Return-to-idle tuning for a 2002 Corvette

[Ken]

!000 idle speed was at the request of the owner.
The 1100 rolling idle was one of the attempts at stopping the stalling issue, which had no effect so I left it at that.

[NSFW]

My car idles at 750, and a rolling idle of 1000 helped a bit. It gives more room for error when things go wrong, because being 300 RPM below 1000 is no big deal, but 300 below 750 is almost stalled. So the closer the rolling idle is to the regular idle, the less it would help.

Not sure it's still useful in my case but I'm also not quite confident enough to remove it yet. The car never comes close to dying anymore but it does dip below the target a little more often than I'd like.

[Highlander]

How does it behave now, does the rpm hang significantly high or does it drop to a suitable level OK?

Yeah it drops to idle speed in expected decel time, doesn't hang or drop too quickly,even if turning when the clutch is pressed.
I'd played around with settings like "stall saver rpm' "throttle follower" "rolling desired idle speed"and "DFCO" with minimal difference.
In the "throttle crack airflow P/N"setting, originally there was no setting in the 3 kph line, was just zero, and the engine speed dropped to 750 rpm when pulling up, then recovered, then as soon as I added that one extra setting , all idle speed dropping problems were solved.
Before playing with any of the cracker settings, it stalled often, even if changing gears too slow at a slow rpm, or not sliding the clutch enough on take off.
it was only when I opened the edited binary with efilive to double check that the XDF entries I added actually worked that I noticed what the extra info stated in efilive regarding throttle follower, that I found that throttle cracker was also meant to be altered / functioning.

Text at bottom left of this screen below.

Ls1_Idle_Fix_03.PNG

For rolling idle to work, throttle cracker must be active. If TC or TF are active, none of your idle functions will work. You can make the car witch to the idle timing table and put say 8 degrees in the 1200 rpm row and 34 on the 800 and you will have a pseudo idle routine for timing that will be controlled by the ecm's interpolation routine.

[NSFW]

Got a new engine, and had to start over from scratch on idle tuning. I managed to get it driveable but still had some surging while coasting to a stop, or if I fixed that I ended up with some stalling instead. Went to a new tuner who made it surge less without stalling, but it was still annoying to have RPM bouncing around so much when coasting to a stop, and while stalling was very rare, it wasn't 100% gone.

Last weekend I got it good enough that I can finally give me wife the keys without feeling an urge to explain why it's always doing weird shit.

Some key things:

* Zero throttle cracker airflow, and/or set cracker to engage at 255mph / disengage at 253mph (so it never engages). As long as cracker is active, the adaptive idle stuff isn't, and that's what causes the surging and the stalling. It took me a while to realize that when the throttle blade is at a fixed angle (because adaptive idle is NOT happening), RPM bounces up and down, due to a combination of big cam and light flywheel. Disabling cracker enables adaptive idle to kick in at any vehicle speed. (For a looong time, I though that the surging meant that the values in my PID airflow tables were too high.)

* Increase the idle hysteresis to 1000 RPM. Apparently, (I'm not 100% sure...) this causes the adaptive idle to kick in as soon as I press the clutch pedal, at 1800 RPM or lower (800 RPM target, plus 1000 RPM hysteresis), rather than waiting until RPM falls down to the target RPM and then jumping into action (at which point it often keeps falling well past the target RPM).

* Lots of throttle follower - 4 g/s at 23% throttle and above. Zero decay when in gear, and pretty fast decay when the clutch pedal is pressed. The idea is to give it enough extra air to slow the return to idle when I press the clutch pedal. I was hoping that the hysteresis change alone would enable the adaptive idle to prevent stalling/stumbling on its own, but it does still need a little bit of help.

So anyway... now when I approach a stop light and push the clutch pedal, RPM drops to 800 and it's perfectly stable. Getting cracker out of the way helped enormously.

The only drawback is that if I push the clutch pedal, put it in neutral and release the clutch before the throttle follower airflow has decayed to zero, it stops decaying, and nonzero follower airflow prevents adaptive idle from taking over, but and it get some mild surging, usually around 1000 or 1100 RPM +/- 100 RPM. But that's rare and it's pretty minor compared the surging that I used to have. And pressing the clutch for a second causes follower to decay to zero and then it idles like it should.

It's vastly better than anything I was able to achieve when I still had throttle cracker enabled, and it only took few iterations to get the follower dialed in. I wish I'd started with this approach.

[antus]

Cool, thanks for shareing. Can you give some idea of what has changed that could have caused those problems to then need that approach? Larger TB? Different IAC?

[NSFW]

Return-to-idle trouble issues are fairly common with...
* aftermarket throttle body (doesn't meet the PCM's assumption of blade-angle vs aperture-size)
* aftermarket cam (runs rougher)
* aftermarket clutch (weighs less, slows down faster)

I have all three, so it it was really bad at first.

I should have mentioned that I also adjusted the table that tells the PCM how far open the throttle body is for a given throttle blade position, to get it to open a little further. There's an "ETC scalar" that has a similar effect, and pretty much everyone adjusts that, but I have a hunch that more people should play with the area/angle table too. The scalar is used to convert from "square millimeters of opening area" to "percentage of opening area" and then the area/angle table is used to convert from "percentage of opening area" to "blade angle."