Adaptive Idle and EAE questions

[shadowplane676]

It has been a couple years since I worked on my MS last. I finally got all the components switched over to my new chassis and got the car on the road. Using my previous tune I was able to get the car running pretty well quite quickly. However, since I also upgraded to a much newer code variant, I was working on properly implementing EAE and the Adaptive idle setup. I have run into a few questions which would help me in dialing in these areas.

Setup:
1994 dodge shadow ES
2001 Stratus RT v6 (SOHC 24v, 3.0L)
MS-II v3.0 running 3.3.2m2 update
Please see attached .msq and log

1) Adaptive idle timing - I'm trying to get this set right, but I can't seem to get it to behave quite the way i'd like. I currently have to set the target rpm to 600 when warmed up for the car to idle around 900-1000 rpm. As far as I can find, I have the base idle screw nearly fully seated and the throttle adjust screw backed off fully, but can't seem to bring the idle down reliably. Should I be shooting for very low advance at idle or some mid-range number which can be advanced or retarded based on the idle target delta? I'm not sure exactly why sometimes it idles around 900 and other times it is pushing 1700rpm idle.

2) EAE - I did a bunch of searching, but haven't quite had the lightbulb kick on for which areas to work on to improve low RPM mid to upper delta-throttle areas. Most throttle transitions above 2000rpm seem to be OK, but if you nail it off the line, it bogs hard, then climbs out of it's hole. I think I need to look an either the adhered or sucked from walls constant curve, but I'm not sure. Are there any better pointers or workflows for tuning EAE?

[Matt Cramer]

You either need less timing or less air to reduce the idle speed. First make sure your timing actually is correct, using a timing light. And make sure the idle valve is going through its entire range and does not stick. You can try increasing the time step size.

First thing about EAE - it needs an absolutely spot-on VE table. You'll want to get the table nailed down before setting EAE on.

[shadowplane676]

Matt,

Thanks for responding. Thinking through the idle issue, I believe the issue lies with additional air rather than timing. I'll re-verify timing, but the toothed wheel to sensor relationship hasn't changed, nor the settings in the MS, so I'm pretty confident that the timing accuracy is good. In monitoring the idle this morning, I observed the idle timing going as low as 8* in an attempt to tame the idle (it was holding in the 1000-1100 range). On the stepper motor, I am sure that it doesn't have more than 255 steps open to closed. I don't know exactly how many steps it is, but 255 gives me just a few "ratcheting" clicks on opening. As I mentioned in my initial post, I have the BISS almost fully seated, which leaves something with the throttle itself. I did make up a throttle cable bracket recently, which might be holding the throttle open just enough to cause the idle to go bonkers. I will check that again to ensure the throttle fully closes.

On EAE, I believe I have a fairly hashed out VE table. It isn't perfect, but I had a darn good starting point when I switched chassis as the motor had been running well for a couple years prior. I did have to lean it out as I believe the fuel pump (walbro 255) is overrunning the 3000GT-VR4 FPR I have installed. Lately, Autotune has made very little, if any changes in the 75% of the map that I drive in (lower right corner doesn't see many records). Did anything jump out at you in the tune that would cause an excessive rich condition on heavy throttle below 2000rpm?

[16vboost]

I'll only comment on the "EAE" problem which I'll now rename "Throttle response" problem.

The first thing I look at is the MAP lag factor. You have yours set to 50. Try setting MAP lag to 99, TPS and RPM lag to 95 or so. This will allow megasquirt to see the change in MAP when it happens and act accordingly.

Next I would make sure to tune the 70-100 kPa x 1000 RPM cells in your VE table. On a quick stab from idle, these are the cells that determine the base fueling just before the RPM climbs up.

Next tune the ignition in the same area of the ignition MAP. I've found that less timing here helps low end torque and should help the off-idle blip.

Then I'd mess with tps-based AE. This is ultimately where you'll get the throttle response you want off idle. (Only after fixing your lags and maps) I like as low a TPS dot threshold as you can get (without false triggers), a short 0.1 accel time, and a reasonable accel taper and end pulse width (something like .5 and 1). The idea is to give a quick shot to kick start EAE. You'll have to read up on the TPS WOT curve and tune accordingly. I like to tune AE so that 1. the response feels good and 2. AFR doesn't change more than +/- 1 initially.

Lastly I'd get into EAE tuning. I find that it helps more for slow transients and on throttle lifts. The first thing I'd change is the sucked from walls curve. Contrary to the default settings (and the documentation) I think it should go from top left to lower right. In other words, lower MAP, more fuel gets sucked, higher MAP, less fuel gets sucked. I have very good results with EAE tuned like this. The rest is by the book.

Good luck,

-Alex

[muythaibxr]

The default settings for EAE were derived from the toyota patent. The idea being that with higher pressure there is more air flow and with higher air flow more aire comes off the walls. However, use whatever works for your engine.

[16vboost]

Thanks for that explanation! I've always wondered where that curve came from. Is it related to the older x-tau compensation? I figured at lower static pressures fuel is more likely to evaporate such as during an over-run condition. I use the RPM-based compensation to account for higher air velocity blowing more of the puddle away. Anyway I know lots of installs work well with the default type curves so I will second that "use whatever works for your engine". For me, flipping the curve solved some EAE oscillation issues at idle.

-Alex

[muythaibxr]

I think the Toyota patent might actually be older than the X-tau one. The basic math behind both is the same, but Toyota feeds in different data than X-tau (Ford).

Lower pressure does cause more evaporation but then so does higher airflow. The only time I would expect the higher pressure to cause less evaporation in the real world than the additional flow adds is during a quick blip. Especially on engines with air velocity problems like a quick blip off idle with ITBs. That brief moment after a blip before the manifold fills with air and the additional flow goes into the engine will be dominated by the decreased evaporation rate.

The toyota patent doesn't really outline things in terms of evaporation, it does things in terms of "added to walls" and "sucked from walls."

I think of it in terms of when a vacuum cleaner is used to clean up water.. if you put the vacuum head on the floor with no room for air to get in, the water may eventually evaporate, but if you actually want the water to get sucked up, you make sure some air is getting by. The effect of air flow is a lot more than the effect of pressure.

Ultimately I think the way your engine works will depend on a huge number of factors. Injection timing, manifold size relative to engine size, air velocity, etc.

I have always tuned EAE with some variation of the default curves at cruise RPMs in top gear with slow to moderate changes in load. Then I go back to idle and fix the RPM curves to get rid of oscillation. So the effective difference may be that on idle I am doing the same as you, but with the RPM curves instead of the main ones.

[shadowplane676]

Thanks for all the info on EAE. I will definitely review it when I get to tuning EAE. I am currently focusing on getting my adaptive idle control tuned as that has the largest bearing on the daily driving portion of the tune at this time. The BISS has been reduced to less than 1 turn open and the throttle stay has been backed down to not hold the throttle open at all so that the IAC stepper motor has more control over the idle air flow. This has fixed some of the random high RPM idle issues so far.

What sort of "general" curve type have people used for idle control? I haven't found any information that has clarified to me if I should have MORE or LESS timing in the "below target RPM" range or MORE or LESS timing in the "above target RPM" range. I currently have a upside down "V" curve with 25-28* timing at the peak (which is also right on 0 variance from the target RPM) and tapering down to the single digits at the base.

[muythaibxr]

I usually use more below the target and less above. I don't vary it that much either.

[16vboost]

The adaptive idle feature is like the icing on the cake. But you need to sort out the batter first. I'd just turn it off for now or set it mildly until you get the pid stuff sorted out.

-Alex

[shadowplane676]

The adaptive idle feature is like the icing on the cake. But you need to sort out the batter first. I'd just turn it off for now or set it mildly until you get the pid stuff sorted out.

-Alex

Not sure what you are trying to say here...

I am now trying to hunt down a setting in my tune that is affecting the idle stepper position during cold start/warmup. (see log, starting about record 2757). Adaptive idle is maintaining idle pretty well, but in cold start/warmup something is overriding the stepper position for a couple seconds until the adaptive algorithm kicks in. I can't seem to figure out what setting is adding steps to the IAC.

Even zipped, the log is over the size threshold, so here is a dropbox link that should work as well https://www.dropbox.com/sh/coed3y9n2p0k ... ZHova?dl=0

[16vboost]

What I meant by the cake analogy was that the adaptive idle advance is capable of fine-tuning the idle once you have everything else set up. The idle valve is primarily responsible for setting the idle speed. Please try turning off adaptive idle advance until you get your idle air sorted out.

It seems that at times, your idle valve is at 150 steps for a 1000 rpm idle and other times it is at 0 steps for the same idle speed. Is there any air getting into the manifold from any other source? Any air leaks? You should be able to go to the idle test mode once warmed up and command the idle valve to any position you want and have the engine repeatably follow in RPM. Maybe your idle valve is loosing steps. You may have to increase your time step size. Maybe the IAC Valve is bad. Confirm that it moves repeatably on the bench using the test mode.

-Alex

[shadowplane676]

16vBoost - Thanks for clarifying. I see what you mean with adaptive idle being more of a "fine tuning" aspect rather than a base-level idle setting (which was what I initially enabled it for as I was/still am having idle issues). I am going to try and see if I have any vacuum leaks around the manifold as I can't count how many times I have had it off and on the car. I did do a throttle swap between the last known running setup and now (60mm to 65mm). It is quite possible that the 65mm TB has leaks in it due to bad seals. I THINK I still have the 60mm TB in one of my boxes of car parts, and if I can find it, I'll swap it and see if that changes the idle characteristics. Luckily, the IAC is the same between TBs, so those settings shouldn't have to be changed.

I also found that in the intervening years since I last worked on the car, there is now a freeware ROM editor and stock ROMs from my donor car available. I was able to pull the IAC step table, target RPM table, and base timing map, which have been added to my tune. The corrected spark table seems to have helped a decent bit as my table was a bit....off to say the least (32* vs 38* stock at cruise for one).

[shadowplane676]

So I have gotten the idle mostly better. However I have a question on where the IAC gets its values from between being at 0 steps during driving and Closed loop kicking in. I have taken to using the "initial values" table to get a smoother idle catch, but when cold I still have an over-rpm condition. Does the MS still use the warmup IAC curve when closed loop control is enabled?

[nismoautoxr]

My advice on your idle quandry is that you go to open loop and get that working reliably. Stepper always on open loop then use idle advance and cl RPM target. That is probably as good as it gets for us stepper IAC guys. THe largest problem I had was that the step count and actual pintle position becomes implausible with all the varying conditions on a drive cycle . THings become quite unpredictable at that point . PWM duty valve guys dont have this problem. We do.

[muythaibxr]

So I have gotten the idle mostly better. However I have a question on where the IAC gets its values from between being at 0 steps during driving and Closed loop kicking in. I have taken to using the "initial values" table to get a smoother idle catch, but when cold I still have an over-rpm condition. Does the MS still use the warmup IAC curve when closed loop control is enabled?

The warmup IAC curve is not used in closed loop.

[16vboost]

When running closed loop, MS uses Closed Loop Idle Target RPM Curve for it's target RPM vs temperature. I have stepper idle working perfectly under all conditions on my car. I think the stepper pintle position is more accurate than with a PWM valve. You know that 75 steps open is always exactly 75 steps open.

-Alex

[shadowplane676]

I do still have a question on my understanding of where the MS is getting position information for the IAC. At the suggestions of others previously, I have deactivated CL control for now and am working on tuning my warmup IAC curve further. This seems to be controlling the idle in a vastly different way even though nothing physically has changed with the motor. I have had to open my Base idle set screw 1.5 turns, up my idle ignition advance from 15* to 25* and open the IAC from 40 steps to 120 steps at 200*. I feel I am missing some fundamental part of what CL control is doing to where these changes are necessary. My thought process/observations are below. Am I missing something? I added a CL indicator to my dashboard 2 weeks ago so I could try and correlate when changes in IAC position happened, whether it was due to CL control or not.

1) Engine start - Uses Cranking IAC position curve
2) Engine run/CL conditions NOT met - Where does this IAC position come from if the IAC warmup curve is not used?
3) Engine run/CL Met - IAC moves to control RPM to target value curve using IAC initial value table, adaptive ignition.
4) Driving - in CL, IAC goes to 0 steps when RPM/TPS goes up, in Open-loop IAC stays open to maintain idle and maintains CLT based steps. (CL = 40steps & 15* advance, Open loop = 120 steps & 20*-25* advance for the same 900-1000 RPM idle)
4) Stopped - RPMs falling towards idle. With CL enabled, IAC opens BEFORE it reaches initial table value conditions. (where is THIS IAC value coming from? my IAC is opening around 2000 RPM and the initial value table only goes up to 1200rpm)
5) Stopped/CL Met - Idle closes in on target RPM.