NefMoto

Lots of great information here.  I've been reading for weeks preparing to perform some deletes on my ECU so that I can achieve readiness after physically removing SAI, N249, EVAP.  I was pretty sure I had all my ducks in a row.  But today after flashing an edited file using Nefmoto ECU Flasher, I could no longer start the car.

I was able to connect to the ECU with Nefmoto with no issues, I was able to read and make a backup of the working flash file.  There seemed to be no communication issues.  I turned the ignition off, and back on again, and reconnected to the ECU, and then initiated a Full Write using the modified bin file.  This file contained only a few edits, and checksums were updated by WinOLS with no issues.

Nefmoto reported that the write succeeded 100% and that it disconnected from the ECU as a final step.  I turned off the ignition, and started the vehicle, and it started just fine.  I turned off the ignition, and attempted to start it again, and it would NOT start.  At this point I decided that it was a bad file, and that I would restore to the backup bin.  However Nefmoto will no longer connect to the ECU.  Neither will VAG-COM.  It seems that KWP2000 can no longer be used.

What went wrong?  I was very careful to follow all instructions... flashed while engine was off, and the battery was connected to a charger.

ECU: 06A906032PL (20th Anniv. 03 VW GTI)
Original bin: GIAC K04 File
Modified bin: GIAC K04 File with sai, evap edits. (6 byte changes)

I know funny things happen with these GIAC files, but the address space I was editing was identical to the stock file.  Beyond that, I was under the impression that you could still connect to the ECU again with Nefmoto even if the file was bad.

sounds like the checksums didn't get updated.  use ME7check - http://www.nefariousmotorsports.com/forum/index.php/topic,447.0title,.html (http://www.nefariousmotorsports.com/forum/index.php/topic,447.0title,.html) to verify your .bin checksums.

if nefmoto won't connect you'll need to connect in bootmode using a galletto 1260 cable.
http://www.nefariousmotorsports.com/wiki/index.php/Galletto_1260_Flashing_Cable (http://www.nefariousmotorsports.com/wiki/index.php/Galletto_1260_Flashing_Cable)

if you post your .bin i'll verify it for you and update checksums if needed.

Yeah, ME7Check reports a checksum even on the original GIAC file.  People have discussed this before on this forum.  I don't really understand why WinOLS doesn't fix this checksum error because it does indeed update checksums when I make changes.  The modified file has 2 checksum errors. 

Looks like I'll have to break out the Galletto tomorrow and work in boot mode.

Is it true, however, that a checksum error will cause the entire ECU to be unresponsive?  I thought it just throws a DTC and the engine won't run?

Thanks for your help.  I have attached:

1.  The stock ECU file.
2.  The GIAC K04 file.
3.  The GIAC K04 file modified with SAI & EVAP deletes.

do you have an .xdf for  06A906032PL?

I was never able to find one for 032 PL, but in my searching I was told that the 032 LP, is nearly identical.  So for TunerPro I've been using the attached 06A906032LP XDF.

crap - sorry i can't do checksums for that .bin - but they definitely need to be corrected.  do you have the full version of winols?

you should be able to reflash your original file w/ galletto then once you can connect to your ecu again reflash it again w/ nefmoto to clear the checksum error (making sure you're choosing the correct memory layout in each).

ALSO - before going to the trouble of getting into bootmode you may try disconnecting the ecu from power for an hour then trying nefmoto again.

Cool, thanks for the advice.  Definitely gonna try the power tip... well, I have the ECU opened up on my desk right now, so it'll certainly be powered off.

So I should be able to flash my original GIAC file back using Galletto?  Or do I need to use the stock flash file?

Regarding memory layouts with the Nefmoto flasher... according to the following wiki information, the 06A906032PL should be using the ME7 AM29F800, however, I am looking at my opened ECU chip-board, and the flash chip has ME7 AM29F800BB clearly stamped on it.

http://nefariousmotorsports.com/wiki/index.php?title=NefMoto_ECU_Flashing_Software#Supported_ECUs

I'm wondering if the failed flash has anything to do with this discrepancy.

you should be able to flash the giac file back just make sure and choose the correct memory layout.  the giac file does show 1 checksum error w/ ME7check but i've seen that w/ giac files before as i do not believe they do an overall checksum update (located at the end of the file) though their files run fine.

remember that in any case you'll have to connect with nefmoto at some point in order to clear the checksum error so you'll be back to normal.  galletto doesn't clear this error.

i've wondered about this as well.  Tony addressed the question of memory layouts a few times - here is one where he states that writing will fail straight away if one chooses the incorrect layout.  the only way to 1/2 write data w/ an incorrect layout is if you choose a 400 layout for an 800 chip. 

http://nefariousmotorsports.com/forum/index.php/topic,258.0.html (http://nefariousmotorsports.com/forum/index.php/topic,258.0.html)

fwiw - my ecu has the BB chip in it but i always use the 29F800 layout and have never had any problems.

also keep in mind that these ecus are pretty sensitive to voltage fluctuations during write procedures.  i am running a lightweight battery for the summer and have had several writes fail when the ecu is in the vehicle b/c the voltage dropped slightly during the write (i was too lazy to hook up the charger).  i've always been able to recover w/ bootmode.

Hmm... good to know.  Have you ever tried writing with the BB layout?

I'm guessing it might be a non issue because the flasher util reported that the modified firmware had been successfully written 100%.  There were absolutely no errors or failures when I wrote the file.

An old story ...
If you patch GIAC file and then write it back car starts only once.
After you switch ignition off programm "crash" serial eeprom (8-pin 95040) content.
Now you must first properly restore eeprom dump to return ECU back to life.

Also, if you ever write an incorrect checksum, you have to disconnect battery for a while and then reconnect.  You should then get comms with Nefmoto.

Rick

Do you have more information on this?  That's terrible news if it's true.

Can the current state of the 95040 be read now? It would be interesting what is written to it! Also is there a definitions file available for the 95040?

There is a basic XDF I believe. A brief search didn't find it but it's here somewhere :)

Here as in Nefmoto? Do you remember what format? (xdf/ols/a2l/damos?)

Yes, an XDF here on NM. Just implementing the checksum method detailed by setzi62.

If you can think of any other search words (other than the obvious) let me know so I can take a look. Thanks!

So I was able to revive my ME7.5 ECU.  I don't really know for sure exactly what was wrong with the flash that I uploaded... (probably bad checksums).  I had a heck of a time trying to get the darn ECU into boot mode in either the car or on a bench.  I wasn't able to initiate a read of the ECU unless I left Pin #24 connected to ground the entire read.  If I pulled it off after 5 seconds of power, I'd get "BOOT MODE INACTIVE".  After a full read, if I viewed the resulting .bin file in a hex editor, it was utter nonsense.  It looked nothing like a legit file.  The entire file was random bits and there were no maps to be found.  It didn't resemble the file I had flashed at all.  I have a feeling leaving pin #24 connected during a read corrupts the output.

When performing a write of a "good" file to the ECU, I was able to get it into boot mode by disconnecting Pin #24 from ground right before the Galletto software initiated the Write sequence.

In any case, I was able to write my original GIAC file back once I figured out boot mode.  There appears to be no problem with the EEPROM contents.  The car started right up.   Strangely I did not have to reflash the file with Nefmoto again to clear any code.  I was told there would be a code that could only be cleared by the Nefmoto flasher after writing with Galletto, but I see no evidence of this in my case.

Thanks again everyone for the help.

That is because the boot mode pin is shared with a Flash pin.

There is a reason you should only hold it to ground *while* powering up the ECU, not after.

Oh but I did.  I tried everything.  The only time the Galletto software didn't tell me that boot mode was inactive was when I left the pin connected to ground while the ECU powered up, and throughout the read.  If I disconnected it before I initiated the read - nothing.  It took me 2 days of trial and error to discover this.

But that makes sense why the resulting .bin after a Galletto read was garbled nonsense, because one of the pins was probably always outputting 0's. ha.

This :)