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 ExtremePSI
Please Support ExtremePSI

Can't Read/Write in EcuFlash *Evo 8 ECU swap*

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.

PFC_CiarlilloM

10+ Year Contributor
234
0
May 22, 2011
Euclid, Ohio
Okay So I am doing the Evo 8 ECU swap in my 2G 96 Eclipse. Pin swaps went fairly easy after I got the hang of removing them. I plugged in my Tactrix 2.0 cable, and reflash adapter.

I started up EcuFlash, and went to read the ROM -failed. Then I attempted to flash over a ROM, and received this;

[20:47:11.788] EcuFlash Version 1.42.2595 [20:47:11.788] OS Version Windows Vista
[20:47:11.788] Qt Version 4.5.0
[20:47:11.818] 56 memory models read.
[20:47:11.828] scanning for metadata models in C:/Program Files (x86)/OpenECU/EcuFlash/rommetadata
[20:47:13.790] 344 ROM metadata models scanned.
[20:47:13.790] checksum module "subarudbw" loaded.
[20:47:13.790] checksum module "mitsucan" loaded.
[20:47:13.790] flashing tool "wrx02" loaded.
[20:47:13.790] flashing tool "wrx04" loaded.
[20:47:13.790] flashing tool "sti04" loaded.
[20:47:13.790] flashing tool "sti05" loaded.
[20:47:13.790] flashing tool "mitsukernel" loaded.
[20:47:13.790] flashing tool "mitsukernelocp" loaded.
[20:47:13.790] flashing tool "mitsubootloader" loaded.
[20:47:13.790] flashing tool "shbootmode" loaded.
[20:47:13.810] flashing tool "shaudmode" loaded.
[20:47:13.810] flashing tool "subarucan" loaded.
[20:47:13.810] flashing tool "mitsucan" loaded.
[20:47:28.138] J2534 API Version: 04.04
[20:47:28.138] J2534 DLL Version: 0.50.2598 Jun 26 2009 15:02:29
[20:47:28.138] Device Firmware Version: 1.11.3298
[20:50:36.648] sending init sequence 1 (0001)
[20:50:39.891] sending init sequence 1 (0003)
[20:50:43.116] sending init sequence 1 (FFFF)
[20:50:45.829] no response to any known code
[20:50:45.829] interface close
[20:51:45.961] J2534 API Version: 04.04
[20:51:45.961] J2534 DLL Version: 0.50.2598 Jun 26 2009 15:02:29
[20:51:45.961] Device Firmware Version: 1.11.3298
[20:51:57.479] sending init sequence 1 (0001)
[20:52:00.699] sending init sequence 1 (0003)
[20:52:03.919] sending init sequence 1 (FFFF)
[20:52:06.632] no response to any known code
[20:52:06.632] interface close
[20:52:15.769] 96530006 (157 tables) inheriting evo7base (73 tables)
[20:52:15.789] 262144 byte image read.
[20:52:20.629] J2534 API Version: 04.04
[20:52:20.629] J2534 DLL Version: 0.50.2598 Jun 26 2009 15:02:29
[20:52:20.629] Device Firmware Version: 1.11.3298
[20:53:04.821] J2534 error [ERR_TIMEOUT].
[20:53:05.321] sending init sequence 1 (0001)
[20:53:05.521] J2534 error [ERR_TIMEOUT].
[20:53:06.041] J2534 error [ERR_TIMEOUT].
[20:53:06.041] interface close

I am assuming it couldn't read because a combination of either/both, A. Computer wasn't set to "High Performance" or B. My flash pin isn't correctly installed.

Flash pin might not be making contact because I had to remove the pin that it came with and swapped it with a smaller pin from one of my OEM wires. I should have checked the front of the plastic harness housing to ensure that it was fully seated. I heard the "click" and couldn't pull it out, but it might not have went in far enough.

Anyways, the not reading isn't my issue. My issue now is that EcuFlash won't let me do anything. It's responsive, but won't let me "Read" <Didnt try "Write"> and if I exit, I get;

"EcuFlash is in the processes of reflashing your ECU. Exiting now would cause your ECU to be rendered permanently unresponsive."


So I am assuming that'd "brick" my ECU correct?

I am not sure why I attempted to "Write" if "Read" didn't work.... stupid me.

The Tactrix 2.0 cable is still lighting up solid blue with flashing white, green, and red colors to the right.



What do I do from here?
 
Last edited:
Just a note here and something out of habit I do before flashing any Evo ecu

Pull the case and take note of the actual board number to make sure you are flashing a compatible Rom and have the ECU you think you do.

I've seen A LOT of switched cases and you can run into issues... .
 
@mike96- I still can't read/write. Sorry for the confusion man. Basically it was around 5:45am, and I had to be at work by 7:00am -and I was 40 minutes out on top of that. So I tried starting the car with whatever ROM/MAP/Tune <presumably Evo 8> was on the ECU. After I got it to start, I logged it for about 5-10 minutes at idle, and then began starting the driving.

@MY1GDSM- That's my next step on a nicer day here since I don't have a garage, and that will help clear up a lot.

I'll work on sharing these logs as well. Ive been up for 36 hours now so I hope this makes sense.
 
Judging by what you have going on here, there's a problem with your reflash port. If the car runs and drives, then your ECU install was likely a success; if Evoscan can read the MUT addresses and datalog then you're getting OBD2 functionality. If the OBD2 port works, you can datalog; it doesn't mean that the flash plug works as it's not required to datalog.

Tuners likely won't let you link the file because of the extension. Try making a copy and changing the extension to .bin as that's an allowed file type. Then we'll just have to change it back to .csv when we download it.

The clicking you hear is a relay, IIRC it's activated (via the OBD2 port) when the ECU goes into diagnostic mode to be flashed. It clicks three times as the three different initialization codes are sent to the flash port (which according to your log, all three fail). This lends that the OBD2 port is functioning correctly, but the flash port is not. Now whether that's the cable failing, the flash pin not making good enough contact, or you got hosed with a case-swapped ECU is the real question.

Step 1) See if your fueling is safe:

You said earlier that the car started and drives, which is a good sign. Typically if it's not flashed to have the immobilizer disabled you'd never get it to turn over, let alone be drivable. It would seem that it is flashable there, but lets see what your fuel trims are doing to make sure your fueling is safe. If the ECU is using the stock Evo scaling you could be putting your motor at risk. Your car will compensate at cruise and idle, but it will not at WOT. Evo injectors are ~20% larger than ours, that can make you dangerously lean even on stock Evos pig-rich fuel maps.

We can't test this directly if you don't have a wideband, so we'll have to log it. I don't see a wide-band in your profile, do you have one? From what I can see in your profile you're running stock injectors, therefore you should be running distinctively lean as the ECU will think the injectors are ~20% larger than they really are.

What are your fuel trims doing at idle (LTFT low)? Warm the car to operating temp and let it idle while watching/logging STFT (O2 feedback), O2 voltage, and LTFT low; if you have a wideband it should be hovering around 14.7; your O2 voltage should jump up and down repeatedly (if it doesn't cycle you won't get into closed loop), and the STFT should hover around 0. If your wideband is very lean or the STFT is far from 0 (more than 10) or maxed at 25, then the injectors are likely not scaled to the stock 450s.

Idle the car for about 20 minutes to let your LTFT set itself; if it's running the stock Evo scalings, the LTFT should be maxed out at 12.5% adding fuel to try and get into closed loop (if it will even go into closed loop) and the STFT should be very positive or even maxed at 25%. If both trims are close to 0, then the ECU had to be flashed previously for our stock values, or came from a car using our sized injectors.

As a side-note, if you idle for 20 minutes and the trims remain at 0 and don't change, something is wrong and you're not going into closed loop. No tune (even the stock one) will stay solidly at 0, STFT should always fluctuate and LTFT should be close to 0 but not at it.

Are you showing any knock counts while driving? Typically if the knock multipliers aren't changed for the 2g knock sensor, you'll have crazy amounts of phantom knock during spool-up/part-throttle. That'd be another sign it's not flashed.

Step 2) See if your flash port/cable are good:

Unfortunately you'll need another 2.0 or 1.3U cable to test this. If you can find someone in your area who flashes their Evo that'd be easiest. I live up in Michigan and have a 2.0 cable that I flash my Talon with (Evo 8 ECU swapped). I could possibly meet you half-way (around Toledo) and test out your cable/see if my known working cable can pull your ROM.

Let us know what you come up with testing the fuel trims, and good luck with getting this figured out.

Edit: I see that the ECU wasn't pre-flashed: removing that. In such a case testing your idle trims will only show you what your fueling is like.
 
Last edited by a moderator:
Uploaded 3 logs that I have so far

1,)https://docs.google.com/open?id=0B5oFjeiE4fiRclZhMXo1eFN1Qjg
"EvoScanDataLog_2012.11.14_05.42.56.csv" was my very first log, and drive. I started the car in desperation at 5:50am because I had to be at work by 7:00am and I was about 40 minutes out at my parents house <nearest garage> It started, and idled pretty rough initially. I am sure the 20-30 degree temp didn't help along with a MAP/ROM/Tune from a different vehicle. Anyways, I was in a hurry that I didn't really log too much other than the basics that seemed to be checked by default ( i.e. TPS, RPM, Knock Sum *little did I know the default "26" address for Knock Sum in EvoScan was for the 97-99 models DSM's* but I also was monitoring Knock Voltage) I pulled away and started slowly accelerating cruising around 20, 25, 30, and 35 checking what I could see on EvoScan for roughly the first 5 miles between <5:50-6:00am> Then the inevitable had to happen, and I had to get on the freeway. I had a nice constant acceleration from 35-60 without getting too high in the RPM range. Once I saw that 60 seemed to be feeling okay, and going with my gut off the readings in EvoScan (complete tuning noob w/ out any experiece), I proceeded to go up to 65 and repeated the steps until I reached roughly 80 ish (I know -Daredevil and really testing my luck on first drive w/out knowing) It seemed to actually pull pretty damn good from 70-80, well maybe accelerate would be a better term. Pull would imply "flooring it" so to speak, and I tried keeping that to a minimum. Anyways, I kept roughly a speed of 70 the remaining 35-40 miles. My car actually seemed to be running better during the trip. I atleast though so, maybe it was the nice crisp air that helped. Drive ended around 6:30am.

2.https://docs.google.com/open?id=0B5oFjeiE4fiROFRXSHdmZWFlNGM
"EvoScanDataLog_2012.11.14_14.42.12.csv" Was a simple drive around the block, and highway. I switched the Knock Sum address from "26" to "3E" to help log knock because I read that it differs from 95-96, and 97-99. Either way, I selected A LOT more options to log this time, and didn't receive any knock most importantly. Drive was roughly around 2:45-3:20pm.

3.https://docs.google.com/open?id=0B5oFjeiE4fiRdF9aWWRPajZMclk
"EvoScanDataLog_2012.11.14_20.23.21.csv" Basically a more relaxed/conservative drive back home from work. Mostly highway with some city roads. Pretty much mirrored the above log in terms of the "options" logged. Trip was from 9:30-9:45P.M.

@Melbowski- I just noticed your reply, but haven't read it yet. I'll read it, and most likely you'll see an auto-merge. Thanks for the detailed reply.

I didn't get my Evo ECU from GSXDNA on here. He simply supplied me with a base ROM that I'd flash over since I have a similar setup as what he had at one point -just something to get me up, and running. I picked up my ECU from EvoM.net forums, but since I am unable to flash the ROM that GSXDNA sent me, I am just running with whatever ROM is on this "hopefully" Evo 8 ECU with stock ROM.

Also I am working on picking up a Wideband. I know it's important, and I shouldn't have started the swap without one as it's really a useful tool, but I should be getting one within the week. I am able to get a AFR reading in EvoScan, is this the same reading a wideband reading would see, or is it just guessing based on readings from other sensors?

On my last log, the AFR that it showed came up 14.7 immediately upon starting, then stayed around mid 14's high 13's. The richest I see it come up on the last log is 11.5, and no leaner then 14.7 FWIW. Again I'm not sure if these are the numbers a wideband would see, or not.

I believe on the last log my fuel trims (low, mid, high) and 02 feedback/STFT come up 0 during the whole trip, and I have it checked to log.

I don't have any knock. I found out in my first log that it was due to having the incorrect address "26" for "Knock Sum" since then I switched it to the proper "3E" or so I read that it's the correct address atleast. No knock recorded in the last two logs assuming the "3E" address is correct. However, in all three logs, knock voltage as well as other knock readings were logged.

I am going to see if for some reason I am experiencing a driver issue with Windows 7 64x by installing Windows XP 32x and trying to flash it. I will also check/redo all my pin swaps, and open up the ECU to grab some numbers to verify that it is a flash-able Evo ECU. My sister actually attends UT, and I am supposed to be going there for a football game next Tuesday (20th) So maybe we could set something up around then if all else fails?

I appreciate the help!
 
Last edited:
I have edited my previous post to reflect that the ECU was not pre-flashed for your car. I misread that, my bad. Please re-read my updated post above.

The wideband is important indeed; I wouldn't suggest the swap without it. However, you're not the first to do so without one. Honestly you should be able to scale the stock injectors reasonably well with the plethora of scaling on EvoM. The stock narrow-band will help you fine-tune the latency via idle and cruise logging once you get the ability to flash the ECU.

That said, the O2 reading you get in Evoscan is from the stock narrow-band sensor. Its range is extremely limited and it is near worthless for tuning because of its small range. Beyond idle and cruise you'll be well out of its range and the ECU will not adjust fueling beyond the stock fuel map.

If the ECU is scaled for stock Evo 560s you could be running dangerously lean at WOT. Do not go into boost until you get those idle logs up to check how far off the scaling is. The only thing that would save your motor if the scaling is for 560s is that the stock Evo maps are very rich under boost. As it stands we don't know what has been done to the maps currently on the ROM; I wouldn't tempt fate.
 
@Melbowski -If I can't end up falling asleep soon *which wouldn't even be right since I've been up since 12:10pm Tuesday working on this swap* I'll go out, and idle my car to see if I can get those fuel numbers in EvoScan that you mentioned, and then we can go from there until I get "reading/writing" working in EcuFlash.

I cut an unused wire with a smaller pin because of the 2.5 Mitsu Flash Adapter having to big of a pin, and soldered them together, and put it in Pin 79. I'll post up some pics tomorrow so you guys can see what I am working with here.

Also I just messaged the previous owner of the Evo ECU to gather more information on it. I.E. confirming that it's an Evo 8 ECU, and asked if it has ever been tuned/flashed
 
I found out in my first log that it was due to having the incorrect address "26" for "Knock Sum" since then I switched it to the proper "3E" or so I read that it's the correct address atleast.

Where did this info come from? I never changed the address in EVOscan to address not being able to logg knocksum and it works just fine.

Also if you haven't yet... You need sleep. 36+ hours, the human body can only take so much before it completely shuts down.

Side note now that @Melbowski mentioned it and jogged my memory I do recall someone else having an issue with either the EVO8 swap or 99ECU swap before. They couldn't flash the ECU, turned out being the pins in the data link connector were either corroded or loose or something. May not be the case seeing how you can at least get a connection with EVOscan but may still be worth looking into.
 
I think you only need to change the knock request to 3E when logging a 95-96 ecu. Since your using an evo ecu you should leave it at the default 26 setting in your evoscan edit data settings.
 
You don't need to change anything to log knock on a 98 DSM black box... I wouldn't think you'd need to change the knock address for the Evo ECU either.

The reflash pin in mine was a bit loose, so I took the pin out and crushed it down a little so it was forced open by the pin. That way I knew the connection was solid. You also need to make sure the voltage is over 12v before attempting to do anything but read from the ECU. There's a couple people who bricked their ECU because of a voltage drop mid-write, and apparently there's no recovering from that. Also, the fact that you're fuel trims aren't moving is a pretty good indication that it's just running in open loop.
 
@mike96- I found it in a thread somewhere on here, but after rereading it, I realized that it didn't apply with the Evo swap. I slept for 9 hours and it's nice enough to work on it but I'm stuck at work!

@blazer1297 That is absolutely correct. Atleast that's what I gathered when I reread the thread when I woke up today after only being up 4 hours instead of 28 ish at that time yesterday LOL.

@h@xtGSX- I haven't been able to get some LTFT because most of my drives have been short unless they log before 15-20 mins. I was planning on idling my car before I got to work for 20 mins and try and see if I could log anything but I was running late.

I attempted to log LTFT while at work but my damn computer died on me! just my luck LOL. I got a STFT reading of 5.xx I BELIEVE before it died though. It was 0 for roughly 7-10 minutes of idling then I believe it rose up to 5.xx. Then my laptop died on me unfortunately.

Also I get the same failed to read/write in Windows XP so it's safe to say it's a wiring issues or maybe a swapped case ECU. I'm leaning more towards a wiring issue but both are easy enough to test.

Will check wiring and grab numbers off the ECU case & board.
 
Last edited:
All you should have to do is look at your STFT. If they start cycling at all after a couple of min of the car running than your ECU is going into closeloop mode. I know I have had an issue before where I did an ECU reset and it actually took almost a week before anything started showing up on my LTFTs. so dont be too shocked it it takes a little while.
 
STFT should be shifting every few seconds. LTFT should be adjusting to that average created by the STFT every 15 minutes. If that isn't happening, you are either never in closed loop, or you have the wiring so messed up the car shouldn't run... So I am assuming the former. I have been using EvoScan for a while, and I am tuning with nothing but ECU Flash, and I make very, very frequent changes to my car... I've learned how to tell what Evoscan should be reading, and what it flakes out on. The fuel trims aren't something I have had a problem reading in ANY version of Evoscan from 2.5 to 2.9... And just as a side note, for DSM use, I have found version 2.7 to be the most stable... Being that you have an Evo ECU, that may mean nothing, but it is something to look into, just in case.
 
@mike96 & H@xtGSX- I'll be taking a look at all my wiring tomorrow, as well as checking/ensuring the pin swaps were correctly made, and I'll thoroughly examine pin 79 too. I'll get back here with my findings either way.

On a side note, I have to read up on open loop, and closed loop. I have SO MUCH to learn, but hey you have to start somewhere, right?

would this error prevent me from reading/writing in EcuFlash w/ a Tactrix 2.0 cable?

'[16:02:41.534] error loading FTD2XX.DLL"
 

Attachments

  • IMAG0160.jpg
    IMAG0160.jpg
    37.1 KB · Views: 113
  • IMAG0161.jpg
    IMAG0161.jpg
    37.2 KB · Views: 106
  • IMAG0169.jpg
    IMAG0169.jpg
    37.7 KB · Views: 93
Last edited:
Because I've tried just about everything. I ensured pin 79 made contact without the yellow harness, and I lightly crushed it a little bit so the ECU pin would have to push it open a little bit to ensure contact on top of resoldering pin 79. I am LOST.

but I just noticed that my connection in EvoScan goes from Green to Yellow every so often and has to autoconnect. So maybe one of my other wires isn't making a solid connection.

Also EvoScan is saying that the latest read ECU is a 01 Lancer7 Ralliart EDM.

So thanks to GSXDNA I was able to read my ECU after using a ECU Unlocker program!!! Finally!

Thanks everyone for all the help! I greatly appreciate everyone's input on my situation, and this is what makes DSMTuners so great.

I flashed this ROM over for a base tune, and will upload the log the first start -30 minute idle.
(working on splitting it into two as it's too big to upload to google documents)

Now onto learning EcuFlash, and EvoScan. I guess I can start with the 172 page Merlin's guide.
 
Last edited:
ROFL

It's your OBD2 port! The connectors inside it are too loose. Very, very common problem. You have to learn how to stick the Tactrix box into it just so, give it a special jiggle, and see if it works... I had to re-pin mine because it was so bad.
 
Shame the price on the 1.3 cable went up so much. Otherwise I would suggest just selling that 2.0 to someone else and purchase the 1.3... You rarely hear of anyone having issues with the 1.3 cable, and when you do it was because some pulled or bent the USB cord to much and broke the wires loose inside.
 
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