Return-to-idle tuning

General Tuning Questions And Discussions
User avatar
NSFW
Posts: 679
Joined: Fri Feb 02, 2018 3:13 pm

Return-to-idle tuning

Post by NSFW »

The relevant bits in my setup:

LS6, so 5.7L / 346ci.
LS2 throttle body
232/234 @ 112 cam
33 pound cluch (RPS BC2, steel flywheel, lightened a bit by Mamo Motorsports)
Long-tube headers
FAST 102 intake manifold

When the clutch was installed, I didn't have any driveability issues. But when the intake and cam and went in, the car would barely idle and would often stall when it tried to return to idle. On the bright side, I got some extra heel-toe practice at every other stop light. But mostly it just sucked.

There are a lot of threads on the net about idle tuning, but not many about return-to-idle tuning, so I figured I'd write up what I've learned so far. The numbers below worked for me, and while I'm sure they won't work for everyone, I think they're a reasonable place to start. Or if nothing else you can use them as a reference when picking your own starting points.

The Cliff's Notes: I tried using cracker and follower to fix it, and while it sucked a lot less, it still sucked. PID tuning helped a lot, and then adding cracker/follower solved the problem. And some hacky tricks made it bearable while I figured out the real fixes. So I'll start with that...

1) Set up some anti-stall hacks.

With these in place, a lot of conditions that would have led to the engine stalling will instead cause the RPM bounce up after dropping a couple hundred RPM below the target. Note that these are just a crutch - when the tune is dialed in, your engine should not actually rely on these. In other words, if your engine is still relying on these, you've got more work to do. But these will make the tuning process a lot less painful.

1a) Set the 0 RPM and 400 RPM columns of the idle spark tables to 33 degrees.
1b) Set the idle underspeed proportional table to 2 g/s at -200 RPM, ramping up to 6 g/s at -300, and 6 g/s at higher errors too.
1c) Set the target idle speed 100 or 200 RPM higher than your goal.
1d) Set the 400 RPM column of the throttle cracker table to 3 g/s.
1e) Enable the built-in stall saver. (It helps a lot, but the other bits described here help even more.)
1f) Enable "rolling idle."

I did a lot of tuning with my idle set at 850, even 900 RPM, and then reduced it 50 RPM at a time as the car got more reliable. It's at 750 now and I'm guessing it will be 700 when I'm completely finished. Being 300 RPM below target is no big deal if your target was 900, but it is on the edge of stalling if your target is 600.

For the stall saver table, I set most of the values be 50 RPM lower than the idle target, but the lowest values are 25 below, and the highest are 100 below.

Rolling idle tells the PCM to target a higher idle speed when vehicle speed is more than 2mph (disabled when it drops below 1mph). I set that at 1000, and added 2g/s of airflow for it. The PCM doesn't seem to use rolling idle reliably, so this isn't as useful as I had hoped, but it does help a little. It seems to be tied to throttle cracker, in that it worked much better when after I re-enabled cracker.

Also consider setting your idle to 1000 RPM when AC is on. That way if you screw something up while tuning, turning on AC might help things until you can reflash. I didn't think of this until I didn't need it anymore, but I think it might have been useful at times.

2) Get a stable idle.

2a) Set the 800 and 1200 RPM columns of the closed-throttle timing tables to 18 degrees.
2b) Set the idle timing correction table to +/- 9 degrees.
2c) Set up the idle airflow table (more below).
2d) Set the idle throttle area factor to match your throttle body.
2e) Set up the IAC steps / area table.

Generally you want your main spark tables to be tuned for best torque (without knocking of course), but that's not the case for the idle tables. For idle, you want best torque when the idle underspeed correction is maxed out - otherwise that underspeed correction can't do its job.

Eventually you'll want to set the desired idle air (aka base running airflow?) based on what the MAF sensor shows the car needing. However, using inflated values in this table may help if your idle really sucks. I had the "warm" values (80F and higher) set at 13 g/s for quite a while, but it's 10 g/s now and might be 9 soon.

Note that if you set the idle airflow values too high, you'll find that the car doesn't slow down much when you lift your foot off of the throttle, so be ready to use the brakes. And then turn it down, because if you get that cruise-control effect you've definitely gone too far.

The "right" way to set up the idle airflow table might be to log RPM and MAF while the engine warms up from a very cold start. I don't have quite that much patience, so I looked at this table as containing two pieces of information: how much the engine needs when it is "warm" (80F and above) and how steeply airflow ramps up at colder temperatures. The warm part is easy. For the cold part, I just adjusted the steepness of the curve based on how a cold start went each day (ambient temps were around 40F at the time). Now that everything else is getting dialed in, I'll set aside some time let the engine warm up from cold with the logger running. And I'll repeat that exercise next winter when temps get closer to 0F, just to get a few more cells set right.

The stock DBW throttle body for my LS6 had a multiplier of 0.02550 and many sources say that the LS2 throttle body is 0.01920. I don't know about any others.

I also don't know anything about IAC tuning because my only LS-powered car uses a DBW throttle. If anyone wants to suggest a link to a good write-up about this, I'll be happy to add one.

3) Turn off throttle cracker and throttle follower by filling the airflow tables with 0.

You'll probably want to add these in again later, but for PID tuning they will hide what's happening in the feedback loop. Get the cracker and follower out of your way so you can see what's going on.

4) Adjust the Proportional, Integral, and Derivative (PID) tables. 

If you still have the stock flywheel, you might be able to skip this step. I wouldn't know. But I am completely certain that this is necessary for a lighter flywheel / clutch / flexplate setup. I couldn't fix return-to-idle using the only cracker/follower features described in the next section. And it seems pretty intuitive that the large corrections that are needed to slow down a heavy flywheel are overkill for a light one. I have to credit Darth_V8r at LS1Tech for pointing that out - while it is obvious in hindsight, I didn't see it myself at first.

PID tuning is really a subject unto itself. If you're not familiar with these terms (pun fully intended) please see this thread:
viewtopic.php?f=28&t=6708

Some things to note about PID tuning for idle and return-to-idle:

4a) You want a 'deadband' in the Proportional table, so that the ECU will use spark corrections for small deviations from the target idle speed. That probably saves some wear-and-tear on the IAC/DBW components, but mostly it keeps the two feedback loops from fighting with each other.

4b) For the Proportional and Integral tables, use smaller values for overspeed than underspeed. If you're having return-to-idle issues, you want the throttle to close slowly when reducing engine RPM, but you want it to open quickly when idle drops more than about 150 RPM below target.

4c) You can use a lot of Derivative to help slow the return to idle. The values I ended up with are about 6x larger than stock. If you go too far with this, you'll notice RPM hesitating on the way down, especially if you lift your foot at high RPM. For example RPM might drop from 5000 to 2000, pause, then settle down to the target.

I mostly made changes to these tables by multiplying all values by 10%. However for the overspeed integral table I also flattened the table by using the same value for 100 RPM and greater. Again, if you're having return-to-idle issues you want the throttle (or IAC) to close slowly, not rapidly.

With some luck (and probably a heavy flywheel), you might be able to get return-to-idle working well with just PID tuning alone. Alas, my car was still hitting the stall hacks from time to time.

Note that there are some PID delays defined in some XDFs, but I think these only affect startup, not subsequent return-to-idle, so they probably don't matter.

5) Add throttle cracker and follower.

Both of these open the IAC (or "virtual IAC" in DBW cars) which causes RPM to drop more slowly, which in turn gives the PID loop more time give the RPM needle a nice comfortable landing at the resting place you've chosen for it.

The follower opens the IAC in proportion to throttle position, and the cracker opens it in proportion to vehicle speed and RPM. The follower may have some relevance to emissions, but I honestly have no idea what GM had in mind for the cracker. But I do know that we can (ab)use both of them to suit our needs.

5a) Fill the cracker and follower airflow tables with 2 g/s (except for the stall hack in the 400 RPM column).
5b) Zero the delays (Why do those even exist?)
5c) Set the decay rates (all of them) to 0.04 g/s.

This is actually where I'm at right now. My anti-stall hacks never kick in, but the return to idle is a bit slower than I'd like, so I'm going to increase the decay rates and reduce the airflow values to find a balance between quick return and flawless return. You might notice that in the stock tune, these tables have some shape to them, whereas I just filled them with the same value everywhere, and that's likely part of the problem. I might be working on these for a while.
Please don't PM me with technical questions - start a thread instead, and send me a link to it. That way I can answer in public, and help other people who have the same question. Thanks!
User avatar
NSFW
Posts: 679
Joined: Fri Feb 02, 2018 3:13 pm

Re: Return-to-idle tuning

Post by NSFW »

Feedback welcome.
Please don't PM me with technical questions - start a thread instead, and send me a link to it. That way I can answer in public, and help other people who have the same question. Thanks!
VK_3800
Posts: 526
Joined: Sun Jul 04, 2010 3:15 pm
cars: SS Torana
Location: NZ

Re: Return-to-idle tuning

Post by VK_3800 »

No useful feedback to offer but interested to see any discussion of the concept in general. I've been chasing similar problems with my Holden 355 with 12P now that it has a reasonable cam (242/252 duration). Previous cam was 228/228 and idled easily so I basically had all the IAC adders backed off to get quick return to idle (since added back in). Not going to be possible in the next few weeks but when I get a chance to do some more testing I will see what parts of your process I can apply in my case.

Its very similar to your description, trying to balance a slow return to idle (and 1200rpm with my current exhaust system is quite obnoxious, thinking about changing it) with stalling, mainly using throttle crack and follower settings. The problem seems to be that while it idles fine in closed loop this isn't consistent with IAC steps varying from 2 to 40+ depending on heat soak etc, and it always learns the lowest figure which causes a stall in a lot of situations.

I'm also struggling to wrap my head around why there are no additional IAC steps unrelated to throttle/transition or don't decay? Its quite easy to coast along to a stop with no throttle for long enough that all these steps have decayed, so even if it does work with the idle hanging high for a bit you have to drive a certain way to trigger it or you get a stall regardless. Something simple like a minimum IAC steps outside closed loop idle seems like it would basically solve the issue as it wouldn't require any sort of transition to make it work.
User avatar
NSFW
Posts: 679
Joined: Fri Feb 02, 2018 3:13 pm

Re: Return-to-idle tuning

Post by NSFW »

The whole time I was fighting this I was wishing for the sort of thing you just described. If the tune had a minimum IAC/throttle value (or minimum airflow) that it would use until RPM dropped to a point where the PID loop could take over, that would just solve the problem. Cracker/follower was the closest thing, but the fixed decay rate means that it doesn't quite align with what we need.

I need to go back and add something about rolling idle - that's another anti-stall hack. It sort of emulates this, but only if you're moving, and it isn't all that reliable. But I think it helped.
Please don't PM me with technical questions - start a thread instead, and send me a link to it. That way I can answer in public, and help other people who have the same question. Thanks!
Ken
Posts: 206
Joined: Tue Dec 17, 2013 1:05 am

Re: Return-to-idle tuning

Post by Ken »

Do the DFCO settings not have an influence on wanting to stall when returning back to idle from decellerating.

I added the dfco entries in the xdf of the last VY I played around with and found that some did have a positive influence to the outcome regarding stalling from decel
DFCO.PNG
DFCO.PNG (11.44 KiB) Viewed 4891 times
User avatar
NSFW
Posts: 679
Joined: Fri Feb 02, 2018 3:13 pm

Re: Return-to-idle tuning

Post by NSFW »

Good point. Turning off DFCO was one of the first things I tried, but it didn't change anything for me. I'm sure it could cause this though. If it doesn't resume fuel delivery quickly enough that's obviously going to be a problem. :)

So it's still something that anybody with a return-to-idle problem should try first. And it's pretty easy to shut off via minimum speed or something like that.
Please don't PM me with technical questions - start a thread instead, and send me a link to it. That way I can answer in public, and help other people who have the same question. Thanks!
VK_3800
Posts: 526
Joined: Sun Jul 04, 2010 3:15 pm
cars: SS Torana
Location: NZ

Re: Return-to-idle tuning

Post by VK_3800 »

I've had DFCO on and off during tuning too, no difference especially since it gets disabled below certain RPM/road speed anyway (I've lowered these from standard due to gearing without issue). I have changed the ignition advance settings here though too so it doesn't retard as far, helps with smoother transitions back to fuel and avoids popping out the exhaust, I guess that also eases the transition to idle advance numbers.

For me I can just watch the IAC steps on screen while driving, it gets decayed to single digits but ideally needs to be 30-ish when coming to a stop. When I reset the ECU it actually works OK for a short while but then the learned steps from hot idle override it and decays down too low again.
User avatar
delcowizzid
Posts: 5493
Joined: Sat Feb 28, 2009 8:38 pm
Location: Wellington NZ
Contact:

Re: Return-to-idle tuning

Post by delcowizzid »

For 12p and 11p a speed sensor makes all the difference to clutch in stalling
If Its Got Gas Or Ass Count Me In.if it cant be fixed with a hammer you have an electrical problem
VK_3800
Posts: 526
Joined: Sun Jul 04, 2010 3:15 pm
cars: SS Torana
Location: NZ

Re: Return-to-idle tuning

Post by VK_3800 »

Yeah speed sensor is all set up and has been working from the start. As soon as closed loop idle takes over its fine, just that getting there (even with no delay) can be tricky as the engine wants more air at times.

If I could increase the throttle stop I would but then it wouldn't be able to drop the idle low enough when hot. Actually that gives me a thought, some carbs used to have a fast idle solenoid that kicked it open a bit for A/C - I wonder if running something like that off a flex table might do the trick for the minimum airflow idea?
User avatar
vlad01
Posts: 7780
Joined: Mon Oct 08, 2012 6:41 pm
cars: VP I S
VP I executive
VP II executive
VP II executive #2
VR II executive
Location: Kyneton, Vic

Re: Return-to-idle tuning

Post by vlad01 »

Only times I had issues is when the throttle bade stop screw wasn't in enough as the min air flow was too little. It would always stall coming to a stop coz the iac couldn't react in time. Opened the blade a bit more and it was flawless.

I have seen people also drill holes in the blade for big cam engines where the blade open requirements were too much that it would trip the tps into an open reading. Having the hole allowed more min air without moving the tps min reading.
I'm the director of VSH (Vlad's Spec Holden), because HSV were doing it ass about.
Post Reply