Reading eeprom from delco PCM, MC68336 based

[durahax]

If you're in Canada I would recommend a USBBDM NT for $65 USD http://www.usbjtag.com. It works great on '0411. The BDM pinout for the 90-Pin header is the same on all GM PCMs I've came across, even from newer controllers with the much smaller pitch 90-Pin header. I do know you need to use a resistor to pull up voltage somewhere to flash, need to look. I would think it would be easy to edit the flash config to suite the flash on your PCM. I have a pogo setup for everything I do, I need to take some videos when I have time. Used an old PCM cut the 90 pins out, soldered pogos in the bdm ports and rigged it to fit in my stand for the TriCore stuff I have. I sold my older 03 Duramax, and now I have a 2008 and 2011 Duramax. So I'm just working on mostly CAN atm. The willem also will do it if you have the 12v one, some are incompatible.

[antus]

I reckon durhax's idea is the go. For reference the GQ-4X is made in canada so would be local delivery from mcumall.com But in your case, yes BDM.

[antus]

Heres a high res pic of the top side of the '0411 pcb.

0411 top side.jpg (3.19 MiB) Viewed 5481 times

[iblackford]

Wow, that looks quite a bit different than mine, has anyone seen something similar to mine?

20150325_193626.jpg (3.45 MiB) Viewed 5455 times

[antus]

Wow, that really is first gen. Delco had not even started manufacturing the chips yet. I wonder if that toshiba chip is the same as 16238528 or 16232947 on the '0411. Does it have the 90 pin header, or does it pre-date that and only have the array you've connected to around the flash?

[festy]

I wonder if that toshiba chip is the same as 16238528 or 16232947 on the '0411.

That Toshiba IC is a 256k 8 bit SRAM.

[iblackford]

I only have the pins around the flash, there are few more pads then needed for the flash, I just ohmed out and connected the ones that I needed.

The PCM actually wraps around the enclosure, one side is the CPU and supporting circuitry, and the other side are all the drivers and signal conditioning circuitry. I'm trying to be very careful not to damage the flex circuitry as it's a pain to fix.

[antus]

I wonder if that toshiba chip is the same as 16238528 or 16232947 on the '0411.

That Toshiba IC is a 256k 8 bit SRAM.

Yes, but on the 0411 and most delco pcms you see delco part numbers and 466 deing delco's manufacturer id. They buy the ip and manufacture the chips with only their own part numbers printed on the ic. So when reverse engineering the pcms we have no reference to the delco part numbers. So if we can cross reference 0411 parts with prior gen with the original brands and part numbers we can figure more out.

[iblackford]

So after family commitments, easter, etc...I'm still working away at this but haven't solved my problem. Current status:

-wired up my test eeprom (in order to build my programmer). I wrote read, write, and erase routines for it. Tested, and working well.
-modified the java code for the meeprogrammer to add support for larger flash chips (also based on arduino, kind of where I started)
-rewrote the serial interface state machine in the arduino (from the meeprogrammer)
-tested reading large blocks of data from my test eeprom
-bought a 2nd pcm (it was cheap!) this one is everything I need it to be, but I can't seem to get it to learn the passlock..I've read my year needs the tech-II...

Still need to do:
- have to rewrite the java meeprogrammer app to match up with my new state machine on the arduino
- have to add support for 16bit data bus (currently I'm using an 8 bit data bus, but I have more than enough pins for a 16bit bus).

Questions:
-Once I can sucessfully read and write the eeprom on the pcm, can I passlock fix the new pcm (does this just mean making the VIN match)?

Thanks for any and all help, and once I get a working eeprom programmer I'll release it to sourceforge or something like that and folks can make their own. I don't have much chip selection at all, so if anyone wants tested support for a particular chip, pm about sending me some samples.

[antus]

I dont know, but i'd suggest copying the whole block of data with the vin as there are likely other things there you'll need. This way you would also inherit the correct checksum for the segment.