$EE major timing anomaly

[steveo]

i've observed this before, my car is affected, and have recently been contacted by a few other tuners that see it.

i need help figuring out what's causing it

the problem seems to be up to +/- 10 degrees of very fast timing surge that occurs during low rpm decelleration or gentle coasting events.

this timing error which modulates so quickly that the log shows it as fast as o2 switching causes the car to buck and surge and act up.

here's an example event



here's the facts i know so far

- in my calibration it seems to start happening during decel at 0% tps consistently around 1500rpm
- variance is anywhere from +/-4 to 10 degrees even in areas where timing is flat to within a few degrees
- disabling DFCO does not affect the problem
- disabling timing adjustments per idle speed error does not affect the problem
- happens in open and closed loop, in any blm cell
- closed tps being flat with the lower map regions of the timing table does not affect the problem
- the timing goes way out of range from any known table

i can't seem to correlate it with any data points other than TPS and RPM or any tables.

my disassembly of the timing code is sketchy and i can't seem to identify anything that way.

bin and csv of above log is attached. this bin runs particularly tame timing. dfco is enabled in this bin, but the effect persists if it's disabled.

im going to get a bin and log from the other person too for comparison

any ideas what im running into?

[dave w]

I wonder if using a different computer with the same flash (except from a different base .bin file) would shed any light on what is happening? I'm thinking an internal electronic component, like a timer or counter has degraded?

dave w

[kur4o]

This the perfect time to try my latest patch.It enables ALDL communication with eside modes 01, 02 and 03. That way we can monitor all spark parameters involved in spark calc and see where it comes from.
I ve already tested, so it`s safe to burn it`s not tested with engine running yet.
Load the patch and make log using this command
E4 57 01 06 xx xx=checksum.
See if its working on runnig engine and i wll give you descriptions on the log data.

Another think to mention is there is spark correction routine on manual transmission cars(on auto bins all parameters are ff). It works only below 1500 rpm, I call it launch control but it might be something else.
The result of this calculation is stored at word_220 on eside.

[Fast355]

This the perfect time to try my latest patch.It enables ALDL communication with eside modes 01, 02 and 03. That way we can monitor all spark parameters involved in spark calc and see where it comes from.
I ve already tested, so it`s safe to burn it`s not tested with engine running yet.
Load the patch and make log using this command
E4 57 01 06 xx xx=checksum.
See if its working on runnig engine and i wll give you descriptions on the log data.

Another think to mention is there is spark correction routine on manual transmission cars(on auto bins all parameters are ff). It works only below 1500 rpm, I call it launch control but it might be something else.
The result of this calculation is stored at word_220 on eside.

The $OD 4.3/Manual and 2.2/Manual setups have some weird spark smoothing routines that also show up in the later Black box OBD2 Vortec computers.

[steveo]

It works only below 1500 rpm, I call it launch control but it might be something else..

that sounds suspcious, that's right around the RPM that this problem occurs. i'll try and find the routine that you're talking about.

This the perfect time to try my latest patch.It enables ALDL communication with eside modes 01, 02 and 03.

i'll install it this weekend!

I wonder if using a different computer with the same flash (except from a different base .bin file) would shed any light on what is happening?

definitely possible. i do have a spare ecm, i could try that.

[steveo]

you know, funny thing is all the assembly i've done has been on an automatic '95 f-car bin. i never even thought to look at the code and unknown table differences.

[steveo]

i do like the odds of this one, kur4o!

right above the table are some constants used to define when the table is used, and those are all different on automatics vs manuals.

on f-body RPM25 matches range of 500-1600 (which is where this table is from) and within a threshold of (i think rpm variance over time?) this table is used..

as you said, zeroed out on automatics, but curiously enough contains different values on a y-body manual vette, which is typical with low rpm manual stuff since it's a lighter car with different gearing. vette uses less intensity and 1400rpm instead of 600rpm.

my first impression is that it's a timing based stall-saver kind of thing.

im going to copy the automatic tables and constants from this region and take a log.

[steveo]

early results are very promising. i can't reproduce the timing surge when zeroing the entire table and its activating min/max rpm constants. however i did notice (maybe it's just my butt dyno) at very low RPM that there was a transitional power loss right off idle that i don't remember before, so now im going to try to figure out how the table works, and how to tune it.

[kur4o]

I made a compilation of the values you can log with msg6
Also fixed some errors on the stream data.
New version of patch is here.
When patch is applied you can fill your own memory stream data with the messages applied 00 to 0d.
So here is the list in txt file.

[kur4o]

steveo

early results are very promising. i can't reproduce the timing surge when zeroing the entire table and its activating min/max rpm constants. however i did notice (maybe it's just my butt dyno) at very low RPM that there was a transitional power loss right off idle that i don't remember before, so now im going to try to figure out how the table works, and how to tune it.

This is what i have on file it might help you out


RESERVED:23DE byte_23DE: fcb 0 ; DATA XREF: sub_48E7+303r
RESERVED:23DE ; VS NVRATIO for LAUNCH control
RESERVED:23DF byte_23DF: fcb 0 ; DATA XREF: sub_48E7+309r
RESERVED:23DF ; VS NVRATIO vs LAUNCH CONTROl
RESERVED:23E0 byte_23E0: fcb 0 ; DATA XREF: sub_48E7+30Fr
RESERVED:23E0 ; VS NVRATIO FOR LAUNCH CONTROL
RESERVED:23E1 byte_23E1: fcb 0 ; DATA XREF: IRQ:loc_59B6r
RESERVED:23E1 ; IRQ:loc_59EDr ...
RESERVED:23E1 ; MIN CALC VALUE FROM LAUNCH CONTROL TABLE 1,2,3
RESERVED:23E1 ; If below use 0
RESERVED:23E2 byte_23E2: fcb 0 ; DATA XREF: IRQ+72Ar
RESERVED:23E2 ; TPS THRESHOLD FOR USING LAUNCH CONTROL TABLE 1 AND 2 VS TABLE 3
RESERVED:23E3 byte_23E3: fcb 0 ; DATA XREF: IRQ+7C5r
RESERVED:23E3 ; MIN TPS TO USE TABLE 3 launch control
RESERVED:23E4 byte_23E4: fcb 0 ; DATA XREF: IRQ+7DBr
RESERVED:23E4 ; VS DELTA TPS FOR LAUNCH CONTROL
RESERVED:23E5 word_23E5: fdb 0 ; DATA XREF: IRQ:loc_5954r
RESERVED:23E5 ; IRQ:loc_596Er ...
RESERVED:23E5 ; SOME COUNTER rel with 23e4
RESERVED:23E7 byte_23E7: fcb 0 ; DATA XREF: IRQ+6FCr
RESERVED:23E7 ; MAX NV RATIO FOR LAUNCH CONTROL
RESERVED:23E8 byte_23E8: fcb $FF ; DATA XREF: IRQ+6DDr
RESERVED:23E8 ; MIN COOLANT FOR LAUNCH CONTROL
RESERVED:23E9 byte_23E9: fcb 0 ; DATA XREF: IRQ+6EFr
RESERVED:23E9 ; VS RPM*25 FOR LAUNCH CONTROL
RESERVED:23EA byte_23EA: fcb 0 ; DATA XREF: IRQ+6F4r
RESERVED:23EA ; VS RPM*25 FOR LAUNCH CONTROL
RESERVED:23EB fcb 0 ; LAUNCH CONTROL TABLE 1
RESERVED:23EC fcb 0
RESERVED:23ED fcb 0
RESERVED:23EE fcb 0
RESERVED:23EF fcb 0
RESERVED:23F0 fcb 0
RESERVED:23F1 fcb 0
RESERVED:23F2 fcb 0
RESERVED:23F3 fcb 0
RESERVED:23F4 fcb 0 ; LAUNCH CONTROL TABLE 2 ??POSS IF AC CLUTCH ON VS TABLE 1??
RESERVED:23F5 fcb 0
RESERVED:23F6 fcb 0
RESERVED:23F7 fcb 0
RESERVED:23F8 fcb 0
RESERVED:23F9 fcb 0
RESERVED:23FA fcb 0
RESERVED:23FB fcb 0
RESERVED:23FC fcb 0
RESERVED:23FD fcb 0 ; LAUNCH CONTROL TABLE 3
RESERVED:23FE fcb 0
RESERVED:23FF fcb 0
RESERVED:2400 fcb 0
RESERVED:2401 fcb 0
RESERVED:2402 fcb 0
RESERVED:2403 fcb 0
RESERVED:2404 fcb 0
RESERVED:2405 fcb 0

[jthompson122183]

The shear intelligence in this thread amazes me. Excellent collaboration!

[steveo]

that confirms my findings this morning and more. i've reviewed the logs after zeroing it all out and confirmed that the problem is solved. so thank you very much for helping me solve that problem so quickly.

i think rather than launch control as you've called it, it's light throttle decel transition, at least that's how it behaves. perhaps in a stock configuration the cam in that operating range 'likes' that kind of crap and it smooths transition to idle, but on my car (and others) it seems to cause some serious surging.

zeroing the RPM @ 23E9 and 23EA seems to be the best way to kill it off.. but im going to try the corvette manual table too (seems to be about half the intensity)

[1project2many]

What does fuel do during this period? Any chance the long decel has cooled the cat and a burst of fuel plus advance is to re-ignite? (sorry, don't have tp on work machine)

[steveo]

oh and a good way to reproduce this problem, if you're affected, is just get rolling up a hill, and pop 'er in neutral while you watch the log. timing will go completely crazy.

[steveo]

What does fuel do during this period?

nothing, afaik... trims are solid, injector pulsewidth is stable, and o2s are switching if in CL.

is an interesting theory but doesn't explain why code is completely disabled in automatic bins