Reading/writing eprom chips on MSA15 ECU.

turbocharged798

Veteran Member
Joined
May 21, 2009
Location
Ellenville, NY
TDI
99.5 black ALH Jetta;09 Gasser Jetta
Hey guys, I have a question. Been wanting to play around with some maps on my 99.5 so I got a William programmer. I put each chip in the programmer and read each chip seperately. Once I had the two files, I used VAG EDC suite to merge the files. I did the LLL file first and HHH file but it wouldn't find the maps so then I did the HHH file first and it worked perfect.I was able to find the bad hot start map and fix it. I also did EGR delete, and massaged the torque map a bit as well. Then I did checksum with a free MSA15 checksum program I found on the net. So then I split the file into two files and wrote each file to a new AM029F010 chips I got off ebay. I didn't know which was the HHH and LLL file so I tried it both ways and neither would allow the ECU to boot.

Any thoughts what I screwed up? How do I know which is the HHH and LLL chips? VAG edcsuit just lists it as file 1 and file2.
 

turbocharged798

Veteran Member
Joined
May 21, 2009
Location
Ellenville, NY
TDI
99.5 black ALH Jetta;09 Gasser Jetta
So edcsuit won't map MSA15 correctly? The hot start map wasn't labeled correctly but I found it after some searching....
 

JFettig

Vendor
Joined
Aug 18, 2010
Location
Blaine, MN
TDI
B5 Passat, 2010 Jetta
edc suite won't work right, it'll recognize a bunch of maps, might even map them right, but you still need checksum corrections which it won't do for sure.
 

JFettig

Vendor
Joined
Aug 18, 2010
Location
Blaine, MN
TDI
B5 Passat, 2010 Jetta
Either you bought junk chips off ebay(with 99.999% of them are), or your checksums aren't working or something else. I have never tried doing a MSA15 ECU in edc suite - I read them and looked at them, but never saved out.

Real AMD chips are next to impossible to come by as they stopped manufacturing them.
 

turbocharged798

Veteran Member
Joined
May 21, 2009
Location
Ellenville, NY
TDI
99.5 black ALH Jetta;09 Gasser Jetta
Yeah the chips are definitely suspect. Gonna try and write the file to new chips again and see if it works. If not, I'll have to retune in winOLS. Seems like I am doing everything right thus far. I also suspect possibly that VAGsuite isn't splitting the files correctly. It did combine the two and found a bunch of maps.
 

Libang

Active member
Joined
Sep 4, 2009
Location
Czech Republic
TDI
TDI power :)
Either you bought junk chips off ebay(with 99.999% of them are), or your checksums aren't working or something else. I have never tried doing a MSA15 ECU in edc suite - I read them and looked at them, but never saved out.

Real AMD chips are next to impossible to come by as they stopped manufacturing them.
Use 39SF010A-70NHE .... 100% working on this ecus . Regards
 

LNXGUY

Top Post Dawg
Joined
Jan 10, 2004
Location
Barrie, Ont, Canada
TDI
'05 Jetta TDI Wagon
I've got a known set of good chips from Mike @ TDTuning (When I had my old 99.5) I've heard they're hard to find, let me know if you're interested. I've also got a spare 99.5 ECU that's been socketed as well. I'm pretty sure I've got a programmer to, lol.
 
Last edited:

Digital Corpus

Top Post Dawg
Joined
Mar 14, 2008
Location
Ontario, California
TDI
'97 B4 w/ 236K mi body, 46K mi soul
Btw, the old chips are 125 ns parts but having faster chips won't be an issue:
http://forums.tdiclub.com/showthread.php?p=4808549

As long as the pinouts match, supply voltages match, and the flash tech is the same (NOR not NAND), they should work just fine. Just make sure your programmer is setup for the different write algorithms if you don't use a 29f-series chip.
 

turbocharged798

Veteran Member
Joined
May 21, 2009
Location
Ellenville, NY
TDI
99.5 black ALH Jetta;09 Gasser Jetta
So I wrote another set of chips(same ebay AMD knockoffs) and used the built in checker for the William software. It read the chip and confirmed it matched the file being loaded into it. We will see if it works.

Do I need a hex address offset on the second chip? That's something that's been bothering me.
 

turbocharged798

Veteran Member
Joined
May 21, 2009
Location
Ellenville, NY
TDI
99.5 black ALH Jetta;09 Gasser Jetta
Ok think found the problem. First off in vagEDCsuite, chip 1 is HHH and chip 2 is LLL. Figured this out by comparing the tuned and original dumps in a hex editor. The original chips are marked so its just a matter of figuring out which file is similar. So that's ironed out.

Second, apperantly the PLCC32 socket on the william is causing issues. When I fully seat the chip in the socket and do a file check, it throws an error usually around address 0x090. I would guess some pins aren't making a good connection and are likely floating. This would cause all sorts of issues if I wrote the chips in this condition. As long as I keep the chip flush with the top of the socket, all seems well.

So, I see nothing now that should stop these chips from working. Checked both several times to make sure they match the original file. Going to give them a shot this weekend and hopefully the ECU boots.
 

Digital Corpus

Top Post Dawg
Joined
Mar 14, 2008
Location
Ontario, California
TDI
'97 B4 w/ 236K mi body, 46K mi soul
I have a standard push-in socket like what's on the ECU that is broken out for a programmer and a Yamaichi IC120-0324 on a breakout board as well. The latter had a bent pin and I had similar issues. However, with careful disassembly, the bent pin was rectified and all if functioning ay-okay now.

I bought a MiniPro programmer, aka TL866CS, since I couldn't seem to justify the cost of a new William programmer. It's smaller, has a case so I don't have to worry about dirt incursion in the field, and is a heck of alot smaller.
 

john.jackson9213

Top Post Dawg
Joined
Mar 14, 2007
Location
Miramar, Ca. (Think Top Gun)
TDI
1996 B4V
D.C.

Whitbread had a problem writing to a couple of sets of chips I was given before I left California. The two spare sets were problems to write to. But as soon as Whitbread used the AMD chips from my old ecu, there were no problems writing to the AMD chips. TDTuning says the VW chips are at best 50/50 and most ebay supplied chips are junk.

Understand the AMD chips are no longer made, so finding chips that work may be a problem. Any thoughts for a fresh supply?
 

Digital Corpus

Top Post Dawg
Joined
Mar 14, 2008
Location
Ontario, California
TDI
'97 B4 w/ 236K mi body, 46K mi soul
D.C.

Whitbread had a problem writing to a couple of sets of chips I was given before I left California. The two spare sets were problems to write to. But as soon as Whitbread used the AMD chips from my old ecu, there were no problems writing to the AMD chips. TDTuning says the VW chips are at best 50/50 and most ebay supplied chips are junk.

Understand the AMD chips are no longer made, so finding chips that work may be a problem. Any thoughts for a fresh supply?
Correlation is not causation.

Brand is not the problem here. Due to the nature of a lot of C.R.A.P. automotive parts here on the forums, brand is engrained into the psychology of a lot of users. Having "genuine AMD" chips is a consequence of that mindset.

Whether it is NOR or NAND flash, it will wear out when it's programed. There are a finite number of times each bit/cell/page can be written. Granted the write lifetime can be 500 or more writes, but they still wear out.

The original chips from your ECU have been written once from the factory and maybe a time or two for tuning after that so though they are old, they are "fresh". Ebay chips are a lottery for how many times they have been written.

Depending on the the socket he was using, the chips can rotate a little. The quick socket you saw me use, the latter one I mention in the post above, is susceptible to this problem. I've had problems writing to the chips because at a cursory glance the chip was seated fine, it was offset or rotated enough to where a pin wasn't making proper contact or a lead shorted two pins together.

I'll buy a handful of chips off mouser from the listing above or from digikey, I'll grab some 39f010's too, and toss them into a continuous program cycle if I can figure a way to script it and see how long they last. Yes, there are still problematic brands in electronics, though electrolytic capacitors are much more sensitive to this than flash chips are.
 

Digital Corpus

Top Post Dawg
Joined
Mar 14, 2008
Location
Ontario, California
TDI
'97 B4 w/ 236K mi body, 46K mi soul
Correction, the [nested] link above, reproduced here, expanded to all if you don't care about below zero operating temps, is for 39SF010 which are pin compatible.

If there is still an obsession for some to buy new, AMD chips, new old stock can be located using this little search engine here. Just run a text search for "AMD" and the number after the hyphen is 125 of smaller. I'm still going to add some of the previously mentioned chips to my next order of parts because I don't doubt their usability.

If peeps want to compare stats, here is a 39SF010 part from Microchip:
http://ww1.microchip.com/downloads/en/DeviceDoc/25022B.pdf
And here is an archived AM29F010 part from AMD:
http://robotics.ee.uwa.edu.au/eyebot/doc/DataSheets/29F010.pdf
 
Last edited:

Digital Corpus

Top Post Dawg
Joined
Mar 14, 2008
Location
Ontario, California
TDI
'97 B4 w/ 236K mi body, 46K mi soul
D.C.

Whitbread had a problem writing to a couple of sets of chips I was given before I left California. The two spare sets were problems to write to. But as soon as Whitbread used the AMD chips from my old ecu, there were no problems writing to the AMD chips. TDTuning says the VW chips are at best 50/50 and most ebay supplied chips are junk.

Understand the AMD chips are no longer made, so finding chips that work may be a problem. Any thoughts for a fresh supply?
Correlation is not causation.
Brand is not the problem here. Due to the nature of a lot of C.R.A.P. automotive parts here on the forums, brand is engrained into the psychology of a lot of users. Having "genuine AMD" chips is a consequence of that mindset.
Whether it is NOR or NAND flash, it will wear out when it's programed. There are a finite number of times each bit/cell/page can be written. Granted the write lifetime can be 500 or more writes, but they still wear out.
The original chips from your ECU have been written once from the factory and maybe a time or two for tuning after that so though they are old, they are "fresh". Ebay chips are a lottery for how many times they have been written.
Depending on the the socket he was using, the chips can rotate a little. The quick socket you saw me use, the latter one I mention in the post above, is susceptible to this problem. I've had problems writing to the chips because at a cursory glance the chip was seated fine, it was offset or rotated enough to where a pin wasn't making proper contact or a lead shorted two pins together.
I'll buy a handful of chips off mouser from the listing above or from digikey, I'll grab some 39f010's too, and toss them into a continuous program cycle if I can figure a way to script it and see how long they last. Yes, there are still problematic brands in electronics, though electrolytic capacitors are much more sensitive to this than flash chips are.
I forgot one other thing. Even though chips are pin compatible and can be read the same, sometimes a manufacturer uses different timings, delays, and voltages for programming. As such, if the programmer was setup to program AMD specific chips and the ones to be programmed do not use the same algorithms, then programing can fail as a result, despite the chips being perfectly fine.
 

turbocharged798

Veteran Member
Joined
May 21, 2009
Location
Ellenville, NY
TDI
99.5 black ALH Jetta;09 Gasser Jetta
So tried the chips, no dice. The I wrote the stock file the chips and still no dice. So that rules out the tune being bad. Then I installed a stock LLL chip and my stock file written HHH chip and it boots. So something with the LLL chip is causing issues. Tried writing it with a 0x020000 offset but the programmer wouldn't take it. I still think its a offset issue...

I am out of ideas at this point. The chips themselves seem to be working fine as I can read back off of them without issues and the HHH chip works fine as well.
 

Digital Corpus

Top Post Dawg
Joined
Mar 14, 2008
Location
Ontario, California
TDI
'97 B4 w/ 236K mi body, 46K mi soul
Writing is a different process than reading the chips. The chips could be a bit worn out or there can be a problem with your programmer too. That's why I bought another one.
 

turbocharged798

Veteran Member
Joined
May 21, 2009
Location
Ellenville, NY
TDI
99.5 black ALH Jetta;09 Gasser Jetta
Yeah but why would the HHH chips work and not the LLL? Every chip matches the original file exactly when read back.

This is seriously frustrating the heck out of me.
 

Digital Corpus

Top Post Dawg
Joined
Mar 14, 2008
Location
Ontario, California
TDI
'97 B4 w/ 236K mi body, 46K mi soul
Yeah but why would the HHH chips work and not the LLL? Every chip matches the original file exactly when read back.

This is seriously frustrating the heck out of me.
Without *seeing* what's going on or trying a different programmer, I cannot give possible reason as to what is the problem.
 

turbocharged798

Veteran Member
Joined
May 21, 2009
Location
Ellenville, NY
TDI
99.5 black ALH Jetta;09 Gasser Jetta
Figured it out! The LLL chip had a bad read when I read off the stock chips due to solder covering the pins. Did the file compare and it came back with an error. Wiggled the chip around in the socket and re-read the chip and varified it several times. Re tuned the file and it works great!!!! ECU booted no problem and car runs perfect. Since my checksum calculator checks against the original file(which was bad), it never caught the problem!

Now onward to attempting a full stage 2 or 3 DIY tune. I have a pretty good idea what to do at this point. For my first tune I just deleted the EGR, fixed the hot start map, and massaged the torque map a bit. I also bumped up the N75 a bit on the high end as I was getting some pretty bad spikes due to the 1019s fueling so heavy. Funny part is now the car is a tiny bit slower due to the lack of boost spikes. Holds a perfect 12-13PSI now. Hehe, this is fun.

Thanks to all who helped. :)
 

Digital Corpus

Top Post Dawg
Joined
Mar 14, 2008
Location
Ontario, California
TDI
'97 B4 w/ 236K mi body, 46K mi soul
Good to hear!

I'll be testing 39SF010 chips in the near future. I've spent the past 48 hrs battling a script kiddy that's been attacking OpenWRT firmware on the house router...
 
Top