You may or may not know, several of us have been collaborating with the final goal of a free. open-source like modified honda ECU supporting boost, vtec, n2o/external relay control, datalogging and pretty much anything else people want. Over the next week or two, I am going to try to add a hackedup WebCVS system and other tools to make collaborative work easier. Collaborative commenting of code dumps, group examination of new subroutines and sharing of ideas should be easier with the new system. Above all, FREE and OPEN exchange of information is encouraged. If the idea of hacking away in the ECU appeals to you, but you are intimidated by it, don't give up. Many of the valuable contributors to the site had little/no microcontroller experience prior to the pgmfi project.
The current consensus is that OBDII is not worth attacking at this time.

So far most work has focused on the 91 CRX Si PM6. Fuel and timing control routines have been isolated and commented. Support for custom fuel and timing maps has been made simple by a series of excel spreadsheets that automaticallyh convert between real world units and hex values to input into tables. Recurving tables for bigger injectors is easy. Datalogging and larger fuel tables to support boost are both in an alpha stage of development. A sound strategy for a 1wire vtec system has been suggested, although no code has been written. Testers and coders are desperately needed.

OBD1 ECUs are considerably harder to understand than OBD0. The microcontroller used does not have any FREE development tools unlike the OBD0 ecus. Glenn Anderson, who aparently wrote the first disassembler for Hondata based on some 66K manuals Doug Macmillan acquired, was kind enough to provide a mostly disassembled, partially commented PW0 dump. Additionally, many of the contributed rom images have been analyzed to determine the size and shape of fuel/timing tables. Current projects are basically writing a assembler and disassembler for the architecture so that further work can be done.

Sorry for the long post here, but I sincerely believe that understanding the Honda computers to the point where they can be used purely as a hardware platform for whatever software routines people care to write is not an unreachable goal. No one seems to be able to devote themselves to this project as their dayjob, so the "two heads are better than one" approach seems to be the only way forwards. Even if you just come and leech some of the many uploaded stock and aftermarket roms, spread awareness of the site.

http://pgmfi.crx-forum.org

Thanks
-D

 

TTT

 

Excellent idea! I've done some coding for the 68HC11, so I'm cool with assembly.

I'll be registering here shortly.

 

Great to hear man! Spread the word!

 

not much help for us OBD2 people, but i'm gonna keep my eye on this one. I'm a C and C++ UNIX programmer, but I've got zero real experience with anything like this. still, code is code, If I can help out I will.

 

looks good....

 

jweller: if you can find enough specifications to build a programmer + software for the 66589, etc. oki micros used in the OBDII ECUs, the DASM we are writing for OBD1 should be relatively easy to modify to support the nx-500 core. Additionally, you can always use an OBDII-OBDI wiring harness and use a OBDI ECU like you would with a hondata system. Seriously, the lack of external eproms and knowledge about the internals of the 665xx series is why OBDII has been ruled out for the time being

Oh yeah, my last assembly experience was 14 years ago hacking my apple II... I sucked at it then, but then again I was only 8 You learn fast, and as a programmer already, you will learn super fast. People are helpful.

eg6ajk: nice seeing you over at the site already!

come on over to http://pgmfi.crx-forum.org !!!!


[Modified by blundar, 10:35 PM 6/26/2002]

 

Sounds like a very interesting project. But you may want to rethink reverse engineering existing ROMs, it can be tricky, both technically and legally. Has anyone thought of building their own ECU from the ground up? I don't think they do anything very complicated, do they? Seems like it would be a bunch of simple calculations and table look-ups. Anyways, I'll be keeping an eye on that board...

 

legality: tell that to hondata and zdyne. Both have been around for years and both are still waiting to get sued. Honda doesn't give a rats *** if ppl make their computers make their cars go faster! They sell more cars in the end!

technical difficulty: its a big, hard project. There is a strong argument for forward engineering a new unit. I see nothing wrong with that. Excellent projects have already been made in this direction. The one that comes to mind first is the megasquirt, however it lacks timing control. If this is the direction you feel is best, there is a forum for standalone engine management stuff, please feel free to contribute.