You use the following information entirely at your own risk; no liability will be accepted by AMC or myself.

ECU definition files

To view or modify the contents of an ECU ROM, ECUFlash requires a definition file.

I have compiled a collection of definition files for 3rd gen Magna/Verada ECUs, downloadable in a ZIP archive: Magna_ECU_Definitions_20151230.zip (http://www.users.on.net/~andymac/Magna_ECU_Definitions_20151230.zip)

The collection also includes several definition files for US market Diamante ECUs.

These definition files need to be added to the ECUFlash installation - see the README.pdf file in the package for more information.

Vehicle types

When attempting to read a ROM image from an ECU, ECUFlash will prompt you with a list of vehicle types.

TH/KH cars have ECUs with 128kB of ROM; select the “Evo 5/6” vehicle type for these ECUs. An OpenPort 2.0 cable is required to flash these ECUs.

TJ onwards cars with manual transmissions have ECUs with 256kB of ROM; select the "Evo 7/8" vehicle type for these ECUs.

TJ/KJ onwards cars with automatic transmissions have ECUs with 512kB of ROM; a custom configuration file (known as a "read template") needs to added to the "read templates" subdirectory in the rommetadata directory of the ECUFlash installation for the whole 512kB ROM image to be readable. A suitable read template is available in this ZIP archive (http://www.users.on.net/~andymac/read_evo7gta.zip); with this read template installed, select the "Evo7 GT-A/01+ Magna AT" vehicle type to read these ECUs. Selecting the “Evo 9” vehicle type may also work, or be required, for many L/W series automatic ECUs.
WARNING: selecting the “Evo 7/8” vehicle type will only read the first 256kB of the ECU ROM; the above definition files will not work with incomplete ROM images – see the README.pdf file in the definition file archive for more information.

Both the 1.3U and 2.0 versions of the OpenPort cable can be used to flash TJ onwards ECUs, or 2001 onwards US market Diamante ECUs.

ROMs with no definition

If you find an ECU for which no definition is included in this collection, you are welcome to contact me for assistance.

You are advised to avoid posting ROMs in a public forum without having obscured the immobiliser code. I understand that the AMC admin team do not want links to ROM files with valid immobiliser codes posted to the AMC forums, nor should links to ROMs that have been modified (e.g. custom tunes) be posted.

ECUs known not to be covered by these definitions

TE/KE (ECUs not reflashable with ECUFlash and no reflashing connector)
TF/KF (pre-reflash connector ECUs, no support for reflash connector fitted manual cars yet either)
TH 3.0 manual
TH/KH 3.5 auto
US market Diamante – model years before 2001

There may also be some as yet unidentified ROM versions in vehicles thought to be covered – sometimes several ROM versions were issued for the same ECU part number.

As ROM image files for ECUs not currently supported become available, support will be added.

Updated 30/12/2015:
long overdue update including:
- revised and expanded documentation
- more ROMs supported
- expanded and reorganised map coverage
- support for neutralising the barometric pressure sensor related CEL when using a J series ECU in an E/F series harness

http://www.aussiemagna.com/forums/showthread.php?t=73939&p=1268878&viewfull=1#post1268878

http://www.aussiemagna.com/forums/showthread.php?t=73939&p=1268878&viewfull=1#post1268878

Andy's are much more up to date and cover TH-TL :)

I am much to lazy to do a compare but I thought it would be nice to share unless there was something I found that he did not.

Also those two files seemed to work on pretty much everything we could find, worked fine for auto and manuals and worked fine for TJ/KJ 3.5L

I am much to lazy to do a compare but I thought it would be nice to share unless there was something I found that he did not.

Also those two files seemed to work on pretty much everything we could find, worked fine for auto and manuals and worked fine for TJ/KJ 3.5L

While the main maps (e.g. fuel and ignition maps) and some simpler structures have stayed in the same location for the Magna/Verada ECUs, the axis tables have moved around a fair bit in the different ROM versions. The AWD and Diamante ROMs also are slightly different. Hence in my collection I created specific definitions for each ROM version (deriving from a base definition).

There are some things in your definitions I've not looked at or verified (and therefore aren't in my definitions), and some things I don't think are correctly labelled.

Thanks though for contributing your definitions.

Not a problem, it has been a while since I looked at them. They were very much in development so I may have been unsure or just given them a weird name.

I figured out most of the positions by comparing the hex on different ROMs.

ZIP archive link in 1st post updated - 2 new definitions added:
- 2001 USDM Diamante (id 87410004)
- TJ Ralliart Magna auto (id 87200002)

ZIP archive link in 1st post updated:
- fix reversed labels in the fan speed threshold tables
- bit 3 in periphery set 0/FAA: re-label to reflect function
- add EGR related maps (Magna/Verada only):
--- EGR solenoid duty cycle (category: Emissions)
--- EGR enabled ignition advance (category: Timing)
- add auto mode shift point maps for all SH2 auto ROMs

The auto shift points are fascinating to compare between the various auto ROMs. Some observations:
- TJ AWD shift points are much the same as the Ralliart auto, though the WOT upshifts are somewhat lower: AWD shifts at 5800RPM, Ralliart would appear to shift at about 6200RPM;
- TL/TW AWD has same upshifts as TJ AWD, but downshifts have been tweaked considerably;
- TJ VRX (5 speed) FWD shift points are noticeably lower than the Ralliart/AWDs;
- all the 4 speeds ROMs seem to have the same shift points, including the USDM Diamantes;
- the 3->2 downshift in the 4 speed ROMs? Mitsu obviously considered 3rd the do just about anything gear in these cars...

All the maps are present in all the ROMs, so the 4 speed ROMs have the 4->5 upshift set to a value beyond reach.

Be aware that the speed values in these maps may not be particularly well calibrated to your car - it seems common for these values to be noticeably lower than speedo/GPS readings (as much as 10% in some cars). Also be aware that there are other functions of the INVECS system which may apply adjustments to the actual shift point speed in certain circumstances.

Before playing with these maps, I'd suggest making use of an Evoscan installation, particularly to determine the TPS values at WOT and FOT (foot off throttle) as these values are important to understanding how the ECU is actually using the tables, and to correlate the speed the ECU is using compared to what you think the speed is (via GPS or speedo). As an example, in my car the TPS values are reported by Evoscan as 12.5% for FOT and around 80% for WOT.

All the maps are present in all the ROMs, so the 4 speed ROMs have the 4->5 upshift set to a value beyond reach.


Does anyone know yet if you can bring this value down and put a 5sp auto in a 4sp car and have it work?

Does anyone know yet if you can bring this value down and put a 5sp auto in a 4sp car and have it work?

That's the rumor, but nobody has tried it yet :P

You would also need to adjust the rest of the values of course otherwise it'll hit the limiter trying to get out of first at WOT :P

Does anyone know yet if you can bring this value down and put a 5sp auto in a 4sp car and have it work?

You would want to check whether any harness mods are required to support the extra gear (& I'm talking about ECU <-> transmission, not the cluster) before trying. I've also gained the impression that you can't mix TJ (& possibly earlier) auto ROMs with TL/TW transmissions or vice versa - the speed sensor signal is different I gather.

If your 4 speed box is okay, you might get nearly as much value simply by improving the shift points - in particular the 3->2 downshifts lead me to think that there are some real drivability gains to be had.

Have never explored the world of ECU's before - never had the chance or ability to do so yet. So excuse the dumb-dunt question...but..

How does this all work?! Im aware of what this/these programs do but what is reqired to connect (what I imagine is) your laptop to your cars' ECU? What other programs are required to run/read? Ive read several posts/threads about these ECU definitions and flashing but havent came across where to start.

Im more switched on with the chassis upgrade/tuning side of things - not so much the computer/ECUs. But if someone can point me towards a post/thread or give me a quick crash course on how and where to get started, that'd be tops. Having the ability to manage the shift-mapping for my 4-sp slushbox is something that I have been wanting to venture into.

Mitch

Evoscan 1.3U Reflashing Cable and a laptop, cable is available at www.limitless.co.nz

Playing with 4sp auto shift maps is fun :) http://s419.photobucket.com/albums/pp271/db_newman/?action=view&current=5c5b3189.mp4

Excellent!! Cheers Life.

Gotta love that 6750 up change!

Is it just a matter of plug and play like HDD or something?

Hi WytWun, thanks for the ECU definitions for ECU Flash, it's just what I've been looking for to reflash my Magna's ECU. I've got a 2001 Magna TJ Sports 5 Speed Auto which has a faulty ECU. I've managed to obtain a second hand ECU from the same type of car and now need to reflash it with my cars immobiliser code. You mentioned that there is a "read template" available separately for my car's ECU Rom and I was wondering if you could let me know where I can obtain this file. I have extracted all your definition files into the Rommetadata directory's subdirectory "Mitsubishi". If I manage to get hold of the "read template" file and save it to the "read template" directory within ECU Flash will I then in your opinion be able to read my ECU's ROM, save it and then reflash it to the new ECU's ROM. Any assistance in this matter would be greatly appreated.

Hi WytWun, thanks for the ECU definitions for ECU Flash, it's just what I've been looking for to reflash my Magna's ECU.

Thanks, its good to see them being used.


I've got a 2001 Magna TJ Sports 5 Speed Auto which has a faulty ECU. I've managed to obtain a second hand ECU from the same type of car and now need to reflash it with my cars immobiliser code. You mentioned that there is a "read template" available separately for my car's ECU Rom and I was wondering if you could let me know where I can obtain this file. I have extracted all your definition files into the Rommetadata directory's subdirectory "Mitsubishi". If I manage to get hold of the "read template" file and save it to the "read template" directory within ECU Flash will I then in your opinion be able to read my ECU's ROM, save it and then reflash it to the new ECU's ROM. Any assistance in this matter would be greatly appreated.

The read template ZIP link is buried in the text of the first post (I'll remedy that next time I update it) but for simplicity: read template ZIP (http://www.users.on.net/~andymac/read_evo7gta.zip)

If your faulty ECU is still readable, then you should be able to read the ROM and write that image to the replacement ECU. I would recommend taking a copy of the ROM from the replacement ECU before overwriting it though, as a backup strategy - if there's a problem with the image from the original ECU you can restore the replacement ECU's image (and see whether the immobiliser code can be transferred or immobiliser disabled pending getting it re-enabled at a dealership).

While I'm reasonably confident your ECU should have a definition, if either ROM isn't covered by the available definitions contact me by email (see profile) for further assistance.

Thanx mate, most helpfull. Harry

WytWun, do you have the cables and hardware to reflash 3rd gen ECUs?

WytWun, do you have the cables and hardware to reflash 3rd gen ECUs?

I do.

did anybody end up finding the difinitions for the TE/KE series? i have all the gear here for my subie, so i am keen to crank some timing into the rada! lol

ps, i know its a really old thread.....

I do.

What tune are you running?
And have you played around much with the maps?

What tune are you running?
And have you played around much with the maps?
I had the car tuned by an experienced tuner in Sydney for Shell E10 (94 RON), on the basis that it was readily available locally and the tune would be fine with 95 or 98 RON PULP in places where I couldn't get the Shell E10. As I have no pretensions to be an engine tuner I've not played with the engine maps. I have however spent some time adjusting the auto shift maps to suit my preference.

did anybody end up finding the difinitions for the TE/KE series? i have all the gear here for my subie, so i am keen to crank some timing into the rada! lol

While it appears that with the right equipment (such as a bench harness and the MMCFlash software) it is possible to reflash the E & early F series ECUs, I've not heard of anyone yet being able to identify the processor architecture which limits the ability to find things in the ROMs via code disassembly. Hence no definitions for these ECUs.

As an alternative, H8 ECUs from mid-1998 onwards F series can be reflashed with an OpenPort 2 cable without too much trouble as long as you have a good power supply. Adding the reflash connector to the E series loom shouldn't be too hard if you source the right bits from a wrecker (though getting pins out of the donor connectors without destroying them is a PITA).

Using an H or J series ECU works but they will throw CELs. I've been able to neutralise the barometric pressure sensor error code in a J series ECU when run in an early TF but its still throwing a speed sensor error code.

What changes have you done to the auto shift map?

Changing to a th/tj ecu would involve new keys and imobilser setup (te/f don't run bem) so probably not worth my while for s daily. It would be cool to be able to tune it like any obd2 suby though, just for fun. Thanks for the reply.

What changes have you done to the auto shift map?

I don't remember all the fine details but the main thrust of the changes was:
- make it a bit less eager to 4->5 at low throttle
- make it a bit more eager to 5->4 at high throttle but not WOT
- make it a bit more eager to 3->2 when wanting to accellerate from low throttle (typically having slowed for an intersection and then wanting to accellerate away without having actually stopped).

The first two work as desired, but the third hasn't been as successful. I think it is because there's some time delay that's having more effect than the speed based switching. I'm still trying to get a handle on the various timers used in both the engine and trans code.

If I could post an image, I'd show the original and modified maps.

Would you be able to reflash my ECU so that it doesn't shift into 1st when coming to near stop?
It's very annoying and jolts the car

Would you be able to reflash my ECU so that it doesn't shift into 1st when coming to near stop?
It's very annoying and jolts the car
While the shift maps could be adjusted, I don't think that doing what you're asking is going to produce the results you expect as I'd be reasonably confident that the problem isn't the shift.

How does a manual downshift from 2nd to 1st between 5 and 10kmh with no throttle feel? If it doesn't create much of a jar then it isn't the shift itself that's the problem but most likely the idle speed controller (ISC).

Changing to a th/tj ecu would involve new keys and imobilser setup (te/f don't run bem) so probably not worth my while for s daily.
True, if you want to keep the immobiliser...

It's not a jolt as such, but rather I can sense that it shifted into 1st while still moving and I have to depress the acceleration very gently for it not to "kick".
It does not feel in anyway abnormal, but it's just not how I'd like it to be.

Also I feel the car keeps gears for too long. Normal to sedate driving sees 3,500rpm upshifts which seems excessive

True, if you want to keep the immobiliser...

so, to install a th ecu, without immob, what involved? any harness swapping or splicing? i can get a th for next to nothing from the wreckers. if its simple plug and play, then connect the obd to disable the immob then im down for that.

It's not a jolt as such, but rather I can sense that it shifted into 1st while still moving and I have to depress the acceleration very gently for it not to "kick".
It does not feel in anyway abnormal, but it's just not how I'd like it to be.
Yes, that kick comes from the ECU starting idle speed management.

The engine part of the ECU treats speed as being in units of 2kmh (according to the logging tools available) whereas the transmission part seems to work with speed in units of 1kmh. The 2->1 downshift is set for "0" speed, which means the shift happens when speed drops below about 1kmh (integer calculations). As speed drops below about 2kmh, with the throttle in the idle position, the
engine code goes into idle mode and activates the ISC. If there's any issue with the ISC, the 2->1 downshift happening almost at the same time makes it seem as though the shift is the problem.
The issue will still be apparent even if the 2-1> shift is avoided as you propose, although it may be a bit less obvious.

The approach I would use, after having had the ISC operation checked, would be to raise the 2->1 downshift speed at idle TPS settings slightly so that the shift happens earlier. I have in fact done this in my car, but not to deal with any ISC issue.


Also I feel the car keeps gears for too long. Normal to sedate driving sees 3,500rpm upshifts which seems excessive

I still think this could be a user calibration issue, though it might be an issue with the accelerator position sensor (as your car is TCL equipped). Some logging would show a bit more of what is happening.

If you'd like to explore the issues further, PM or email (address is in my profile).

so, to install a th ecu, without immob, what involved? any harness swapping or splicing? i can get a th for next to nothing from the wreckers. if its simple plug and play, then connect the obd to disable the immob then im down for that.
If at all possible, I would try and find an ECU from a late (mid-1998 onwards) TF with the reflash connector (the covers are easy to take off these ECUs to check - you're looking for a chip marked MH7202F; 2 of them in an auto ECU). This circumvents CEL issues. Even though the immobiliser bits external to the ECU are different in the E/F to the H & later, there is still an immobiliser "switch" in the reflashable F (& H) ECUs which can be used to deal with a mismatch. For either reflashable F or H ECUs you will need an OpenPort 2.0 cable and the addition of the reflash connector (which can be done easily without a full loom change - it is only 1 wire going straight into an ECU pin).

If the car is automatic, there may be issues with the gearbox - I've seen posts by MadMagna saying the E gearboxes are not compatible with F, H & J cars but I have no idea whether this is just a loom issue or not. I've seen a J auto ECU with immobiliser disabled used in an early TF without functional issue, though the CEL is permanently on (I might be able to deal with this in a J ECU, but am not in a position to do so for an H ECU).

My little beastie is a KH 3.5L auto, so do I get to miss out on the fun of fiddling with the engine mapping?

I'm running LPG and have access to 94 E10, so was hoping to throw a bit more timing at it, as well as fiddle the shift point of the trans.
It seems to hold the gears a bit long on WOT as well as at minimal throttle (which I'm lead to believe may be a TPS related issue) but be reluctant to shift down at ~50% throttle. Was hoping to make the 3 -> 2 or perhaps even the 2 -> 1 shift happen a bit earlier to help with the 'squirt' in traffic. Having the sink the boot right in to get a shift is annoying and not that crash hot for economy.

Would be nice to get an idea of what's possible on my car to make a decision if it's worth splashing the cash.

My little beastie is a KH 3.5L auto, so do I get to miss out on the fun of fiddling with the engine mapping?

I'm running LPG and have access to 94 E10, so was hoping to throw a bit more timing at it, as well as fiddle the shift point of the trans.
It seems to hold the gears a bit long on WOT as well as at minimal throttle (which I'm lead to believe may be a TPS related issue) but be reluctant to shift down at ~50% throttle. Was hoping to make the 3 -> 2 or perhaps even the 2 -> 1 shift happen a bit earlier to help with the 'squirt' in traffic. Having the sink the boot right in to get a shift is annoying and not that crash hot for economy.

Would be nice to get an idea of what's possible on my car to make a decision if it's worth splashing the cash.
H series ECUs are reflashable with an OpenPort 2.0, but at the moment the publicly available definitions are very rudimentary (enough to fiddle with fuel and ignition maps but not much else). It is much harder to fiddle with the auto trans settings in these ECUs too, as the engine and trans code isn't integrated the same way it is in later ECUs - as a consequence there aren't any shift map definitions publicly available.

Swapping in a J series ECU makes life easier. While the easiest approach is to neutralise the immobiliser, I believe I know how to copy the necessary immobiliser info across from an H series ECU to a J series ECU and make it work if a copy of the original ROM from the car is available, though I've not been able to test this as I don't have access to an H series car.

Making downshifts happen more quickly is a pain. I haven't had time to really take apart the trans code to try and figure out why it takes of the order of half a second after applying throttle for the 3->2 shift (in a 5 speed AWD) to occur in the sort of circumstances you mention. There's one particular intersection I use with a give-way and frequently I only slow to walking pace before taking the turn in which case the 3->2 shift is a PITA :( I usually use the tippy downshift, but that corner is in an area frequented by pedestrians and I tend not to do so there...

In the 4 speeds, the reason for the reluctance to shift 3->2 is obvious from the TPS vs speed shift map... :nuts:

If you want to look at some ROMs, a little googling should turn up sources and you can install and use ECUFlash with these definitions to view them without having an OpenPort cable (you just can't flash any modified ROM).

Okay, sounds like there are some options for me.

I must admit I am reluctant to modify the old girl, since she is upwards of half a million kays and I dont want to disturb her so far into old age!
Ill have a look around and have a fiddle with the software and see whats what - thanks :)

How have you been progressing with the TH definitions, as I know someone with a dyno and cables. And would like to play with the ecu :)

Hi WytWun, I have a 2003 TL V6 auto and just need to clarify if I can delete the immobilizer check in the ECU, I live in the bush and would just like one less problem to deal with.
I am at the moment trying to find the no start issue, I have fuel but no spark, even though I have rebuilt the dissy with new coil, module, cap and leads, I have checked trouble codes and none show except if I disconnect the dissy then it throws up 44, disconnect battery, reconnect and all codes are clear.
I would appreciate any help you can offer.
Cheers.
rossi.

Hi WytWun
PM sent

Thanks WytWun, reply sent.

Okay - question time.

Bearing in mind you're not an engine tuner, I'm not going to hold you to any of this. I just want to ask questions to get my head around things, not so much "do" anything.

Can you "copy & paste" the best parts from one ECU to the next? For example (and I'm making stuff up here); Ralliart MAF, VRX shift points, Verada AC settings all combined in one ECU.

You mentioned the stock Magna MAF is limited. Is there a similar vintage Mitsubishi MAF that could be modified to work on the Magna? Might be a bit of a stretch asking you that, loose speculation would be fine. Just looking to discuss/get my head around the lean out you mentioned.

The auto shift points thing is something I'm interested in. I'd like to see the car shift a lot lower at WOT, and maybe use any other gear besides third for everything.

I'm going to do a bit of research to interpret the data I get from my Evoscan cable. The lean out should be a visible thing?

Autospeed.com.au have had articles about electronic modifications to engine sensors, to trick the ECU into certain actions.
Could the MAF signal be modified to bring readings into alignment with expectations?

Can you "copy & paste" the best parts from one ECU to the next? For example (and I'm making stuff up here); Ralliart MAF, VRX shift points, Verada AC settings all combined in one ECU.
Sure, that is the object of using ECUFlash and the ROM definition files I've created for it.

You can download ECUFlash and install the definitions without a cable. A little searching will yield copies of various Magna ROMs which can then be explored. You can even modify ROM files, but without a reflash enabled cable you won't be able to read the existing ROM or write modified files back to the ECU.


You mentioned the stock Magna MAF is limited. Is there a similar vintage Mitsubishi MAF that could be modified to work on the Magna? Might be a bit of a stretch asking you that, loose speculation would be fine. Just looking to discuss/get my head around the lean out you mentioned.
The MAF itself is not limited (at least not in practical terms for an NA engine) - rather the table in the ROM that I mentioned is used to place a limit on the MAF reading that is then used for the fuel calculations. The enleanment arises when the engine is pulling sufficient air for the MAF reading (the MAF signal is actually a variable frequency pulse train) to rise beyond the values in the table; the values in the table are then used instead of the actual MAF reading and the amount of fuel delivered is less than actually needed. This is only a problem when the ECU is in open loop mode (which is usually the case in the circumstances noted); in closed loop mode the oxygen sensor signal would compensate. Adjusting the table values upwards raises the limits and the enleanment disappears.

I don't know why the ROM code works this way - I can only speculate that it is a safety mechanism to deal with one particular form of MAF failure that the ROM developers were concerned about.


The auto shift points thing is something I'm interested in. I'd like to see the car shift a lot lower at WOT, and maybe use any other gear besides third for everything.
The stock shift point maps in the TJ AWD are identical to the shift point maps in the Ralliart auto with the exception of the WOT upshifts, where the AWD shifts slightly earlier (i.e. at lower RPM).

There are in fact several sets of shift point maps, but at this stage I only understand the operation of the set used under acceleration or rising "load". Adjusting this known set of maps covers most of the refinement that drivers usually want, though by no means all.

There are also other criteria, including rates of change of various parameters, used to "refine" the selected shift which aren't yet identified.

Shift pressure and shift timing is also specified in maps, but I just haven't got to trying to sort those out. The torque converter lockup management is something I'm really looking for - even with my mildly modded 74, the torque converter lockup behaviour annoys me.


I'm going to do a bit of research to interpret the data I get from my Evoscan cable. The lean out should be a visible thing?
The way this issue is most easily identified is with a wideband oxygen sensor. With the right calculations I think it could be identified from an Evoscan log but it wouldn't be trivially obvious just from looking at the log :(.

I have the functions and table values in my notes; I'll try and dig them out for you.

Autospeed.com.au have had articles about electronic modifications to engine sensors, to trick the ECU into certain actions.
Could the MAF signal be modified to bring readings into alignment with expectations?
In this case I don't believe so. In this case trying to increase the sensor response (to increase fuel flow by simulating increased airflow) will just result in hitting the limit earlier and make the problem worse. Reducing the sensor response will also result in enleanment (in open loop mode) because the airflow reading will be less than the real airflow and less fuel will be delivered accordingly.

You can download ECUFlash and install the definitions without a cable. A little searching will yield copies of various Magna ROMs which can then be explored. You can even modify ROM files, but without a reflash enabled cable you won't be able to read the existing ROM or write modified files back to the ECU.

So, I can experiment - look - without uploading to my car? The cable I have is only as 1.3D, which is fine for now, can I read/download my ROM with this cable?

In my search for a guinea pig ROM I found a few other threads worth reading. Most of which reaffirms my decision to not modify anything. Just look.


The MAF itself is not limited (at least not in practical terms for an NA engine) - rather the table in the ROM that I mentioned is used to place a limit on the MAF reading that is then used for the fuel calculations. The enleanment arises when the engine is pulling sufficient air for the MAF reading (the MAF signal is actually a variable frequency pulse train) to rise beyond the values in the table; the values in the table are then used instead of the actual MAF reading and the amount of fuel delivered is less than actually needed. This is only a problem when the ECU is in open loop mode (which is usually the case in the circumstances noted); in closed loop mode the oxygen sensor signal would compensate. Adjusting the table values upwards raises the limits and the enleanment disappears.

The ECU listens to the MAF until things go over a limit, then it references this table?

The oxygen sensor overrides the use of this table - closed loop? Sounds like a good argument for replacing the oxygen sensor a the specified intervals.


The stock shift point maps in the TJ AWD are identical to the shift point maps in the Ralliart auto with the exception of the WOT upshifts, where the AWD shifts slightly earlier (i.e. at lower RPM).

There are in fact several sets of shift point maps, but at this stage I only understand the operation of the set used under acceleration or rising "load". Adjusting this known set of maps covers most of the refinement that drivers usually want, though by no means all.

There are also other criteria, including rates of change of various parameters, used to "refine" the selected shift which aren't yet identified.

Shift pressure and shift timing is also specified in maps, but I just haven't got to trying to sort those out. The torque converter lockup management is something I'm really looking for - even with my mildly modded 74, the torque converter lockup behaviour annoys me.


If I were to mod anything this would be it. At least just to try it - lowering the WOT upshift RPM. The rest I could live with.


The way this issue is most easily identified is with a wideband oxygen sensor. With the right calculations I think it could be identified from an Evoscan log but it wouldn't be trivially obvious just from looking at the log .

I'm leaning toward a postal tune more and more, so I wanted to know if sending logs/data would help the tuner. It wouldnt be me interpreting the data.


I have the functions and table values in my notes; I'll try and dig them out for you.

Okay... that would be like giving a monkey a machine gun. I'm keen though. More so if I could at least read the ROM from my own car.

So, I can experiment - look - without uploading to my car?
Yes.


The cable I have is only as 1.3D, which is fine for now, can I read/download my ROM with this cable?
I'm lead to believe that it may be possible to use a 1.3D for flashing in conjunction with an ancient version of ECUFlash but it involves manually jumpering the reflash connector to +12V.

A 1.3R cable can be used with Evoscan for reflashing, but Evoscan won't read the whole of an auto J series ROM (it will read the whole of many, but not all, L/W series ROMs).

Either 1.3U or 2.0 cables work well for reflashing.


The ECU listens to the MAF until things go over a limit, then it references this table?
Yes.


The oxygen sensor overrides the use of this table - closed loop? Sounds like a good argument for replacing the oxygen sensor a the specified intervals.
Unless the fuel trim limits are reached (which I understand are +/- 25% of nominal stoich), in closed loop mode a lot of issues get compensated for.

The long term fuel trims are used in open loop fuel calculations, so a failing oxygen sensor does tend to mitigate this particular enleanment issue. It burns a hole in your wallet much more than it helps though... unless you tend to spend a lot of time at WOT ;)


If I were to mod anything this would be it. At least just to try it - lowering the WOT upshift RPM. The rest I could live with.
It is also worth tweaking the 3->2 WOT downshift downwards a bit and maybe the 4->3 WOT downshift too, though the stock J series shift maps aren't as annoying as the L/W series equivalents in this respect.


I'm leaning toward a postal tune more and more, so I wanted to know if sending logs/data would help the tuner. It wouldnt be me interpreting the data.
I'd talk to the tuner first - I doubt that most would want to spend the time going through logs unless they specifically request them (and specify what is logged).

I'll PM you with some other info.

WytWun, you spoke earlier (2-3 posts ago) about a wideband oxy sensor....and potential benefits
QUOTE "The way this issue is most easily identified is with a wideband oxygen sensor. With the right calculations I think it could be identified from an Evoscan log but it wouldn't be trivially obvious just from looking at the log "

As I am contemplating changing mine (oxy sensor) at the big 200k service, is there a particular model that has this capability (as a retail, ready to go purchase?) and may produce a wider set of number for the ECU to digest, thus giving some fuel burn advantage?

good question jdisnow

WytWun, you spoke earlier (2-3 posts ago) about a wideband oxy sensor....and potential benefits
QUOTE "The way this issue is most easily identified is with a wideband oxygen sensor. With the right calculations I think it could be identified from an Evoscan log but it wouldn't be trivially obvious just from looking at the log "

As I am contemplating changing mine (oxy sensor) at the big 200k service, is there a particular model that has this capability (as a retail, ready to go purchase?) and may produce a wider set of number for the ECU to digest, thus giving some fuel burn advantage?
Quite a number of wideband oxygen sensor systems have narrowband simulation signal outputs which could be fairly easily wired in place of the standard narrowband sensor and the wideband sensor output can be wired to an unused sensor input on the ECU from where it can be easily logged with Evoscan.

If you don't feel the need for a special gauge to show off the sensor and would settle just for ECU based logging, the least expensive wideband sensor system I've come across is the Spartan 2 sensor kit from 14point7.com (http://www.14point7.com/). I'm not aware of any plug and play solutions I'm afraid, though I think the Spartan kit comes the closest even though it would require some wiring.

The stock ECU code is not capable of using a wideband sensor input for closed loop fuel management, however there are patches to add this capability floating around the Evo community that I have plans to adapt to Magna ECUs - not sure when I'll get to this, though it is something I want myself...

:ninja:

I have made available a long overdue update to the definition files via the link in the first post of this thread (http://www.aussiemagna.com/forums/showthread.php?t=89584&p=1413699&viewfull=1#post1413699).

The update includes:
- revised and expanded documentation
- more ROMs supported
- expanded and reorganised map coverage

The documentation describes how to make some common configuration changes with these definitions, including neutralising the barometric pressure sensor related CEL when using a J series ECU in an E/F series harness.

There are in fact several sets of shift point maps, but at this stage I only understand the operation of the set used under acceleration or rising "load". Adjusting this known set of maps covers most of the refinement that drivers usually want, though by no means all.

There are also other criteria, including rates of change of various parameters, used to "refine" the selected shift which aren't yet identified.

Shift pressure and shift timing is also specified in maps, but I just haven't got to trying to sort those out. The torque converter lockup management is something I'm really looking for - even with my mildly modded 74, the torque converter lockup behaviour annoys me.

Hi Wytrun, did you progress any of this?
My only issue in my KL AWD is at high speed (>80 kmh) when I put my foot down (presumably WOT) the transmission shifts down from 5 to 3 and RPM peaks without much improvement in speed. I'd prefer it didnt shift below 4th under those conditions.
So everytime I go to overtake a slow vehicle at that speed the engine screams at me. Its really difficult to manage it in those circumstances as you want to move quickly without wasting time with the foot trying to find the sweet spot.

My only issue in my KL AWD is at high speed (>80 kmh) when I put my foot down (presumably WOT) the transmission shifts down from 5 to 3 and RPM peaks without much improvement in speed. I'd prefer it didnt shift below 4th under those conditions.

Believe it or not you are actually getting dropped to 2nd, not 3rd, which is why the engine bounces off redline when this happens (check the relative gear ratios if you don't believe me). 3rd would only have you at around 4000-4500rpm in this situation, which is actually where you want to be for maximum acceleration (as this is just above nominal peak torque), with the 3->4 WOT upshift not until over 130kmh.

If you open your ROM (85490001 for an L/W AWD) and look at the 3->2 shift map, WOT(*) downshift is set for 149kmh - when all other 5 speed ROMs have this shift set for 88kmh :nuts:

To fix this, just change the WOT cells in the 3->2 shift map to a more reasonable value - I would recommend starting with 80-85kmh as I found in my TJ AWD that I could very occasionally trigger this behaviour with the standard value of 88kmh.

The L/W AWD 4->3 WOT shift is also quite different to the other 5 speed ROMs but causes much less problem. There are improvements that could be made to this map to make to reduce the likelihood of dropping to 3rd at less than WOT, but I wouldn't worry about it until you are no longer hitting 2nd...

Note that the speed of the shift point is only approximate so you should be prepared to take several iterations to get things the way you want them. I've found sometimes quite small changes can change the driving dynamic quite noticeably. I would also advise that you test any changes in as controlled a manner as you can manage before needing to rely on them in the daily cut'n'thrust driving.

(*): WOT means TPS is >= 78% for this purpose

EDIT: if your ECU uses the Evo 9 flashing protocol, be aware that reflashing will wipe all learned behaviour (fuel trims, auto shift parameters etc).

If you open your ROM (85490001 for an L/W AWD) and look at the 3->2 shift map, WOT(*) downshift is set for 149kmh - when all other 5 speed ROMs have this shift set for 88kmh :nuts:

To fix this, just change the WOT cells in the 3->2 shift map to a more reasonable value - I would recommend starting with 80-85kmh as I found in my TJ AWD that I could very occasionally trigger this behaviour with the standard value of 88kmh.

The L/W AWD 4->3 WOT shift is also quite different to the other 5 speed ROMs but causes much less problem. There are improvements that could be made to this map to make to reduce the likelihood of dropping to 3rd at less than WOT, but I wouldn't worry about it until you are no longer hitting 2nd...
Thanks Wytwun, I will try this asap.


EDIT: if your ECU uses the Evo 9 flashing protocol, be aware that reflashing will wipe all learned behaviour (fuel trims, auto shift parameters etc).
Does that mean that the evo9 also writes every field? I thought it only wrote actual changes? How many write cycles are available then, typically?

Does that mean that the evo9 also writes every field? I thought it only wrote actual changes? How many write cycles are available then, typically?
It only writes the actual changes made to the ROM into the flash. What I'm referring to is the operating variables that are stored in RAM - these actually get retained while the car is switched off because the RAM is battery backed (which is why an ECU reset normally requires battery disconnection). The Evo 9 flashing protocol wipes the RAM as part of the flashing process, which the earlier Evo 7/8 protocol didn't. This can be both blessing and curse.

H series ECUs are reflashable with an OpenPort 2.0, but at the moment the publicly available definitions are very rudimentary (enough to fiddle with fuel and ignition maps but not much else). It is much harder to fiddle with the auto trans settings in these ECUs too, as the engine and trans code isn't integrated the same way it is in later ECUs - as a consequence there aren't any shift map definitions publicly available.

Swapping in a J series ECU makes life easier. While the easiest approach is to neutralise the immobiliser, I believe I know how to copy the necessary immobiliser info across from an H series ECU to a J series ECU and make it work if a copy of the original ROM from the car is available, though I've not been able to test this as I don't have access to an H series car.

Making downshifts happen more quickly is a pain. I haven't had time to really take apart the trans code to try and figure out why it takes of the order of half a second after applying throttle for the 3->2 shift (in a 5 speed AWD) to occur in the sort of circumstances you mention. There's one particular intersection I use with a give-way and frequently I only slow to walking pace before taking the turn in which case the 3->2 shift is a PITA :( I usually use the tippy downshift, but that corner is in an area frequented by pedestrians and I tend not to do so there...

In the 4 speeds, the reason for the reluctance to shift 3->2 is obvious from the TPS vs speed shift map... :nuts:

If you want to look at some ROMs, a little googling should turn up sources and you can install and use ECUFlash with these definitions to view them without having an OpenPort cable (you just can't flash any modified ROM).

G'day Wytwun
I've just come across your post with a lot of interest. I have just purchased the Openport 2.0 interface to be able to download the rom from my KH verada in order to retrieve the immobiliser codes and to copy the code onto a KJ ecu. As being new to this type of tuning/programming is it possible to ask for some basic assitance in terms of definitions to be able to read my rom.

Regards
Strav

I've just come across your post with a lot of interest. I have just purchased the Openport 2.0 interface to be able to download the rom from my KH verada in order to retrieve the immobiliser codes and to copy the code onto a KJ ecu. As being new to this type of tuning/programming is it possible to ask for some basic assitance in terms of definitions to be able to read my rom.

So far, I still don't know whether transferring the immobiliser code from an H series ROM to a J series ROM will actually work as I've never been able to have this properly tested. If it does work, the info in section 7.2 of the README pdf included with the definition package describes what should be copied.

I'm also only aware of one 3.5 ROM for the H series (so far) - if you're unlucky and your ROM doesn't match that definition, contact me by email with a copy of the ROM and I should be able to create a definition without too much trouble. You don't actually need a definition file to read a H series ROM - just use the Evo 5/6 vehicle type from the list ECUFlash will show (J/L/W series auto ECUs may require a custom vehicle type definition, aka read template).

With the OP 2.0, I hope you also got the Mitsubishi reflashing extension as well (Tactrix sell this separately, whereas Evoscan includes it in the package as I recall). It is possible to make a substitute but the correct connector is so much easier to use...

So far, I still don't know whether transferring the immobiliser code from an H series ROM to a J series ROM will actually work as I've never been able to have this properly tested. If it does work, the info in section 7.2 of the README pdf included with the definition package describes what should be copied.

I'm also only aware of one 3.5 ROM for the H series (so far) - if you're unlucky and your ROM doesn't match that definition, contact me by email with a copy of the ROM and I should be able to create a definition without too much trouble. You don't actually need a definition file to read a H series ROM - just use the Evo 5/6 vehicle type from the list ECUFlash will show (J/L/W series auto ECUs may require a custom vehicle type definition, aka read template).

With the OP 2.0, I hope you also got the Mitsubishi reflashing extension as well (Tactrix sell this separately, whereas Evoscan includes it in the package as I recall). It is possible to make a substitute but the correct connector is so much easier to use...

Thank you for your quick reply. In my case when I plugged in OP2 and selected 5/6, it kept asking for definitions so I installed the definitions you put up and the one it went to was the TH manual. That didnt bother me as all I'm after at the moment are just my immobiliser code to enable the transfer. Am I right in thinking that the code that I can see now even though it used a manual definition is actually from my ecu and not picked up from other definitions. I have sent you my rom just in case and thank you for your assistance.

Thank you for your quick reply. In my case when I plugged in OP2 and selected 5/6, it kept asking for definitions so I installed the definitions you put up and the one it went to was the TH manual. That didnt bother me as all I'm after at the moment are just my immobiliser code to enable the transfer. Am I right in thinking that the code that I can see now even though it used a manual definition is actually from my ecu and not picked up from other definitions. I have sent you my rom just in case and thank you for your assistance.
Yes, if ECUFlash will read the ROM from the ECU and then immediately display various values without complaints about missing definitions then you should be good to go.

Your success with opening the ROM from your ECU indicates that the 3.5 manual ROM I built the definition from is the same as the engine ROM from a H series 3.5 auto ECU, which I thought would most likely be the case.

Oops :censored: :headbange - the link to the downloadable definition package in the first post was actually to the old set of definitions :redface:

Now fixed...

Oops :censored: :headbange - the link to the downloadable definition package in the first post was actually to the old set of definitions :redface:

Now fixed...

A big thumbs up

Oops :censored: :headbange - the link to the downloadable definition package in the first post was actually to the old set of definitions :redface:

Now fixed...

I should add if you are using my ROM to add to the definitions database then by all means do so.

Hey WytWun... would there be any chance of you looking at a leggy rom and seeing if you could add a transmission definition to it? would be great to be able to adjust some trans values lol

Hey WytWun... would there be any chance of you looking at a leggy rom and seeing if you could add a transmission definition to it? would be great to be able to adjust some trans values lol
Do you know what sort of ECU (H8 or SH2) is in the car? I've been under the impression that the Legnum ECUs were all H8, even in post 2000 cars. H8 auto ECUs actually have 2 separate ROMs: the first is for the engine and the second for the trans. Each ROM is 128kB in size. The standard reflashing cable only works with the engine ROM; getting at the trans ROM requires a wiring mod to the reflash connector wiring to swap the reflash signal between the two wired pins when desired. There is a reasonable chance of finding the speed based shift maps in an H8 trans ROM if you can read it.

If the ECU is an SH2, the auto ROM will be 512kB and that I can easily deal with. You may need the custom read template mentioned in the first post to read the whole ROM in that case.

The wiring mod to the reflash connector wiring mentioned in the previous post may also permit access to the trans ROM in H and late F series Magna/Verada ECUs, provided that there are two wires running to the reflash connector.

Auto cars with SH2 ECUs only have one wired pin in the reflash connector as the trans code is in the same ROM as the engine code.

Already have some auto stuff defined, can define more if required.

http://www.ozvr4.com/threads/official-ozvr4-definition-disassembly-thread.11840/

Anyone have a manual ralliart tune adjusted for 98 Ron available?