wideband logging

[steveo]

haha yeah I figured as much, was just a funny thought

[steveo]

ok, so.. im thinking that's probably it, after poking around some more.

it's added/subtracted to a value stored at L0197 via a short subroutine, depending on a status byte. it stores the result in L0196 a bit later on.

L0197 seems to be calculated earlier on from a large subroutine concerning fueling calculations that i don't quite understand right now; but i betcha it's blm/int/corrcl control input?

why add/subtract, though? i don't get that. perhaps this is a matter of the base afr being stored in L0196, and scaled up/down a fixed amount according to BLM?

i assume L0196 is a prime fueling calc, seems a bunch of AFR values end up getting dumped there, and also notice that's the byte that's compared to 0xFF to branch for fuel cut

[RobertISaar]

don't suppose you want to stick the questioned values into one of the datastreams that come from the E-side to see what they're doing while the engine is running?

i don't remember if EE primarily communicates through the E or T side though, either way, they can both chatter directly over ALDL.

[steveo]

that's a great idea! i'd be up for that

i do have to do a flash in the next few days to fine tune some other stuff anyway, and i could log it for a bit.

can you help me figure how to patch it? that might even become permanent, if L0196 is what it think it is, it would be really useful to have in the datastream all the time.

i think it's all T-side for datastream from what i've seen, this all seems to be done on the e-side. the comms between the boards part still trips me up a lot.

oil temp is a good one to override on EE i think, since f-bodies spit it out but dont really need it...

for now, i have to sleep

[steveo]

sorry for all the noob questions here, i think i'm getting it.

over coffee i managed to figure out some of the park/neutral and idle table selection routines,

i had a goal a while ago for my first actual code patch, trying to get p/n switch behavior in bins without the automatic trans bit set, just to enable selection between the two idle target tables. in my case, i could have 'show off the lope' as well as 'sleeper mode and/or wife is in the car' idle configurations.

it looks like there's only a single branch that checks the transmission type in that entire routine, and it seems to skip a small section of further branching if it's a manual trans:

557E	F6 20 26        	ldaB	b_2026_b1_6spd_b6_auto
5581	C5 40           	bitB	#%01000000
5583	27 0C           	beq	@87

reading the 68HC11 specs, i'm guessing i could just patch that 0x27 BEQ with 0x21 BRN, and it'll never branch.

is this a good way to skip a branch in this platform without screwing around too much? seems that way, since it would preserve the jump address, and make it a one-byte patch, right?

[RobertISaar]

looks like the DS files only show communication with the F4 side, E-side appear to be E4, i'll have to work out how the streams are setup in it.... for some reason, it looks like a whole bunch of tiny streams, which may work out to your advantage.

[RobertISaar]

BRN seems like a good idea.

[steveo]

cheers thanks, i'll give it a go! i'll look at the datastream stuff too but i think its a bit over my head right now.

before i get too far into this, im going to go on a campaign for a spare ecm donation

i think my flash is probably about due to expire if i dont brick it myself from noobcode

if i dont find one, im going to have to break my ol' honda out of storage just so i can get to work if something goes south