Printable View
Today I discovered that all of the available $8F XDFs are missing one very key compnent, Mask ID.
This is important to have access to when trying to emulate as this item when changed to "AA" will disable checksum and keep the ECM out of (partial) LHM.
Anyone have a full XDF, I'm 99% sure I used to have one, but can't seem to locate it right now.
Related question, is $8F Mask ID located at 0x08? Hex editor does show a value of "8F" at that location.
That's probably the mask id... hope it's in here.
I'll have to look through that, mask ID was not defined in the hack that is posted in the $8F thread.
So far I have discovered that when using $8F in a '7749, the injectors need to be paired to one injector driver, same pins as a '7730. If 3 injectors are on one driver and the other 3 injectors on the other driver, one driver is not being triggered and so causes the engine to be lean, since 3 cylinders are not actually running and just pumping air.
The F28 table (BPC vs EGR) seems to have no effect on fuel delivery at idle. The only way I was able to lean out my fuel delivery was to adjust the fuel injector size, but needed to be set much larger than what my injectors actually are to be close, and still didn't run smooth.
The manual/automatic flag is not correct, at least it doesn't change the trans logic on it's own. My shift light is currently a TCC. This would support the information that I've read for years that $8F natively can not run manual trans logic. It's only recently that I've even seen mention of this flag or that $8F is capable of manual trans logic.
Hmmm...
you MAY be able to run 8F with dual injector drivers, but i've never tried it, nor do i have a 7749 handy to test the code on.
if BPC vs EGR isn't effecting fueling but the injector size scalar is, it's because you're in alpha-n mode during idle.
the shift light code... i can probably look again, but the code appears to be there to support a manual trans.
i got fed up with the current 8F XDF and did my own hack and fixed quite a few errors along the way. i know there was a certain table that even had the wrong datatype in the columns/rows, i want to say it was boost related too.
if you use IDA, i'll post the database file i have, otherwise i can do a ASM export if you want to look at it.
IDA?
I've been finding huge differences in XDFs from all sources. A major difference being in the main VE table, so it's really hard to trust any of them.
Alpha-N on an SD code? :confused:
IDA Pro. kind of..... pricey, but i'm not saying there aren't alternatives to many choices in life.
and yes, the 8F application is pretty well misunderstood enough to where there are a lot of different interpretations of it out there, some more correct than others but even my work isn't 100%. especially when GM totally screwed the code in certain portions.
and yes, "idle fuel" is actually alpha-n. there is a flag to disable it in certain situations, but to remove it completely you have to change the ~8 enterance/exit scalars or patch the code to even jump past the possibility. all of the speed-density MPFI 60V6 masks actually have the capability and a lot of them use it as well. combine that with the fact that there is an idle VE table as well and you can quickly start adjusting the wrong values and run into an endless loop of "why isn't this working?".
oh, more fun facts: idle fuel also partially disabled closed loop operation.
it does allow for a smoother, more consistent idle by doing that if it's well dialed in, but other times it causes some odd things to happen, especially on a non-stock application.
i took a look and oddly enough, the space for shift light stuff was allocated and populated, but none of the code exists to support it. odd considering that 8F was built off of 6D, which has functional shift-light values. the only portion of it that is active in 8F are the 1st and 2nd gear lower n/v window values, and even then, they're only used for gear-specific EGR DC multiplication.
if you're looking to use the shift light for something a bit more exciting than how the factory did, that's fairly simple to patch in.
I have read that the shift light tables are not properly used, or implemented.
Right now it seems that I can't get proper manual settings, because my shift light acts like a TCC currently, or at least, not like a proper shift light anyway.
It seems this might not be a good choice for the car I had considered it for. I wanted to test it on my car to get used to it with an engine I know. :/
Might have to go back to $58 like I had tried before. The '7148 that it originally had ($31T) didn't seem to respond well to changes, and the 160 baud datalogging made it painful to see results to changes. I'm also tempted to keep the original ECM, and give tuning it another go.
i'm not a fan of 8F at all... not only is the code badly put together, the factory calibrations are total crap too.
it's not done yet, but i am in the process of "translating" AZRC(which IIRC, is the last 8F BIN released) into nAst1, but i need to get wastegate control done first. i also removed the ability to control the EVRV EGR early on, so that would need fixed as well.
How big is nAst1 now? Do we need to go to a full 512 Kb EEPROM yet? lol
The more I look at this ($8F) again, the more I remember why I didn't use it previously.
i don't like the way IDA produces ASM files, so i made it as a LST. easier to read and better labeling.
anyways, i did find quite a few values that are just completely unused, including the baro update stuff.
i've still kept it down to a 32KB BIN, but that is a serious limitation at the moment.Quote:
Originally Posted by Six_Shooter
Well, I think I'm going to forget about $8F for now, and concentrate on a couple other masks. I may even give nAst1 a go, once the boost control is figured out. ;)
i've been trying to steer people away from 8F, when possible, but when it comes to GM and boost with an OBD1 ECM... there just weren't that many options to begin with, and the ones that did exist seemed pretty much thrown together, with the exception of the later L67 code.... that is actually VERY well thought out, to the point of confusing.