RomRaider

Donations are welcome!

Here's an XML file that defines ALL the tables in the ROM in their RAW state. You can use it to locate the tables of interest and refine a usable set of defs from it.
Fuel tables will be in the 3D - uint8 section. In Raw form the cells will be the % off of 14.7, i.e.: 14.7/(1+x)

Dale

Edit: fixed axis endian value to big

So one would look start searching at tables with similar y and x elements like 15x14?

That's really cool!

Any chance you could make the default scalings for axes just be float values with no conversions? I checked a dozen or so tables at random and that always made the values more intelligible. Same is true for the table content when the table has 'unknown' data type. I guess this may not apply to 16-bit ROMs, but for the 32-bit ROMs I'm pretty sure this is going to be more usable.

Turbofied, I think the first step would be to find a D2UH001M ROM (05 Liberty 3.0R). Find the fuel table in there, then look for tables with the same dimensions in your ROM.

A web search for "D2UH001M" didn't turn anything up. Try asking on various Subaru boards if anyone has a stock ROM from their car that they're willing to share.

So i have 1 table

I taking the more simple route and just trying all the tables in ecuflash

1 of the 15x20 tables is a fuel table all the values make sense are are interpolated in a tuning fashion.

now just a matter of working out why there is 5...

im thinking map a and b then fail safe a and b...

Yes, you will probably find that two of them are set richer - those would be the fail-safe tables.
That would account for 4 tables. I have no idea about the 5th.

Ive goone through 5 15x14 tables they are all timing

and 5 20x15 tables and 1 is fuel

still searching but should get results this way id imagine


found two 14x10 tables same sizes as the 3.0r liberty tables looking more promising this time

Here is the two tables.

Anyone see a problem with it starting at 3200?

the liberty and legacy tables are only 14x10 tables...

Since the 3200 values are 14.7 subaru wouldn't make it bigger is the values below it are stoich also.

the maps make perfect sence just need confirmation on the 3200 bottom...

I just opened another h6 file with a predefined definition and it starts at 3200 so it must be right.


Seems to be a slow but fairly sure way to get results if you have a similar base map to get table sizes. then just keep plugging them into ecuflash and evalute map data and trends to allocate what table goes where.


Thanks for help that raw dump was great how is it performed?

There are a lot of ecu's trying to keep stoic below 3000 RPM.
I saw the same for earlier H6.
And this is not a good thing for moderate octane gas.

Thanks for the confirmation

Yeah well the reason im doing this is because this stock car is detonating on australia's best fuel, on lower grades its even worse. robbing so much power running them so lean also!

Well that was very successful.

I have all of the 8 timing tables and 2 advance tables.

They are the same as the 09 legacy.

Can anyone explian why some timing tables go to 4400 max and some go all the way to 6800

this is related to my definition its related to most of the romraider and ecu flash defs...

Can i post the xml up her for someone to check?

4400 RPM limited table are probably failsafe.

European H6 cars sold here in Russia love 102 RON badly that is not available at all.

There is no problem to enrich from 0.7 g\rev the lowest RPM row upto 5-7% (14.0-13.8 AFR) in order improve bottom end.

And completely redesign ignition tables in order to "downgrade engine" for 98 RON available.

There is no scaling applied to any axis. Axis in the 32bit ROM are all in float so what you see is the value from the ROM represented in decimal format. The only thing that is scaled is the unit8 and 16 bit table data. The scaling is straight out of the ROM, the same values used by the ECU, you can see the formula in the Table Properties window.
What you will see for instance with the ROM table axis is:
Temperature is in Celsius
Percentage is 0.0 - 1.0
Pressure is in mmHg
Time base might be in microseconds but I have not confirm that yet.

So you are after the open loop and timing tables then... Because there is what appears to be 4 closed loop compensation tables too in the 3D - unit16 section size 9x13.

Out of those Timing tables I suspect there are A & B (possibly C & D) tables and the ECU uses all of one map, all of the other map or a ratio between the two. Think of the two maps as a 3 dimension cube two maps thick. A on the left and B on the right. The ECU can take any value along the two axis RPM and Load around the outside and any cell value between the lower bounds of the A map and the upper bounds of the B map. In other ECU's already defined there is an extended logging parameter that gives the ratio between the two maps. I believe it is Map Ratio...

Are we having fun yet?

Well i found the maf scaling, avcs A+B, and both o2 scaling also because they have stand out elements which are easy to search.

I have 2 open loop fuel tables and the timing tables are the same as 09's

A,B,C,D,E,F,G,H

4 full scale and 4 fail safe (upto 4800 it seems)

there is also advance A and B. A for fail safe and B for full scale

I was wondering why it would have so many timing tables your explanation makes sence because the SI drive i thought only effected the throttle tables. So 3D interpolation of variables sounds rad!

I have found plenty to be able to tune so i will probably stop there. had more then enough fun for a day or two

1 more thing for the sake of learning, in that raw file where are the single element tables like injector scaling? the minimum i found were 2 element tables