PDA

View Full Version : Moving forward with ECU FLash



scientist
11-10-2010, 01:37 PM
Like I suggested on OZVR4, to help simplify matters and speed up development on the VR4 roms, we need to decide on a universal ROM ID.

Since the roms can work on all ecus with a few definition changes it would be simple to deploy.

It would help eliminate the issues of addresses and do things similar to how Ceddy has set it up for the Evos.

pbaron
11-10-2010, 07:29 PM
Would you be looking to maintain this information in both forums or just here on CVR4, how do we get this moving forward as I am keen, cheers.

Wodjno
11-10-2010, 08:26 PM
Sounds like a brilliant idea /yes

Even though i have no idea what it means /help

scientist
11-10-2010, 10:19 PM
Right now i only have the following rom ids in my possession...but don't know how well the definition files i have are.

Since things are all over this forum

I have:
23810003
20030011
20030013

Adam.Findlay
12-10-2010, 01:46 AM
so you are trying to deploy a complete definition file? so every aspect of the ecu can be altered.
that would be fantastic as we could get so more from the car with a complet definition. I will be folowing this very closely

pbaron
12-10-2010, 02:08 AM
After agreeing on a base image and custom ROM ID. Basically we pick a base ROM (7202 manual), work on producing an accurate definition file for this base and then everyone applies it to their ecu regardless of whether you are running a 7202 or 7203. That way everyone is working on deciphering the same base image rather than having to get the definition files correct for lots of different images, if there is no problem with applying and running the image on any 7202 or 7203 ecu then this is the way to go IMO.

so you are trying to deploy a complete definition file? so every aspect of the ecu can be altered.
that would be fantastic as we could get so more from the car with a complet definition. I will be folowing this very closely

scientist
12-10-2010, 03:19 AM
There will not be a problem. You can flash a 7202 rom to a 7203 ecu. You just have to make sure you set the right mem model in the definition file.

So at moment...which romid is the most defined?

scott.mohekey
12-10-2010, 03:24 AM
Just subscribing to this thread.

Patryn999
12-10-2010, 03:43 AM
Must admit, I would be very tempted to use the newest ROM as it might have bug fixes/new implementations or something in it.

When I was making up a definition file for the EM24xx (can't remember off the top of my head) there was a byte offset for most tables, and some tables were moved, but I could find all the tables/locations from the 7202 definition on the newer ROM.

As for the most defined, probably have to ask merlin that one ;) But my guess would be the EM2004.

scientist
12-10-2010, 04:58 AM
So 23810004 it is.

So here is what i'm going to do

238x000y

for x:
1 = Manual
2 = Auto

y:
1 = 7202
2 = 7203

These will all just be modified versions of the 23810004 rom



So in the end we will have:

23820002 would make the rom work on a 7203 Automatic car.
and 23810001 would work on a 7202 Manual car etc.

Will start working on a draft to start tommorow

pbaron
12-10-2010, 07:04 AM
OK, who has a copy of an unmolestered 23810004 rom that they can upload here or send me? Cheers.

scientist
12-10-2010, 12:48 PM
I have a few 238 roms that have been posted...but not sure if they are molested or not

foxdie
12-10-2010, 03:26 PM
Sounds like a brilliant idea /yes

Even though i have no idea what it means /help

At a guess mate, they're talking about having a stock definition file (ROM) for our cars, devoid of specifics such as whether the car;
Is Automatic or Manual
Has Traction Control
Has AYC
Has a 7202 or 7203 ECU
It would be a base for making fuelling maps, fuel cut levels, rev limiter etc etc.

Then when the tuning is done, it would be trivial to re-add those features above back in again. It would mean that all of us could reap the same benefits and not be wholly reliant on having one-off unique tuned ROMs for each variant of the EC5x VR-4 :)

scientist
12-10-2010, 04:47 PM
basically what he said.

But I will be doing testing to make sure it doesn't loose any features on certain vehicles or throw any check lights.

I have a 2001 Auto VR4 I can do testing on and a 97 manual. Both have AYC and the 2001 has trac

Patryn999
12-10-2010, 09:46 PM
Yeah, I have (what I believe to be) a unmolested 2381 auto rom on my PC at home. Correct me if I'm wrong, but I was of the understanding that the difference between ECU code for auto/manual was just an entry in the periphery bits indicating which map set to use?

Question is we require a place to store the rom/definition. And a sufficient interface for storing the .xml def, if we plan on working on it/discovering more stuff. I'd suggest an SVN on my PC at home (have a laptop running a mailserver for canterbury climbing club, so always-on is not an issue), but I'm not sure how many people here would be familiar with a version control system (definition developers only).

On a side note I had a small email chat to Ceddy, who's done the modifications to the Evo roms to add SD, CEL on knock and a bunch of other stuff. Turns out becauce the compiler & linker is not ideal he did it all in a hex editor... might start looking at the code jumping to known locations (such as timing tables, etc) and see if the code is similar to evo, in which case we may be able to port some of the mods across to the VR4.

Hopefully :P The latter suggestion could all turn out to be considerably more difficult than imagined.

scott.mohekey
12-10-2010, 09:49 PM
I've had a fairly in depth play at disassembling the roms with IdaPro, and found it to be VERY helpful. It doesn't help as far as compiling stuff goes however, but it does have a hex editor built in (from memory).

I second the idea of version control, but I would suggest using http://bitbucket.org/.

foxdie
12-10-2010, 10:08 PM
That or Assembla (http://offers.assembla.com/free-subversion-hosting/) that offers unlimited users and unlimited repos with a 2GB limit spread across all repos you create.

I can personally vouch for the service as we use it for hosting web application code that we need to share with our clients directly.

As the likelihood of reaching that 2GB limit is highly unlikely when you're working with 1-2MB files, this may be a viable option for you guys :)

Nutter_John
12-10-2010, 10:11 PM
Ok a question , what happens when someone uploads a corrupt file into the repositry and people start to use it , this then blows up someones engine

where is the quality control and change control processes

scott.mohekey
12-10-2010, 10:13 PM
What we can do is have 'branches' where there is a development branch that we work on and a 'stable' branch that people use who are NOT developing things. Bit Bucket also allows file releases/downloads which means we could put known working/stable files up in a separate section to avoid confusion.

Jason, I would still recommend bitbucket, simply because it uses mercurial over subversion.

Patryn999
12-10-2010, 10:16 PM
The idea being that only the developers would use the repository, and we (hopefully) would pay a little more attention to details such as what was changed.
Probably would also require people to sign up, so we know who changed what/etc. (elimintates most random trolls).

A stable (and tested) release would be posted on the forum(s)? by one of us for other people to use. (Or as scott suggested, on third party site).

I don't see it as any less dangerous than the current system we have of everyone willy-nilly handing out definition files.

Edit: Too slow in my proof reading... Scott won the race.

foxdie
12-10-2010, 10:23 PM
Good point.

And one thing these versioning systems typically allow is the ability to compare differences, and even "blame" someone for something :)

Nutter_John
12-10-2010, 10:24 PM
Well playing devils advocate the current system does not imply that any of it has been tested , where as the system your suggesting is offering tested versions .

By the way I am not knocking the efforts that your trying to do , I am making sure that all understand what the risks are .

Patryn999
12-10-2010, 10:48 PM
Sounds like you've used BitBucket before Scott, I dug a little but couldn't find out if a public repository (hence unlimited users) can have write restrictions. It'd pretty silly if that wasn't the case, but documentation stating that seems to be non-existant.


Yeah I understand the devils advocate position, but in any scenario we will always have to release things with an "at your own risk" clause. We're doing this not-for-profit, so we can't afford to do investigations on peoples bricked ECU's to determine whether or not it was our fault. Editing your ECU can be dangerous, and we will have to put a disclaimer somewhere stating that.

But we'll now make sure that we do put something in ;)

scott.mohekey
12-10-2010, 10:58 PM
Carl, what you can do here is have the development repository a private one, and have the public one frequently pulled from the private one.

scott.mohekey
12-10-2010, 10:59 PM
Actually, you can define 'readers' and 'writers', people with read access and people with write access.

Patryn999
12-10-2010, 11:09 PM
Was just thinking about the 5 user limit on private repo's. But then we're probably not likely to get 5 of us actually working on the definition file are we?

scott.mohekey
12-10-2010, 11:37 PM
Well, with the discovery of readers and writers, it doesn't need to be a private repo.

Patryn999
13-10-2010, 06:27 AM
Guess we wait for Scientist's (sorry dude, don't know your name) opinion and start putting something together in the way of original ROM's and base .xml's to start with?

jasone
13-10-2010, 08:24 AM
if somebody tells me how to i will retrieve what you need from my car it's a 1997 auto. Willing to help if i can.

Nutter_John
13-10-2010, 09:16 AM
There is nothing for you to retrieve Jason as yours is a 7201 and none flashable , the only option for 7201 is one of kinkyafros' and nutter_johns reflashed ecus

Pugme
13-10-2010, 10:30 AM
as john knows, i have a spare 7203 for some reason out of my coty if that helps anyone!

scientist
13-10-2010, 05:51 PM
All great suggestions as it creates a nice general location to obtain the roms + mods.

taylor
14-10-2010, 01:13 AM
Been meaning to jump in here an comment.

Has anyone been able to get in contact with valmes and get a copy of his def file? I have compared his modded rom vs my stock rom, using the latest definition from ozvr4.

There is alot of defined address in the file however it is unknown as to what they do.

Comparing Valmes rom to mine I noticed a lot of the unkown address had been changed.

My experience with tuning is getting the Russian tune flashing it and using map tracer in Evo Scan tuning out the knock and with the assistance of a wideband tuning the AFR.

It is my goal to try and up the boost to 12-14psi on the factory solenoid.

Have many people used map tracer in conjunction with 2-byte load to tune their cars? it makes the process so much easier, I could get a fairly stable tune in under an hour I reckon.

Would it be worthwhile doing a quick write up on map-tracer/2byte and posting up the modded def files to get it to work in evoscan?

Sorry if thats a little off topic but Im all for jumping on the bandwagon to keep the whole tuning scene "open" as a lot of people put the hard yards in working on the def file but are reluctant to share due to a few people who decide to use their hard work to make money.

As I said above, if we had Valmes definition file I think that would be the best place to start, with the 7202 ecu, just my two cents anyway

Nutter_John
14-10-2010, 01:54 AM
IIRC Vlames work is nearly 3 years old .

This map tracer are you doing it with an AFR gauge as well as the calculated load cell , or are you just tweaking and hoping .

When we tuned the 7201 images this was not done on a road it was done on a rolling road as we are in complated control of the load cell , you can also tell if the tweak you are making is good bad or indifferent . We are running at 17psi on the stock boost control and it holds around 12 psi by 6500 rpm .

A simple observation , how many people with 7202 / 7203 ecu's will actually do anything to them , they may neither have the skill nor the time to do anything. I can only think of one person in the UK that has used a 7202 to tune there car by Benh@Eurospec

scott.mohekey
14-10-2010, 01:56 AM
There's already at least three people with 7202/3s here in Christchurch that have tuned them, as well as a good few in Australia. There will be another few here in Christchurch before long too.

Nutter_John
14-10-2010, 01:58 AM
Ok Thanks Scott , so what xml defs did they use ? as surely they are now proven to work and have dyno results to back up there tunes

scott.mohekey
14-10-2010, 02:01 AM
I can't comment on what def files they've been using. And as far as I know, no one here has yet done anything more than road tuning with a wideband. I believe some dyno tuning has occurred in Australia, and I know that we're planning a dyno day here towards the end of the year.

Nutter_John
14-10-2010, 02:16 AM
Ok once again thanks Scott , so not to be condersending to the people who have tuned there cars using the xml defs all they have done is used there bum dyno's to say something is better or worse ? (being a devil's advocate here )

To pick up on a point Taylor made about keeping the whole tunning scene "open" , how does this play with the likes of the 7201 which we decided to do in a closed method so we could recoop the money we invested in gaining the knowledge , same goes if I spent £500 on a dyno getting all the info known for a 7202 should it be open ?

If you want to make the XML open source then you need to fund it's growth or people will turn round and keep the info to them selves , Hell I can fully understand why they would

scott.mohekey
14-10-2010, 02:31 AM
I think what Taylor is aiming to do is get a group of people working together to develop an xml def file that covers most, if not all features of the roms. I know that you've done a lot of work to uncover this information for the 7201, and can understand you wanting to keep it closed, given how much money you spent on dyno time to get where you got to. That's not the only way to uncover the various tables etc available in the roms however. It's entirely possible to disassemble the code etc and take a more analytically driven approach rather than a test driven approach.

What I'm trying to get at here, is that the xml def file itself is what I THINK we're trying to fully define here, and we'd like to do it in as open manner as possible. We're not so much talking about getting working tunes for people to flash straight to their ECUs. Instead, we're looking at making it possible for people with the resources and knowledge to do their own tunes, to do so with a richer set of adjustable tables etc.

Sure, people will take the xml def file and do seat of the pants tuning, but that's totally within their rights. I agree, a proper dyno tune is far more sane.

Edit: Everywhere I've said Taylor, I mean Taylor, scientist, etc, etc..

taylor
14-10-2010, 02:57 AM
IIRC Vlames work is nearly 3 years old .

This map tracer are you doing it with an AFR gauge as well as the calculated load cell , or are you just tweaking and hoping .

When we tuned the 7201 images this was not done on a road it was done on a rolling road as we are in complated control of the load cell , you can also tell if the tweak you are making is good bad or indifferent . We are running at 17psi on the stock boost control and it holds around 12 psi by 6500 rpm .

A simple observation , how many people with 7202 / 7203 ecu's will actually do anything to them , they may neither have the skill nor the time to do anything. I can only think of one person in the UK that has used a 7202 to tune there car by Benh@Eurospec
Its a 2byte load value, it from what I can tell is the the value the ECU uses to calculate load.

we have been tuning by using valmes maps and scaling them back to our gas in NZ, 98. this has been on the road but, we are going to tweak it on a rolling road.

By logging 2byte load and knock and RPM we can compared the logged knock and trace to where in the table it is knocking and pull the timing back a few degrees, so no, not change and prey, as from experience the Map tracing is damn near spot on.

As we are able to edit the maps in maptracer, it means that we are only using ecuflash for writing the rom to the ECU.

Theres no reason why you cant do this the opposite way of course, push the timing until you get knock and back off, but i like to know that Im scaling something I know works, and then can have more of a play on the Dyno.

I think being part of a club, that we all put in a little effort to benefit everyone else, and hope that by following ethics of being in a club, people dont screw us over, after all my own intentions are to write up my ways to get more and more people into tuning yourself, afterall the reason I joined CVR4 was so that i could get information and help to do all the work on my car myself.

I completely understand the need to recoup cost, but i feel there are other ways of doing this, and I hope that one day when you have broke even, and made a bit of money, you share some of the info you have gained, so that other, especially those on the other side of the world can benefit.

taylor
14-10-2010, 03:00 AM
Here is a screen cap of maptracer. The log I have used is only a sample, there generally is more data available.

38235

scientist
14-10-2010, 03:48 PM
You can have the approach ceddy n the evo guys do....Donationware

Eurospec
14-10-2010, 05:46 PM
Before i start, i appologise i haven't read the whole thread- so if i am saying something that has been covered, then my appologies.

None of the stuff i am going to say is meant as a criticism, just my personal observations.

Firstly the current XML defs, the 'route maps' of the ECU if you want, i think the first thing that is needed is for these to be stableised. There are loads floating in the ether. Some seem good, some are not as good. Without stable defs you might think you know what you are editing, but you will actually be changing something else. If ecu flash is to have a future then there needs to be some form of version control, or way of publishing a known good def for people to use. Maybe it could be hosted here on the forum, but there is a whole issue around checking its validity prior to being declared known good. I can help in that process but since i have no idea how to write a Rom Def, all i could do is validate.

Secondly, the main dissadvantage of ecu flash is that it is exactly that, a flash process. This limits the tuning pottential that it has. If you could live map it, by holding each individual load cell and adjusting the maps live, then that would give a true tuning ability to the mapper. For example in its current guise, if you tried an injector swap, it would be extremely time consuming to tune anything other than wastegate and above. As a result the likelyhood is that everything else would be sub optimal.

Finally, when we get to the heady heights of having to rescale the afm, there may be an issue there. In the evo community i note that they are now using the secondary maps sensors fitted to the later evos to dimension the load maps. If this can be done it will open up the opportunities beyond whatever is the ceiling of the stock maf.

Hope that helps,

Cheers,

Ben.

scientist
14-10-2010, 07:16 PM
My main goal with this thing is to get things ordered to a point where it would be easier to start building Custom features into the roms.

At the moment things are all over the place without definite control. Even the current rom naming convention is off.

eg some people use the EM2005 etc as a rom name...but so far i've come across 3 rom ids that still have EM2005.

I'm also not good at building def files but i plan to dedicate a lot of my time into learning and getting things done....including trying to understand how to build new subroutines

AderC
14-10-2010, 07:29 PM
There's already at least three people with 7202/3s here in Christchurch that have tuned them, as well as a good few in Australia. There will be another few here in Christchurch before long too.

and I've fiddled with my 7202 too...

bradc
14-10-2010, 07:47 PM
There are 15 or so people in Aus that have done tuning on the 7202 using a guy called Merlin who is experienced in tuning evo's

scott.mohekey
14-10-2010, 08:35 PM
Before i start, i appologise i haven't read the whole thread- so if i am saying something that has been covered, then my appologies.

Ben, you're pretty much on the ball with everything you've said, except for the additional aim of modifying the code itself.

There are essentially two goals to all of this. Stabilize and rationalize the xml def files, and enhance the ecu code through modification.

Version control is essential in both of these, and to that end I've set up https://bitbucket.org/smohekey/vr4-ecu-hack. The name of the repository can be changed from vr4-ecu-hack quite happily. If you want to work on this, please sign up on bitbucket and then either pm me your username or leave it in a message in this thread.

Next step is to discuss how we want to organise all of this.

Patryn999
14-10-2010, 09:37 PM
Thats me signed up. Same username as on here.

I've started looking at the assembly, just to get a feel of things. It looks like the memory general layout matches the evo 5/6, so I've gotten my hands on Ceddy's latest Evo 5/6 mod package, I'll see if I can find some of the code dealing with fueling (using locations given in his .xml) and see if the code pattern is similar to something in our roms. Its a start, which might give a general indication as to whether or not we can take this shortcut, or have to do all the grunt-work ourselves.

The bitbucket site also includes a wiki, so we may be able to leverage that for discussion/planning/notes/etc?

scott.mohekey
14-10-2010, 10:08 PM
I have a bunch of disassembly done already somewhere. I THINK it survived my last laptop reinstall. I'll have a look tonight.

pbaron
14-10-2010, 11:30 PM
Hey Scott I have just signed up with the same username as on here, cheers Pierre

scott.mohekey
14-10-2010, 11:41 PM
I've added both Carl and Pierre. We could start messing with the wiki on there, but I think its probably better if we start laying out a plan in this thread.

MackTheKnife
15-10-2010, 12:18 AM
Hey Scott, signed up too with same name. Happy to contribute to costs if BitBucket bandwidth is pushed.

Am only starting out (did first logs on way to work this morning), but hope to contribute to this.

scott.mohekey
15-10-2010, 01:05 AM
Added MackTheKnife.

taylor
15-10-2010, 01:19 AM
Signed up Scott. Can u add me in with write access and il put my work till date in.

Are you still limited by users?

scott.mohekey
15-10-2010, 01:21 AM
It's a public repository, so no, I don't think there's any limits to the number of users. What name did you use to sign up with? 'taylor' isn't found.

scott.mohekey
15-10-2010, 01:23 AM
Nevermind, found you.

scientist
15-10-2010, 01:49 AM
signed up

scott.mohekey
15-10-2010, 01:52 AM
Added scientist.

merlin
15-10-2010, 03:28 AM
well I think I have signed-up, as merlin_tuning

found the original.idb file, is that an IdaPro dissassembly file?
Winblows diddnt know what to do with it.

scott.mohekey
15-10-2010, 03:32 AM
Yeah, that's an IdaPro disassembly library. It's got what I've worked out so far.

Shtiv
15-10-2010, 03:40 AM
Forgive me if I have read this wrong as I only read the first page in a hurry via a link from ozvr4 but I thought you were talking about making a standard modified ROM which to me has all sorts of issues but you were talking about standardised xml or definition files. I'm all for it, when I have more than 5 mins of time I'll be reading through it all and signing up to the bitbucket site.

Also would be good for people to post the addresses they have tested and found to be incorrect, saves the rest of us doing them again. This has some serious xml acceleration potential.

BTW the EM2428 ECU Patryn999 refers to is the same except all the scale addresses are moved by a certain number of points (512 from memory) so it's pretty easy to redefine if anyone has the need....

merlin
15-10-2010, 03:41 AM
Scot, I dont have IdaPro, do I need it to open and read the file or will some other app work?

scott.mohekey
15-10-2010, 03:50 AM
It will only work for IdaPro as far as I know. I'm going to look at exporting it into a more useful, and open format.

scott.mohekey
15-10-2010, 03:51 AM
Added merlin_tuning.

scientist
15-10-2010, 04:26 AM
So i'm guessing we're moving forward with the rom id: 20030011?

scott.mohekey
15-10-2010, 04:29 AM
I don't know? Is there a newer one? That's the first I found on my desktop!

Patryn999
15-10-2010, 04:40 AM
I remember some of the tables moving by more/less than others. It was a long time ago and I was sort of rushing it, so someone might want to do a diff 2003 vs newer with a decent diff tool, and see if its just movement or theres some additional code there.

If no one's done it by the time I get home I'll have a quick nosey, just got to find where I put my unmolested rom...

Patryn999
15-10-2010, 06:29 AM
The code for the EM2381 does appear to be a little different. Some fuctions have changed, but it could just be a compiler optimistaion thing (large one I inspected had removed a jump and added in code block directly).

There doesn't appear to be any new large blocks, so it was either minor tweeks here and there, or (more probably) a few funcitons got inlined by the compiler. Chances are the compiler they used for the 2381 was newer than the 200x version.

scott.mohekey
15-10-2010, 06:38 AM
So I guess we need to compare the 2381 and the 200x with the evo ones, and then pick the one that is closer to them.

Patryn999
15-10-2010, 08:02 AM
I added some disassembled roms (evo 5 GSR, 2004 & 2381) so that people can use a diff tool to spot the difference. They are different enough that I'd require a better tool than the one I'm currently running to choose :P

Shtiv
15-10-2010, 11:22 PM
Hi Scott,

Signed up, username = Transformance

Can you save the disassembly as an older ida version (I only have version 4.9)?

Oh and my car has been tuned a few times on the road and a dyno with a flashed 7202....

http://ozvr4.com/forums/showthread.php?9868-Slow-and-Steady-Peformance-Mods-and-Results

scott.mohekey
16-10-2010, 12:35 AM
Added Transformance.

I'll look into adding my annotations from Ida into the EM2004 asm file that Carl uploaded.

That's another thing, can we all put our real names into our bitbucket profiles. Just makes it a bit easier to communicate I find.

Fully
16-10-2010, 04:25 AM
Ive added myself I think =)

Ive being doing some work on retarding the timing in relation to the boost pressure.

I'll zip it now.

scott.mohekey
16-10-2010, 04:27 AM
Cool Shane, I've added you as a writer.

Shtiv
16-10-2010, 12:09 PM
How do I add the xml I am using to bitbucket? Never used that site before and I have no idea? Do I even have the correct level of access?

taylor
16-10-2010, 12:34 PM
Scott, would it be possible to set a naming convention for some of the addresses in the def file. If we are to set this, it should allow the new version on evo scan, 2.7, to use ecu flashes definitions, instead of having to write separate ones for evoscan.

In turn this would mean if the person has ecu flash and evoscan, asking as they hav one of our "released definitions" then it will be seamless being able to load our rom into evoscan.

I hope that made sence

scott.mohekey
16-10-2010, 10:39 PM
I thought that evo scan used a completely different xml format? There's nothing to stop us using the same names for tables etc as what evo scan uses though.

Shtiv
19-10-2010, 12:29 AM
The 7203 and most of the 7202 Chipsets have the EM200x code, the 2428 code seemed to be only on the very early 7202 ecu's (late 98 from what I have seen) and I have never seen the 2381 code yet. Point is I reckon the 2004 code is the latest and greatest....

scott.mohekey
19-10-2010, 01:30 AM
Agreed. Has anyone confirmed however whether a 2004 will work on an ecu that had a 2428?

Shtiv
19-10-2010, 01:49 AM
I tuned an EM2428 on the weekend but I wasn't game to drop a 2004 ROM on, I didn't have a spare ECU and it wasn't my car so sorry, lost opportunity....

scott.mohekey
19-10-2010, 01:53 AM
How often do cars with EM2482 ecus come up?

Shtiv
19-10-2010, 02:18 AM
I have seen 3 '98 facelifts and every one of them has been an EM2428 ECU....

scott.mohekey
19-10-2010, 03:59 AM
If anyone has an EM2428 and doesn't mind trying to flash an unmodified EM2004 rom on to it, that would be very helpful in determining whether we can concentrate on a single rom or not.

Kenneth
19-10-2010, 04:51 AM
Yes, the EM2004 ROM goes onto an ECU with EM2428 from factory. When I started the thread about burning my 2002 ECU code onto the 1999 Auto ECU, these were the ROM codes.

scott.mohekey
19-10-2010, 04:54 AM
Cool, so is the EM2004 the most recent of all the roms we have access to? If so, we should move forward with that one.

pbaron
19-10-2010, 06:33 AM
I have been doing some searching looking for later roms than the EM2004/5 ones I have from my 7202's. I have a copy of an EM2428 with internal ID 23810003 and an EM2381 with internal ID 23810002, these have been uploaded. Unsure though on whether the 2381 and 2428 roms have been molestered.

Feel free to move them around or remove as necessary.

If someone has a stock 7203 that they can download the rom from and upload here then this is probably the way we should be going.

Preferrably we will have examples of manual and auto 7203 roms to compare.

scott.mohekey
19-10-2010, 07:35 AM
I've got some roms on my laptop that I downloaded from somewhere. They're named as follows:

* MD340289(EM0004).bin
* MD340289(EM0005).bin
* MD340289(EM2005).bin
* MD340289(EM2005)_tuned.bin
* MD359138(EM2428).bin

I haven't opened them to see what they are. The last three are pretty self explanatory, but the first two.. I'm not sure.

pbaron
19-10-2010, 07:46 AM
The EM0004/5 ones are from 7201, I would have thought the EM0004 was the manual one and should have MD340288 as the part number.

The EM2005 tuned rom, I have a similarly named one that is 160kb in size, too big to fit.

I think we really need someone to step up with pristine late 7203 roms, then we can make a start on defining a complete xml def for them.

scott.mohekey
19-10-2010, 07:46 AM
I've just uploaded the H8/500 hardware manual and programming manual to the bitbucket site. You can find them in the download section.

pbaron
19-10-2010, 07:49 AM
Kenneth, do you have a stock 2002 rom that you can make available to us, cheers.

Yes, the EM2004 ROM goes onto an ECU with EM2428 from factory. When I started the thread about burning my 2002 ECU code onto the 1999 Auto ECU, these were the ROM codes.

CANDEE
19-10-2010, 08:03 AM
Scott can you add me as well? Same username. :)

bradc
19-10-2010, 08:05 AM
errr, you may want to edit that Jeremy, it currently says: "Scott can you ass me as well"

scott.mohekey
19-10-2010, 08:12 AM
Yeah.. I can ADD you if you want.. I'm not sure if my fiancee would be keen on my assing people.

You've been added.

CANDEE
19-10-2010, 08:20 AM
Damn Dell keyboards... lol enough said... lol

Kenneth
19-10-2010, 08:57 AM
lol

Yep, I have the original ROM

2002 Manual ROM (http://www.clubvr4.com/forum/uploaded/1403/1287475035.zip)

pbaron
20-10-2010, 06:59 AM
Just comparing Kenneths 2002 EM2004 rom to my 1997/8 EM2004 rom there are only 9 bytes different, 6 of those are in amongst the rom id info so I think will have little if any impact on the operation, 1 more byte is located just after the speed cut data with the last 2 bytes being at rom addresses 0x0A01 and 0x0A03 these map to xml def addresses of 0x010A01 and 0x010A03.

I guess from this I am a little surprised that there is so little difference between the roms after 4+ years.

The ecu part number appears to be embedded in the rom located in the 4 bytes starting at address 0x0237. That puts Kenneths as MR578135 and mine as MD340288.

Based on this it would seem that early EM2004/7202 code should run just fine on a 7203 ecu.

This has changed my thinking a little on what roms we should be standardising on.

The most well defined xml def files currently available are for EM2004/5 7202 ecus which are basically using most everything from vr4base. If Mitsi are happy using mostly this same code in 2002 ecu's then we should probably pick these as a base.

What say you all?

scott.mohekey
20-10-2010, 07:10 AM
The reason that addresses like 0x0A01 map to 0x010A01 in the xml file is because the rom code copies a large chunk of itself to memory address 0x010000 at start up, creating a shadow copy of all the tables etc. It does this because the h8 cpu uses a paged memory model, with addresses being made up of a page number and an address offset. In order to address a piece of memory in a page, the code must first set the page register to the page number and then reference the memory via the offset. So, when I say that the ecu code puts a copy of the tables etc at address 0x010000, what I'm really saying is that its putting a copy in page 0x01 at offset 0000, which is the same offset that they have in page 0x00. The net result of this is that code can use program counter relative operands to access the memory in the 0x01 page (where most of the run time code is), rather than page relative addresses to the 0x00 page.

End of tangent.

As for which rom to use. It looks like you've identified that there is effectively two differences between the two roms, they could be rom version identifiers in the MUT table at a guess. If that were the case, there isn't really any difference between the roms at all and we can use either.

pbaron
20-10-2010, 07:23 AM
My thinking on using the 7202 roms is that I/we have pristine auto and manual roms to work with, we only have a single 2002 rom at present.

At some point we need to agree on what our base is going to be. Prior to settling on the 7202 code though we will need to have it confirmed that a 7203 can be flashed with the chosen base code and still operate as expected.

If I had a 7203 I would certainly offer to be the guinea pig, if someone will sell me a 7203 for a reasonable sum I would be the guinea pig.

Any offers?

scientist
21-10-2010, 01:34 AM
Like is said...on the evo side of thing they are using the RA 7202 code modified Ceddy on the 7203

Shtiv
21-10-2010, 07:31 AM
Can someone point me as to where to start with bitbucket? I'm lost with it

scott.mohekey
21-10-2010, 07:49 AM
http://confluence.atlassian.com/display/BITBUCKET/Bitbucket+101

Patryn999
21-10-2010, 08:25 AM
If you're a windows user and more familar with a GUI, you can try TortoiseHg (and still read that tutorial, as it gives good background info).

pbaron
29-10-2010, 05:55 AM
We are a week on with no visible action.

What do we need to do make a decision on the base images to work with? EM2004 for manual and EM2005 for auto seems like the best shot, Scientist what say you?

There were a couple of other threads recently, one on here and one on OzVR4 asking about boost control using the stock solenoid, this is probably the next thing on my wish list to identify and get working. I can advise that the current entries in the def file for "Boost Desired Engine Load" had no discernable affect with my EM2004 image, nor did tweaking the WGDC entries. I have put the MBC back on as running with standard boost just does not do it for me any more.

Cheers

scott.mohekey
29-10-2010, 06:14 AM
I've been working on a disassembler for the h8 cpu. Some of the code is up in the bitbucket repository already. My main reason for doing this is IDA isn't free and it's h8/500 support isn't nearly as good as it's support for other cpus.

Patryn999
29-10-2010, 11:28 AM
Yeah sorry, I've also been somewhat lax as I spent last weekend fixing an oil leak in the front rocker cover and replacing the front turbo. I'll have a read of what Scott's done tommorow morning, and see if I can add anything to value to the endeavour by utilising the grey matter tommorow.

scott.mohekey
29-10-2010, 11:46 AM
I've got a bunch more stuff to commit yet. As it stands the disassembler is very much incomplete.

scientist
29-10-2010, 11:58 PM
some of the other H8 guys seems to have done their stuff with Hex using ECU edit

scientist
30-10-2010, 12:00 AM
We are a week on with no visible action.

What do we need to do make a decision on the base images to work with? EM2004 for manual and EM2005 for auto seems like the best shot, Scientist what say you?

There were a couple of other threads recently, one on here and one on OzVR4 asking about boost control using the stock solenoid, this is probably the next thing on my wish list to identify and get working. I can advise that the current entries in the def file for "Boost Desired Engine Load" had no discernable affect with my EM2004 image, nor did tweaking the WGDC entries. I have put the MBC back on as running with standard boost just does not do it for me any more.

Cheers

I'll compare the data on an Auto rom vs the Manual rom. anyone has tried driving an Auto with the Manual rom on it to see if it works? If it does it eliminates another un-necessary break

Shtiv
30-10-2010, 01:47 AM
I've been meaning to test a few things and then I have to understand bitbucket better before I post my stuff up. Pbaron, which addresses did you use for boost control? I have the following I was about to test:

C and D would be the hi octane ones I would have thought....

<table name="Boost Desired Engine Load - A (Untested)" category="Limits" address="12228" type="2D" scaling="Load8">
<table name="RPM" address="2d232" type="Y Axis" elements="17" scaling="RPM"/>
</table>

<table name="Boost Desired Engine Load - B (Untested)" category="Limits" address="12254" type="2D" scaling="Load8">
<table name="RPM" address="2d232" type="Y Axis" elements="17" scaling="RPM"/>
</table>


<table name="Boost Desired Engine Load - C (Untested)" category="Limits" address="1226a" type="2D" scaling="Load8">
<table name="RPM" address="2d232" type="Y Axis" elements="17" scaling="RPM"/>
</table>


<table name="Boost Desired Engine Load - D (Untested)" category="Limits" address="12280" type="2D" scaling="Load8">
<table name="RPM" address="2d232" type="Y Axis" elements="17" scaling="RPM"/>
</table>


<table name="WGDC - A (Untested)" category="Limits" address="121c8" type="2D" scaling="WGDuty">
<table name="RPM" address="2d232" type="Y Axis" elements="17" scaling="RPM"/>
</table>


<table name="WGDC - B (Untested)" category="Limits" address="121de" type="2D" scaling="WGDuty">
<table name="RPM" address="2d232" type="Y Axis" elements="17" scaling="RPM"/>
</table>

<table name="WGDC - C (Untested)" category="Limits" address="121f4" type="2D" scaling="WGDuty">
<table name="RPM" address="2d232" type="Y Axis" elements="17" scaling="RPM"/>
</table>


<table name="WGDC - D (Untested)" category="Limits" address="1220a" type="2D" scaling="WGDuty">
<table name="RPM" address="2d232" type="Y Axis" elements="17" scaling="RPM"/>
</table>

I am trying to find lean spool, boost control and thermofan control, so if anyone already knows those....

pbaron
30-10-2010, 01:57 AM
Hey Steve,

I tried the following:

<table name="Boost Desired Engine Load - A (Untested)" category="Limits" address="12228" type="2D" scaling="Load8">
<table name="RPM" address="2d232" type="Y Axis" elements="17" scaling="RPM"/>
</table>

<table name="Boost Desired Engine Load - B (Untested)" category="Limits" address="12254" type="2D" scaling="Load8">
<table name="RPM" address="2d232" type="Y Axis" elements="17" scaling="RPM"/>
</table>


<table name="Boost Desired Engine Load - C (Untested)" category="Limits" address="1226a" type="2D" scaling="Load8">
<table name="RPM" address="2d232" type="Y Axis" elements="17" scaling="RPM"/>
</table>


<table name="Boost Desired Engine Load - D (Untested)" category="Limits" address="12280" type="2D" scaling="Load8">
<table name="RPM" address="2d232" type="Y Axis" elements="17" scaling="RPM"/>
</table>


<table name="WGDC - A (Untested)" category="Limits" address="121c8" type="2D" scaling="WGDuty">
<table name="RPM" address="2d232" type="Y Axis" elements="17" scaling="RPM"/>
</table>


<table name="WGDC - B (Untested)" category="Limits" address="121de" type="2D" scaling="WGDuty">
<table name="RPM" address="2d232" type="Y Axis" elements="17" scaling="RPM"/>
</table>

Had not seen WGDC C and D addresses before, like I mentioned, none of the above BDEL addresses had any effect on operation with my image, there wasn't anything too technical about my approach, I just pushed the settings up to 220 from 3000rpm to 7000rpm for all of them.
The WGDC entries I changed from 0 to 100 again with no effect.


I've been meaning to test a few things and then I have to understand bitbucket better before I post my stuff up. Pbaron, which addresses did you use for boost control? I have the following I was about to test:

Shtiv
30-10-2010, 11:11 AM
well that's good and bad news.... Bad in that I still have to disconnect my ebc and test them tomorrow, good in that maybe the WGDC C and D will work!

Patryn999
30-10-2010, 11:49 PM
Yeah, when I was looking the wastegate tables at 0121F0 and 012206 looked most promising, but as I broke my stock solinoid I cannot test them :P (tables for EcuFlash will the +4 bytes, as my indexes include header bytes).

Just thought I'd post this for those that haven't figured it out already, tables look like this:

* A byte (or word, for word-sized tables) for the number of dimensions: 2 = 2D, 3 = 3D.
* A byte (or word, for word-sized tables) for a value "added" to all values returned from the table.
* A long word for the position on the X-axis.
* (For 3D) a long word for the position on the Y-axis in a 3D table.
* (For 3D) a byte (or word, for word-sized tables) for the length of each row in a 3D table.
* A series of words or bytes containing the table data.

The provision is there for word-length tables, however I have yet to find any. The helpful part is matching the axis indexes, as this can help trying to figure out WTF new tables do. i.e. the timing tables request F00E and F010, which seem to be RPM and Load respectively. Then one can change the scaling selector for the other axis in EcuFlash and list the ones that come out with numbers that look semi legit (further limits possibilies) and try to go from there.

I'm still only investigating, but I suspect there are more than one RPM/LOAD etc values depending on the index length (i.e. 12, 17 etc) so that each is actually an index into a table rather than a raw RPM value, and that these are all updated when the RPM is actually read by ECU. But this is only a guess as of now, as only just starting to try and figure it out.

My current suspicions (in ram). Cut leading zero's to save my fingers.
F00E - RPM17
F012 - RPM12?
F010 - Load19
F016 - Temperature8?

scott.mohekey
31-10-2010, 12:01 AM
Carl, I have some stock solenoids floating around if you want one. Also, can you pm me your msn address (if you use msn), or your facebook username.

Shtiv
31-10-2010, 10:03 AM
So.... Disconnect my EBC and 3 port solenoid, reconnected the factory boost controller with factory settings for the above maps. The car ran at around 8 psi above 2500RPM. maxxed out the duty cycle maps and the engine load maps and got the same result, though crap, they don't work. Then I though what does the car run without the solenoid (it's bleed branch back to the intake blocked) and it ran about 7.5psi.

So I thought, hmmm not much difference, why would the factory even bother with the solenoid and this is with the restrictor removed.

So I tidied up the plumbing as I had done it as a temporary thing with long pipe runs and I thought maybe these were causing a restriction through the bleed. Fixed all them up and got....

8psi again. So I think my solenoid is pharked. so I'll log the voltage to it and see if that makes a difference but it can't be today now....

End result, I proved nothing.

Shtiv
05-11-2010, 10:28 AM
Hi All, so.... The boost control maps work, the ones I posted above! :)

We got around 15psi (maybe more available yet, don't know) with the restrictor removed.

So here's the story, I tried again in my car swapping another solenoid into it and again got 8psi ish.... grrrrr....

So swapped them back thinking, there must be something else wrong with my car. Hang on a sec, right next to me is another Legnum with a 7202 ECU, mind if I drop a new map in it and we go for a drive round the block? Nope. So off we go and I say, don't go too hard as I've maxxed out the duty cycle and desired engine loads and bamm, 15psi on the gauge (I had a MAP logger going but I haven't downloaded them yet)! This was with the restrictor removed, but at least 15psi is up for grabs which is probably nough for a lot of people! (No not enough for me but I have a 3 port to plumb in again so I don't really care).

It was a rush so really little testing, only enough to know it worked and I changed all of the 4 tabled for WGDC and BDEL

Hotwire
09-11-2010, 01:10 AM
I'll compare the data on an Auto rom vs the Manual rom. anyone has tried driving an Auto with the Manual rom on it to see if it works? If it does it eliminates another un-necessary break

I have a manual EM2004 ECU in my '98 facelift auto. Zero problems thus far here.

Shtiv
09-11-2010, 07:42 AM
One of the guys on OZVR4 had a manual ROM on his auto ECU in a manual PFL (was 7201 ECU) car. It ran fine but he is suspect of it now. All very anecdotal though.

Patryn999
09-11-2010, 09:54 AM
Stuffed around with adam's car this evening, got sick of him bitching about only making wastegate pressure :P Talk about a bit of a think-storm. First off none of the tables made any affect, however it was a bit sus as he was only making 6-7psi (wastegate) all the time. Had multimeter on solinoid - its fine. Actuates with voltage and can blow air through it so wasn't that. Attached LED to it, and got flashing so solenoid was working.

However he had replaced stock pipe-work, and I rembmered there is a little restrictor in the red-tipped high pressure pipe that comes just before throttle body - luckily he had spare engine, so pulled one off that and sure enough, it exists down by the tee. Turns out that without this the air pressure saturates the solonoid, and you only make wastegate pressure (makes sense as solinoid apatures are not big).

Tried editing maps (is EM2005 on 7203) and WGDC D at 1220A definately changes boost. However its a bit erratic, so I'm wondering if the ECU goes something like open/closed loop - i.e. uses the WGDC on sharp accel, and then reverts to closed loop with boost desired engine load? Just a thought, didn't really have enough time and gas to test it so maybe we will this weekend.

EDIT: WGDC would bring up initial and peak boost till about 5000-5500, but the longer you were on boost the better it seemed to track to 10psi. Which is why I hypothesise that it has closed loop feature? Could just be comepletely wrong thou ;)

Hotwire
09-11-2010, 01:47 PM
Can you elaborate a little more on this restrictor? Any photo's of said restrictor?
I reverted my car back to factory solenoid tonight to test this out, but was stuck with only wastegate pressure :S

Patryn999
09-11-2010, 09:49 PM
I can explain it a bit better, but as I do not have stock piping anymore someone else (i.e. adam) will have to take a pic. This is all from memory when I cut up my pipework after doing intercooler & EBC (what? I'm curious!).

Anyway on the stock pipe the high pressure pipe connects from elbow by throttle body, to a "T" connector just by solinoid (Has red ring on elbow-side). About an inch back from the end opposite the red end is a restrictor inside the pipe, the ID appears to be quite small, as I'd say about 0.5-1mm dia, and in actuality is a small bit of brass that has been molded into a plastic form that fits within the pipe.

You can feel where the restrictor is in the pipe by just squeezing it, as it (being plastic) is rock solid.

So I think you won't really get much more in the way of detail unless someone was willing to cut theirs to pieces :D. On the otherhand it should be quite easy to duplicate, as "T" intersections are about $2 from any hardware/garden store, and you can make your own restrictor on one side with some epoxy and a bit of wire/toothpic/smallcylindricalobject.

Hotwire
09-11-2010, 10:35 PM
Thanks for that description.
Based on what you've said, I have attached a pic of the factory boost control with the line coloured red and highlighted where I believe you are saying the restrictor is, can you confirm this is correct?

Thanks
Lee
38961

Patryn999
10-11-2010, 01:42 AM
Correct. Nice search-fu BTW, went for a quick search myself for pic to demonstrate but had to get back to work before I could find one ;)

Hotwire
10-11-2010, 04:42 AM
excellent - will check out the hose tonight.
As a side, how the hell do you get a multimeter down to the solenoid with all the piping there?! got me farked trying to test it with the car running.....

Patryn999
10-11-2010, 10:26 AM
Lol, bit of a three stage process. First measured with multimeter - resistance about 35ohms, which seemed to be in the realm of legit for a solenoid. Then blew on it as I energised coil (aka put 12V across it) which showed that it opened. Insulation had broken off wires just before they entered plug, so lightly soldered a couple of wires on and hooked up an LED+560ohm resistor (Once side of solenoid wiring is permenantly grounded). LED gets pinned under window wiper and you go for a blat - flashes if solenoid is opening.

Hotwire
10-11-2010, 02:57 PM
Well I can confirm that my car is missing the red-tipped hose, BUT an ozvr4 member around the corner from me is giving me the factory hose off his car tomorrow avo so fingers crossed can get it operating properly.
However tonight I did wire up the output of my Blitz Dual SBC to a spare input of the ECU so will be able to log MAP in correlation to RPM/Load so should be able to get a good understanding of how the boost control is ;)

But this is all a digression from the original thread... feel free to move these posts to another/new thread moderators ;)

Shtiv
11-11-2010, 06:14 AM
Awesome, I'll add some form of restriction to mine, sort that out and then I can read up on bitbucket

Hotwire
11-11-2010, 01:21 PM
Boost control discussion continued here:
http://www.clubvr4.com/forum/showthread.php?53046-Boost-Control-via-Factory-ECU

MackTheKnife
16-11-2010, 08:17 AM
Finally got around to flashing my 7203. Was halted by a busted clutch.

Will do some testing on my commute and see what I can come up with.

Shtiv
18-11-2010, 11:10 AM
Sorry I have done bugger all in the last few days.... Haven't even read that bitbucket stuff....

MackTheKnife
22-11-2010, 06:31 AM
Alright, tracked down Max Retard limit thanks to my own vanity and missing the pops and bangs all my other performance cars had. The Scaling is a little bit different thanks to EcuFlash being an order of operations Nazi:

<scaling name="RetardHex8" units="Max Retard Degrees - hex" toexpr="x" frexpr="x" format="%02X" min="0" max="255" inc="1" storagetype="uint8" endian="big"/>
<scaling name="Retarduint16" units="MaxRetardDegrees" toexpr="-((x-203)/3+10)" frexpr="(-x-10)*3+203" format="%.1f" min="-200" max="255" inc="0.3" storagetype="uint16" endian="big"/>


<table name="Ignition Max Retard Degrees - decimal (to be tested)" category="Ignition Parameters" address="1051c" type="1D" level="1" scaling="Retarduint16"/>

<table name="Ignition Max Retard Degrees - hex (to be tested)" category="Ignition Parameters" address="1051d" type="1D" level="1" scaling="RetardHex8"/>

I'm going to test tonight with a -15* retard on low loads over 4000 RPM and will log and see if I get any sexy flame action out the pipe. If I log more than the -10* stock settling, then we're looking good.

Shtiv
22-11-2010, 07:31 AM
I don't let ecu flash edit my xml or it makes it all look like crap.... but apart from the change in the order of operations I have the same scales and get -10.0 as the stock setting in decimal

MackTheKnife
22-11-2010, 10:02 AM
Cool, I had a slightly different frexpression which played silly buggers. Glad to hear you have the same sums. Tested on the way home and got -10*. Lots of knock from the rich patch so put mix back to stock and tried again. -10* again. Going to give it afew days to see if there is any learning function by the ecu as it did seem to have to figure out how to idle after the battery was out for the clutch job.

At the very least I'm learning a bit about mechanics with this car.

Shtiv
22-11-2010, 11:18 AM
I tried it and no go, so who knows what you are actually changing, change it back until we know what we are changing.

MackTheKnife
22-11-2010, 11:53 PM
Yeah, at least it's tested. I can rule it out.

As far as pops and bangs on gear changes and down shifts are concerned, 10* is enough for me.

Kenneth
24-02-2011, 08:38 AM
Not sure if it is just my definition is wrong, thought I would share just in case it was a well spread isssue.

One of the Periphery addresses is incorrect
in my definition they are 10272,10282,10292

The ECU disassembly says they are 10262,10272,10282

The 10272 appears to be the correct table/address in EcuFlash because bit 7 matches the high octane disable flag. Not sure about the others yet.

Shtiv
26-02-2011, 02:20 PM
Kenneth,

All the defs I have seen have the perpheries at 10272,10282,10292, I'd guess it's a widespread problem....

I have the following and I do know that the disable lean spool function does not work with these periphery addresses.... Are you saying my 10272 should be 10262 etce, notice that the periphery00 is referenced as 10262 after the tables, but then I don't have a table for that....

<!-- *** ECU PERIPHERY *** -->


<table name="ECU Periphery0 (FAA) Bits" category="ECU Periphery" address="10272" type="2D" level="2" scaling="blobbits">
<table name="Periphery Bit Number" type="Static Y Axis" flipy="true" elements="16">
<data>bit.15 .</data>
<data>bit.14 .</data>
<data>bit.13 .</data>
<data>bit.12 .</data>
<data>bit.11, 0=Disable EGR function .</data>
<data>bit.10, 0=Disable WEIRD UNKNOWN CRUISE IGN RETARD .</data>
<data>bit.09, 0=Disable WARMUP IGNITION RETARD function .</data>
<data>bit.08 .</data>
<data>bit.07, 1=Enable High Oct Ign Map Lookup .</data>
<data>bit.06, 1=Enable WG solenoid? Enable Speed Limit Test.</data>
<data>bit.05 .</data>
<data>bit.04, 0=Disable CLOSED LOOP function .</data>
<data>bit.03, 0=Disable IMMOBILIZER function .</data>
<data>bit.02 .</data>
<data>bit.01, 0=Disable REAR O2 SENSOR function .</data>
<data>bit.00 .</data>
</table>
</table>

<table name="ECU Periphery1 (FBA) Bits" category="ECU Periphery" address="10282" type="2D" level="2" scaling="blobbits">
<table name="Periphery Bit Number" type="Static Y Axis" flipy="true" elements="16">
<data>bit.15 .</data>
<data>bit.14 .</data>
<data>bit.13 .</data>
<data>bit.12 .</data>
<data>bit.11 .</data>
<data>bit.10 .</data>
<data>bit.09 .</data>
<data>bit.08 .</data>
<data>bit.07 .</data>
<data>bit.06 .</data>
<data>bit.05 .</data>
<data>bit.04 .</data>
<data>bit.03 .</data>
<data>bit.02 .</data>
<data>bit.01 .</data>
<data>bit.00, 0=Disable LEAN SPOOL (Believe Incorrect) .</data>
</table>
</table>

<table name="ECU Periphery2 (FCA) Bits" category="ECU Periphery" address="10292" type="2D" level="2" scaling="blobbits">
<table name="Periphery Bit Number" type="Static Y Axis" flipy="true" elements="16">
<data>bit.15, P0443 P0446 P0403 P0243 P0090 P2263 .</data>
<data>bit.14, P0450 P0451 P0452 P0453 Evap Pres. Sensor .</data>
<data>bit.13, P0441 Evap Incorrect Purge Flow .</data>
<data>bit.12 .</data>
<data>bit.11, P0031 P0032 P0037 P0038 O2 HEATER CKT .</data>
<data>bit.10, P1400 MAP Sensor .</data>
<data>bit.09, P0401 EGR Test .</data>
<data>bit.08, P0506 P0507 Idle Test .</data>
<data>bit.07, P0170 P0171 P0172 .</data>
<data>bit.06, P0134 Front O2 Sensor No Activity .</data>
<data>bit.05, P0300 set bits 4 and 5 to 0 .</data>
<data>bit.04, P0300 set bits 4 and 5 to 0 .</data>
<data>bit.03, P0132 P0136 O2 Sensor Circuit .</data>
<data>bit.02, P0133 P0159 .</data>
<data>bit.01, 0=Disable P0421 WARMUP CAT EFFICIENCY LOW .</data>
<data>bit.00 .</data>
</table>
</table>

<table name="ECU Periphery00 Hex" category="ECU Periphery" address="10262" type="1D" scaling="Hex16"/>

<table name="ECU Periphery0 Hex" category="ECU Periphery" address="10272" type="1D" scaling="Hex16"/>

<table name="ECU Periphery1 Hex" category="ECU Periphery" address="10282" type="1D" scaling="Hex16"/>

<table name="ECU Periphery2 Hex" category="ECU Periphery" address="10292" type="1D" scaling="Hex16"/>

<table name="ECU Periphery3 Hex" category="ECU Periphery" address="102a2" type="1D" scaling="Hex16"/>




On another note, it's time for me to get serious about disassembly, what software are you using? A licence for IDA pro seems damned expensive, anyone got any other options?

wintertidenz
27-02-2011, 01:50 AM
Sent you a PM.

Will be getting onto trying to crack into the 7201 soon, the 7202's and 7203's are getting more expensive and hard to find. Once I pull the code I'll share it on here and we can start building definitions for this one too.

mattnz
27-02-2011, 02:11 AM
Might be clearer to use
tags around the code :)

Kenneth
27-02-2011, 06:52 AM
Not 100% sure about the Periphery sorry, just that it does look like the definitions are incorrect. I'll have to verify it against a known map such as an Evo setup.

I have some bad news about the Lean Spool. I am not fully through decoding it, but it does look to be disabled. From what I have looked at today, there are stubs for it, but no data. I can get the scales and load the maps from the ECU pointers (i.e. same as the evo, but slightly different addresses in the image) but the data is bollox.

Shtiv
01-03-2011, 01:19 PM
Are you sure about lean spool because if I run a quick ramp time on a dyno (hub dyno) I see AFR's get leaner by up to about 1-1.5 points vs when i run it at a static speed, and it goes a bit flaky around 3000RPM which I assume to be the start of lean spool....

Davezj
01-03-2011, 02:34 PM
random i know,
but has anyone thought of doing a low emissions MOT map to gaurentee you pass the test.
Just an option.

kinkyafro
01-03-2011, 09:12 PM
random i know,
but has anyone thought of doing a low emissions MOT map to gaurentee you pass the test.
Just an option.

Thought about it and decided that it couldn't be done without using a wideband to provide a false signal to the existing narrowband sensor as the car will be in closed loop during slow idle/fast idle tests. Even if somone figured out getting the car out of closed loop changing the fuel from stoich would if anything increase the chance of a failiure. So basically it seems a non-starter.

Davezj
01-03-2011, 09:31 PM
i see you have given it some thought.

is it true that the standard ECU can add and remove up to 18% of the available fuel at any time based on lots of peramiters.

Kenneth
01-03-2011, 09:57 PM
I'll double check. I assume you have turned on lean spool in the Periphery section? Can't remember if it is on or off by default.

I'll also double check my map addressing, if it is out then the data would look rubbish too.

Kenneth
01-03-2011, 09:58 PM
As far as I am aware, you can modify the open loop thresholds or just turn of closed loop altogether.


Thought about it and decided that it couldn't be done without using a wideband to provide a false signal to the existing narrowband sensor as the car will be in closed loop during slow idle/fast idle tests. Even if somone figured out getting the car out of closed loop changing the fuel from stoich would if anything increase the chance of a failiure. So basically it seems a non-starter.

Shtiv
02-03-2011, 04:43 AM
Dave - My car is running completely open loop at the moment and yes you can change the extent to where the closed loop runs, so you could make it run stoich longer simply by making it run closed loop for more of the map, we have a car over here about to undergo an emissions test (just got defected) and it's runing hard pipes and a big intercooler etc so we were thinking to do exactly that to ensure it gets through.

Kenneth - re: the open loop thresholds, yes they can be changed
re: the lean spool, by the periiphery addresses I have the default is for lean spool to be turned off but are my periphery addresses even correct?

Kenneth
02-03-2011, 07:47 AM
I believe the Periphery definition is correct for lean spool.

Unfortunately, here is the lookup data for the Lean Spool trailing time (rich side) and Lean Spool Load



seg001:11744 .byte 2
seg001:11745 .byte 0
seg001:11746 .byte 0xF0 ! *
seg001:11747 .byte 0xE
seg001:11748 .byte 2
seg001:11749 .byte 0
seg001:1174A .byte 0xF0 ! *
seg001:1174B .byte 0xE
seg001:1174C .byte 2
seg001:1174D .byte 0
seg001:1174E .byte 0xF0 ! *
seg001:1174F .byte 0xE


Lean Spool Trailing Time is at 11744 and Lean Spool Load is at 1174C. As you can see, these are 2d table headers with no data.

The equivalent lean spool trailing time (rich) from the Evo is



seg001:11DA0 .byte 2
seg001:11DA1 .byte 0
seg001:11DA2 .byte 0xF0 ! *
seg001:11DA3 .byte 0xCE ! +
seg001:11DA4 .byte 9
seg001:11DA5 .byte 9
seg001:11DA6 .byte 9
seg001:11DA7 .byte 9
seg001:11DA8 .byte 4
seg001:11DA9 .byte 3
seg001:11DAA .byte 3
seg001:11DAB .byte 2
seg001:11DAC .byte 2
seg001:11DAD .byte 1
seg001:11DAE .byte 1
seg001:11DAF .byte 1
seg001:11DB0 .byte 1
seg001:11DB1 .byte 1

Shtiv
02-03-2011, 12:50 PM
Well that's weird, I definitely ran leaner on a quick ramp power run than when it was at constant load, I think this needs a bit of investigation, it makes sense why nothing I did made any difference though!

Davezj
02-03-2011, 06:00 PM
Is merlin's ecu flash guide still available on here i have searched for posts by merlin and there are none. I think i have a copy of it some where just wondering everybody is still using it.
Are there any errors in it that need updating.

Shtiv
02-03-2011, 08:52 PM
It's on ozvr4, nt sure if it's on here, but yes there are a few errors to be updated in terms of addresses

Davezj
02-03-2011, 09:28 PM
It's on ozvr4, nt sure if it's on here, but yes there are a few errors to be updated in terms of addresses

Ok so has anybody fixed the errors and reposted the guide back on to the site?

phosty
02-03-2011, 11:20 PM
According to contributors to the evolutionm.net forum lean spool is described as:


The lean spool was designed to lean the turbo spool area not to make more HP from a faster combustion, but to deliver more potential turbine driving energy increasing boost build during load increase only (like 100% up to 280%), not after peak where RPMs wind up to redline

The only thing I see when I go WOT is the AFRs dip to 10-11. The only leaning out behaviour I have seen is when fast cruising (80+mph) with as little throttle as possible to maintain that speed. Then I consistently see AFRs from 15 - 15.5 which really helps the mpg on the motorway - I naively thought that was what 'lean spool' meant until I read the evo forums in more detail. Oddly cruising at 70mph it just sits at stoich (or wobbles from 14.5 - 14.9 to keep the cat happy).

If the lean spool code points to empty tables then that would seem to match what I get from on-road testing. Obviously your dyno would allow to explore other regions in a more controlled manner but I never see leaning out on WOT - the needle goes low and stays low! The fact I get cruise AFRs greater than the highest given in any of the octane maps suggest to me there is some 'lean cruise' code - maybe limited to a max throttle position or Load?

Has anybody else experienced the same effect?

btw I have tried frigging the closed loop setpoint using a WB02 to pass the MOT and while you can get it to move the operating point a little after a certain point the ecu just ignores further tweaks. I guess that there is a max adjustment the closed loop code is permitted to make. I was sooooo close to getting it to pass with a decat....

Shtiv
03-03-2011, 12:31 AM
I'd agree, on the road I haven't seen it do this, only on the dyno, wonder what causes that, it is a different (better) wideband so I'll have to look into it.... I wouldn't be surprised if an innovate wideband ca't pick up the lean spool whereas the dyno uses a motec plm so.....

Anyway, I have my closed loop section turned off above 1000RPM to run leaner in what is the normally closed loop section. In that area the maps are all over the place and the result is much leaner than the target, worst at really low loads (like below 40%). My guess is at 70mph you'd be in the closed loop region and at 80mph you are just into the open loop section before the table starts getting richer.... So if the table is after 14.7 you will probably get 15.3-15.5 and even leaner below 40%

phosty
03-03-2011, 01:19 AM
....I have my closed loop section turned off above 1000RPM to run leaner in what is the normally closed loop section.

To be clear, to achieve this you have set the values in the Open Loop Load tables 114c6 & 114d6 (7202 rom) to read 0% Load above 1000rpm?

Shtiv
03-03-2011, 02:59 AM
Yep re: the adresses, exactly that but I am only aiming for low 15's AFRs I saw 16s and low 17s wile playing but I don't know how it goes the other way....

and Dave, no one has re posted the guide to my knowledge, it's only in PDF format and to touch someone elses work would be a bit rude, you could ask Dave (Merlin) but he's probably been too busy.... I can send you our current xml files if you like too, pm me if you want them.

Kenneth
03-03-2011, 05:10 AM
The fact is that there isn't lean spool. At least not how the Evo does it, there COULD be some other adjustment somewhere which does something.

Also consider this, if you are setting up the main AFR tables to give you a specific AFR at WOT, then are you compensating for enrichment and whatnot in the main (high octane) tables, therefore causing a leaner than expected condition when enrichment is no longer being applied?

At this stage we don't know enough about how and in what order things are being done to make a good guess on why it is happening. We will get there though.

Shtiv
03-03-2011, 08:57 AM
That we will.... :)

Anderz
03-03-2011, 09:21 AM
According to contributors to the evolutionm.net forum lean spool is described as:



The only thing I see when I go WOT is the AFRs dip to 10-11. The only leaning out behaviour I have seen is when fast cruising (80+mph) with as little throttle as possible to maintain that speed. Then I consistently see AFRs from 15 - 15.5 which really helps the mpg on the motorway - I naively thought that was what 'lean spool' meant until I read the evo forums in more detail. Oddly cruising at 70mph it just sits at stoich (or wobbles from 14.5 - 14.9 to keep the cat happy).

If the lean spool code points to empty tables then that would seem to match what I get from on-road testing. Obviously your dyno would allow to explore other regions in a more controlled manner but I never see leaning out on WOT - the needle goes low and stays low! The fact I get cruise AFRs greater than the highest given in any of the octane maps suggest to me there is some 'lean cruise' code - maybe limited to a max throttle position or Load?

Has anybody else experienced the same effect?


I experience the exact same thing, have not yet checked at which loads it occures, but if driving @3000 rpm in 5:th with boost gauge reading -0.1-0 bar I get AFR:s around 15:0-15.5. This has to be the point when it changes from closed loop to open loop. If you change throttle position slightly AFR:s will return to reading from maps or closed loop

This can also occur when mantaining cruise speed on motorway ~120 km/h, but slight change in load or throttle will result in higher AFR:s
At first I suspected the fuel pump control, that full voltage to the fuel pump was applied too late, but as others get the same it is probably something in the ECU
BTW does anyone know when full voltage is sent to the pump, is it just reduced on idle?

Shtiv
03-03-2011, 11:34 AM
atmospheric pressure, (0 boost) is usually about 80% load. For a stock map, 80% load is where the closed loop stops at 3000RPM, at which point the factory target ratio is 14.1 so I wouldn't be surprised to see around 15:1 actual AFR.

No idea on when the fuel pump gets different voltages.... Easy to test though

lateshow
03-03-2011, 04:46 PM
I too exprience the same. When driving in 4th or 5th and accelerating on little boost or in zero boost i get 15.X with my PLX. One another question, my PLX says that im doing 14.2-14.6 in closed loop most of the time, on rare occasions it goes over 15. I'm running too rich and propably spending too much fuel?
Shtiv:

What value did you modify to "remove" the closed loop. I can only modify LOAD when it goes off the closed loop to open loop? Is fuel consumption getting any smaller. With my auto I can't get under 11 litres per 100km in the wintertime driving mostly on motorways 100km/h.

Shtiv
04-03-2011, 11:54 AM
Just set the load for when open loop comes on to 0% in each rev range if you want to remove it. A word of warning though, with my map set at 14.7 I was getting AFR's of leaner than 16:1 so be careful how far you go.

As for fuel econonmy, it's a little better (about 2 or 3% so not much) but then mine runs at 14.7-14.9 in closed loop anyway, if it was 14.2-14.6 I guess you'd see more improvement. My best tank was 8.7L/100km but that's driving pretty carefully, rarely go above 10 though. Most of my driving is freeway/highway. Changing my spark plugs more often would save more fuel to be honest.

Davezj
18-03-2011, 03:10 PM
Has anyone played with the "fuel premier enrich" table in the fueling section of the map.
does anyone know how it interacts with other tables.

it seems to only go as low as 6% at a temp of 86 degree or there abouts, i don't exactly remember.

lateshow
29-04-2011, 09:24 AM
Just managed to get the rom out of a Mine's tuning 7202F ecu. Seems rather stupid ( too little boost, too much advance) but here you go!

bradc
30-04-2011, 12:18 AM
That is interesting. Remember they have 100 octane over there that allows them to run a tune like that.

lateshow
30-04-2011, 08:14 AM
Yes, the car the rom sits in tends to knock a bit. I said to the owner thath he shouldn't go "full boost" too much now. The ecu has a mine's sticker in it. :)

german_VR4
02-05-2011, 11:04 PM
to get a better economy i do it this way:

- modified the map to hold the car up to 120% Load in "Closed Loop".
- installed a wideband with programmable output
- used the programmable output-function to simulate narrowband lambda for ECU
- modified the output-signal from 14,7 to 15,5

now i have the car at 15,5 all time in closed loop. works very fine and as a extra it is regulated by ECU.

information:
the EGT´s are raised up to 850°C at closed-loop at around 150-160km/h on the german motorway.
by this reason the spark-plugs are changed to NGK BKR8EIX to have security at this point


greetings
Tobias


http://www.tarmac-gravel-snow.de/index.php/legnum-vr4

Shtiv
03-05-2011, 04:49 AM
I saw a Japanese tuned ROM on the weekend (Educe I think it was), it was crap for the car it was in, I don't know what is done to the car this is from but assuming not too much, I'd guess it runs rich as hell, knocks like crazy at high loads and doesn't fix the knock itself when it does....

Wodjno
03-05-2011, 09:52 AM
to get a better economy i do it this way:

- modified the map to hold the car up to 120% Load in "Closed Loop".
- installed a wideband with programmable output
- used the programmable output-function to simulate narrowband lambda for ECU
- modified the output-signal from 14,7 to 15,5

now i have the car at 15,5 all time in closed loop. works very fine and as a extra it is regulated by ECU.

information:
the EGT´s are raised up to 850°C at closed-loop at around 150-160km/h on the german motorway.
by this reason the spark-plugs are changed to NGK BKR8EIX to have security at this point


greetings
Tobias


http://www.tarmac-gravel-snow.de/index.php/legnum-vr4

What fuel usage difference do you see running at 15.5 AFR..

Cheers

Wodj

german_VR4
03-05-2011, 07:51 PM
for fuel economy i do it this way:
- force ECU to be in closed loop up to ECU Load 120%
- install wideband lambda with programmable output
- use prog. output to simulate narrow band for ECU
- modify output-values for progammable output (from 14.7 to 15.6)


Car runs now 15.6 all the time in closed loop and that regulated by ECU.
EGT´s raised up around 120°C at 150-160 km/h at the german autobahn.
for security i changed the plugs to NGK BKR8EIX.

greetings
Tobias

http://www.tarmac-gravel-snow.de/index.php/legnum-vr4

Mitsiman
08-05-2011, 06:33 AM
With the help of Acamus and reprogramming myself, I believe we have all the addresses worked out for the boost mapping. We need to do some testing of the boost limit maps as well as work out exactly under what circumstances the different boost maps work but I believe its now all here to work with.

I have included in the attachments the full definition files I am working with. I would recommend that people first backup the ones they have and use all the ones I have here, including the new VR4 base map as well

If people can do some testing and give me some feedback, then we can clean these up some more but i believe at this stage these are probably the most comprehensive definition files available to date.

I have uploaded the definition files to the following address

http://www.evoscan.com/roms/mitsubishi/index.php

The new base file will be put in the right position soon but for the moment you wil have to download it seperatly.

Please let me know any information you can find, if anyone has hte reported 2 byte load addresses then let me know and I will update accordingly

TwinTurbo
13-05-2011, 08:33 AM
Hi Guys

can anyone tell me correct address of Stationary RPM Limit for Galant VR4 Manual, EM2004

Shtiv
15-05-2011, 12:32 PM
As far as I know there isn't one....

TwinTurbo
16-05-2011, 12:00 PM
so you mean that we can not set Stationary Rev Limit on VR4 Manuals ?

http://geekmapped.com/forums/showthread.php?p=23917#post23917

which one of you should I believe ? :D

ok I will check that address today on my car and post results here

Hotwire
31-05-2011, 03:14 AM
Just set the load for when open loop comes on to 0% in each rev range if you want to remove it. A word of warning though, with my map set at 14.7 I was getting AFR's of leaner than 16:1 so be careful how far you go.

As for fuel econonmy, it's a little better (about 2 or 3% so not much) but then mine runs at 14.7-14.9 in closed loop anyway, if it was 14.2-14.6 I guess you'd see more improvement. My best tank was 8.7L/100km but that's driving pretty carefully, rarely go above 10 though. Most of my driving is freeway/highway. Changing my spark plugs more often would save more fuel to be honest.

Hey steve,
I've recently got a Wideband O2 and i've noticed that at cruise it sits consistently around the high 15s, low 16s. Stationary it can head up into the 20s (I presume as not enough exhaust is passing the sensor). Is this normal? under full load i've seen 11s, which corresponds to the AFR Map when I log that.
I was initially concerned that my wideband was innacurate, but now i'm thinking the car just isnt keeping at target 14.7, and also the exhaust outlet measured values would definitiely be more lean than pre-cat values. Does this seem sound logic to you?

Mitsiman
31-05-2011, 03:31 AM
That is normal, to get a truly accurate reading of the O2 sensor you need to have it mounted as far forward as possible, we normally palce it in front of hte cat converter after the Y Pipe join.

Also remember that the AFR Values on the rom mapping are not actual targeted AFR's, they are hypothetical values and the tuning will not correspond exactly with them.

taylor
31-05-2011, 03:37 AM
Hey steve,
I've recently got a Wideband O2 and i've noticed that at cruise it sits consistently around the high 15s, low 16s. Stationary it can head up into the 20s (I presume as not enough exhaust is passing the sensor). Is this normal? under full load i've seen 11s, which corresponds to the AFR Map when I log that.
I was initially concerned that my wideband was innacurate, but now i'm thinking the car just isnt keeping at target 14.7, and also the exhaust outlet measured values would definitiely be more lean than pre-cat values. Does this seem sound logic to you?

What wide band are you using?

You should be getting similar readings at idle as you get at cruise, as it will be running closed loop both times.

I had an issue where the simulated narrowband was inverted to what it should be, and was running far too lean, sometimes 20:1. Then it would get to a point where it would go beyond the closed loop limits and it would drop out to open loop again.

If you hook the 0-5v out to the rear o2 input on the ecu and log it with the formula: 10*x/255 + 10 (assuming the 0-5v is reading between 10-20:1)

You should then see while dipping the accelerator a bit that when the rear o2 reading is below 14:1, the narrow band input should be about .9v, and should drop down to 0v anything above 15.5:1 should drop to 0v. this is assuming your wideband is simulating a 14.7 narrowband output, however the logic is the same for a different ratio, just the switching point will change

Hotwire
31-05-2011, 03:45 AM
Sorry should have clarified that is an Innovate MTX in the innovate exhaust clamp at the rear of the car. I dont want to permenantly mount it, as i want to be able to take it between my Liberty GT and my Legnum.

Therefore when idling, there would be very little flow past the sniffer pipe at the rear of the car, and due to the shape of my exhaust tip (4" tip on 3" exhaust) Im pretty certain there is more around the sniffer pipe than when the exhaust gasses are pushing out.

Thanks for the feedback Dave, the theory seemed correct just wannted confirmation ;)

Mitsiman
31-05-2011, 04:02 AM
I would recomend doing what we do - get an exhaust shop to weld an O2 bung on the exhasut closer up the front of the vehicle, and then you can blank it off when not in use. That way you can do your testing and remove when no longer needed. Best way to get accurate data. Reading at the tailpipe has a delayed signal and is not necessarily the exact AFR's that are running at that rpm point

Hotwire
31-05-2011, 04:09 AM
Yeah I noticed the time delay, about a second or so, but was just accomodating that in my head. Interesting point about the bung being welded on, will consider it, however what do you do with cars you tune on the dyno? Surely you dont weld a bung on every car you tune to connect in the wideband....

Mitsiman
31-05-2011, 04:33 AM
That is correct every vehicle we tune, we charge a small $40.00 charge to weld an extra O2 bung onto the exhaust for tuning if they are still running stock ECU's. That way I can tune with 100% confidence that the AFR V RPM scaling is correct, no delays.

Hotwire
31-05-2011, 05:03 AM
OK well there ya go. Not a bad idea.

Shtiv
05-06-2011, 12:39 PM
Hey Lee, the time delay can be two things, at light loads the exhaust gases will take forever in the tailpipe and an innovate controller isn't the fastest thing around either. Put a motec PLM in the tailpipe next to it and you'll see massive differences in response time, but yeah a bung is the go especially for low load stuff. You also find some tailpipe/tip designs mean you can't get an innovate clamp in the airstream properly, again worse at low loads....

In summary you have to work around a being an innovate and b having it in the tailpipe (unless you put a bung in and you probably got one with the innovate kit)

Mitsiman
15-06-2011, 01:04 PM
For those interested, and I hope a few, I have been working hard the last 24 hours on the roms, and after finding more stuff in the forums, found the excellent vr4base1 file in another discussion, added more stuff into it that was missing and updated some addresses I went through all the xml's I had. I did a tune today on a VR4 Galant and it all worked perfectly. Perfect boost control, ignition mapping, fuel mapping, boost limiters everything. The car went out tuned at 285hp at all four wheels at 15 psi boost with stock boost system. Check the following link for a dyno graph

http://www.rpw.com.au/shop/index.php?option=com_content&view=article&id=420:-mitsi-galant-vr4-awd-6a13-twin-turbo-285hp-atw&catid=55:mitsubishi-galantlegnum-models&Itemid=40

Anyway, I have re uploaded the definition files to the Evoscan database following link

http://www.evoscan.com/roms/mitsubishi/index.php

All the hard work from all the various sources including Steve Knight Racing in South Australia (I spent several hours on the phone with him yesterday with some key data), guys on this forum, the geek mapped forum, Acamus (Spent around 2 hours with him going over code late last night as well) are finally all coming together. Hopefully with the further development work being done on the base file will start to pay off soon.

I have fully tested the 20030013 file with my car but would be great if we had some people doing the same on the other files. Just to be certain.
Looking forward to doing more of these vehicles now.

lathiat
15-06-2011, 01:42 PM
Good to hear Mitsiman, someone has recently put together a denso coil on plug setup for the VR4 using denso coils that runs (fully installable having custom harnesses with mitsi to denso plugs etc, plus milled plates to hold them, etc) about $750 which might help with those misfire issues - other people have reported those exact same problems even with newer plugs/leads. I'd be curious to know if you can fix that enough for the stock boost system purely with standard new plugs/leads..

My car will be up for tuning in the next couple months just getting things done in preparation - got the clutch & intercooler sorted now, need to sort exhaust/front pipes, 3port boost solenoid, etc... [blue legnum had you do a base run for me the other week]

Shtiv
15-06-2011, 02:49 PM
Hi Dave, some parts of those xmls look like a bit of my work, small world in vr4 ecu hacking circles.... I haven't gone through your xml in detail but I will and i'll give you a list of what i know works for sure if you can do the same for me. I have tuned a heap of these including 7202, 7203 and EM2428type 7202 ECU's and in autos and manuals. I even came across some 7203 ecu's with EM2428 code on them but I think that was a previous tuner attempting to lock the ecu as both those cars had been flashed previously. Assuming you're willing once we have combined our knowledge I will clean up the scaling names etc and repost them up as messy xml's just s&%t me.

lateshow
15-06-2011, 04:07 PM
Mitsiman: Why does your dyno paper state "power flywheel" and "torque flywheel", doesn't that mean same as from the engine? 285 from wheels would mean ~350 from engine... and 411 from the wheels would mean over 500Nm from engine...

Edit. We gained 310 bhp and 414Nm from the engine with 0.8 bar boost pressure (full exhaust, no fmic)

foxdie
15-06-2011, 04:09 PM
Mitsiman: Why does your dyno paper state "power flywheel" and "torque flywheel", doesn't that mean same as from the engine? 285 from wheels would mean ~350 from engine... and 411 from the wheels would mean over 500Nm from engine...

Yeah I was curious about this :)

Mitsiman
15-06-2011, 05:19 PM
The xml's are a combination from every source I could find, so yes code may be messy. By all means if you clean it up send me the cleaned versions and I will re upload them

My dyno is a hub dyno so ti can read hub horsepower aka slightly higher than wheel horsepower, or flywheel output> I don't use correction factors so flywheel or wheel horsepower are identical on mine (Correction factor is set to 1.0 so they read the same).

I normally set horsepower to wheel horsepower, but torque has to remain flywheel horsepower as the scaling is different and people get confused on the torque side of things.

At the end of the day, the main thing is the vehicle in question has gained around 105hp at the wheels, doesn't really matter from where the gain is measured from. But I will update the website to reflect that.

Mitsiman
15-06-2011, 05:23 PM
Hi Dave, some parts of those xmls look like a bit of my work, small world in vr4 ecu hacking circles.... I haven't gone through your xml in detail but I will and i'll give you a list of what i know works for sure if you can do the same for me. I have tuned a heap of these including 7202, 7203 and EM2428type 7202 ECU's and in autos and manuals. I even came across some 7203 ecu's with EM2428 code on them but I think that was a previous tuner attempting to lock the ecu as both those cars had been flashed previously. Assuming you're willing once we have combined our knowledge I will clean up the scaling names etc and repost them up as messy xml's just s&%t me.

No problems - I can confirm 100% though that on the 20030011 & 13 roms that the fuel, ignition and boost maps all work perfectly. I didn't need to mess with idle or other stuff like that. The core tuning components appear to be working fine and I have every beleif that the other roms will do the same as well.

I did most of the recoding at midnight last night so yeh the xml coding is probably a little hacked, but I also did it through ecuflash rather than direct coding.

Anything you want tested throw my way for sure, happy to do what I can. At the end of the day its only to all our benefits on this one.

lateshow
15-06-2011, 07:29 PM
But I'm still curious that what would be the actual peak values for this car. In Finland we have such dynos that measure the powertrains resistance, it is somewhere around 20% in manual 4wd mitsubishi. So could the peak power be then somewhere around 350? Because if we want actually do some(international) comparison we must speak on the same level :) It seems that in Australia you use that kind of dynos? Was it so that stock car makes about 185 atw there? Here stock autos do about 260 even facelift (flywheel) and manuals deliver about 280 as they should.

You must have upped the ignition value in 160 Load column few degrees in the upper RPM area to get such good max power. :)


The xml's are a combination from every source I could find, so yes code may be messy. By all means if you clean it up send me the cleaned versions and I will re upload them

My dyno is a hub dyno so ti can read hub horsepower aka slightly higher than wheel horsepower, or flywheel output> I don't use correction factors so flywheel or wheel horsepower are identical on mine (Correction factor is set to 1.0 so they read the same).

I normally set horsepower to wheel horsepower, but torque has to remain flywheel horsepower as the scaling is different and people get confused on the torque side of things.

By the way, ID love to see boost curve or AFR curve if you've got one!

At the end of the day, the main thing is the vehicle in question has gained around 105hp at the wheels, doesn't really matter from where the gain is measured from. But I will update the website to reflect that.

Mitsiman
15-06-2011, 07:44 PM
I can only guess in Australia there are three main Dynos, there is the Dynapack Hub Dyno which most reputable tuners use these days, there is the Mainline Roller Dyno, definitely the best roller dyno on the market and then there's the dyno dynamics roller unit, much cheaper with less options compared to the mainline dynos

Timing wise yes, this vehicle was running in the 180 - 190 load levels and running around 18 - 21 degrees timing if I recall. No knock at all, not even 1 count. But I wouldn't say that would be the norm for every vehicle.

Kenneth
16-06-2011, 12:20 AM
Thats interesting, the base xml looks exactly like mine, even down to the naming of items :P

Shtiv
16-06-2011, 03:08 AM
Yeah I noticed all my untested/unsure/incorrect comments that Kenneth and I had been sharing made it in ;) In fact I could trace some of those back to another guy in Brisbane....

Assuming all the addresses are the same as mine (as I haven't gone through it) I can tell you that
ignition maps work, fuel maps work, boost control works, boost limits work, boost cut delay works, air temp vs timing works, cat retard maps work, speed limits work, one of the dwell timing works (I think it was your first one but it's my 2nd one), open loop tables work

as there is no lean spool this is enough for most to tune.

I would say the people to thank for this are AndyF and a few others from geekmapped, merlin from ozvr4 and here, ninjadaniel from ozvr4, kenneth (above), steve knight from racetech, hotwire from here and ozvr4, patryn999 from here, oh and me.

To explain the dyno numbers a bit the flywheel numbers on a dynapac dyno have a correction if you know it but most leave it at 1 so it doesn't alter the number from what power is measured at the hubs. The reference to flywheel mean that the torque has been calculated referenced to the engine speed, the torque at the wheels is obviously different due to the diff and gear ratios. so if you knew it was a 24% drivetrain loss then you could multiply both numbers by 124% and get engine torque and power.

I have used dynapac hub dynos and they are very safe, consistent and once the car is on them they are quick to use, just start in 4th and away you go.

I also use a mainline roller dyno which is also consistent but has the usual benefits and downsides of a roller dyno.

Dyno dynamics dynos are generally less consistent.

lathiat
16-06-2011, 07:04 AM
Here's my car on Mitsiman's dyno

Un-tuned 7201

http://lathiat.net/files/dyno-legnum.pdf


Only mods are some kind of catback exhaust, a terriblly small looking cat converter (for aussie compliance) and a manual bleed valve up as high as it will go (since wound back a snip as we found a small boost spike @ 4200rpm / 14psi was causing knock).. unfortunately didn't do a run on stock boost. Having a couple manuals & autos done totally stock would help for comparisons I guess.

Mine ran ~355NM / 172kW

lateshow
16-06-2011, 11:02 AM
I thought that this applies: engine power * 0,8 = hub power -> So instead of multiplying with 1,2 you divide by 0,8 which gives you actually a bigger result. It power train loses 20% of power it must multiply the power by 0,8. What do you think guys?

Now we have everything we need for normal ecutuning, great thanks to those who had been involved in this!

Shtiv
16-06-2011, 12:36 PM
Each dyno is different even dynapac vs dynapac so best bet is to get a few stock versions of the car you are talking about and assume the quoted figures are correct, then you know the multiplier for that dyno with that drivetrain.

EG on the hub dyno i use I got 211kW, I was told to multiply by about 1.24 based on other evos, which gave me 262kW at the flywheel. The mainline roller I use has a much bigger multiplier (1.42 for our cars I'm told) which worked out to 267kW with a few differences between the two states of tune so I would have expected a very slight gain, whatta ya know. ps dividing by 0.8 is multiplying by 1.25, very close to my 1.24 advice....

That all said I don't compare cars on different dyno's. If you want comparison at least stick with the same dyno and the same settings, same day ideally. Really dynos should be used as a tuning tool and to see if you are going forwards or backwards or as a comparison between different cars but keep all the runs on the same dyno at least and pref on the same day.

so back on topic I'll try to combine and clean up the xmls over the next little while....

Shtiv
16-06-2011, 01:57 PM
Got Motivated, have gone through the scalings but now I have to do the tedious bit and check/rename scalings etc within the rest of the xml. As for the xml so far I have:

2byte logging info added

Made to suit 7202 ECU, change H8539F to H8539FA for 7203 ECU's (mine is 7202,'ll do a 7203 at the end for those of you with this ecu)

AFR16 Scaling to and from expressions were equal, change to match AFR8 scaling

Deleted Redundant AirflowGrams scalings

Made both Aiflow Hz scalings to single decimal place increments (8bit was to whole numbers, 16bit was to 0.2)

Added Blobbits scaling

Added x-1 function to coil dwell time, not sure this is right but it means the scale works, need to check there isn't a 1ms correction from address???? (have to lok this up later) all values are 0.96ms....

Added Coil dwell time offset scaling incase above is correct

Deleted redundant dynamic boost scale

Deleted Enrich128_8 scaling

Deleted redundant ISC steps scaling

Deleted redundant injector Latency evo

Deleted Redundant Injector size scalings

Deleted Redundant Load 16 bit scaling

Deleted Redundant Load 8 bit scalings

Deleted redundant Percent (128) scaling

Deleted Redundant Percent (255) scaling

Deleted Redundant RPM scalings

Deleted Redundant Speed Scalings

Deleted Redundant Temp Scalings

Deleted Redundant Time Scalings

Added Time(s) 8 bit scaling

Deleted Redundant Voltage scaling

Mitsiman
17-06-2011, 05:23 AM
Looking good once you have cleaned it up to your satisfaction let me know and I will rehost it up on the site.

Shtiv
20-06-2011, 01:27 PM
Did a bit more tonight, I now have everything combined into one file and only one of everything, I still have to clean up all the scaling references within the actal addresses and I'll chat to a few others offline about how we would all like to categorise each address before I go any further so we agree on a naming convention of where things should be.

This may take a week or two methinks

zentac
23-06-2011, 06:30 PM
Good to hear Mitsiman, someone has recently put together a denso coil on plug setup for the VR4 using denso coils that runs (fully installable having custom harnesses with mitsi to denso plugs etc, plus milled plates to hold them, etc) about $750 which might help with those misfire issues - other people have reported those exact same problems even with newer plugs/leads. I'd be curious to know if you can fix that enough for the stock boost system purely with standard new plugs/leads..

My car will be up for tuning in the next couple months just getting things done in preparation - got the clutch & intercooler sorted now, need to sort exhaust/front pipes, 3port boost solenoid, etc... [blue legnum had you do a base run for me the other week]

Cant really see the point, good standard coil packs with the proper grade plugs will take you well into the 600's

Shtiv
24-06-2011, 03:17 PM
On a daily driver the plugs don't last that long with more than 15psi, many people over here are getting 10-15k from a set and then having to replace, and yes this is with new coilpacks and decent leads. On a car that doesn't do massive miles though I agree there would be little benefit.

zentac
24-06-2011, 04:55 PM
On a daily driver the plugs don't last that long with more than 15psi, many people over here are getting 10-15k from a set and then having to replace, and yes this is with new coilpacks and decent leads. On a car that doesn't do massive miles though I agree there would be little benefit.

what rate plugs are you using?

Shtiv
24-06-2011, 10:55 PM
tried bkr6eix (don't handle the boost), bkr7eix, bkr8eix, and the copper equivalents (the coppers only lasted about a week with big boost!), currently running denso IK22's and have some vk20's to go in when these die but currently running coil on plug on all cylinders with denso coilpacks and they seem to be lasting better (up to almost 10000km and they haven't missed at all with big boost)

Nutter_John
24-06-2011, 11:00 PM
I ran ik24's at 1.2 bar boost for the 25,000 miles with no issues at all , have run bkr7's at 1 bar for 20,000 miles and no issues

strange that you guys down under seem to get quick plug wear and spark blow out

what fuel octane do you guys use down there

CANDEE
25-06-2011, 12:05 AM
Most of us are running BP Ultimate John, which is 97.5 octane. We dont get your 99 octane v-power over this side of the world...

wintertidenz
25-06-2011, 01:09 AM
Steve - have you tried the double platinum Bosch plugs to see if they last longer? They are the FR5DPP222 from memory.

I'm running the WR7DPX platinum plus ones which seem okay on increased boost.

zentac
25-06-2011, 07:19 AM
if your running big boost you want ik24's (depends what your classing as big boost) I've even got a set of ik27's that I was going to use.

evo's run 500bhp on standard coilpacks all day long if you get the correct plug in. VR4's should be the same sounds like you've just been using the wrong plugs.

Ryan
25-06-2011, 08:31 AM
Most of us are running BP Ultimate John, which is 97.5 octane. We dont get your 99 octane v-power over this side of the world...

Mobil 8000 claims to be "minimum 97.5RON" on the pump... BP 98, says nothing apart from the usual marketing. Just curious to know where your 97.5 comes from?

Shtiv
25-06-2011, 10:52 AM
I don't know but it's not just me, it's everyone here, once the Denso IK's go then the VK's are a platinum electrode so hoping they go well but like I said with COP no issues so far anyways....

The big boost I was talking about is 19ish psi but everyone here running more than 15 has issues. Anyway enough of the tangent we have COP and it's fixed the problem.

NLduffyNL
25-06-2011, 11:47 AM
What problems do you guys get because ur talking about the sparkplugs?

Shtiv
25-06-2011, 12:15 PM
misfires (spark blowout) at high boost

NLduffyNL
25-06-2011, 12:40 PM
Ohh ok didnt read it good sorry. Ive got bkr7EIX, driving it at 1boost and no problems (driving on Ron98)

MackTheKnife
08-07-2011, 03:23 AM
I can confirm the address 1036c does change the Speed Limiter on my car. I was pulling away from an Exige on the main straight of Eastern Creek Raceway a month or so ago and he had an GPS indicated speed of 200 km/h so definately well over the 185km/h speed limiter.

Just noticed it is in the XML as untested.

Strangely my datalogger and the speedo maxed out at 188km/h, should have used a GPS myself.

Shtiv
11-07-2011, 11:37 AM
Yeah I learned that a little while ago on the dyno, my speedo went almost vertical but the dyno was reading just over 200. I found a switch in my car to enable/disable the ecu input, it was disabled, now it's back on for the launch control. Good to see you pulling away from an exige in a station wagon :)

Mitsiman
11-07-2011, 12:36 PM
How is the cleaning up of the definition file going shtiv? Looking forward to getting a clean up version. Once you do i will go through and retest all the features on another vehicle on the dyno.

MackTheKnife
12-07-2011, 12:14 AM
Yeah I learned that a little while ago on the dyno, my speedo went almost vertical but the dyno was reading just over 200. I found a switch in my car to enable/disable the ecu input, it was disabled, now it's back on for the launch control. Good to see you pulling away from an exige in a station wagon :)

Off topic: The Exige was the non supercharged version, I was only pulling away because I got out of the last turn way ahead. The Exige has alot of aero work so he could take turn 1 at 180km/h. The most the Leggie could do was 140 km/h in current trim, any quicker it either understeered or oversteered depending on how I was feeding the throttle.

Back on topic: I hunted for a switch on the speed signal wire and I definately don't have one, so can definately confirm the address. I can only move it higher so can't really test any better and I can't quite hit 350km/h (where it's set now).

Shtiv
20-07-2011, 02:09 PM
Well I haven't hit my 250 limiter setpoint yet, when I do I'll just flick the switch inside my console (that should be fun at 250) haha!

Sorry about the cleanup, I've been doing what feels like a million projects at once, I'm spendingsome more time on it now....

Shtiv
20-07-2011, 05:02 PM
Hmmm, this is going to take longer than I thought, I'm going through the scalings and there are a heap of wrong ones in there (wrong increments) so I have to fix all that up too.... At least it's not just cleaning now but correcting also....

Mitsiman
21-07-2011, 01:55 AM
well let me know if there is antying I can do to help

As for hitting load 250 I don't think that would even be possible on a Legnum unless you welded the wastegate actuators shut

foxdie
21-07-2011, 09:26 AM
As for hitting load 250 I don't think that would even be possible on a Legnum unless you welded the wastegate actuators shut

I thought he was referring to 250KPH?

Kenneth
21-07-2011, 10:37 PM
Hitting 250 load is not only possible, I have done it.


As for hitting load 250 I don't think that would even be possible on a Legnum unless you welded the wastegate actuators shut

Mitsiman
22-07-2011, 02:27 AM
off the top of my head 250 load would be around the 20 - 22 psi range? Didn't think the stock turbo's had that much in them

Kenneth
22-07-2011, 02:40 AM
Yeah, I had some smoke on deceleration for a while after that... seems to have cleared a bit now though. (Or at least I haven't noticed it)

It was partially by accident to be honest, I used the red ended hose (the one with the restrictor pill in it) on the wrong side of the BCS during a blonde moment.

Mitsiman
22-07-2011, 02:50 AM
Yeah, I had some smoke on deceleration for a while after that... seems to have cleared a bit now though. (Or at least I haven't noticed it)

It was partially by accident to be honest, I used the red ended hose (the one with the restrictor pill in it) on the wrong side of the BCS during a blonde moment.

Classic - hopefully no long term damage :)

Kenneth
22-07-2011, 03:16 AM
Got spares, so not too worried. Main worry is the time and inconvenience of doing part swaps.

I have a turbo upgrade in the pipe line also, so not really worried about the life span of those turbos.

So long as you can get rid of detonation, I don't think that standard turbos have enough in them to do the engine much harm.

lateshow
22-07-2011, 04:45 PM
Hitting 250 means about 1.5 bar boost pressure, i have hit 240 sometimes. But it's just a brief moment. I'm not too worried about those little bastards, if they get ruined I'll be replacing them with bigger ones. Biggest load in the max power that i've seen is somewhere around 160 -180.

foxdie
25-07-2011, 08:34 PM
This is related to previous posts in this thread but OT to the thread as a whole, for that I apologise, but I must know, how are you guys getting the Map Tracer in EvoScan to work?

I can get the table to populate by selecting "Load Calculated" and "RPM" respectively at the top, but I can't get it to load the table below. I try selecting my ROM and nothing happens, I've tried copying vr4base1.xml to EvoScans ROM folder (My Documents\Evoscan v2.7\ROMS or something like that) to no avail, copying all my defs and the ROM from EcuFlash's folders to the same, I even tried following instructions from EvolutionM on removing all definitions in vr4base1.xml apart from load/rpm scaling and fuelling/ignition tables, using both the samples and my own CSV logs, still can't get the blasted thing to work.

Despite what the above path says, I'm actually using EvoScan v2.8.0005.

foxdie
26-07-2011, 08:23 PM
Nobody has it working? :/

taylor
26-07-2011, 11:52 PM
Pm me your email and Ill send you bits you need. I offered to do a write up a while back but noone seemed interested.

scott.mohekey
26-07-2011, 11:59 PM
Pm me your email and Ill send you bits you need. I offered to do a write up a while back but noone seemed interested.

You should do a write up.

taylor
27-07-2011, 12:01 AM
You should do a write up.

You should find a front diff for the v8

scott.mohekey
27-07-2011, 12:03 AM
You should find a front diff for the v8

You should GET a v8.

taylor
27-07-2011, 12:11 AM
You should get a 4g63. If I wanted a heavy lump of steel in my engine bay ill go buy a ship anchor. Plus no one likes a copyfag

scott.mohekey
27-07-2011, 12:12 AM
You should get a 4g63. If I wanted a heavy lump of steel in my engine bay ill go buy a ship anchor. Plus no one likes a copyfag

The v8 is all aluminium and actually lighter than the v6..


Sorry for the thread hijack guys.

foxdie
27-07-2011, 09:48 AM
PMSL! You're like Nick Mann and Richy Rich when they get together :D

PM sent, if you want once I've got it working I'll write up a clear and concise howto document crediting yourself and whomever else involved :thumbsup:

Shtiv
11-08-2011, 02:45 PM
Update, I was cleaning up the base xml file, I have decided I'm doing it arse about and I've now made it take 20x as long as it should. So! I'm starting again....