[lilsmokey]

I have been searching this site and the internet and read lots of things with some confusion. What I am wondering is if there is a way to use the quarterhorse to read the base strategy? Or how does someone go about disassembling this? I wish to do this for many reasons. I have many toys and I wish to be able to adapt this eec-iv to many of my toys. I know it is possible. I have nothing but time here to research and adapt this. I think It would be a great benifit to many.
also I am curious as to what can be eliminated in strats through disassembly and programming. As I am in school now for Computer programming and have many years of backround in racing and tuning with basic carb ignition setup. I just wondered If could delete things to manage the outcome of things to operate similar to a standalone? Any replies or comments would be greatly appreciated.

[decipha]

you would extract the bin and disassemble that, pretty tedious work

i would recommend against the older eec4's as the 8061 has less documentation then the newer ecu's

however, you wouldn't necessarily need to do any disassembly for what your trying to do, the more common ecu's are already broken out pretty well and have the ability to be adapted to just about anything

[lilsmokey]

well I have a couple eec right now I guess what I am really looking for is a manner to program remove or add functions I guess. the most Prevelant catch codes in pick a parts around here a the GSALI but I have a Cbaza and also and older speed density eec from a 87 Mustnag gt but cant remember the code. think it is an LUXO or something like that. I guess to be more specific I was interested in learning a direct way to eliminate things that effect other things like auto trans when using a manual and to be quite honest I would like to build a whole injection system from stuff I have around that can tune for a 57 chevy with a 383 that my dad was working on when killed in a car accident. I know It can be Done and I can do it. I just need some good Ideas of the how and where. Any help or Ideas would be awesome! Thanks in advance.

[decipha]

as for actually rewritting the code that the ecu uses, that will probably be pretty tedious

the only person that i know of thats done anything remotely near that extent would be derek fenwick, good luck trying to get any tidbits from him, i tried once but in his own words he 'isn't much of a teacher'

as for removing function thats as simple as nulling it, but adding... uhh can't help ya

i have no idea how involved you are with the newer ford ecu's but once you get involved in them deeply you'll realize just about anything you would want to do can be done, just may not be black and white, you may need to do things different from the norm but in the end you still achieve the same objective

[lilsmokey]

Thanks for that reply. Tedious I can handle. I do tedious day in and out. finding info on where to look and how to look at rewriting a program may be what I am looking for in the end result. But for now I would be happy with just learning as much as I can about them. I have been playing with this for a little over a year and seem to have really gotten not much further than where I have begun from. That is frustrating. Having put a whole car together from parts I have been told simply would not work was one feet. but getting it tuned and doing what I want it to has been another. For example. I tune the car It has no egr and nothing to do with emissions at all. My 87 that was built in about the same manner with the same gear and tranny and motor with a stop speed density computer got 30 mpg at 65 on hiway. Other day I drove 350 miles in my 95 with various tunes and although the BE said I was getting around 29-31 I was only getting 19-20 the hegos were jumping back and forth and the Lambse was between 15.2 and 16.4 and the timing steady at 40 so I just dont know what I have done wrong to throw the mileage off so far. the other things I can work on slowly trying to pick up here and there. but I have to say even in stock form without the quarter horse in my X2p runs better and gets better mileage than my Cbaza with a tune so am thoroughly confused. I guess with all that and my tune I was wondering if there was a way to delete all the automatic trans stuff in the cbaza j4j1 I have and would that make a difference to how it will perform the load percieved and the milage it gets?

[cgrey8]

If you want a challenge, you got one with just disassembling Ford EEC code so you can patch in whatever modifications you want. Until I "tried" to get into it, I was naive in thinking that a disassembler could be written to completely disassemble a BIN. But as I got into it and with help from others here, I saw why the disassemblers we have just can't do the job without help from a directive file that YOU have to develop yourself. The gist of why disassemblers tie their shoelaces together is that Ford embeds data where the disassemblers expect execution code to be. It was done to be more efficient. Instead of loading registers with the data the subroutine wants, the subroutines just expected the data it was looking for to come immediately after the "gosub" instruction. Once in the subroutine, the subroutine pull the return pointer the processor puts on the stack off, use it to access the data it is looking for, update the "real" return pointer, and put it back on the stack. Once back on the stack, the return pulls the modified return pointer off the stack and jumps to where it is supposed to. But the disassemblers don't realize this is going on because they don't actually execute the code the way the EEC processor does. So to create a real disassembler, what you really need is a processor simulator to truly execute the code, keep up with the register values as the code is manipulating them, and keep up with what's being put on and off the stack. With a real simulator-based disassembler, you should be able to get a clean disassembly without the headaches.

[lilsmokey]

So simulator based disassembler let me make sure I am up to speed here just so I make sure am following correctly you need either a program that can simulate where these directions are going and coming from as if the engine was running at all the different variations along the way and all the while the disassembler would read the code to follow it from where it begins to ends. I guess that is kinda the logic that I follow is that if it is pulling the thing apart so as it traces it piece by piece then it would function better as a complete disassembly of the code. Is that kinda what you you meant by simulator?

[cgrey8]

That's pretty much it. I keep saying I'm going to write such a thing. I just never get around to doing it. One thing that has happened since my last investigation into this is I had to write a C# tool for a project I did at work. I went from knowing nothing about C# or .NET to having 90% of what I needed written and working in 4 days. I was blown away by how easy .NET development has gotten. The hard part about this project would be programming in all the opcodes, the order of their operands, and associating them to their corresponding mnemonics so you can spit out disassembly. The rest is just data manipulation, state condition saving, and keeping up with which areas of the BIN has been walked & which areas have not so when the simulator hits a branch, it knows it's walked the pass thru condition and next time needs to perform the jump instead. It'd also be nice if the disassembler could accept notes about what certain offsets are as they are learned so it could update its disassembly code with descriptive info. For instance, associate certain offsets with ACT, ECT, etc. And of course the disassembler needs to know where the vector table is so it can handle disassembling each interrupt AS interrupt routines, not as background loop code.

[lilsmokey]

Allright now that is something that I can understand. although intitially when I started into this. I was just a gearhead with lots of computer backround this tuning and tweaking this way has brought a great cant put down stay up thinking interest. I agree that the programming part is getting easier as time goes by
and I also know that this is one of the major reasons I went back to school to add more programming knowledge to the stuff that I had studied and worked on on my own. As you were talking about a simulator through programming which I can understand initially when I started this deep holed project I have gotten into I found myself wanting to build a some what electronic and mechanical simulator for the eec to plug into to control and change the parameters that the eec was given. But I can see how it would be a good program based thing to write. with that said my next question would be can the quarterhorse be used as a means to do this? or would there be other things that would have to be attached to the eec to make it function?

[decipha]

haha now thats funny

if memory serves... chris had the same basic idea of turning the eec into a basic i/o device with a different OS

i would think you'd be better off just making your own ecu and going from there with it

[cgrey8]

What I envisioned was an FPGA that contained all the smarts of the control system and turned the EEC's processor into nothing but an I/O coprocessor. This would allow for really flexible control designs and customization. FPGA efforts at my office have moved slower than hoped. But I think we've finally made the decision to use them. However your point about just building your own EEC has merit. You can combine the power of a high speed FPGA with very cheap ARM-based A/D devices in quite affordable packages. The cost would NOT be in the hardware. In either case of redesigning an aftermarket EEC replacement or reusing the stock Ford EEC as a coprocessor, the cost and time would be in the software and firmware architecture development.

[decipha]

i think the hardest part of that is actually making the program the operating system, if you split it and tried to do it off an OS interface, i would think there would be too much lag time from when the ecu actually powered up til the time it computed an output, not to mention that wouldn't be very efficient

I've often though about it and even contemplated making one just for fun

[cougarnut281]

I guess I don't see the point in this exercise. Megasquirt is an open source efi system and from my understanding you could do what you want very easily. Much easier than trying to reverse an EEC. Or am I missing something? I admit I am not a programmer by any stretch of the imagination.

[cgrey8]

I don't know anything about MegaSquirt. But I didn't realize their software was open source. I also don't know anything about the hardware MegaSquirt is based on.

As for OS, no this would not be based on an OS. Or if it was, it'd be an RTOS which is very light weight and can launch and be up-n-running in a millisecond. Besides until the FPGA's processor is up and running, you could have the FPGA feed the EEC No-Op instructions just to keep it busy. The hardware logic in the FPGA could handle that without too much trouble.

[cougarnut281]

http://www.bgsoflex.com/megasquirt.html

This is mostly outdated info but it is the most in depth and concentrated source of it I know.

[lilsmokey]

sure the megasuirt is a nice Idea and I even looked into it. they sell a real nice kit that you can put together for around 300 plus dollars and you can easily download software to tune it. here are the draw backs that I saw 1) ford has sequential fuel injection. this adds to better performance all the way around. Mega squirt is a batch fire system. 2)I am not trying to slam megasquirt cause may be what I have to use on My dads 57 chevy project but. The simple fact that I can see it is a Glorified electronic carberator. limited spark tuning for the most part. and if you want you fuel map close you spend another 300$ to get a digital wideband so that you can properly adjust it. 3) I think decipha put it simply on one of his posts I read If I can tune this ford eec to work on anything I want and can go to the salvage yard and buy one for 10$ to 20$ than I am money ahead. For instance. Last night after doing some repair work I somehow managed to short a positive key on wire to ground which smoked my cbaza computer. so I put a Gsali I had laying here in it. cost 10$ now lets see someone smoke a stand alone like this purely by accident and not watch them cry.

[decipha]

sounds like me

i smoked an a9l and a j3 chip in a clients car once, pulled the j3 with the ignition on

swapped in a spare gsali ecu i had laying around made my pin/tune changes and off he went in no time


I also had a qh get pulled out a cbaza ecu in a clients car once, he dry hooked leaving a light and yanked the front tires, i slid back in my seat and hit my head on the roll cage somewhere in the process the usb cable for the qh got wrapped around my shoe and pulled out the ecu, it was hanging down not mounted up in the kick panel because i wasn't leaving the qh in there, surprisingly the ecu survived, it was funny though, when the qh pulled out she started puffing black smoke out the tailpipes, thank god nothing bad happened, since then i put duck tape on all my chips so they won't pull out, also aids in sealing the case as well at the j3 port

[cougarnut281]

1) ford has sequential fuel injection. this adds to better performance all the way around. Mega squirt is a batch fire system.

That has never really bothered me. The only benefits to sequential are cleaner emissions and better idle/low speed manners with huge injectors. I have yet to see a sequential system make more power or run any faster than batch fire.

2) limited spark tuning for the most part. and if you want you fuel map close you spend another 300$ to get a digital wideband so that you can properly adjust it.

You should be buying a wideband regardless of the engine management system so that is a moot point. What makes the spark tuning limited?

3) I think decipha put it simply on one of his posts I read If I can tune this ford eec to work on anything I want and can go to the salvage yard and buy one for 10$ to 20$ than I am money ahead. For instance. Last night after doing some repair work I somehow managed to short a positive key on wire to ground which smoked my cbaza computer. so I put a Gsali I had laying here in it. cost 10$ now lets see someone smoke a stand alone like this purely by accident and not watch them cry.

That is the whole point behind megasquirt as well, you can put it on anything. So what would stop you from getting replacement parts at radio shack to fix the megasquirt? That is the whole point behind it, you can fix, change, and modify the code and hardware.

[cgrey8]

With the exception of batch fire only mode, you are really starting to sell me on the MegaSquirt. If the code is open source, I wonder if it can be adapted to a J3-chip integraetd FGPA processor? Using their code as a starting point might be a fast-track way to get such a device up and running.

[lilsmokey]

they sell an adapter for to piggyback the megasquirt onto a j3 port. I looked into all these options when building my junkyard dog. car that I own. I hope no one is getting me wrong here I think that megasquirt has some major strong points in its use. I just believe that The developement of it to do everything that everyone wants it to do will take some time. for the most part I see the megasquirt very similar in use and way it works other than the batch fire issue to one ford has had since 87 or before. It uses a map sensor no mass air flow. the only other issues I have come across in the reading and research is how do you get it to run an IAC? I know that it is a very good piece of equipment. I just not sure I see a point in using it in a ford when millions have been spent to develope and update the eec in these vehichles. Yes I like to reverse engineer things so that I can repurpose them for as many things as possible. Here was my point about the difference in cost. It was stated you can go to a local radio shack and buy the parts to fix. The simple fact is no You can't I tried to build one. Sure they can order the parts to some degree but for the most part I found that you had to go to specialty shops usually online just to get the parts to build a Megasuirt. Now lets say you had a program that you could run do dissassemble and find all the areas you needed in any given ford computer code. would it not be as easy to program it to do everything save time of putting this computer together yourself and the hunting of the parts. to just spend 10-30 on a computer at the scrap yard?

[cgrey8]

Those are becoming very specific questions about how the MegaSquirt system works. Start those questions as their own threads so they can be tracked separately than what this thread originally started as.

[cougarnut281]

I hope no one is getting me wrong here I think that megasquirt has some major strong points in its use.

Not at all, everyone is entitled to there opinion, and I agree that in most cases switching from an eec to megasquirt is a step backwards. I just think that when starting from scratch on say a studabaker motor the megasquirt has a distinct advantage.

As for the radio shack issue, I did not know that and I think it sucks.

[bw5500]

....regarding Megasquirt, you need to look at the MS3.
It is significantly more advanced.
http://www.ms3efi.com/product.html

I have abandoned the dual eec-IV with dual edis-6 in favor of an AEM series 2 for the V12 project.
The AEM fires 12 sequential injectors, will push 6 twin coil packs, (or 12 paired pencil coils,) and control my 4L80E transmission.

However I do plan to run the A9L on my 302 and use the extra edis-8 that I have on it.

[teal95]

In the EECTuner group of Yahoo I have a directory (Assembler Tools) that has the tools I used in the past. I also have several disassemblies (either mine, sent to me or that I found on the internet) if you want to see something that someone else did.

steve

[decipha]

big STEVE what up man? I thought you lived in florida?

[bender460]

they sell an adapter for to piggyback the megasquirt onto a j3 port. I looked into all these options when building my junkyard dog. car that I own. I hope no one is getting me wrong here I think that megasquirt has some major strong points in its use. I just believe that The developement of it to do everything that everyone wants it to do will take some time. for the most part I see the megasquirt very similar in use and way it works other than the batch fire issue to one ford has had since 87 or before. It uses a map sensor no mass air flow. the only other issues I have come across in the reading and research is how do you get it to run an IAC? I know that it is a very good piece of equipment. I just not sure I see a point in using it in a ford when millions have been spent to develope and update the eec in these vehichles. Yes I like to reverse engineer things so that I can repurpose them for as many things as possible. Here was my point about the difference in cost. It was stated you can go to a local radio shack and buy the parts to fix. The simple fact is no You can't I tried to build one. Sure they can order the parts to some degree but for the most part I found that you had to go to specialty shops usually online just to get the parts to build a Megasuirt. Now lets say you had a program that you could run do dissassemble and find all the areas you needed in any given ford computer code. would it not be as easy to program it to do everything save time of putting this computer together yourself and the hunting of the parts. to just spend 10-30 on a computer at the scrap yard?

Piggyback the megasquirt onto a j3 port?? Please tell me more..