Dies on Return to Idle

[SecondhandSnake]

As soon as you tip out, RPMs will drop to around 1100 RPM, hang for a split second, then it dies. It's pretty consistent, and it's always at that point. Air and fuel are dialed in. I was able to manually manipulate it down to hot idle with the throttle and it will happily live there, STFT switching within +/-1%. But if you give it some throttle and let out, it comes down, hangs, dies. It's a big pain with a manual when it wants to die coming to every stop. I've looked through the logs and there doesn't seem to be anything that precipitates such a sudden drop. It tries to recover with IAC duty cycle and fuel, but usually fails.

As far as settings I'm relatively close to the settings EFI dyno has on his site http://www.efidynotuning.com/dashpot.htm

I assume it's dashpot related since I can manually manipulate it to live at low idle and it's happy there. Best I can figure is to decrease the decay rate at lower IAC flow even further, and/or add more to the dashpot preposition to that region. Maybe adjust the Max Clip RPM if it looks like it's bumping against that limit in that region?

EDIT: I looked into idle air integrator. It was showing a pretty high value so I added to it. Loaded it up and tried again. Surged a little but still showed a positive integrator. When blipping the throttle it would come down slow, hit 1000RPM, bounce up to 1200 RPM, then come down and die. Added even more. Smooth idle, but still died in the same fashion after blipping the throttle. Integrator was near zero.

I'm totally at a loss on this one.

 

[Bullitt5566]

I've had the same problem for well over two years. I have Comp Cams XE270 w/10 degrees overlap. I've replaced just about everything and tried everything in the tune. I found that if I set the throttle stop by setting the base idle using the disconnect IAC/TP and set by the lowest stable idle method. The base idle is 850 and the lowest idle with IAC is around 900. This method screws up the dashpot settings and getting from driving to idle the car will hang and die. The same problem you described, hangs at 1000-1200 and drops to idle uncontrolled. In fact if you watch the dashpot, the idle will hang until the dashpot completely decays then drops uncontrolled. However if I set the throttle stop to stock (all the way closed) the idle is completely dependent on IAC, but the dashport settings and return to idle is much better, no hanging. But, the idle will drop too low and the engine will stall at times. Even if I set the idle at 900 rpm.

I keep thinking the problem is entirely the tune, cams, draw through MAF, returnless fuel system, but then I see cars with my exact setup and they don't have the idle problems I do. If it was tune related, I'd think even an amateur would be able to stumble across the correct settings after so much time.

So I'm back to a mechanical issue or the cams. I'm getting ready to pull the comp cams and put the stock back in, that seems to fix the issue or more correctly mask the issue, because they have a much larger margin for error.

Other ideas are crankshaft or camshaft sensors only because I've checked, replaced, or set everything else. In the tune the RPM seems to jump around. Maybe this is normal.

I think you'll find that the hanging idle when coming to a stop can be tuned using the minimum dashport when in low gears. If you lower this at the lowest setting the engine will be allowed to idle down lower when coming to a stop. I've seen some places say to set this to 0. This will absolutely cause the engine to die when coming to a stop.

Let me know what you find. I'd really like to get this problem fixed.

 

[SecondhandSnake]

Thanks for the help. I'm waiting for the weather to break to test my theory. I had some positive results in trying to dial in ISC neutral idle air. But it kept bugging me those values seemed way too high.

Then I checked my IAC transfer function- still the stock 00 2V values. D'oh. I'm going to try throwing the stock 03 4V values in it and see if that helps clear it up, at least as a starting point. It's definitely ISC neutral air related, at least in my case. That would explain why it would idle fine from the last tuner with the same configuration (sans MAF); he knew enough to update those tables, I didn't even think of it.

Regarding your case I think it's going to need a lot more ISC air and bump up the idle a bit. That seems to be standard practice with bigger cams (and to be expected, it's moving more air at idle). I've also seen people with big cams pull a lot of timing around idle to increase load and make it a little more stable, but I don't know if I buy into that theory.

It's a shame there isn't a better way to calibrate the IAC transfer function; you just have to log ISC integrator and get what you get. It would be nice if there was a way to do a slow steady ramp or sweep through idle values to calibrate it.

 

[Bullitt5566]

I've trimmed the ISC Integrator over and over. I wish it were that easy. They are dead on at 0 and it still doesn't return to idle.

As far as the ISC transfer function. How are you setting that up? What do you mean by calibrate? My understanding is that you use the IAC function that matches your setup. I have tried reducing the entire table by 1/3 thinking the function was off because the supercharger was pushing air into the intake at idle. It didn't make a difference. Well other than throwing the IAC integrator values off.

Can you setup livewire to log the actual IAC values in a chart like you would fuel pump voltage, then update the transfer IAC table from that? I may try that this afternoon.

 

[SecondhandSnake]

You might try purposely increasing your ISC neutral idle air above what it should be (integrator goes negative) for your target idle speed and below. That more than helped in my initial situation, and I know a few people that do a similar thing with spark timing in megasquirt, so when it starts getting just below your target idle it forces it up, possibly even before the derivative is big enough to cause the PCM to make adjustments. It's kind of a patch for the PCM not "catching" it fast enough.

My mention of calibrating wasn't for the transfer function but for the neutral idle air table. You could in theory do it but it would take some calculations. You'd need to establish base airflow through the throttle (0% IAC DC), and subtract that from your total airflow, indexing that to IAC DC. Of course load would also impact it, but it doesn't seem to be a big factor on the 03/04 4V calibrations.

What I was referring to is it would be nice if you could set the desired idle to slowly sweep from say 650 RPM up to 1200 RPM, log the integrator the whole time and you'd be golden. But I'm not aware of a way to force that functionality, at least with SCT.

 

[Bullitt5566]

I understand what you are saying. I'll try that with the IAC integrator. I'm already doing something similar with the spark like you said. I does seem to help.

I just don't understand that some setups seem to work fine and other don't. I have a friend with almost the same exact setup I have and he has no problems at all.

Just for testing I'm removing the SC and going N/A with the cams. Then I'm putting the stock cams back in with the SC. I'm also going to do a leakdown test to see if it is a mechanical issue. I rebuilt the engine so maybe the rings didn't seat.

 

[SecondhandSnake]

Latest update:
Ran it with the Cobra ISC tables. Didn't die all the time, saved itself a few times, but still died on occasion.

Then I moved on to cold start and man what a mess that was. Still had lots of surge. When it cold starts it surges around 800-1200, very rich (~10AFR), which is to be expected with the load reading 0.8-1.6. Meanwhile desired idle I've jacked up to 1600 RPM, and it's just not getting there. It smooths out once you open the throttle, but idles like crap there. It seems like it's just not getting enough air to be able to idle up where it wants. ISC DC is around 80% there.

What didn't help:
Smoothing out the MAF curve in that region. It's ugly, but apparently it is dialed in. I wondered if this was the cause because there's a pretty steep drop off below cold start high idle.

Flattening out the cold fuel table. With the load reading >1, it wants to run reeeeeeally rich.

What helped a little:
Adjusting the throttle stop screw out around a full turn did increase the idle (though surging) at least a little closer to desired.

Old revisions were able to idle up and then slowly come down, very well controlled. What was the difference in those tunes? I wondered. That high idle region was waaaaaay leaner in the old revisions than this one. This meant potentially an extra 1.5-2lb/min of air compared to fuel. So we're back to the theory of not enough air.

I know this region is dialed in with the most recent revision because the old ones would buck like mad cruising there, and this one holds the AFR right in check in that region.

So what are the next steps?
More throttle stop adjustment.

Increasing values in the ISC neutral idle air table. Before I changed to the correct 03 tables I did find dumping in a bunch of ISC air helped.

 

[Bullitt5566]

Honestly. I would start over from a known good point. I don't know about you but sometimes I get way off in the weeds and have to start over.

First, I'd put the IAC valve table back to stock. I think that adjusting the IAC integrator table compensates for any errors in the IAC valve table.
Second, I'd put the MAF transfer function back to stock in the idle region. I've tried to mess with it before, but with bigger cams it is almost impossible to get a good reading at idle and it is calibrated from factory and SCT provides those values.
Third, for the throttle stop. Set it like you hear else where. Disconnect the IAC and TP sensor. Se the screw to the lowest setting the engine will idle it. Turn the engine off, reconnect and start it. Datalog and dial in the IAC integrator table so the error is close to 0. Once you get a good idle update the air around the throttle body, remember you opened it TB more. Stock is .4. Datalog the MAF counts and IAC DC at idle whatever isn't going through the IAC is going around the TB. So if you are pulling 1.00 lb/min and IACDC is 30% then 60% is going around the TB. So .6.
Last, back to the dashpot. I think the hanging idle is cause by the minimum clip settings, try setting the lowest setting to 0 and see what happens.

 

[SecondhandSnake]

As is the ISC neutral idle air and transfer functions are stock - for an 03 Cobra, as it should be now.

MAF transfer function...that one isn't very feasible. I'm running an HPX-E and down low it doesn't match defaults very well. It's in a less than ideal spot, welded into the pipe, but I don't have any flexibility there. I don't think that's the issue.

I am tempted to reset the throttle stop. Looking at the stock tables, for a 750 RPM idle it should be commanding around 60% DC. That seems to match what I have, though everyone seems to set theirs at a much lower DC.

As is the integrator is very positive upon startup, between 0.15-0.40, ergo why I want to dump more air in via ISC neutral idle air. Right now going off the top of my head I think the air through the TB value is set at 0.525. Is the stock 03 value definitely 0.40?

Also for calculating flow through the TB- wouldn't you use the IAC transfer function to translate duty cycle to mass flow?

I think dashpot is functioning as intended. It's when it hands off from dashpot to idle control that it falls on its face, so it's all in the realm of idle air control.

I think we're both on the same page that the issue is calibrating the idle air control, both electronically and physically, i.e. ISC neutral idle air, air flow through throttle, and throttle stop.

EDIT:
Found a couple of points in a recent log with a steady idle. Doing the math to calculate air bypassed it's in the ballpark of 0.23lb/min, vs the calibration value of 0.525lb/min. That would explain the large ISC integrator values trying to compensate for that. Sounds like I need to increase the throttle stop and adjust the calibration to match.

 

[gt347mustang]

Anybody having this issue in Colorado, specifically Denver area? I'd like to have a shot at tuning your cars.

I've got a few different setups to idle well after local shops couldn't figure out what's going on. Idle and startup seem to be neglected in most tunes I see. I try to focus on thees since they cause the most headache and make the car unpleasant to drive.

 

[Bullitt5566]

Correct for calculating the airflow around the TB. Use MAF counts to find the airflow in lbs and whatever isn't going through the IAC is going around the TB.

As far as the IAC integrator. It should be 0 at all rpms at a warm steady idle. In Livewire setup a datalog table (YouTube will show how) and datalog the average IAC integrator at warm idle and at the RPMs in the SCT IAC integrator tables. If the ECU is commanding more air add that value to the IAC idle table, negative subtract.

I have noticed that if you set the TB throttle stop to stock the dashpot values line up much better with the airflow required. The dashpot airflow values are in addition to what you are commanding for idle. For example a dashpot setting of 300rpm and .1 lbs means that for an rpm of idle+300 rpm add .1 lbs to the IAC idle value. When you crank the throttle stop open you are adding a ton more air and the dashport values will be way off, typically it pushes RPM up., causes hanging idle, etc.

As far as the falling on its face on the transition from drive to idle. I suspect the minimum moving dashpot settings are commanding too much air, forcing the RPM too far up. So when the ECU transitions from drive to idle the rpm drop too fast. You can datalog the dashpot, I bet RPM hangs until the dashpot decays to 0 then drops. Go into the dashpot minimum clip in low gears and set the lowest setting (300 rpm I think) to 0 and see if this helps.

Adjusting this is a huge PIA. It is tricky getting a balance between too much dashpot and hanging idle and too little and stalling when stopping.

 

[SecondhandSnake]

Not sure if you saw the edit, but it does look like the problem is the discrepancy between actual airflow through the TB and the calibration value. (0.23lb/min actual vs. cal value of 0.525 lb/min) Can you confirm the stock value in the 03 Cobra tune, as far as what I should be targeting?

That would explain the positive IAC integrator I'm seeing in my logs.

I realize my dashpot is overly conservative right now, but that's as a bandaid until I get the idle sorted out.

I don't think it's minimum clip in low gears, or dashpot related. It does it the same stationary (VS = 0) as it does in motion. I think the dashpot is ok, but as soon as it's transitioning to idle control it's dropping off because of the aforementioned throttle air issue; it's expecting more air than it's getting.

It's a cold week so we'll see when I get around to it if the adjustments will do the trick. Unless I hear otherwise I'm going to target 0.4lb/min through the TB both physically and in the tune. Fingers crossed the IAC integrator will be next to zero with that.

I do appreciate the help.

 

[decipha]

it sounds like all of yall's issues can be fixed with tuning dashpot

first, as the write up says

"Dashpot relies on the idle airflow modeling to be accurate, thus any dashpot tuning must be done AFTER Idle Air is dialed in, which must be done after Idle Fuel is dialed in "

make sure your good there first and at idle the iscdty is not stupid high, anything 35% and less is fine, if its higher than that you need to open the throttle stop to give it more air to idle correctly, some cobras specifically the 96-98s will develop a surge below 45% iscdty at idle, I think its something to do with the intake harmonics that causes it but I've witnessed it on numerous occasions

anywho, make sure IPSIBR (idle airmass correction) is near 0 at a warm stable idle, if its not dial in your idle air

Now to do a quick dashpot adjustment, for simplicity, go set FN882A the maximum dashpot clip to all 2.3s for all rpms, that way dashpot isn't clipped by rpm

make sure the DASPTK scalar is maxed out to 0.062, that's the maximum dashpot clip based on relative throttle position

DASCTL is a huge scalar, it is the maximum amount of dashpot for idle feedback to be permitted, this should be set high enough that you have quite a bit of dashpot left for idle to stabilize before it completely expires, a value of 0.35 is a good value for this for 4v engines, make sure this isn't set too high though, anything in the 0.60 to 0.80 range can cause a hard start stall in some ecu strategies as it freaks out

MINMPH is the minimum mph scalar that idle feedback is allowed to run, on the 03-04 cobras they are very very very air sensitive down low and on decel, in a stock 03/04 cobra tune they are around 25mph, that's a good value for this scalar on those engines

now the biggest function for dashpot is the decay rate function, you want it to decay at a controlled and steady rate but not too fast, and not too slow, ideally too slow is less problematic than too fast, I often have the decay slow up top to stabilize the engine first, fast in the mid, and very slow down low, something like this for example

16.00 0.0100
1.500 0.0100
0.750 0.0400
0.500 0.0050
0.400 0.0020
0.350 0.0007
0.000 0.0007
0.000 0.0007
0.000 0.0007

for a perfect dashpot decay, you want it to hang about 300-500 rpm above idle for just a second then decay down, if the rpm's fall too low and then have to come back up that means you have it decaying too fast, the decay should always catch the rpm in that +300-500 rpm range and bring it down to idle slowly there after

also you should set the decay very slowly at and below the dasctl value since its in idle feedback control so decel to idle spark control can stabilize slowly and smooth

 

[Bullitt5566]

We need a TunerPro to SCT translation guide. I've troubleshot this problem for a while. Honestly I think mine is a mechanical problem because at some point I'd actually fix it if it were a tuning issue. But to describe the problem. When coming to an idle the dashpot will bring the RPM down to almost idle then it will bump back up a little bit say 1100 rpm and when the dashpot expires the engine will drop uncontrolled to the idle speed. Sounds like a clip minimum clip to me. I have the bumpy bump camshafts for sale now. Done with it.

 

[gt347mustang]

We need a TunerPro to SCT translation guide. I've troubleshot this problem for a while. Honestly I think mine is a mechanical problem because at some point I'd actually fix it if it were a tuning issue. But to describe the problem. When coming to an idle the dashpot will bring the RPM down to almost idle then it will bump back up a little bit say 1100 rpm and when the dashpot expires the engine will drop uncontrolled to the idle speed. Sounds like a clip minimum clip to me. I have the bumpy bump camshafts for sale now. Done with it.

What cams and how much?

 

[Bullitt5566]

Not the right place but Comp XE270AH, $450

 

[decipha]

bullit, send me your mtf file and i'll take a look

decipha at efidynotuning dot com

 

[SecondhandSnake]

decipha- I've actually been using dashpot values based on your guide for awhile.

Finally got a chance to put together a cal and take it for a run to get some data. After looking through the data, it looks like I overshot the adjustment, down to where my IACDC is too low. It did drive remarkably better though. Maybe it's the engineer in me, but I'm just as happy to get bad data, because with interpolation two wrongs can make a right.

I've got the data for the base throttle stop position; IACDC, and flow through the TB. I've got the same for my second position. I pick the target of ~45% IACDC, and interpolate between the two points to determine how many turns of the stop I need, as well as how much airflow I should see through the TB. It may not be right on the money due to using linear interpolation, some MAF noise and rounding error, but it ought to be pretty damn close, at least to the point it's easily driveable. I think I'll also dial the dashpot preposition and decay rate to be a little more aggressive now that I don't need a bandaid to keep it from dropping too quick.

EDIT:

Still pretty garbage at startup. Only likes it with the adjustment screw way out.

Any chance I'm running into an issue with "Max Mass Air Allowed At Idle?" Stock setting is 1.3 on it. It seems like it's capping my ISCDC around that airflow.

I think it's dialed in when hot. The ISC neutral idle air table is pretty spot on for my tested hot idle values, and matches pretty closely the stock 03 Cobra table. The airflow through the TB should be at least in the ballpark.

Now I think I've got another spot to tune; the startup ISC airmass multiplier. I just stumbled across Decipha's cold startup guide. I hadn't even thought of needing to change that table from stock. Needless to say, it's going to want a lot more air on cold start than it currently has.

Only drawback is I'm using SCT Advantage so there's a painful amount of trial, error, reflash, and datalogging. No live adjustment for me on cold start. Feels bad, man.

DOUBLE EDIT:
Used the cold start data with the throttle body adjusted way out to calculate the multiplier. After a lot of interpolation and calculations, the cell for startup around 60 deg F ECT should be 3, vs. the default table that starts at 1.6@0s and tapers to 1.15@30s. It was getting half as much as it needed. That would explain things. I'll update the table, as well as stretch out the time running from 30s to 120s on the y axis.

 

[SecondhandSnake]

Update: I did the above. Cold start and idle is dead on. Hot idle is absolutely perfect. Had a couple stumbles but I was datalogging so I can establish where on the ISC multiplier table I need to tweak. I suspect it's a couple cells I didn't modify.

Lessons here: There's a lot more to idle quality than just the ISC neutral and drive idle air tables like a lot of material says. Make sure your throttle body bypass air is correct, your throttle stop is set to achieve the correct IACDC, and the multiplier table can have a much bigger impact than you'd expect; especially if you only have idle issues in one region and not another.

 

[Bullitt5566]

I wanted to provide another update on this issue. I put the cams back in and worked on the tune again. The dashpot settings help, but I don't think they are the problem. I set the tune to idle in open loop and then leaned out the A/F by 10%. This helped a lot but the problem is still there. The engine stumbles when returning to idle.

I data logged Air/fuel from the wideband. I noticed that at idle the A/F is ~1.05 and open loop. A quick rev and the ecu goes closed loop and A/F is ~1.0. On return to idle the A/F goes dead rich ~.80 or more until the ecu goes open loop and the A/F recovers. If I force open loop and check the A/F everything is spot on at ~1.0 throughout the rpm range.

The key has been to keep the engine in open loop at idle. I've set it up in two places. Open loop RPM and TP, and Open loop at low loads. The latter required changing a time and vehicle speed scalars to 0. So basically the ecu drops into open loop when closed TP and RPM below 1200 and when load is below .20 returning to idle.

This masks the problem by making the transition to open loop idle faster, but it still doesn't answer why it spikes pig rich when transitioning to idle.