pcmhacking.net

ttvn is the expert on that half of the operation

you must clamp it when its oriented correctly so it presses into the solder and contacts properly

as advised by Newbie on here i have successfully swapped them using a full on heat gun from bunnings

they can take an incredible amount of heat, amazes me honestly

i was so nervous the first one i did

i suggest pulling apart an old computer or something to practise on first to get the confidence up

i practised on an old video card out of pc and used sergical tools for pulling the chips off

you could just hold it vertical and heat evenly till it falls off too

Essential programmable features to turn on for the elm327 for testing.

AT PP 01 SV 00 - HEADERS
AT PP 01 ON - HEADERS ON AT START UP
AT PP 02 SV 00 - ALLOW LONG MESSAGES
AT PP 02 ON - ALLOW LONG MESSAGES ON AT START UP
AT PP 0C SV 23 - 115200 BAUD
AT PP 0C ON - 115200 BAUD ON AT START UP

AT Z

If you need to return you ELM to normal, type AT PP FF OFF or if your in more serious trouble refer to pages 54 and 55 of the ELM327 manual.

Gas soldering iron with a hot blowing tip works well. Work one side at a time and lift gently with a flat blade. Put something on top if you are concerned as a heatsink but I've pulled off over 100 chips and never cooked one surprisingly

But too much heat can lift the tracks/pads after a couple of pull offs

vn5000 thanks for the follow up on the $AE structure (6C 10 F0).

Have been doing a lot of reading/searching and found some interesting things. First according to ELM http://elmelectronics.com/obdtips.html they don't support J2411 single wire CAN or J2534 pass through. Now short of saying i'm getting a little confused with all the (J) protocols, this is typical as they are the J-protocals holy grail for gm re flash/programming for the LS1 and other engines. It's not the end of the road... ELM327 DOES support J1939, it just so happens that J1939 has many (maybe identical) similarity's to single "CAN" 4x it just uses different PINS on OBD2 obviously. Companies make J1939 to OBD2 converters for just this reason ? ...

This information is quite preliminary and much more needs to be looked into, it might be the ELM firmware that needs to be examined, maybe jumpers installed in the IC, or OBD2 pins on the vehicle switched around (or fast internal circuitry switching).

Comments appreciated

-UKNOMEPRK

You dont need 4x to flash. It just takes 4x longer without it!

antus wrote:You dont need 4x to flash. It just takes 4x longer without it!

1. if we could devise a way that "takes 4x longer" to reprogram the PCM I'd be ok with that.
2. According to http://www.akmcables.com/obdii.htm <--(typo here caused the edit sorry) the people who made the cable for LS1edit a "flash" is not possible without 4X.

I think you might be thinking of "bit banging" which is a good choice if it can be assembled. 4X (from reading) seems to somehow (probably by A0, A1) send the 12v signal to the processor to accept CAN high speed. Either way I know for sure that dumping the bin will be easier than a FLASH or REPROGRAM. Definitely the cart before the horse here, but this still needs to be addressed.

The 0411 is not a CAN bus pcm. Its J1850 VPW.

Although I cant remember where I saw 4x was not required. But at this stage I still believe it. It'll come to light as we go further with it. Keep in mind that the flash code is user supplied and runs from ram, so I dont see what would limit you to 4x if your writing the software for both sides of the process. Its essentially an embedded system and you can do what you want with it after its unlocked and running your code.

Wait WHAT! isn't J2411 4x Single wire CAN for reprogramming? and J1850 for diagnostics?

No its not can. Its 4x VPW. Its the same protocol, just 4x as fast. Can bus came later. As I understand it at this stage the boot sector has the comms code in it, and when you set a flag the ecu resets, the bootloader inits the hardware then in this case will jump to your code in ram. Then usually the pc software requests 4x and things continue. I dont see any technical reason why you cant leave it in 1x and still do the flash, provided the code in ram calls the right comms functions in the boot loader and the pc also knows whats going on. I wonder if dimented24x7 is monitoring this thread? I know hes looked in to the comms and the bootloader.