bez_bashni

I am thinking on the way to do this relativly simle without total reconstruction of the flash.
However i dont know if it can work .

The CONCEPT for the discussion.

Hardware mods :

1. Remove the mav.
2. Put the 2 bar map -wire it to the stock barometric sensor.
3. Put the AIR temperature sensor between intercooler and manifold.
4. a. Open the ecu, get the cranc signal from the comparator and wire it to the mav signal.
5 b Or put he external comparator and put the cranck signal to the MAV input.

So what do we have: MAV frequency- Linear with rpm so MAV FREQ/RPM =CONSTANT

Firmaware -
1.Constant number in MAF calibration table.
2. Flash VE Tables into the place of fuiel maps..

I dont know exactly how mitsu ecu calculates fuel pulse - i assumed it is done like described here:
http://www.gems.co.uk/downloads/Manu...0draft%20R.pdf

If it multiplies the resulting value the temperature related coeff and pressure related coeff - it will work. If its adds temperature trim and pressure trim - sorry - i wont work...

Just an idea...

MalibuJack

iTrader: (5)

I have heard some rumor that now that the Roms are available, there will be attempts at rewriting the ECU's code itself, so even though it might not be possible to do it your way, adding/changing functionality may not be impossible in the near future..

FWIW I'd be interested in seeing if this would work.. But as far as I know, the air temp and air pressure are used to baseline the MAF input to calculate Air Mass from Air Volume, so it likely wouldn't work the way you would need it to.

jcsbanks

I did a conversion before the days of MAFTPRO on an L-Jetro Power FC for the Subaru (you can't buy a D-Jetro version).

I used an Atmel AVR Mega 8.

The tacho output or a feed to an ignitor module were passed through a resistor and a shunt diode into a logic input and I used a timer to count the period between pulses. Inverting this gave RPM.

An Apexi 3.5 bar (approx) MAP sensor was used and the output was fed into an ADC input.

The code in its simplest form multiplied MAP and RPM to give a high frequency PWM output which was put through a simple low pass filter and fed into the ECU as a simulated MAF signal.

In the ECU compensations were made for VE in the airflow calibration tables which were rewritten to also do a transfer function that mirrored one in the MAF simulator to give more resolution at idle (square root then squared).

Later compensations were added to the project to give crisp throttle response as the MAF sensor would give an inlet manifold filling spike during throttle tip which the MAP sensor doesn't.

There was an air temperature compensation available, but we found it best not to use it because when the air temperature increased the mixture would richen and the ignition retard in a safe manner. The effect was less than expected possibly because the fuel also heats up at high temperatures.

This setup has been used on the fastest manual Impreza in the world which has recently done 9.3 at 160, and over 210 mph on a 2.3 engine. I used it on my 11 sec daily driven Impreza. The reason I developed this was that there were no speed density ECUs that had the manners I wanted, and the MAF sensor didn't like the 4" ported shroud GT30R (or bigger) inlets. We could have gone to blow through MAF conversions as an alternative.

For the Evo it would need the RC filter moved from the MAF output, also the transfer function is linear I believe, and it would be a frequency output instead of PWM.

Whilst easy to visualise all these steps, the development time to get this to work is immense.

MalibuJack

iTrader: (5)

There are already devices for maf translators available to do what your describing, either with speed density or a different type of meter.. No rewriting of the ECU's code is necessary.. I have an analog version for my car using a ford cobra sensor, but will likely get a digital version soon..

I would strongly suggest staying away from any sort of speed density setup (Even this isn't truely speed density) and stick with using higher flow capacity MAF sensors with the stock ECU.

bez_bashni

I also planned to build the translator on a 25$ Atmel AVR atmega 169 butterfly kit (i already made boostcontroller and datalogging on it)-but the idea is to decrease data conversions and use stock ecu potential .
I'm planning to make some test to determine if it works
InjectorPulse~Load~Freq/RPM*Baro/Temp + Latency(Voltage)+Acceltrim
or InjectorPulse~Load~Freq/rpm+ Latency(Voltage)+ Trim(Temp)+Trim(Baro)+Acceltrim

The second idea is to use stock mav sensor in pair with 2 bar map in blowthru config.
Map into the manifold BEFORE throttle- and multiple the mav reading by pressure.