Announcement

Collapse
No announcement yet.

Connect to 3.1 v6 (L82) MAF, likely A263 stream

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • Connect to 3.1 v6 (L82) MAF, likely A263 stream

    Summary
    I'm trying to connect (for maintenance not tuning) to my 1995 Olds Cutlass Ciera, 3.1 v6 (L82) with MAF, likely the A263 protocol. I have the ALDL documents, a laptop, Tunercat Pro (the nagware version), a $1FA.ads file for A261 protocol, several line converter schematics, max232s in the mail, a self built 2-transistor converter, and computer programming experience. I've run into some problems, has anyone had experience with the A263 protocol? Skip down to "Problem".
    Greeting/Intro
    Hello to the community. I'm not into tuning, and because of my strained finances, I probably never will. I am just a broke geek (computer geek) with an internet connection, who used the net to fix their car and now to find and fix all the other problems with the car. After all repairs are done (months from now?) I may do some "modding" such as timers to roll with windows, a lowend car-puter, custom alarm system (apparently Atmel sells chips).

    Last week I broke etiquette and PM'd 'pocket rocket' about their post of making an aldl cable. He seemed knowledgeable and talkative, so I broke etiquette this one time, and asked via PM what color the wires from the car were. Sorry, pocket. His response was quick and useful, and he suggested I post here about the car, since someone may have already had success with the A263.

    The Car
    My car is one of those "Bastard cars" from 1995. Its a 1995 Oldsmobile Cutlass Ciera, 3.1L v6 (L82) with MAF sensor. Its ODB, and apparently it uses the A263 protocol. Up until May, I didn't know much about cars other than how a generic engine ran from when I did a middle school project ~1993. I studied up and replaced the gaskets (head, LIM, UIM, Throttle Body, etc) on the car when yellow "snot" appeared in the oil. I then found that the mechanic of the previous owner did a lot of hack jobs on the car, and I also found many broken/damaged things. Broken vacuum lines, radiator fan wired to battery, camshaft position sensor wires damaged at the sensor, exhaust manifold bolts broken below the line, and many more problems. Some were beyond me and I had to leave as they were (drilling out the broken bolts), others I did fix or will as soon as I have more cash (cam sensor). To help isolate any other problems, I wanted to ask the computer what it saw wrong, and that brings me to why I am here...

    Problem
    I built the 2transistor converter, but the $1FA.ads file wouldn't work. I then learned that mine is a different protocol (A263), so I rewrote the ADS file. Still no connection. I threw together a C# program to read from the serial port, log the information, and allow me to send a few commands (mode 08, and a message request). Hooking it up while off, the program saw a few null bytes (the number varied each time). Turning the key to on, and the screen flooded with data. For my chatty ECM, that is perfectly normal. I tried sending a few Mode 8 requests, but it ignored it. I turned the key off and saved the captured data. Either the connection was out of sync, or my EBCM (ABS is device $F9) is ranting with weird messages of 1 to 137 bytes in length.

    I didn't process the log with a PERL script yet to see if they are valid messages (checking the stated length and then calculating a checksum to see if they match), so it may be jibberish. Also, I logged at 9600 as some sources say you can use 9600 in place 8192. I will try again later today at 8192 baud.

    Some people say that the 2 transistor circuit is crap, and it fails to connect more than it works. So perhaps when the max232s arrive everything will be ok. Has anyone had any experiences with the A263 protocol -- regardless of success?

    The ALDL files listed protocols for airbag and ABS computers too. Once I can talk to the PCM, I'm eventually going to see if I can get information from them too.

  • #2
    Are you running tunercat or tunerpro?

    The 1FA def for tunerpro is not the right one for your car. it's a bit miss-labeled and works on all the 3100's that have no MAF and are flash based. On the other non-maf cars you send a disable coms command, the ECM shuts up, then you send a connect command.
    sigpic New 2010 project (click image)
    1994 3100 BERETTA. 200,000+ miles
    16.0 1/4 mile when stock. Now ???
    Original L82 Longblock
    with LA1, LX9, LX5 parts
    Manifold-back 2.5" SS Mandrel Exhaust. Hardware is SS too.

    Comment


    • #3
      *checks quickly* Oh. Well apparently I was mixing the two up. I had heard both names and thought it was different flavors (lite, pro, etc) or a name change. I'm using TunerPro RT (http://tunerpro.markmansur.com/).

      I tried the A261 1FA before I learned the MAF version was A263. I did a quick rewrite of the ADS (I changed the commands, but I didn't rewrite the table strings, etc), and it still wouldn't connect. And yes, I was trying to send Mode 8 requests (the "shut up" command for usurping the PCMs master role).

      What I forgot to mention, is that I tried having TunerPro RT talk to my logger I made, and programmed my logger to respond to Mode 1 requests with the correct message (but filled with 0 or some specified number as the codes data) and it still would not connect. (Yes, I made sure to calculate the correct checksum for my outgoing message)

      The previous owners "mechanic" was shifty. I hope he didn't swap out the previous computer for one that ignored the DTC. (I read in a few forums that people have encountered computers that will not respond and need to be replaced.) He repaired a front end collision, but changed parts all over the car. He even _REMOVED_ the trunk release (no more solenoid, no more button, just wires).

      Comment


      • #4
        Interesting. I processed the data captured a 9600, and found several packets (length and checksums are valid), but the device IDs are invalid and the lengths are way too big. That's kinda strange that jibberish actually "looks" valid. Anyone tying to make their own scan software should keep this in mind.
        Code:
        Packet:  Dev[$01]  Length[$A0/ 75]  Mode[$00]  @   587
        Packet:  Dev[$80]  Length[$C0/107]  Mode[$A0]  @   701
        Packet:  Dev[$2F]  Length[$FF/170]  Mode[$9F]  @   941
        Packet:  Dev[$DA]  Length[$7F/ 42]  Mode[$DB]  @  1611
        Packet:  Dev[$FC]  Length[$9F/ 74]  Mode[$DE]  @  1744
        Packet:  Dev[$00]  Length[$FE/169]  Mode[$FE]  @  1950
        Packet:  Dev[$FF]  Length[$7F/ 42]  Mode[$FF]  @  2190
        Packet:  Dev[$FF]  Length[$FF/170]  Mode[$01]  @  4401
        Packet:  Dev[$00]  Length[$80/ 43]  Mode[$80]  @  7226

        Comment


        • #5
          Hmm. All i know is my 3100 (not the same code as yours of course) won't log right in TP RT. It will play in real time fine, but if you log, it gets messed up and the playback doesn't work as there is chunks missing out of it. The 3100 ECM that I have is the only known one to do that. However ScannerPro (beta) will record just fine on my ECM.

          Dunno, keep trying and checkig the aldlstuff files, that's what I did to expand the 3100 def they had and fix a few things.
          sigpic New 2010 project (click image)
          1994 3100 BERETTA. 200,000+ miles
          16.0 1/4 mile when stock. Now ???
          Original L82 Longblock
          with LA1, LX9, LX5 parts
          Manifold-back 2.5" SS Mandrel Exhaust. Hardware is SS too.

          Comment

          Working...
          X