Dear AMC members, I wish to share my current ECU ROM file that has been working fine on my NA car (A manual TJ Sports ECU ROM ID 9832000) for the past 2 months but still being NA, there is no way to test if the 100+ load cells are being used.

With some help from posts about ECU disassemby at evolutionm.net and plenty of reading of the SH2 assembly language, which is used on the magna and EVO ECU, I have managed to implement bigger fuel and timing maps.

The maps now exceed the 100% load limit, thus should allow the ecu to be used in boosted applications. I have compared the lookup routines between the evolution and our ecu's and it is identical, so it should look at values above 100% load.

So unless there is any other load limit implemented somewhere else in the ECU ,which I could not find then all should be well.

I don't have much time these days to carry on experimenting with the ECU, but would still like to post my work so that if there are any the other members who are interested and willing to experiment, will take this to the next level.

http://img.photobucket.com/albums/v199/charandeepinhouse/bigmap.jpg


HERE IS THE ROM AND XML:


http://www.fileserve.com/file/Cgnt8J9/MAGNA_BIG_MAP.rar


THE ROM POSTED IS ONLY FOR EXPERIMENTAL PURPOSES. PLEASE USE AT YOUR OWN RISK AND BE PREPARED TO PROPERLY MONITOR FUELING AND TIMMING AS I WILL NOT BE RESPOSNSBLE IF ANY DAMAGE OCCURS.

Ahh nice, you wouldnt know of a way to give the ECU a way to measure boost??

I'm curious as to how you've actually implemented the maps.

I've also been experimenting with big maps in an NA car (a TJ AWD), and while patching the ROM table/axis definitions works for most parts of the map, I've discovered a problem with one particular part of the map (low revs/high load) where spurious ignition advance is being added and causing knock. I'm still chasing the source of the problem as time permits.

Ahh nice, you wouldnt know of a way to give the ECU a way to measure boost??

I'm not quite sure what you're after here? The ECU code can deal with boost just by looking at the MAF output (which is all the Evo ECU really does for the fuel/ignition maps).

So the Evo ecu's dont have a provision to measure boost at all? How is it possible to measure more than 100% load also? This is what I dont understand.

How can more than 100% load be measured by the ecu without a provision to measure boost?

So the Evo ecu's dont have a provision to measure boost at all? How is it possible to measure more than 100% load also? This is what I dont understand.

How can more than 100% load be measured by the ecu without a provision to measure boost?

The MAF effectively measures air mass flowed per unit time. From the measured air flow rate the ECU computes how much air per cylinder stroke has been inducted and from the AFR computes how much fuel to inject, then opens the injector(s) for an appropriate time to deliver it.

A turbo or supercharger just pumps more air through the MAF, which is duly measured. The later 3rd gen MAFs and Evo MAFs include a barometric sensor which is only used to adjust the computed air mass for changes in atmospheric pressure (*).

Hence the actual intake boost (or vacuum) doesn't need to be directly measured.

You'll only run into problems when you hit 100% injector duty cycle as you can't deliver any more fuel to match the airflow (**).

The Mitsu MAFs can read considerably higher volumes of air than are ever flowed in a stock engine, though the ECU calibration for the MAF is usually optimised for the normal range. Evo people do adjust the MAF calibration (sometimes called MAF scaling) with wilder builds.

ECUs which rely on measuring boost/vacuum don't use a MAF, but they still end up calculating the mass of air in the cylinder (using what are known as speed density calculations) - they just do it differently.

(*) with after market FI (turbo or SC) its generally better to suck through the MAF than blow through it as a) the stock airflow calibration is done for standard atmospheric pressure and b) the stock baro compensation curves aren't really up to dealing with boosted "atmospheric" readings. If you redo the calibration, blow through can apparently work well (it can work without calibration changes, but not as well).

(**) there is also a MAF limit table that needs to be doctored too; without doing this the ECU will decide the MAF has gone troppo and shut down the engine.

Dear AMC members, I wish to share my current ECU ROM file that has been working fine on my NA car (A manual TJ Sports ECU ROM ID 9832000) for the past 2 months but still being NA, there is no way to test if the 100+ load cells are being used.

With some help from posts about ECU disassemby at evolutionm.net and plenty of reading of the SH2 assembly language, which is used on the magna and EVO ECU, I have managed to implement bigger fuel and timing maps.

Having now looked at your ROM, I see how you have implemented this - the same way as I have. Please be aware that from my experience this approach is incomplete. The problem may not show up in some (even many) usage patterns, but I have a log that shows up to 20 degrees more ignition advance than called for by the ignition map at a load/rpm point where there are no other active adjustments (EGR etc). It is probable that specific map sizes may make this issue more or less obvious, and the conditions likely to provoke it may be more easily encountered in an auto (which I have) than a manual. And yes, the extra advance did cause knock which was quite audible to me.

PLEASE BE EXTREMELY CAREFUL AND ALERT IF YOU USE THIS!


The maps now exceed the 100% load limit, thus should allow the ecu to be used in boosted applications. I have compared the lookup routines between the evolution and our ecu's and it is identical, so it should look at values above 100% load.

So unless there is any other load limit implemented somewhere else in the ECU ,which I could not find then all should be well.

There is an RPM dependent MAF limit which needs to be increased; without doing so the ECU will consider the MAF faulty when the measured airflow increases above the limit and shut the engine down with a CEL.


I don't have much time these days to carry on experimenting with the ECU, but would still like to post my work so that if there are any the other members who are interested and willing to experiment, will take this to the next level.

{...}

THE ROM POSTED IS ONLY FOR EXPERIMENTAL PURPOSES. PLEASE USE AT YOUR OWN RISK AND BE PREPARED TO PROPERLY MONITOR FUELING AND TIMMING AS I WILL NOT BE RESPOSNSBLE IF ANY DAMAGE OCCURS.

You are to be commended for releasing your work, but I consider it most unwise to have released a functionally modified ROM without changing the ROM ID - this is a recipe for confusion, impedes comparison with a stock version of the ROM and could result in someone making an expensive mistake despite your warning.

So the Evo ecu's dont have a provision to measure boost at all? How is it possible to measure more than 100% load also? This is what I dont understand.

How can more than 100% load be measured by the ecu without a provision to measure boost?

Hmmm... I realise I didn't quite answer your questions...

I understand that Evos (at least the 8 onwards) do have a MAP (Manifold Absolute Pressure) sensor, its just not used for fuel management in the stock ROMs - the MAF is still the basis for fueling.

As for the 100% load issue, don't get too hung up on it meaning the transition from vacuum to boost - its mostly just a way of referencing the nominal maximum airflow into the engine in the absence of any intake augmentation - a normalised value, rather than an absolute value. Boost just happens to be the only way to get beyond this threshold.

I'm curious as to how you've actually implemented the maps.

I've also been experimenting with big maps in an NA car (a TJ AWD), and while patching the ROM table/axis definitions works for most parts of the map, I've discovered a problem with one particular part of the map (low revs/high load) where spurious ignition advance is being added and causing knock. I'm still chasing the source of the problem as time permits.

If you examine my ROM and XML Definition file, you will notice, I have had to implement two other maps that reference the same RPM and LOAD Axis as the Timming maps.

This other maps that had to be made bigger as well are: Low Coolant Temp Timing Trim and EGR Enabled Advance. This could be the reason for your ignition issues.

So the Evo ecu's dont have a provision to measure boost at all? How is it possible to measure more than 100% load also? This is what I dont understand.

How can more than 100% load be measured by the ecu without a provision to measure boost?

The 100% load is just the maximum air/rpm that a certain sized engine can ingest naturally at fully open throttle.

Theoretically you can calculate the maximum amount of air an NA engine can ingest. i.e The maximum a magna 6g74 3.5liter engine can ingest is 3.5 liters of air+fuel every four strokes at 100% Volumetric efficiency(VE). This theoretical maximum can be considered 100% load and any measurent above this is boost.

Atmospheric pressure is 14.7 psi so cylinder pressure at 100% load at 100%VE is 14.7psi.

If we were to get to the 150% load cell, then cylinder pressure would be 150/100 X 14.7 = 22.05 psi absolute pressure, which is 7.35 psi boost (22.05 - 14.7) for our puposes.

The MAF, measures airflow in HZ and uses the Temp and Baro sensors to determine air density, all this is used to calculate the volume of air entering the engine every revolution.

for our purposes LOAD = CALCULATED AIR MASS / RPM (there may be some other factors as well)

If you examine my ROM and XML Definition file, you will notice, I have had to implement two other maps that reference the same RPM and LOAD Axis as the Timming maps.

This other maps that had to be made bigger as well are: Low Coolant Temp Timing Trim and EGR Enabled Advance. This could be the reason for your ignition issues.

I had those maps enlarged too. I'm still looking for whatever I missed at this point...

Very interested guys, keep up the good work!