At 08:00 AM 2/22/99 -0800, you wrote:
>DML,
>I get the feeling most of these aftermarket computer guys are struggling and
>confused... JET told me my R/T has a knock sensor (it doesnt), Hennessy
>thinks JET is made by Superchips and will turn your check engine light on
>(it doesnt), and Mopar seems to forget what year it is (Im beginning to
>relate to that one)... Speed brains seems interesting, but Im not that fond
>of the idea of using a mass air flow sensor in my truck (kinda obvious and
>has its own issues), and Hypertech doesnt seem to know what a Mopar
>is...Arent there any game computer science guys out there that would like to
>take on a reverse engineering project?? Great senior thesis with real cash
>at the end!!
>BKB
I'd like to try to make a new computer some day... I've got the equipment
to burn microcontrollers and a C compiler that generates code for Microchip's
line of microcontrollers. So, we've got the tools, all it would take is
getting the right code on the chip(s) and wiring the thing up...
I don't know anything about how CC did the Dak computers, but I'm assuming
they're using code protection, which makes reverse engineering extremely
difficult. You can't just pull the code off the chip, so (as far as I know)
what you have to do is follow all the leads to see what sensors they're going
to, then find the specs on the sensors (voltage range), and apply varying
voltages to that pin while monitoring all of the others. You *might* be
able to extrapolate the way the computer is thinking by doing that. (I don't
have any formal training in this area, so I could be completely off the
wall; the above is how I'd do it though.) Anyway, a time consuming, difficult,
expensive process.
Anyway, Accel sells a fuel injection kit that you can use to install
fuel injection on any vehicle, and the general theories about EFI are
available (the diy_efi mailing list is an example), so it seems almost
like it would be easier to start from scratch. Use the pre-existing
sensors and injectors to get a fuel injection system up and running,
then add all the other stuff. You could probably run the custom computer
in tandem with the PCM, with the custom computer intercepting all signals
to and from the PCM. Anything that hasn't been implemented by the custom
unit could be "passed through" so the PCM won't generate an error code
for a sensor it doesn't see... One problem with this method is the OBD-II
tests. The PCM will sometimes purposely run the engine lean or rich to
test the O2 sensors, so if the PCM sprays extra fuel and the engine
doesn't go rich, it'll probably cry "foul". One way around that might be
to build an "engine simulator", which would be a computer that interfaces
with the PCM, and feeds it sensor signals to simulate a properly working
vehicle. Problem with that is if you know enough to do that, you can
probably build an entirely new PCM anyway... :-P
On the pre-'97 trucks, the guages and stuff aren't run by the computer,
so starting from scratch would probably be fairly simple; you just need
to get the fuel injection up and running, then add a few other little
things. I have a feeling it'd be a lot harder on the new trucks.
Once I swap out my 318 for a 360, I've been thinking of putting the
318 on an engine stand and using it to develop a computer. The problem
with this though is that I barely have time to do the simple piddly
little mods that I want to do to my Dak, let alone tackle a project
of this magnatude!
If anyone wants to collaberate on such a project though, I might
be interested in participating, even if its just to butt in with
my opinions... :-)
-Jon-
.--- stei0302@cs.fredonia.edu ------------------------------------.
| Affiliations: DoD, EAA, MP Race Team, NMA, SPA, USUA. RP-SEL |
| '96 Dodge Dakota v8 SLT CC (14.58@93.55), '96 Kolb FireFly 447 |
`----------------------- http://www.cs.fredonia.edu/~stei0302/ ---'
This archive was generated by hypermail 2b29 : Fri Jun 20 2003 - 12:12:43 EDT