The Top DSM Community on the Web

For 1990-1999 Mitsubishi Eclipse, Eagle Talon, Plymouth Laser, and Galant VR-4 Owners. Log in to remove most ads.

Please Support Morrison Fabrications
Please Support Morrison Fabrication

ECUflash EvoScan and EcuFlash tuning

This site may earn a commission from merchant
affiliate links, including eBay, Amazon, and others.

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

Thanks Ill give it a try. I would rather run windows 98 if I could because the computer seemed a little quicker running it.
 
Alright, have and update for everyone. The 16G turbos are appartently on backorder according to RRE, so I just went with a T-28 for now. I'm dropping my car off at RRE next monday (10/27) to get the turbo and the Evo VIII ecu installed.

I will keep you posted with updates as they are working on the car.
 
An update on the early 1997 model year ECU.

EvoScan seams to work perfectly. The car is a August 1996 production date, 1997 model Talon AWD. The date rate bounces from 80-100/sec. 80-100 what? I don't know. . . Megabites, bites, bits, breaks. . . Anyway, this is with 20 perameters being logged simultaneously. I didn't start the car. I'm not done with a headgasket swap. But I couldn't wait. I just plugged in the cable, turned the ignition to 'on', and the laptop began downloading the drivers. Then I clicked start DataLogger and she began recording. Injector scaling was preset to 450cc, baudrate 15625, device setup: Mitsubishi, Ecu: EFI.

I can play back the log. It takes me to the saved logs folder in the evoscan folder. But when I go to that folder independent of clicking 'start playback', the file is not there. I can't drag and drop the log somewhere else, when the folder does show it (when I click 'start playback'). I didn't have 'log to file' checked in the 'Logging' dropdown menu when I logged this data. Could this be the reason why? I wanted to upload the CSV file for you guys.
 
I've tried so many different ways to upload this kind of file to show on here too and can't seem to find a good host for the file. Anyone else want take a swing at it?

:dsm:
 
Well mine isn't a loading-onto-the-server-problem. I can't find the file after I log it. But I can replay the log. ?

The log files come out CSV (comma separated values). So you can open up the file in notepad, wordpad, or even MSword or other typing program and then copy/paste right on to a post you're creating here. It probably would be good to go through the log and only paste the string that you have a question about. Since CSV can get very long. But I can't open up the file. Or even see it. I can jsut replay it in the logger and graphing utility.
 
The 100 is the samples per second ;)

Concerning CSV, you can probably enter it here as it's only a text file, but there was a problem with regional settings, if the decimal divider is , or . I don't remember which, the EvoScan doesn't want to read it back without some fiddling.

Good to know the pre-97 ECUs are loggable too :) Although I had problems with some older ECUs when I tried the same settings.
 
I think 'log to file' should do the trick, lots of times it is easier to hold the data in an area/file/format that can't be messed with by the user as a means of stopping things from breaking and making it easier to write the program. If you've ever noticed importing or exporting files from just about any application is extremely unreliable compared to the sort of everyday operations that you might use your computer for.
 
Well mine isn't a loading-onto-the-server-problem. I can't find the file after I log it. But I can replay the log. ?

The log files come out CSV (comma separated values). So you can open up the file in notepad, wordpad, or even MSword or other typing program and then copy/paste right on to a post you're creating here. It probably would be good to go through the log and only paste the string that you have a question about. Since CSV can get very long. But I can't open up the file. Or even see it. I can jsut replay it in the logger and graphing utility.

Well, I meant a way to actually post the file, so we can open it up in EvoScan. Kind of like what the DSMlink owners do.

:dsm:
 
I did a log last night (on someone elses car thats around 425whp or so) and it seemed a little slow. It could be my computer though, I am currently waiting on more memory I ordered. This was on a 1995 GSX with an eprom ECU. The info wouldnt show up on the screen unless I was recording the data??? anyone know why? Did I have the settings jacked up. I had it on hybrid/3000gt.... and on EFI.
 
Well mine isn't a loading-onto-the-server-problem. I can't find the file after I log it. But I can replay the log. ?

The log files come out CSV (comma separated values). So you can open up the file in notepad, wordpad, or even MSword or other typing program and then copy/paste right on to a post you're creating here. It probably would be good to go through the log and only paste the string that you have a question about. Since CSV can get very long. But I can't open up the file. Or even see it. I can jsut replay it in the logger and graphing utility.

Others have had problems finding log files with Windows Vista. If you are using Vista there is a service pack from Microsoft that is downloadable that should repair the problem.
I had problems finding my logs too, and I could find them after running searches. I did not do the service pack though. Under logging I go into the "Open log file in graph", but right click on the one I want and open it with Microsoft Works Spreadsheet.
 

Attachments

  • open log.jpg
    open log.jpg
    58.7 KB · Views: 762
I did a log last night (on someone elses car thats around 425whp or so) and it seemed a little slow. It could be my computer though, I am currently waiting on more memory I ordered. This was on a 1995 GSX with an eprom ECU. The info wouldnt show up on the screen unless I was recording the data??? anyone know why? Did I have the settings jacked up. I had it on hybrid/3000gt.... and on EFI.

Evoscan will not display anything unless you are logging. Also if you only log the essential data items it will log faster. In other words the less data items you are logging the more samples you will get up to a point. For example. I currently get 150 samples per second logging only around 8 data items. If I log 20 - 25 I only get somewhere around 60-80. Im using my corperate laptop with a core duo .
 
Others have had problems finding log files with Windows Vista. If you are using Vista there is a service pack from Microsoft that is downloadable that should repair the problem.
I had problems finding my logs too, and I could find them after running searches. I did not do the service pack though. Under logging I go into the "Open log file in graph", but right click on the one I want and open it with Microsoft Works Spreadsheet.

I run vista on my personal laptop and have no problems finding the logs when I actually use it to log. I always goto LOGGING > Explorer - Open csv Log File Folder

:thumb:
 
Others have had problems finding log files with Windows Vista. If you are using Vista there is a service pack from Microsoft that is downloadable that should repair the problem.
I had problems finding my logs too, and I could find them after running searches. I did not do the service pack though. Under logging I go into the "Open log file in graph", but right click on the one I want and open it with Microsoft Works Spreadsheet.

I've tried exactly what works for you, but it doesn't work. The only thing that works: open up the saved logs folder by clicking 'playback log', delete the file, go to the recycle bin and copy/paste the file the desktop. Then I can open it up. . .

Do you have a link to the service pack?

Here's the log file. Click on 'free user' to down load the file. Remeber, I didn't even have the car running.
 
I've tried exactly what works for you, but it doesn't work. The only thing that works: open up the saved logs folder by clicking 'playback log', delete the file, go to the recycle bin and copy/paste the file the desktop. Then I can open it up. . .

Do you have a link to the service pack?

Here's the log file. Click on 'free user' to down load the file. Remeber, I didn't even have the car running.

I dont think I ever downloaded a service pack. Maybe its because I have Vista Ultimate.
 
Thanks! Yep, I can see the file just fine when I click on 'compatibility files' at the the top of the vista window of the saved datalogs. . .

Ok this is a 1997 model year (aug 1996 production date) Eagle Talon AWD turbo. Interestingly, the log I have shows the tps at 0%, air temp at 77*F when it was about 66*F that day, 66*F coolant temp which should be about right, and MAF air temp scaled is at 172.4*F. . . Which is clearly wrong since the car wasn't even running. . . Target idle is also reported at 1216rpms when the car is stock, so target should be 750ish. Clearly the formulas are wrong. The speed with the car on blocks is 263 mphLOL . Do I have to wait for evoscan updates to get this to read accurately? Does someone have the accurate formulas for the logable variables? Or will I just have to trial and error figure out what the formulas are? Does anyone have a suggestion as to how to insure the knock sum is reading accurately :) And any suggestions on a procedures to generate or test the forumlas for any of the logables.

The injector scaling is defaulted to 450, which is good.

EDIT:
-I'm a douch. This is a MAFT car, so the 77*F air temp is about right since the MAFT locks the air temp input. However, I thought that it locked it at 80*F though, not 77*F (but that's not too much difference). . .
-Apparently I'm not really supposed to pay much attention to the MAF air temps scaled perameter. . .
-The target rpm COULD be 1200rpms because the coolant temp is reading cold, so the ecu is in cold start mode. Would this make sense?. . .
-The evoscan notes say that the ability to log speed from the ecu is for 98 or newer cars. So this explains the speed sensor issue. So I guess we won't be getting any 70-90 times to log how well certain upgrades are doing :( .

So does anyone know if the 1997 ecu spits out a correct knock sum with evoscan? Does anyone know if the 1997 calculates knock sum the same as 1998-1999 ecus? If so then, the formula in the evoscan software should be the same, correct?
 
Thanks! Yep, I can see the file just fine when I click on 'compatibility files' at the the top of the vista window of the saved datalogs. . .

Ok this is a 1997 model year (aug 1996 production date) Eagle Talon AWD turbo. Interestingly, the log I have shows the tps at 0%, air temp at 77*F when it was about 66*F that day, 66*F coolant temp which should be about right, and MAF air temp scaled is at 172.4*F. . . Which is clearly wrong since the car wasn't even running. . . Target idle is also reported at 1216rpms when the car is stock, so target should be 750ish. Clearly the formulas are wrong. The speed with the car on blocks is 263 mphLOL . Do I have to wait for evoscan updates to get this to read accurately? Does someone have the accurate formulas for the logable variables? Or will I just have to trial and error figure out what the formulas are? Does anyone have a suggestion as to how to insure the knock sum is reading accurately :) And any suggestions on a procedures to generate or test the forumlas for any of the logables.

The injector scaling is defaulted to 450, which is good.

EDIT:
-I'm a douch. This is a MAFT car, so the 77*F air temp is about right since the MAFT locks the air temp input. However, I thought that it locked it at 80*F though, not 77*F (but that's not too much difference). . .
-Apparently I'm not really supposed to pay much attention to the MAF air temps scaled perameter. . .
-The target rpm COULD be 1200rpms because the coolant temp is reading cold, so the ecu is in cold start mode. Would this make sense?. . .
-The evoscan notes say that the ability to log speed from the ecu is for 98 or newer cars. So this explains the speed sensor issue. So I guess we won't be getting any 70-90 times to log how well certain upgrades are doing :( .

So does anyone know if the 1997 ecu spits out a correct knock sum with evoscan? Does anyone know if the 1997 calculates knock sum the same as 1998-1999 ecus? If so then, the formula in the evoscan software should be the same, correct?

Well, the accuracy of formulas is a bit unclear. I noticed some weirdness as well, and actually one 97 (late) car, as well as mine with plastic ECU, are showing speed output incorrectly as well.

Then again, I guess the only way to get the real formulas is either from the dealer or if someone manages to disassemble the ROM. :mad:

BTW, there is a new EcuFlash (1.40) version, has anyone tried it yet?

And now some bad news, I might have been wrong about extended mode 4, and there seems to be an evidence, with which I am starting to agree, that our ECUs are in single chip mode 7 (H8 ECU Definition Information Thread - Club3G Forums).

If the information is correct, then we might have some problems... Although specification says "When pin settings are made for mode 2, 4, or 7 and 12 V is applied to the VPP pin, flash memory can be programmed or erased. See section 18, "Flash Memory" for details."

On the bright side, all maps/axis can be found really easily as far as I know ;)

Also note that if we have our chips in mode 7, that would mean we can write and then read back our custom ROMs fully.

The areas would be divided as follows if we have mode 7:
[0x0-0x1ff] (0x200 bytes / ROM) -> (IVT) interrupt and exception handling table should be here as far as I understand from spec, mirrored to page 0 in flash
[0x200-0x3fff] (0x3E00 bytes / ROM) -> functions and other stuff should be written here, mirrored to page 0 in flash
[0x4000-0xEE7F] (0xAE80 bytes / inaccessible) -> inaccessible area in mode 7, but this is a good thing, because it cannot be read back when reading ROM
[0xEE80-0xFE7F] (0x1000 bytes / RAM) -> the content is junk and defined at runtime, so we don't care about the initial values
[0xFE80-0xFFFF] (0x180 bytes / sort of RAM) -> it's the registers whose initial values at boot up we know from CPU specification
[0x10000-0x1FFFF] (0x10000 bytes / ROM) -> IVT and ROM is mirrored here, but also the code we need to write, would write and read back as page 0 in flash
[0x20000-0x2FFFF] (0x10000 bytes / ROM) -> only code and data here, would write and read as page 1 in flash
This is what I understood from the H8/539F specification.

Mode 7 guarantees that if we are able to write our own ECU code and write it into ROM, we would be able to read it back fully as well. In mode 4 we would have problems with area [0x4000-0xEE7F]. We couldn't erase it or write data to it, and it might be possible that it would cause us problems.
 
Last edited by a moderator:
Evoscan will not display anything unless you are logging. Also if you only log the essential data items it will log faster. In other words the less data items you are logging the more samples you will get up to a point. For example. I currently get 150 samples per second logging only around 8 data items. If I log 20 - 25 I only get somewhere around 60-80. Im using my corperate laptop with a core duo .

Thanks. I thought I did somthing wrong when I had it plugged in and wasnt getting any readings. Also I did have all of the data items selected the only thing I messed with was switching the check from DSM to Hybrid/3000gt/blahblahblah.
Thanks for the reply
 
Ok, there is one more thought that came to my mind.

As far as we know, there is some sort of lockout problems that doesn't allow the ROM to be erased. That could be caused by booting in incorrect mode as well.

The specification at page 520 says:
Example of User Program Mode Execution Procedure:
...
Activate the CPU with a reset. (Activation from user program mode is possible by applying 12 V to the Vpp pin during the reset, i.e. while the RES pin is low*2.))
...

That means you can only enter programming mode if the +12V are applied to a programming pin or something.

I'm not too good with all the electronic stuff, so can anyone confirm that we do have 12V applied to a programming pin if the white plug is connected?

If we don't have, that might be the problem. The programming code supplied in specification relies on +12V being supplied at bootup as well. If the key has been turned, and there is no +12V, the ECU boots and it's too late to reprogram it.
 
Yea I'll check when i get to my garage tomorrow

Ok, there is one more thought that came to my mind.

As far as we know, there is some sort of lockout problems that doesn't allow the ROM to be erased. That could be caused by booting in incorrect mode as well.

The specification at page 520 says:
Example of User Program Mode Execution Procedure:
...
Activate the CPU with a reset. (Activation from user program mode is possible by applying 12 V to the Vpp pin during the reset, i.e. while the RES pin is low*2.))
...

That means you can only enter programming mode if the +12V are applied to a programming pin or something.

I'm not too good with all the electronic stuff, so can anyone confirm that we do have 12V applied to a programming pin if the white plug is connected?

If we don't have, that might be the problem. The programming code supplied in specification relies on +12V being supplied at bootup as well. If the key has been turned, and there is no +12V, the ECU boots and it's too late to reprogram it.
 
Any word on the new beta version of ECUFlash that will flash H8?


EDIT: EcuFlash - OpenECU this was just posted up today.
Information seems scarce, even on the openecu boards. There's practically no mention of the 16v requirements other than in the change log.

Is it possible to modify a cable or acquire a cable that is capable of the 16v output needed to flash H8's?
 
Looks like it still doesn't work.
we will not be able to release a version of EcuFlash for the Mitsubishi H8 processor family of ECUs that works with the Openport 1.3. The reason being that the Openport 1.3 power supply circuit that generates the ~16V the ECU needs to supply the regulated reflashing voltage to the CPU does not have a sufficient current capacity, causing this voltage to drop to levels insufficient to safely flash the CPU (the CPU goes into a protection mode to prevent you from even trying). The Openport 1.3 does generate a voltage large enough to cause the processor to startup in boot mode, which is why you could read and compare in past versions, but unfortunately, it just isn't usable for reflashing. Later processors like the SH2 and M32R generate this reflash voltage internally, and have minimal current requirements, which is why the Openport 1.3 works with the Evo 7/8/9 without a problem.
 
Support Vendors who Support the DSM Community
Boosted Fabrication ECM Tuning ExtremePSI Fuel Injector Clinic Innovation Products Jacks Transmissions JNZ Tuning Kiggly Racing Morrison Fabrications MyMitsubishiStore.com RixRacing RockAuto RTM Racing STM Tuned

Latest posts

Build Thread Updates

Vendor Updates

Latest Classifieds

Back
Top