Help with PID controls for IAC.

[Ben.]

Short version: I want my IAC to stay open with throttle application, then slowly taper back down to idle when lifting off.

Long version: Hi guys, I use HPtuners but have found this forum to generally be more helpful than the HPT forum, so I thought I'd join up and ask here!

I'm having some trouble around idle. It doesn't want to catch the idle after giving it a rev, I've got it pretty close, however it's still not 'right'. I've scanned a mates car for live data, 1 bar MAF enhanced, car is AUTO, unknown cam specs, however from listening to it at idle, it sounds as though it has more overlap than mine. His tune still retains closed loop fueling at idle, almost stock PID's for idle control (just slightly softened Integral values) and fairly aggressive under and overspeed timing. (9* each way). It runs and drives great.

My tune is 3bar SD, has fairly hacked up PID's, open loop fueling at idle, and idle spark tables cut right back to stop idle swings. It idles rock steady at 800, but if I give it a rev, on return to idle it drops below target idle speed and then picks back up. It also has a bit of oscillation when I'm slowly letting the clutch out with small throttle openings. I'm confident my Base Running Airflow is very close, I did the Russ K idle config a while back. My cam is a 227/223@115 for reference.

Now, in the logs of mine and my mates car, I notice my mates IAC position goes up with small throttle applications, and SLOWLY comes back down to idle nicely, mine does weird things with light throttle like going to 0, then spiking back up, then back to 0. With the tune I have now, if I give it a rev, the IAC goes to near 0, then spikes up to about 80 just as its returning to idle, before settling back into its normal idle range values (38ish steps).

What is causing our two IAC's to act so differently? I've tried all manner of PID configurations, from all stock, to slightly softening each individually, to softening all of them, tried increasing them etc, tried adding in throttle followers, tried running massive amounts of Base Running air flow to try and make it catch really high, then use the PIDs to slowly bring it back down but nothing seems to work. Anything I do results in it dropping to 400ish RPM, then springing back up. Any help would be greatly appreciated.

Cheers, Ben

[joecar]

Hi Ben,

Firstly, given 2 different cam specs, each engine (assuming same intake manifolds and same throttle bodies) flows a different amount of air (at idle and anywhere else).

Secondly, the idle and throttle parameters/tables determine when and how (damping factor) the return to idle occurs... this in turn determines how the IAC behaves (in order to bring engine speed down) and how much IAC and spark timing interact...


I would suggest going back to stock, getting each VE table correct, and then tuning for idle.


( also, make sure ignition timing has enough range to control idle )

[Ben.]

Thanks for the reply joe, I appreciate you putting in the time considering I can't provide you with a log or tune you can read. My VE table is dialed in and I have timing where I want it. I did this by tuning for best vacuum and smoothest idle the old fashioned way. I have a pretty sound understanding of basic PID principles, just not how the PCM implements them. I have tried going back to stock, it all goes pear shaped very quickly.

To simplify my question as much as possible: As it stands now, with throttle application, my IAC steps falls to 0, is there any way to keep this at around 40 (where I idle) or greater? Without adding in throttle follower, as this causes other issues.

[Ben.]

No one with anything to ad?

Is there anything apart from throttle follower I can do to keep MY IAC open? is there a minimum IAC value to set anywhere that maybe I'm missing?

[ScarabEpic22]

Hi Ben,
Not sure if these will be directly applicable, but I bet with a little work you can make them work with HPT. I don't deal that much with LS1B/P59 PCMs (more of a Gen IV guy that all have DBW) so cant help a ton with the IAC steps. RAFIG PDF Thread (https://forum.efilive.com/showthread.php?26435-RAFIG-PDF-Includes-all-of-SSpdmon-idle-info) Idle Tuning Helpful Info Inside (https://forum.efilive.com/showthread.php?4661-Idle-Tuning-Helpful-Info-Inside)

[joecar]

Hi Ben, sorry, I missed your last post.



Yes, I agree with Erik, read this thread: Idle-Tuning-Helpful-Info-Inside



(https://forum.efilive.com/showthread.php?4661-Idle-Tuning-Helpful-Info-Inside)

[joecar]

If your IAC shows zero counts, try to open the throttle blade a little using the stop screw (you will have to reset the TPS).

[Ben.]

Nah, it actually idles at about 40 counts, but drops to zero with throttle application, so if I'm taking off from the lights, I give it light throttle, the revs quickly come up to about 1500-1700, then the IAC closes and it drops down to about 1100 revs and the IAC intermitently spikes back up to 40 or so, then down to 0. It makes it really weird to drive. However, it only does this if I have some integral effect acting on the IAC. If I zero out all my integral tables, and just use the proportional, the IAC seems to stay at around 40 counts (where it idles) all the time, even with throttle application.

I've come to the conclusion now that the PCM is having trouble deciding if the car is in an idle state or not. With small amounts of throttle, it seems to reduce IAC and spark timing in an attempt to get it back down to 800rpm (idle).

Thanks for the input guys.

[joecar]

Hmmm, interesting... does it think it has too much air when this happens...?

[Ben.]

The light bulb above my head just came on after reading that joecar. This is quite embarassing for me at this point as I remember reading somewhere on here about dynair vs desired or actual or something similar (for the life of me, I can't find it now). Coupled to this that I only got my scanner working properly yesterday, I can now finally see that my 'dynamic air flow' is a fair way off desired airflow, and I suspect this is a big part of my problem.

How close should my desired vs dynamic air values be?

[joecar]

They should be quite close.

Idle tuning is one of the more difficult tasks.

Post new logs and scteenshots as you proceed.

[Ben.]

Made a bit more progress yet again. I absolutely raped, pillaged, stole from and set fire to my "IAC steps vs effective area" to bring my desired airflow into line with my dynamic airflow at idle. This did very little, but in my mind was a step in the right direction. By this I don't mean fudging the IAC map, I'm certain it was the wrong way to go about it, but bringing dynamic and desired into line has to be doing good things for me.

After that I started adding throttle follower back in, and now my IAC is doing exactly as I wanted it to, and it returns to idle at a reasonable rate now. It's not too slow but it doesn't plummet straight passed desired RPM either. At this point my Proportional, Integral and Derivative values are not too far from stock, as opposed to the hacked up mess they were before. This has also helped with another issue, which was hunting when engaging the clutch, as the PID's before were hacked up to catch the idle after a free rev, they seemed to struggle with finding the right rpm when under load.

Now the issue I'm facing is (and has been half the battle all along) oscillations just off idle. As the idle is slowly coming down, it gets to about 1200rpm and starts oscillating up and down, maybe 5 times before coming to a stable idle. It also does this when starting the car, basically under the same conditions, from the added airflow from cranking, it slowly dissipates, and as revs drop to around 1200rpm, it starts oscillating a few times before coming to a stop.

It used to do it before with light throttle, which it still does it now, just much less. The only way I can get rid of it is to drown it in fuel till it starts to bog down, which smooths it out. Tried all manner of fuel and timing, nothing seems to fix it. I don't have a wideband fitted to the car, but I know I'm pretty close, as I can use Bi-Directional controls to take and add fuel in most places and feel where it goes rich and lean. I have actually noticed in a couple of other tunes that the tuner has done something similar, throwing lots of fuel into one or two cells to richen it up (I remember exactly because it was also in the 1200rpm column). Going to try that tomorrow, laptop died today before I had a chance to have a crack.

[statesman]

Now the issue I'm facing is (and has been half the battle all along) oscillations just off idle. As the idle is slowly coming down, it gets to about 1200rpm and starts oscillating up and down, maybe 5 times before coming to a stable idle. It also does this when starting the car, basically under the same conditions, from the added airflow from cranking, it slowly dissipates, and as revs drop to around 1200rpm, it starts oscillating a few times before coming to a stop.

It used to do it before with light throttle, which it still does it now, just much less. The only way I can get rid of it is to drown it in fuel till it starts to bog down, which smooths it out. Tried all manner of fuel and timing, nothing seems to fix it. I don't have a wideband fitted to the car, but I know I'm pretty close, as I can use Bi-Directional controls to take and add fuel in most places and feel where it goes rich and lean. I have actually noticed in a couple of other tunes that the tuner has done something similar, throwing lots of fuel into one or two cells to richen it up (I remember exactly because it was also in the 1200rpm column). Going to try that tomorrow, laptop died today before I had a chance to have a crack.

Someone over on LS1Tech thought it was a good idea to do an open loop idle instead of learning how to tune closed loop idle properly... and now everyone's doing it. :doh2:

[joecar]

Made a bit more progress yet again. I absolutely raped, pillaged, stole from and set fire to my "IAC steps vs effective area" to bring my desired airflow into line with my dynamic airflow at idle. This did very little, but in my mind was a step in the right direction. By this I don't mean fudging the IAC map, I'm certain it was the wrong way to go about it, but bringing dynamic and desired into line has to be doing good things for me.

After that I started adding throttle follower back in, and now my IAC is doing exactly as I wanted it to, and it returns to idle at a reasonable rate now. It's not too slow but it doesn't plummet straight passed desired RPM either. At this point my Proportional, Integral and Derivative values are not too far from stock, as opposed to the hacked up mess they were before. This has also helped with another issue, which was hunting when engaging the clutch, as the PID's before were hacked up to catch the idle after a free rev, they seemed to struggle with finding the right rpm when under load.

Now the issue I'm facing is (and has been half the battle all along) oscillations just off idle. As the idle is slowly coming down, it gets to about 1200rpm and starts oscillating up and down, maybe 5 times before coming to a stable idle. It also does this when starting the car, basically under the same conditions, from the added airflow from cranking, it slowly dissipates, and as revs drop to around 1200rpm, it starts oscillating a few times before coming to a stop.

It used to do it before with light throttle, which it still does it now, just much less. The only way I can get rid of it is to drown it in fuel till it starts to bog down, which smooths it out. Tried all manner of fuel and timing, nothing seems to fix it. I don't have a wideband fitted to the car, but I know I'm pretty close, as I can use Bi-Directional controls to take and add fuel in most places and feel where it goes rich and lean. I have actually noticed in a couple of other tunes that the tuner has done something similar, throwing lots of fuel into one or two cells to richen it up (I remember exactly because it was also in the 1200rpm column). Going to try that tomorrow, laptop died today before I had a chance to have a crack.Ben,

Just playing with it is helpful as it gives you more insight to how it works :cheers: ... if only we had more time to spend :doh2:

Post your tune file so we can have a closer look (can you also post your stock tune file).

[Ben.]

Someone over on LS1Tech thought it was a good idea to do an open loop idle instead of learning how to tune closed loop idle properly... and now everyone's doing it. :doh2:

Agreed, I know closed loop can work with cams much bigger than mine, and it is indeed on my list of things to do, however I want to get it running better without all the corrections taking place, then ad them back in. If you can provide any helpful hints or tips for closed loop fueling at idle, I'll gladly listen. I was actually playing with it a little a while ago, by changing my switching points I could get a decent idle (once it settled) but if it was swinging back to its lean side it would stumble if you hit the throttle at that moment. So I'd decided that for now, I'd let it idle a bit rich in open loop, and use the factory switching points to maintain fuel economy when it's being driven. When I get back to it, my next approach will be to leave the switching points and play with the 'closed loop proportional idle' settings, and use this as you would any other Proportional and Integral terms, trying to keep the o2 voltage pretty centered rather than swinging each side of its target value. Anything else you can suggest before I head this way?


Ben,

Just playing with it is helpful as it gives you more insight to how it works :cheers: ... if only we had more time to spend :doh2:

Post your tune file so we can have a closer look (can you also post your stock tune file).

100% agree, I'm taking this as valuable learning. A lesson learned hard is one not soon forgotten. I bought this car already 'done', it just ran like crap when I got it, much worse than it is now. The guy that tuned it is very well respected in Australia so I was telling myself there must be something small in there somewhere, but as time has gone on, I'm pretty confident some of the main input data is skewed. I would love to licence a stock tune and start fresh, I'm sure this would be a better place to start, however I don't want to spend more money on licences and dyno time, I'll just try to salvage this one for now. All the high load stuff was done on the dyno, just told myself I'd sort out the idle and 'less critical' things in my own time.

I doing my tuning with 'the other brand' so I can't post those here. I'm just here because you guys seem to be (and have been) much more helpful than 'those guys' haha.

[joecar]

You can post .hpt files here for reference purposes, I don't have any problem with that (and we're forever learning more, especially idle).

[joecar]

I'm finding that to save time/effort in fixing already-tuned tunes (containing hundreds of changes), I find that we always end up going back to stock, inserting correct injector data, redoing VE, and then we find idle is fairly reasonable just needing minor tweaking (which also means we don't learn any hard-earned idle lessons...)... and the tune file now has just the changes it needs, no more and no less.

[Ben.]

That may just happen in the not too distant future, but I'm going to try and persevere with this for now. If I was keeping the car long term, that's most likely what I'd do, but I may be selling it in the near future and just want to get it driving nice, preferably without spending all the time and money on licences and dyno time. Although I know it won't be 'right', if it drives good, then I'm ok with that, even if the numbers in the tune look odd. Bandaid fixes all the way!

I've got a couple more things to try (hopefully later today) and will re-assess from there.

Here is my current working tune, it is a HPT file, I just changed the extension to .tun so that it would let me upload it, don't know if it will open in EFILive:
19697

[Ben.]

19698
This is where I got to today. It's catching it high enough, but goes into a funny oscillation while its at it. I've got more timing in it now, at idle regions and just above idle, this extra timing at around 1200rpm really helped it catch. I had taken a heap of timing out of this area a while ago to combat some pretty severe bucking on decel. Also added a bunch of fuel right at 1200rpm also to smooth it out a bit.

Kept the log pretty simple, just 3 separate free revving events. I know it's a different scanner to what you guys are used to, but it's pretty straight forward, I'm sure most of you can decipher the data there.

You can see the IAC is doing what I want it to, if anything a little slow but it's following the right pattern at least. Any hints on what I should log to track down that oscillation? Apart from the all too obvious wideband which I don't have in the car?

[GMPX]

Here is my current working tune, it is a HPT file, I just changed the extension to .tun so that it would let me upload it, don't know if it will open in EFILive:
Oddly enough you can't open a file created in HPTuners with EFILive :ermm:

[Ben.]

Funny how that is... I thought there was an outside chance that the tunes were a somewhat generic binary file that either software package could recognize.

[joecar]

That may just happen in the not too distant future, but I'm going to try and persevere with this for now. If I was keeping the car long term, that's most likely what I'd do, but I may be selling it in the near future and just want to get it driving nice, preferably without spending all the time and money on licences and dyno time. Although I know it won't be 'right', if it drives good, then I'm ok with that, even if the numbers in the tune look odd. Bandaid fixes all the way!

I've got a couple more things to try (hopefully later today) and will re-assess from there.

Here is my current working tune, it is a HPT file, I just changed the extension to .tun so that it would let me upload it, don't know if it will open in EFILive:
19697You can enclose it in a zip file.

[joecar]

19698
This is where I got to today. It's catching it high enough, but goes into a funny oscillation while its at it. I've got more timing in it now, at idle regions and just above idle, this extra timing at around 1200rpm really helped it catch. I had taken a heap of timing out of this area a while ago to combat some pretty severe bucking on decel. Also added a bunch of fuel right at 1200rpm also to smooth it out a bit.

Kept the log pretty simple, just 3 separate free revving events. I know it's a different scanner to what you guys are used to, but it's pretty straight forward, I'm sure most of you can decipher the data there.

You can see the IAC is doing what I want it to, if anything a little slow but it's following the right pattern at least. Any hints on what I should log to track down that oscillation? Apart from the all too obvious wideband which I don't have in the car?Yes, I can decipher it, I'm analyzing it.

[Ben.]

Bit the bullet and put a wideband on it. set my entire EQ ratio map to 1, and power enrich to 1.15 so that I can just target a global A:F of 12.8ish.

VE map is pretty good, made a couple of small touch ups, but nothing significant. One thing I have noticed is that it goes slightly lean as it's returning to idle, around 14.2:1, it's at this point that it shoots back up and gets itself in line before settling at idle with a 12.8A:F ratio. Normal idle is around 40-45kpa, but it was going lean with slightly more vacuum, down to 30-35ish kpa.

So as a test procedure, I used real time tuning to smash 20% more fuel in at low vacuum up to 1200rpm, and the thing was perfect! Gave it a few revs, stayed at 12.8 from any RPM down to 1100 where it would now catch, then slowly return down to idle (it's catching high due to a raised Base RAF and my slow returning IAC, as shown in the log above) Brilliant, I'll just add more fuel into it there and be done with it. Made the same changes to my VE, flashed the tune, no deal. Does the same thing as before. Went back into real time tuning to ad more in, same deal, doesn't matter what I do I can't get it to stay rich on return to idle, don't know why it worked the first time, but that's the problem at this point. Monitoring my injector pulsewidth it's normally around 1.3-1.4ms at idle, it now shows 1.5ms the whole time rpm is falling back to idle, even when it goes lean.

So, in short, what can I do to keep it a bit rich on return to idle?

[joecar]

If DFCO is enabled, disable it and try.

[joecar]

Does HPT do compare of current tune to previous tune, if so can you show some screenshots of what you did...

[Ben.]

it does, but I've probably gone through 30 odd tunes today, most of which I've now deleted.

I thought I'd go back to basics. I 0'd out my idle overspeed and underspeed tables, started the car and went about finding a good stable idle, problem is, I can't get it to idle steadily no matter what I do. I will use bi-directional controls to lock the IAC at about 30 steps, lock timing to 15 degrees (tried more and less also), but still no luck. I've taken away every possibly variable, fixed air delivery, fixed timing, and just slightly correcting fueling, but it just gets into an oscillation of 600-1200rpm. during this the wideband is pretty steady at 12.5-13, doesn't move around a great deal, however I think this might be due to placement.

I remember trying a similar thing a year or so ago when I first got HPT, with the same result. At the time I asked a fairly well respected tuner and his response was "yeah, that's just what they do", but I can't help but think I should be able to get it perfectly stable with the above scenario. So I'd concluded that my VE map is the issue, or my injector data was no good, tried referncing some other tunes for data but haven't had a good chance to try it yet.

From your experience with LS1 PCM's in particular, is there any reason it SHOULDN'T idle stable with the IAC and timing locked to reasonable numbers?

[joecar]

...

From your experience with LS1 PCM's in particular, is there any reason it SHOULDN'T idle stable with the IAC and timing locked to reasonable numbers?
Nope, no reason.

[Ben.]

So, decided I couldn't possibly be this bad at tuning. Started pointing at things in the engine bay and swearing at them, pulling it apart and finding something to blame. Sure enough, there was a few days of removing and unnecessarily refitting my intake 4 times. Long story short, injector flow was all over the shop, injector seals were cactus, intake manifold gaskets had seen better days and in the process of me displaying my masculine strength to those around me by severely over tightening my intake, cracked it in several places.

The vac leaks were so bad that I decided vac leaks couldn't be a problem in the first place because I physically couldn't physically spray enough carby cleaner around the intake to seal it. I specifically remember having one can in each hand and absolutely flooding up and down each side of the intake and it making no difference, so clearly vacuum leaks weren't the problem

Anywho, I've now got it 99% right after undoing all the things I did to get it to run the way it was before. Closed loop, 700rpm idle, and if I lock the IAC and timing I can bring it down and have a stable idle at 700rpm with nothing funny going on. The only issue I'm battling now is, still, an oscillation on return to idle. HOWEVER, it's only when running closed loop, if I set it to open loop, even with my VE map dialled in around 14.5:1, it runs perfectly inline with what my IAC is doing, catching high then slowly coming down.

Is there any way to get it to ignore O2 feedback on overun? I've tried stabbing a bit of extra fuel into my VE map on its return path to idle, which helped slightly, I'm worried about putting to much in, and having the ECM trying to pull to much fuel as it gets into my actual idle zone. Any thoughts?

[joecar]

good job :cheers:

reinforces that physical issues can't be tuned around, and that some issues are not always easily apparent.