Junk, I don't really like to speak for Bruce (Grumpy), but the issue that he is raising is something that kind of bothers the both us: Guys come to this board, we take the time to help them learn how to burn eproms and optimize their tuning. We get them through their "hard times" with the "learning curve". Then, once they've gotten the "handle on it", they leave and never help any of the other "newbies".

It's kind of a selfish attitude where they only stay with the board long enough to overcome their "initial problem". Then, once they've gotten the answer they are seeking, they don't bother posting anymore or helping others. Worst, some then turn around and start to purport themselves as "eprom burning experts" and start to sell eproms to "unsuspecting newbies" who still are under the misconception that you can "buy a good eprom" without any "hands on testing". Then, a couple of months later, we get these nasty messages on our Boards complaining how "so and so" ripped them off with a custom eprom they ordered from some "expert" from this Board.

Bruce, Tim and I have a "mission statemet". 1) To introduce as many people to eprom burning as possible so they can get the optimal results from their engine for max performance (and best gas mileage too) at the LEAST COST. 2) To develop a pool of expertise that can help others wanting to get into eprom burning. And 3) To have this pool work together to work together to share information and knowledge and possibly develop some interesting "tools" to help EVERYONE at the least cost possible. Bruce's (and his other "dwarfs") have done that on other things (most noteably the DIY WB O2 sensor).

I hope you understand why Bruce and myself feel kind of "burned" when someone comes to the board, "picks our brains" and then disappears once they've learnt what they want. Bruce, Tim and I (plus others do remain to help others after they have learnt) could all easily "work together" and probably setup a nice little business if we chose to (and make a good buck too). I get a half-dozen e-mails a week from people requesting eproms. I always refuse to burn eproms (because I don't have hands on access to their car, so I would be no better than all the others). But I do help them with their eprom burning if they get the equipment.

I beleve in the old adage from the Bible "give a man a fish, and he eats for a day. Teach him to fish, and he eats forever".

PS: I am going to delete this part of the post from mine, yours and Bruce's postings in a couple of days as I REALLY want to have an "untainted" post regarding Source Code that can eventually be used for a FAQ. You can post a response to clarify. But don't take it to heart when I delete/edit the portions which don't relate to "developing Source Code".

 

Overall, I have learned some stuff and have helped some others out.......I thought that is what this board is about.

That is what this board is about. Grumpy and others are concerned about those who learn from this and other boards but withhold knowledge when they have an answer for comercial and other reasons.

This is not my job, I do it for fun and enjoy sharing ideas with people that are interested. Correct me if I am wrong, but you seem very serious about this stuff. I am not....just for fun.

Most are serious about this being just for fun. If this becomes you job (for profit) then there are more than a few who will be upset.

If you choose not to participate then fine but remember that the information that allowed you to make the source code is available because someone put aside their ego, their fear, and their greed. No one person hacks a computer by himself without any help. Having a hack or a circuit drawing or even information on what processor is used will shave months off of your project. People gave that information out to further your hobby. Try to give back at least what you get.

John

 

Please don't get me wrong-I agree 110% with you about the folk who "purport themselves as eprom experts." However I'm sure that there are quite a few folk like myself who e-mail direct to a newbie thats asked a question and gotten no response- then I'll try to help them out as much as I can up to my limits. So no, you won't see all the help given on this board.
I can't speak for everyone, but personally I don't always post here and instead ask or answer direct via e-mail to avoid some sarcastic remarks........I have great respect for the knowledgable people here, but the "learning enviroment" is lacking at times; which can be especially intimidating for the newbies. Just my $0.02 and not meant as any sort of flame-more of constructive critizism.

 

Personally I don't give a **** what anybody does with the information on this board. I just hope that some of the smarter members don't leave, then it would be back to square one and we'de all have to spend 10x more time learning what source code can and can't do.
You want to grab the info, keep it to yourself, sell that info...I hope you don't make a profit, I hope it's a waste of your time and don't expect anybody to help you unless you sign an agreement that splits your "buisness." There is a reason people go to school but this is NOT a school, it's a hobbiest learning envirnment. If you want to do buisness good luck
Back to the source code.
Anybody have any papers or links to source code information? The assembly code, how to read it, what everything means.
I'd love to help but I need to know the path that you guys took to learn about source code. Is there a book, info on the internet? if so, where, what's the name?
It's really easy, through me the information and I'll do it so long as it doesn't require me to pay somebody $$$ .

 

Glenn,
I appreciate your words. As usual, they are well thought out and to the point.
Maybe it was misunderstood what my burning experience is. I am not a newbie to prom burning and I am not trying to steal anyone's ideas. I have thought about and tried to build a lot of the things that are talked about on this post. Quite frankly, a lot of the ideas floating around here are old ideas that are rehashed.
I wasn't trying to upset anyone......if I just wanted to grab info. I would read things and never reply with my thoughts. No one would know or care.
I agree Glenn, all of this non-topic stuff should get trashed. I don't know what else to say......I feel like a MAF.
J

 

Again, back on topic. I just thought I'd let everyone know that I just started programming my TI-86 calculator in assembly. I know this has nothing to do with the gm ecms but the ti-calc is the same idea, it uses a microprocessor which I'm learning about (on my own) and www.ti-calc.org has provided lots of information. I've even got absoft proFortranMP which lets me even look at what the *.inc files are doing. It's a LOT of information and my printer isn't working . I just started doing the assembly code programming today but I'm a quick learner. I'll try my hardest to understand everything but I really need some help for when I have questions. Last year if you had asked me to teach myself assembly language I would have laughed so hard, but I guess this is the future.
Some ideas I might play around with is using my graph-link to send aldl data stream information and displaying all that good stuff. BTW, the TI-calculators have both 8-bit "a" and 16-bit "af"
and I have absolutely NO CLUE what that means, lol.
Anybody know any good books about assembly source code?
My first real question: Does all the assembly code just go onto the eprom? Can you just use jump and label commands and place the code anywhere or does it have to be in a specific location?
Grumpy, how's that for doing things on my own and I'm only a month into programming in fortran yet all my instructors say I'm doing "too much." Another funny quote from my TA yesterday, "I don't like coming over to this table because you always have hard questions that are above and beyond this course." Good confidence boast, now I just need some more good reading material and somebody to help me hack the 8746 .

 

It sounds like you are taking a FORTRAN course and learning assembly at the same time. That's a good thing. Normally, you write your FORTRAN code and then compile it. A lot of things happen during the compilation process, but the end result is machine code that runs on the Uprocessor. Assembly code is an intermediate step. You can think of it like this: FORTRAN code ---> assembly ----> machine code. If you hack your ECM .bin file (machine code), you can generate assembly code. You can't turn the assembly into FORTRAN code though.
Adding code........for the most part all of the code is in the EPROM. This is the stuff you want to study. You can insert more code, but it must be inserted where there is no present code.
Cheers,
J

 

Some ideas I might play around with is using my graph-link to send aldl data stream information and displaying all that good stuff. BTW, the TI-calculators have both 8-bit "a" and 16-bit "af" and I have absolutely NO CLUE what that means, lol.

This may be to simple an answer but 8 bits make a byte and 16 bits make a word. #$00 - #$FF are bytes and addresses for example are words.

Anybody know any good books about assembly source code?

Search for source code. There were some threads that gave links to some basic explanations and to the hc11 pink book. Get the pink book and read it. It will answer most questions about how the processor works and what the op codes mean. If you have trouble with specific op codes we can discuss them here.

My first real question: Does all the assembly code just go onto the eprom? Can you just use jump and label commands and place the code anywhere or does it have to be in a specific location?

In C3 computers they also have a ROM portion. One thing you can do when you have all the source code is to eliminate the ROM. Most other GM computers are all on EPROM. The pink book will explain the other options.

somebody to help me hack the 8746 .

You already have the Hack that will give you information on what most of the RAM is. This will make your source code go a lot faster than without it. Use the Hack as a tool to understand what is going on then dig deeper into exactly what it is doing. I think Glenn is right about the do it yourself approach. Regardless of how many people help you will still have to read and understand what is happening for yourself. The only thing you might save is wear on your keyboard

HTH

John

 

Awesome information, thanks. Now if only I could find an assembly debuger for the source code . My calculator has a compiler that runs in dos and it tells you where there is a problem. What do you think I should remove from my 4k bin to make room for custom code? I'm scrapping EGR, scrapping the air diverter solinoid, and want room for a shift light and fan controls. I'm also getting rid of a/c this summer but how much room does the a/c code actually take up?
Any ideas would be appreciated.

 

Programming 101 is on the DIY EFI website. I found it worthwhile reading.

 

If your working on the 8746, or some C3, some of the code is on the printed Circuit Board ROM. So you have to read that also.
Ludis has some ROM stuff that's he's found posted on his site, and explains how to do it.

With a C3 to develope your own code, means eleiiminating code to make room for your own.

On the 148 I'm working on the code has been removed from the ROM, and incorporated into a larger eprom. So now instead of a 2732 I use a 27128, and about 1/4 of that is empty so I have tons of room to work in.

There is some PID for IAC for A/C on.
There is room to clear out with removing the EGR, and don't forget the CCP.

My original push on you guys doing this was ref the P4 ecms that use a memcal, those already have all the code on them.

 

I'm following this thread and keeping notes, never got into doing code other than some "ladder logic". I'm going to run a '730 and "learn" all I can, then jump back into turbocharging my toy using the '749. The more i've read the more I realized trying to turbo my setup from the start is too much at one time. Did get a copy of the 'pink' book from my boss today. Going to hit the bookstore this weekend for more info... I'll slip quietly back to my desk now...Bob

 

There was another thread recently on "Source Code" and programming the ECM where I gave links to ALL the reference material you need.

https://www.thirdgen.org/techbb2/sho...ht=source+code

 

Thanks for the link Glenn. Hehe "lite" reading.. Just got my AUJP Memcal from Flow chevy today and craigs adapters on its way. Need to finish off the Y-pipe, shorty headers, and 3 inch cat and go over the harness again, (its a reworked '165 harness),need to check a few things, and its time to fire.. Learning how to get beyond just the EPROM sounds like too much fun not to learn it. ohh the options...will be here, reading....Bob

 

The one word you didn't mention is cheap.
Lots of $2-5K ones that will.

Dual SCRAM, and a processor to synch things.

 

If you want to use the dual port ram architecture. Do not use the dual SCRAM (sync RAM), use dual port async RAM.

 

I may be guilty of lurking, but not on purpose. My concern is the spread of misinformation through ignorance. I have seen some rampant cases on this board. Being an engineer I have access to tools that have let me ramble my way around in the sorce code and program files, but I have yet to make a definitive yes or no on many things because of my trile and error methods. The best I have been able to do so far is correct and modify ECU files for winbin, so people who are using these files blindly will have a better chance at not screwing up there cars. I think we need to create a controlled thirdgen.org library so those who know what they are doing can upload realiable knowledge that beginners can use without fear of authenticity. I would be happyto help proof the uploads as well as include al of my own programming and files. Sorry if I got to off the subject, but this lurking thing got me thinking and I feel somewhat guilty. Thank you for all the help guys.

 

That's the point of my Final Answer postings.

The trouble with too easy to access info is that it kind of breeds the too lazy to do more then hit reply crowd.

If you were to look at the noise to info., ratio from when the list went from private to public, and then follow it's history to date, it's readily apparent what's going on, and some of it aint too good.

What we need are people to participate in developing new stuff.

There are just a few guys that really understand code, and their time would be best spent hand holding others thru that. The math and names are quote unquote known, and we need to go where no man has gone. Rehashing stuff is just rehashing. Folks will quick enough figure out what the real truths are as they actually spend time tuning, and find out how personal of thing tuning is, er can be.

Just my 2 cents

 

I see your point grumpy and that my be the price that has to be paid in order for use to expand the hobby.

My main problem with using or working on the gm p4 series ecms is the fact that it all boils down to two low side drivers for the injectors. To gain control of the system maybe it would be in our best interest to either figure out how to add two to six more drivers or move on to the LS1 based ECMs. The LS1 ECMs are fully sequential leading me to believe they contain more drivers for the injectors. I have not personally hacked one yet so I cannot confirm this.

 

I've run my car in SEFI and batch.
It runs best in batch.

Feature Creep doesn't mean better.
Need to start wtih the basics and build from there.

Most all the P4 use one driver, the reason for two leads is redundancy and trace loading in the ecm.

BTW, it's more then just slightly better in batch mode.

 

Once again I agree with you Grumpy> It is one driver I knew this why I put two,,, I guess my mind slips away when I am at work. One driver is grounded in the ecm witch drives all the injectors through a single PWM output.
As for sequential fuel injection. I do not see that as feature creep, but as new path that is being forged. I don't think we have readily available ECMs that are cheap enough and fast enough to do fully sequential, but there are plenty that can fire two injectors at a time while on cylinder is on its powerstroke and the other its dead stroke. This method has worked well for me when I tune electromotive fuel injection setups. My latest project, an 89 countach responded very well to this type of sequential injector firing.
I agree the batch works well. I myself am suprised at the result and thus that is why I continue to mess with my P4 ECM. I guess I will forge ahead and aquire an LS1 ECM and venture to adapt it to my GTA. This way I can say I tried it and give my honest opinion wether it is better or not.
I read your posts Grumpy and I see the history you have here with this website and others. I also see your vast experience in tuning these ECMs. I hope you are not to discouraged by this board and the people who visit. I would enjoy coming here once an a while and shoving around ideas with you,,,if your available and not to "Grumpy" about the idea.
If you don't mind me asking how did you accomplish you SEFI?

 

Well were all intitled to our opinions.
But the GNs had SEFI back in 87 with the 148 and that was a C3 ecm. Using the SEFI seems to have let them clean things up enough to get bu without an Air Pump.

The latest GMs are all suffering from the same error as the mini rams, but just not quiet as bad, so they have to use the SEFI to trim out the injectors for emission reasons (I might be wrong here, but I doubt it).

The GNs were SEFI, and defaulted to Batch when the cam senosr died, so I do about the same thing.
So rather then going from batch to sefi I went SEFI to batch.

Using Peak and Holds, I'm running 55#/hr Injectors, and it looks like I can still run in batch mode with 72s, which is what I want to do next.

With 55s I idle better then a stocker, and I can see no reason the 72 wouldn't either. In the case of a v8 the 72s would put you at well over 1,000 HP. If your running over a 1,000 HP then the SEFI might really be a help, or if you want to run a poorly designed manifold.

Manifolds are the key now, since the manufacturers are stuck in the 80s manifld technology, and GM hasn't gotten wind yet of what they need to do with the new ones yet.

Don't mean to sound arrogant, just they way I type.

 

Wow! That is quite a feat. I have not looked at the GN ECM in depth,but it is my guess that if you wanted to run 72 you solder in some silver fly wires that will carry the load better than the trace or traces as the case may be. The only problem I see in doing this is this may slightly change the output of the driver.
After studying some different fuel injection methods last night. I can see why the batch firing may give better results. To put it simply batch fire will mimic a well tuned carbuerator that can calibrate itself on the fly.
I look at it this way. Take for example you run two tpi ecms and one batch fires the odd cylinders and the other batch fires the even cylinders. Now you have effectively gained some headroom on your driver by reducing its load. Next you have increased the speed of your ecm response. Hopefully allowing you to bring timing and RPMs up and now you can better monitor your engine.
Your closed lood adjustments will be more resolute as it corrects each bank and not an overall reading.

 

The traces are fine, if you look at the TBI computers that have an even higher current load, they are the same.

Me thinks your over thinking things.

The advantage on batch over SEFI is firing the injectors twice as often, which get you your better atomization. Atomization is better then vaporization in so far as it displaces less O2 from the air.

The closed loop corrections are about meaningless is some reguards since the ecm just toggles over 14.7 to average that AFR.

Trouble with 2 ecms is that thay can't share sensors, they use pull up resistors, and doubling the load fubar's them.

Just getting the code to use a full time baro, and then use the MAT as a MAT would do wonders for the timing and fuel control.

Toss in a 2-3 bar code change and things would be really neat.

The problems with big injectors are all at idle. you simply run out of time. All the injectors start getting flakey at 1 msec operating times.

 

Why would 2 bar and 3 bar be an improvement unless you are using boost. I have used the GN map sensor in other applications as a two bar. Is it used on the GN ecm as a two bar. If there is room it would not be hard to add the code to p4 ecm.
1. Need to know absolutely where the code for the ve tables resides.
2. signals coming from the map sensor when it is under boost.

I realize this is simplifing it some. I have not had any official involvement with GM ECMs, but I have done Chrysler tuninig at Roush and I am currently doing ABLE controllers for CATs medium wheel loaders.

 

I think you will find the eprom address listing at on the diy efi site

www.diy-efi.org/gmecm/papers/prog_101.html

Davecs1 I think reading through some of the pages on that site will teach you a lot about GM ecu's if you have not had much involvement with them.
Guys using supercharges with high boost will need 3 bar map sensors. These are absolute pressures ie 1 bar (=14psi I think) is atmospheric pressure so a 3 bar will allow boost up to 28 psi which should be more than enough.

While I was looking at the diy efi site I noticed in the project section someone has developed a pic driven 4 *20 digit LCD screen showing the info from the cars data socket. The pic software is for a 1227747 ecu MAP and my car is a MAF 1227165. Just wandering while you experts are together if anyone has writen the pic software for my car?

Just want to say I'm learning a great deal from you guys keep going. I dont understand about source code but good luck trying to work it out.

I'm new to GM cars but I think I have worked out that Craig Moates is the scan software expert, ZaphodB does the Win Bin definition files and Grumpy knows more than the rest of us put together!

Cheers

Simon

 

"Spyware is ANY SOFTWARE which employs a user's Internet connection in the background (the so-called "backchannel") without their knowledge or explicit permission."

What you described grumpy almost sounds like trojans where someone has remote CONTROLL of a computer. Spyware is a simple (and arguably illegal) annoyance that advertisers usually use to report back where you browse, and what ads you click on so they can generate more "similar" advertisements thinking that it will be more effective.

But yes ad-Aware is a great spyware removale program.

As far as trojan removal most common virus scanning software also scans for trojans.

 

I have been tinkering with the DHC11 disassmbler for the Motorola processor and it has been great for telling me were space is on the chip and how some addresses are called, but it lacks the ability to breakdown the code into legible edit. We use the same processor in our EMS three monitoring unit for wheel loaders and have a codewright program we use to edit and produce software. GM's version for their P4 would make things much easier. With that I could go to town. My fellow collegues and car buddies think I should just use what I have to edit parameters and let it be. Their explanation being that I haven't met the limitations of the information I already have. And they call themselves car guys

 

Yeah, there is a lot of stuff you can do without hacking the code. I never found the need until one day I was bored so I checked it out. If you look into it you'll seem some interesting stuff. You'll see stuff and wonder why they did it that way. I think that is when someone gets into modifying it.
J

 

One thing that puzzles me is on the AUJP.bin they are always running some sort of enrichment or choke even when the engine reaches temp. I have been slowly dialing this back and adding more to the lower VE tables. To tell the truth I haven't been able to tell that much of a difference except that the BLMs are more stable and my throttle on performance is smoother, enough so that my wife asked if I had did somthing because the car seem to run smoother.

So why did GM use temp enrichment all the way up into the operating temps?

 

To make the code as versatile as possible. Often they used just chunks of code and dirived a calibration from that. Then the drivibility guys went out and drove the cars to see what they needed to make them drivibile. Then they EPA tested the cars, and they went back out to the field. Then depending on the repatching retested by the EPA. Remember most all the 3rd gen software is the better part of 15 years old now. They also test the cars in Alaska, and the Mojave desert. While you might not need the code in any reasonable situation, the code was tested to be reliable everywhere.

 

Hi,

Sounds like it make more sense to write a custom board to control/read from all the sensors/devices that utilizes a laptop, or even a palm pilot.

Writting a program to run everything should be fairly trival, if someone is good enough at hardware to build the unit..

You wouldn't even need to thread it, a non-blocking select() based proggy would work just fine..

Depending on how configurable you want the unit, there could be two basic approaches:

1) Pointers + structures to modify values on the fly.
2) A pre-defined "failsafe" mode, and a script-based config for
value definitions, and custom loops.


I've seen the programs (like tunercat) and, as good as it is to some people, I think its just soo limitted.

-- Joe

 

They are already available. Check out aftermarket ECMs or the EFI332 project. The code is not trivial. Well, not to me anyway.

J

 

Look around for the Megasquirt, they're going to be doing another batch, and for about $120 you can build your own pasic ecm, fuel only. But, that that's getting off topic.

 

Bruce,

1227165, 1227730, etc use 6801 CPU right??

-- Joe

 

Motorola 68HC11

 

Hey Glen,

Hrmm. I wonder what the HC processor is.. I look more into that.

I dissassembled the ARAP code i've been messing with in tunercat yesterday.. Reminds me of my 6502 days on the C64..

Who else is working in the source on this bin? I think a lot of you guys are using 730, and 749 these days.. hrmm.

-- Joe

 

Hey all!

I just found this thread, and you can count me in! I am a programmer, and have experience on pcs and some mainframe. I've worked with assembler a bit, and am ready to get started on my ecm source code. I will be trying to hack my 16197427 ecm due to the fact that my vehicle is running the 4L60E, among other things.
That said, in exchange for all this great information and assistance, I will be more than eager and willing to help anyone out in any way I can with similar endeavors. The things I want out of this are 1) a boost aware ecm, 2) a much better grasp of assembler, and 3) to help others along the path.

Now all I have to do is figure out where my instructions start. Much reading to do...

EDIT: Does my ecm even use the same 68HC11 processor that is being used elsewhere, or am I totally in a different ballpark here?

 

I"m going to reply to this topic now after reading it and going through some things. Now I don't know much about this whole thing I am just learning on it especially the source code. I would be glad to help on this anyway I can, expecially after I learn it. So far source code is fairly easy to understand as I Read through the various links given out. I can't wait to get to the meat of this and start actually getting in on this project. Be something good to leave as my legacy or something to that affect (Not quite old enough to be worrying about my legacy tho. lol) THanks and keep all the good thinking coming it's really really helpful!

 

Just as an update.

My GN is now running on *my* Source Code. My meaning I had alot of help from a friend of mine. It is now a 3 Bar MAP system, instead of MAF, and has all the little tuning tables, that I consider necessary. There were some hiccups, and alot of time went into tuning, but the results are 10x more then I expected. The car is just incredible..

I'm also working on doing the same thing code wise with another ecm. Again, have had help. But, doing alot more of it myself this time. Well at least so far.

While tuning is grand, it's a whole new level when you have the code that you want doing what you want when you want it.

ie I just use the IAC for the choke function and idle speed control. No Throttle follower.

No routines running in the background working against me.

Full Open loop, the code ain't even there. And using the WB have the engine running at all the AFR/VE/PRM levels that makes the engine happiest.

No mystery code looking for the HU running.
etc, etc, etc, etc,

A team of guys should be able to knock something like this out in just a couple months, and with enough players and organization maybe weeks. Dunno about ya'll but having a junkyard ecm that is some ways better then a $2K+ aftermarket ecms is just kinda exciting to me.

 

ok guys I'm new to all this stuff, but highly interested... would someone please inform me what the "source code" and all this other stuff is? It's all greek to me... If someone would take the time that is, I'd appreciate it, and be willing to work with people on developing it.

 

The ecm uses a .bin file to get some if not all the instructions on what to do. On the ecms that use a memcal, all the code is on the memcal.

Source code is in machine and human language so that you can read and understand what the code is actually doing.

A sample of source code lookes like this,

;==============================
; POWER UP OR RESET INITIALIZATION =
;==============================

INIT LDS #STKPTR ; INITIALIZE STACK POINTER
JSR ECUDELAY ; TURN OFF MALF LIGHT
LDX #$000A ; RAM IS OK (JSR CHECKS RAM)
STX ECUCR
LDX #ECU7

The ; seperates the human stuff from the processor stuff. So the ecm reads what's on the left hand side, and the human can read what's on the right hand side. Once you have the code so that it assembles, and is fully commented you can edit and change what the code actually does. ie like the oem ecm controls when the A/C is enabled, with your own patch you can say use that some output to turn on an extra injector, or cooling devise for an intercooler, water injection, or change the TCC code so that the TCC acts like a 5 th speed.

 

If you can get a disassembly of what ever ecm you want to work on it will have pointers in it to various spots in the Table Switch, and settings part of the code. Going thru and listing all the corresponding locations with names, tells you what is going on in each routine in the code, and makes life alot easier. Especially if you have some partcially commented hacs and can compare notes.

 

This stuff is good. I'm going to go back through this thread and look at the links that are here. I am a total newbie on the ECM tuning stuff, and I think I would rather go into it knowing it from the bottom up! I wanna learn the assembly code! From there, I have many things in mind. I'm thinking about a way to put aside the MAF vs. MAP debate aside, such as using both to check each other. This may not be a good idea, or even possible, but I need to get to a point that I can mess with it! I have a really good grasp on the concept of the way the ECM "thinks" as I do repair cars for a living. I have the theory in hand, I just need to learn more behind how it;s actually programmed to think.

For the time being, I may be seen lurking about this board, not saying, or even askiong much, but I do plan to get in to this stuff, and I will be very willing to share what I do.

Is there any place that all the stuff talked about in this thread is posted? I think it would be simpler to learn this stuff if people posted the stuff they learn on a website in a structured form. Yes, I'm looking for the easy way out of this, but in reality, anything to do with the source code level is not easy!

Go ahead and yell at me now!

 

I and others have done it, but if you want, feel free to do it.

Most others went the 165-730 swap, and I went with changing the code, and used EVERYTHING else intact. I in this case meaning with help.

Building an ecm bench first would be an excellent first step. Might be interesting to see what a stock 165 does on the bench. To date I don't know of anyone that's done that.

But, if you get involved in the MAF, MAP debate, be ready for lots of name calling, and misquotes. Even when you objectively point out the system differences, you'll get labeled.

The GM code runs in a bunch of routines, and loops, it can be very time consuming to figure out what's really going on.

Having Source Code is a good thing. Once you get to that stage the world is your's.

 

First of all thirdgen.org rules!!!
C3-ECM 91 305 TBI
Been reading alot on ecms and how they work. Can't get grumpy's tutorial on ecms, the link seems to be broke. If some one could email it to me jimbo@jimbo.kicks-***.org that would be cool.

Is there a flow chart on how the computer adjust's control the engine. Any extra info links would be awsome.

I have a number of PIC controllers floating around my house and been thinking of using one emulate the eprom. This little computer would be able to make on the fly changes to the ECM. It should be able to capture data from ecm's bus. But need to know the memory locations that the thing stores them. With some realtime anayzing of the computers memory locations, someone should be able to deduce what memory locations is storing what.

Now once the whole system is understood it should be possible to completly replace the ecm's processer and just use the peripheral chips mainly the I/O and analog convertor to make your own ecm.

Cost of this would be about $20-$30 for the processor and some wiring, a programmer for a pic chip runs about $50.

Now you will need at least 22 io lines, but 32 would give you room for growth. A fast 20mhz pic processor should be able to control your engine even if you write the code in basic. (man would that be cool! You could hook a computer to this unit via a serial port to gather much faster than 1600 buad output and once you adjust the pic processor the way you want it you could interface a lcd screen and small keypad to change engine running habits like "economy", "Horspower" ect...

I can make the thing, but need to learn exactly how the ecm functions and have not found much info searching everywhere. So any links would be appreciated.

Technical info
97 ($61) 1989-90 5.0 V8 TBI “E” L03
1991-92 5.0 V8 TBI “E” L03 (“F” body)
1989-90 5.7 V8 TBI “7” L05 1228746 2732 $D000 $D004 $D002:$D003 ANLU, AXKS

http://www.ppgia.pucpr.br/~afonso/do...s/nmc27c64.pdf

12 address lines
8 data lines (output)
2 input (chip enable)(output enable)
So your pic processor would need 22 io lines + at least 4k memory 8k -16k would allow backup of the ecm code if it crashes.

A processor that would be suited to this is a Pic Micro 16876 24 Input/Output lines
http://www.kronosrobotics.com/DataSheets/16f87x.pdf
or a better solution would be a 40 pin package 10 bit a/d converter.
http://www.microchip.com/download/li...xx2/39564b.pdf

 

Screw it. Im making a new thread. I seem to be missing one thing that i cant find. Im going to gather up some info on the bin file i want to disassemble then ask the question for the answer i need to know.