 internal combustion engine. Only turbo jet engines work that way and they are called
 continuous combustion engines.

 The ignition is not there so you must provide that. In fact it won't even work unless
 you provide a tach signal of some sort. You must use it with one of several stand
 alone ignition systems that will work and that put out a tach signal. These are GM
 HEI, GM DIS, Ford EDIS, Ford TFI, Bosch 0 227 100 124 and MSD 6A. IMHO the MSD6 is
 the best of the bunch but expensive. All of these are documented.

 No direct coil control (dwell) is available on MS 1 Version 2.2 and that includes
 late LS1 and RX8 coils with built in high current switches. Of course there is no
 electronic way of leaning the engine from the cockpit. You could rig a mechanical way
 of changing the pressure setting on the fuel pressure regulator to affect leaning.

 The documentation frankly is a mess with megasquirt. Basically what has happened
 over the years is one kludge is piled on top of the last one as they have added
 features. There is a lot of redundancy in the html docs (many web sites) so the whole
 mess of dozens of html files  has becomes extremely confusing.

 The the top of the line hardware is also a mess with daughter cards and adapters
 piled on the original MS II using the MC9S12 on daughter card instead of the
 MC68HC908. The whole mess is in dire need of separate USB mother boards tailored to
 the needs of the customer. Say 4, 6 and eight cylinders.

 A documentation trick I use is set up Word doc files on various subjects such a USB
 and dwell among others and cut and paste the pertinent info from the many redundant
 html files.

 I aim to fix all this with a simple board specifically designed for the rotary with
 just the features we need.  USB and RX8 or LS1 coil support (dwell) and cockpit leaning.

 Paul Lamar


  I am watching with interest Paul. Glad someone can sort this!!.
 BTW Paul are you intending to use an O2 sensor. The 4 gas exhaust
analyzer is supposedly lead safe, cheaper and can be used for best
ignition timing indications, but I have no idea If it is compatible
with your planned set-up. Thanks Paul
Graeme Riley

There will be an input for 0-to 5V O2 input but the megasquirt software only uses it
for initial tuning. It does not use it for run time feedback control.

"Since the MegaSquirt® EGO input (pin #23 on the DB37 connector) will accept a 0 to 5
volt signal, no changes are needed to the MegaSquirt® EFI controller hardware, and
you can simply change the switch point (2.50 for stoich.), and sensor type (since NB
and WB have opposite slopes to their response), and use it to tune."

"The MegaSquirtâ„¢ EFI controller EGO correction algorithm treats a WB O2
sensor/controller as if it were a narrow band sensor with a different voltage and
slope (see diagram), but does not take advantage of the fact that it can accurately
report AFR away from stoichiometric. In this sense it takes limited advantage."


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.