Jump to content
Ultimate Subaru Message Board

Ok, so remind me again (ECU reflashing)


 Share

Recommended Posts

I know this was discussed before...there were several ideas but no solutions/real answers...

 

I need to be able to reprogram/reflash the stock ECU's on 87~89 turbo ECUs. Specifically my 87 RX.

 

What needs to happen?

 

The past threads i've read mentioned pulling the stock rom, yanking a chip off the ecu mainboard and soldering on a RAM-ROM chip....then dumping the stock ROM on it....then, you can start editing it.

 

 

Anyone have more details?

Link to comment
Share on other sites

MegaSquit would be easier and more tunable. Likely cheaper as well.

 

Editing the code of the stock ECU isn't something you will be able to do without extensive computer programming experience. What you dump off the rom is in pure assembly code and would have to be dissasembled and then properly understood before you go editing stuff.

 

GD

Link to comment
Share on other sites

Ummm...NO.

 

After playing with it for over a year, I'm extremely disappointed with it.

 

I am absolutely done with standalones unless there is no other solution. Right now, I have several solutions with using a stock ECU...if I have to, I'll go back to stock injectors/turbo and running 12psi.

Link to comment
Share on other sites

Nothing is 'wrong' with it....except that there were no engineers that spent hundreds of thousands of man hours with it to make it run the car PROPERLY....

 

All I know is that whats on the car is very inconsistent. One drive cycle it works fine, the next its trash. The MS as no way is seeing where the engine is in its cycle....it just guesses. Its too slow/unable to be programmed to use the stock CAS to it fullest extent like the stock ecu has. The stock ECU can see all 720 degrees of one engine cycle, in a 2 degree resolution. Thats dammed good. The MS only gets 4 pulses that signify the engine is running. It as no way to see if the engine is nearing TDC 1, or how much longer until TDC 1. No one will help me or guide me in making the MS read the fine resolution on the CAS wheel. if i could do that, then the MS would be CONSISTANT....THEN I could tune it to run the car properly. However, there is no support for it since I am the only one running it this way. Plus I've done everything in all the books, had professional tuners look at it, and got really creative outside the box...and...the accell enrichment is ************. There's no way around it, it refuses to work correctly.

 

Oh, did I mention the knock control sucks?

 

There is EDIS-4. I have everything I need. However, there's no clean way to install it. It just creates more clutter in the engine bay. But...its another bandaid and I'm sure it'll still have issues.

 

With the stock ECU, the car starts up every time with no protest, cold starts, hot starts, warm starts, upside down and sideways. The enleanment and enrichments work perfectly. Plus, the knock control is PHENOMENAL over what the MS has.

 

Either way, if I cant get a stock EA82T ECU to work...I'll just use a 2002 WRX ECU to run the car. At least I can reflash that one.

Link to comment
Share on other sites

Wasn't it Garner that used a modified crank pulley to run an EDIS CAS?

 

At any rate, you have the (well commented) code for MS, and you don't have the code for the stock ECU. Dissasembling the stock code would be as much/more work than just getting dirty with the MS code and writing your own routine to handle the EA82T's stock CAS.

 

....I might be persuaded to assist you with the MS code - I have some intrest in using MS on an SPFI EA81 in the future, which uses the same CAS.

 

GD

Link to comment
Share on other sites

I did fab up a crank pulley with the edis wheel and make a mount for the CAS....all thats left is wiring and redo the MS.

 

As for the code...I'd still rather use the stock ECU if I'm going to go thru the trouble of sorting code and doing all that...as the stock ecu already runs the engine perfectly.

Link to comment
Share on other sites

As for the code...I'd still rather use the stock ECU if I'm going to go thru the trouble of sorting code and doing all that...as the stock ecu already runs the engine perfectly.

 

I have no idea what experience you have with software engineering, but I have a lot and I can tell you that working with the MS code would be a LOT easier than trying to dissasemble the stock ECU's code. You don't have the original code so you would have to sort through thousands of lines of assembly code and spend probably hundreds of hours decoding it to get anything meaningful out of it. No thanks - I would much rather have a community of developers to assist me and well documented code/utilities at my disposal. Adding support to the MS code base for the EA82 CAS both helps you understand MS better, and helps the MS community as well.

 

GD

Link to comment
Share on other sites

I have no idea what experience you have with software engineering, but I have a lot and I can tell you that working with the MS code would be a LOT easier than trying to dissasemble the stock ECU's code. You don't have the original code so you would have to sort through thousands of lines of assembly code and spend probably hundreds of hours decoding it to get anything meaningful out of it. No thanks - I would much rather have a community of developers to assist me and well documented code/utilities at my disposal. Adding support to the MS code base for the EA82 CAS both helps you understand MS better, and helps the MS community as well.

 

GD

 

GD is right, period.... the project you have in mind in "just editing the stock ROM code" involves decompiling the entire program that is written onto the ROM, then interpreting that (keeping in mind that it was written by Japanese engineers at Hitachi) and THEN, you start fiddling with it, immediately violating the intergrity with which it runs the car so smoothly.

 

I assume that you've talked issues out at the megasquirt forums?? MS was INVENTED specifically to make a computer that you could do exactly what you are talking about.. fiddle with the source code, in an end-user friendly fashion. I understand the problems you have with its limitations, but I think you fail to fully grasp what you are talking about trying to do with the stock ECU...

 

You wouldn't QUITE have to take the code down to pure binary, but assembler is close enough to raw machine code that, unless you have EXTENSIVE auto-computer engineering experience, it may as well be 11001010010100101010101001010, yanno?

 

Aplogies if I am coming across as condescending.. and also apologies if you DO know what you are getting into by asking this.

Link to comment
Share on other sites

I've talked about issues before on there, and was told that I am wrong no matter what. However when I tried said things, it made it better.

 

I have also talked in person to other M'squirters and since they dont see the issue they automatically assume its the tuning.

 

With over a year with this and no real success, I'm done with it.

Link to comment
Share on other sites

possible to get a different standalone that could read the CAS signal properly? I've been working with a Link LEM G3 on my buddies LGT project this summer, and it's very well setup. with separate maps for cold start, etc. it can even control A/C.

 

might it be possible to set it up to read the EA82 CAS?

Link to comment
Share on other sites

If you already have the EDIS wheel fabbed up then why not give that a try?

 

I understand what you mean about MS - it's open source and that's rough on people that can't code. But there's enough of us that can - we should be able to figure it out if you are willing to work with us enough.

 

Perhaps locate the person that coded that section of MS (if you can) and ask them for assistance. There must be something slightly different about the EA82's Hitachi disty than theirs. I would bet that a run on the o-scope would tell you for sure and give a good idea of how to change the code to make it work with the stock CAS.

 

GD

Link to comment
Share on other sites

I can tell you that working with the MS code would be a LOT easier than trying to dissasemble the stock ECU's code.
as a flight software engineer for two NASA satellites i will add that GD is understating the truth, he's saying it rather nicely. i know people that have the equipment and capabilities to do this, but their time is worth bank. you'll be hard pressed to get the support you need to do this without massive amounts of money or very lucky in the friends department. have you tried contacting companies that do this? i talked to one in Australia that reprograms soob ECU's, and they would work with this older stuff for around a grand, but i don't remember what all they would do or what kind of flexibility they could give you, but it might be worth talking to someone like that. hope you figure something out....i'm tired of seeing "MS isn't cutting it" reports. i've always been tempted, but until they're reliable i'm not going that route.
Link to comment
Share on other sites

well having been the happy recipiant of a free megasquirt (MS 1 on a 2.2 board) I'm gonna try getting it working using the 360 slot wheel in the cas by covering one of the slits and setting it up as a '360-1 generic wheel with optical pickup on the camshaft' and driving wasted spark directly from ms we'll see if it'll work..if that doesn't work I might just go fuel only and put some dummy loads on the stock ecu so it thinks it is driving real injectors and drive ms off a shielded tach signal to control the actual injectors (450cc dsm injectors or 550cc rx-7, both of these are 'bosch style injector and should be very compatible with our stock setup simply requiring a length of rubber fuel injection hose and 8 efi hose clamps using stock retainers and seals to mount them) that still leaves us with the maf and all stock sensors..I have no idea how the stock ecu will react when the maf starts reading its max flow limits, ie fuel cut..is it just a fuel cut or is it a fuel cut/timing retard?

what are the communities thoughts?

Kaz

Link to comment
Share on other sites

It is my understanding (from the FSM) that the tdc signals and the 360 slits are read by two different optical picups, since it is my intention to block one of the slits in the 360 slit section of the wheel and actually ignore the 4 90 degree slits that the stock ecu uses to determine TDC for all four cylinders, I can set tdc1 arbitrarily (almost) depending on where I put the 'missing slit', how the distributor is cocked in it's mount and where I tell ms tdc1 is relative to the missing slit.

Am I wrong in my assumptions/data sources?

Kaz

Link to comment
Share on other sites

PMI,

 

Here's how I think I would do it.

 

Cover one of the slits, then MS will know where it is after one rotation of the disty. Then, in software, you can move TDC for each cylinder around, since you know where the covered slit is, and where that is in relation to the engine cycle.

 

Then, you wouldn't have to move the covered slit around to change your timing; just do it all in software. You can even adjust the disty itself to change the timing.

 

-Dave

Link to comment
Share on other sites

With the disty set to 20* BTDC, the MS can successfully move the timing from 5* to ~45*.

 

All you'd need to do is make the #1 reference on the 360 side of things 'wider' so the MS can see it....and make sure the disty is timed correctly to the engine. Then set the rest up.

Link to comment
Share on other sites

With the disty set to 20* BTDC, the MS can successfully move the timing from 5* to ~45*.

 

All you'd need to do is make the #1 reference on the 360 side of things 'wider' so the MS can see it....and make sure the disty is timed correctly to the engine. Then set the rest up.

then do ett

Link to comment
Share on other sites

What about Steveman09 and his E-Manage piggyback.

 

From what he said, he had great success with it.

 

The car idled cold, hot whatever using the stock ecu, knock retard worked and he used the emanage to extend his fuel and ignition maps.

 

Isnt this exactly what you are looking for?

 

 

I plan on using Emanage after i install my TF035

Link to comment
Share on other sites

something I've been thinking about is using a SAFC

 

Some people have tried and the biggest problem I know of is that they work at +5v and the MAF works at +10v, but that's no real problem, you could make a voltage divider to get the signal to the SAFC at +5v and after the SAFC did it's thing with it ramp it up again with an amp to the needed +10v to the ECU.

 

If this is done with the correct parts and filters you'll have no (noticeable) signal loss.

 

The stock ECU will still control idle, cold enrichment, knock control and all ignition timing. (as always with an SAFC)

Link to comment
Share on other sites

Please sign in to comment

You will be able to leave a comment after signing in



Sign In Now
 Share

×
×
  • Create New...