ECMlink Car wants to stall at cold idle, full speed density setup

[Kryndon]

So this is my first real winter with having a full Speed Density setup. I managed to make a really good tune towards the end of Autumn where ambient air temps regularly hovered around 68F. Now as temps got lower, around 35F ambient, I have noticed the car has issues trying to stall after a cold start, in neutral until it warms up to about 140F coolant temp. After that, and once closed loop is fully functioning, it idles great, good AFR, good airflowper/rev and all that jazz. But it just doesn't like being cold.

Keep in mind this was not happening prior to ambient temps dropping down and I have not done anything else to the car so all mechanical and electrical issues are non factors. I am suspecting I need to tweak the SDtempweighting to make the calculations 100% entirely based on the IAT's temperature input, as there is no need for it to be basing its fuel calcs off coolant temp. Should I just put all those cells to 1 completely ignoring the CTS? Or is there something else wrong in my log? Don't pay attention to the 0.59/0.61v TPS value, I adjusted it after saving that log and it's not a factor. Also I added more fuel to my idle range in the SD table but it again made no impact, so I dropped my global deadtime a tad to get my AFR back in range.

What do you guys think could be causing a stall only during the first few minutes? Log attached, appreciate any advice.

 

[Kryndon]

I tried tweaking the CoolantTempFuelAdj to give it more fuel when cold and also tweaked the FastIdleMTISCPos a tad. Made no real difference. I just compared both cold start logs and the car attempts to stall EXACTLY 77 seconds after cold start. Also, both times it is EXACTLY 25 seconds after going into closed loop that it stalls. This is ridiculous.

What could I be overlooking?

 

[We're on Boost]

Well, your car goes into closed loop pretty early, coolant temp is only 60 or 62 degrees at that point. Then in closed loop it tries to run stoic and it's probably just not ready for that yet. Your combined fuel trims in there are negative about 15 to 20 percent so closed loop is pulling fuel that you need during warm-up. I guess I would try to keep it in open loop longer. I don't even know how you do that. But you could try it out just manually with the "Lock in Open Loop" button on the Misc tab. See if it does any good.

My car after a cold start with coolant at 34 deg F goes into closed loop when the coolant reaches about 87 deg F which is about 120 seconds after it fires, and it seems ok at that point.

There's something else though. Your LC2 numbers and your InjOn times are not doing what I would expect out at about 155sec and 169sec in the second log. Those are places where the LC2 is showing way too lean. But then you follow it to the right to where it goes rich again and the InjOn times are actually less there than they were when the LC2 was showing Lean. Doesn't seem possible.
Finally after coolant temp reaches 109 deg F it all straightens out and seems ok to the end.

I'm kinda lousy with DSMlink. Probably someone else will come up with a better help.

 

[Stapl3]

I had a car that did this. When coolant temps hit the low 130s it would relax and be fine. What helped a ton was putting the timing in the top left corner of the timing table at 5. So 1250 to .05, those 9 cells, would have a value of 5. That's sort of hacky so I decided to use the secondary tables to switch over around that coolant temp of low 130s, so it would then revert to the 'non-5' timing table and all was good.

The real problem was the throttle plate not being cracked open at idle. Stock calls for the throttle to be cracked open 15/16th turn on the switch. Maybe yours is and it wants a little more - I don't know the pro/con of getting away from that value. But definitely check this.

 

[Kryndon]

I had a car that did this. When coolant temps hit the low 130s it would relax and be fine. What helped a ton was putting the timing in the top left corner of the timing table at 5. So 1250 to .05, those 9 cells, would have a value of 5. That's sort of hacky so I decided to use the secondary tables to switch over around that coolant temp of low 130s, so it would then revert to the 'non-5' timing table and all was good.

The real problem was the throttle plate not being cracked open at idle. Stock calls for the throttle to be cracked open 15/16th turn on the switch. Maybe yours is and it wants a little more - I don't know the pro/con of getting away from that value. But definitely check this.

You might be onto something. I also noticed that, prior to this problem appearing, I had set my ISC to perfectly hover around 30-32 at idle accessories off 800RPM, but after the issue started, I noticed it now rests at 38, which means it's trying to open up to compensate for lack of air?

Usually the way the FSM explains the TPS adjustment, you are meant to put a 0.45mm feeler gauge between the SAS and throttle stop which simulates it being slightly pushed open. However, it also requires the throttle cable tension to be set properly. I had both of those re-checked and done last Summer, but I now remember that a month ago I had to loosen my throttle cable bracket a bit one night because it made the idle rise to about 1500 on its own, guessing the cable starts pulling more onto the valve due to heat expansion.

I will recheck my throttle cable again. @We're on Boost you made me doubt that thing about when it enters closed loop but I just went through all my cold start logs ever since I have ECMlink and even way back on January 15th, 2022 when the car was running the stock MAF, at a 42*F coolant temp cold start, it entered closed loop exactly at 62*F again, just like it's doing now. And back then I didn't have such an issue. Now I'm curious to see how long it takes other people's cars to enter closed loop at cold ambient and coolant temps, but I don't see why mine would be any different, since there isn't a setting to tweak this. We can only tweak when it enters open/closed loop based on load factors but not temperature.

The mystery continues...

 

[1990TSIAWDTALON]

Wish I would have logged yesterdays start on my 92 Talon. I hadn't started it in a couple weeks so I jumped in, cold, around 35 maybe, and she lit then died. I restarted her and she took off at a good fast idle of 1300ish, 11.5 AFRs that slowly climb to stoich. My LC2 takes about 20 seconds or more to "heat" up but I just walked away and did my little piddly stuff in the shop while it ran. I asked my son to finally go shut her off (she is on E and I don't need to burn it all idling). He said it was at normal temp and 14.7afrs. BUT....I wasn't logging so nothing to show you to compare with. I did do a cold start for another member but it wasn't near as cold. I will add that log, let me find it so you can all hash and thrash on it, but it works (needs some WOT love currently but with a broken trans, I don't do much WOTing).

 

[Kryndon]

Wish I would have logged yesterdays start on my 92 Talon. I hadn't started it in a couple weeks so I jumped in, cold, around 35 maybe, and she lit then died. I restarted her and she took off at a good fast idle of 1300ish, 11.5 AFRs that slowly climb to stoich. My LC2 takes about 20 seconds or more to "heat" up but I just walked away and did my little piddly stuff in the shop while it ran. I asked my son to finally go shut her off (she is on E and I don't need to burn it all idling). He said it was at normal temp and 14.7afrs. BUT....I wasn't logging so nothing to show you to compare with. I did do a cold start for another member but it wasn't near as cold. I will add that log, let me find it so you can all hash and thrash on it, but it works (needs some WOT love currently but with a broken trans, I don't do much WOTing).

330 seconds until it enters closed loop? Now what the hell is up with that!

 

[Stapl3]

I also think it could be from your openloopthreshold values. Lock it in openloop before a cold start and see what happens to confirm. I would bet it doesn't get wonky after 62*, maybe much later or not at all. I don't think the issue has anything to do with the coolant temperature, rather that's the temp where it happens to switches to closedloop.

 

[We're on Boost]

Have one more tid-bit to throw at the closed loop question. In the 1997 FSM there is a chart on page 13A-113 that seems to show that the "Normal Condition" is OPEN-LOOP when coolant temperature is Less than 86 deg F.
This is specified at 2,500 rpm though, and the chart is for the Non-Turbo 2 Liter.
I can't find a chart like this for the Turbo 2 Liter, or for the 1g.
This "fuel control status" item is near the bottom of the page.

Another strategy I've seen for other makes is to combine 2 or 3 things, like coolant temp combined with waiting some number of seconds after O2 voltage goes over half a volt.
I notice in Marty's log that he is simulating O2 from the wideband and his O2 voltage went over half a volt at only 21 seconds into the log, and his coolant temps are coming up normally. So it's not as if it couldn't have gone into closed loop earlier. But whatever, we aren't trying to protect a cat anymore in most cases.

You must be logged in to view this image or video.

 

[1990TSIAWDTALON]

330 seconds until it enters closed loop? Now what the hell is up with that!

I pulled another "No Touch Start" log yesterday. It fires right up and I walk away. Looking at my log, it apears to not go into closed loop until the WB starts cycling around stoich and I simulate NB. The tune (E85, SD) was setup by my tuner but due to mechanical issues, we never got finished (and its been worked around some also, so it is no way perfect),
Here is another No Touch Start log to look at. It was very nice yesterday so it wasn't "cold". Maybe it will shed some light on others having to "pedal" their cars for a minute or so when it IS COLD. I never have to do that.
Marty

 

[Kryndon]

Thanks Marty. That's interesting, in that last log it took 100 seconds to enter closed loop at around 86*F. However, you mentioned you're simulating NB. Now I'm starting to suspect if this behavior could be due to a weird front O2 going bad, lazy or something else? I'm still running a stock (non-OEM) front O2 and have my LC2 placed just under the intake manifold. I wonder if it's worth simulating NB from my LC2.

 

[1990TSIAWDTALON]

I think you could try it and if you don't like it just reverse it or replace the front O2 sensor. I am trying to help even tho my tunes aren't finished, they do work pretty good. Next time its COLD I will go get a log for you/anyone viewing this thread to compare with.

 

[steve]

The 1G ECU doesn't try to enter closed loop until the coolant hits 87F and if you are already over that when you start there is a timer but I forget how long. This is based on reading the code and observation.

I never paid attention to the 2G warmup but ECMTuning says the ECT must be 51F before it starts trying. I suspect it also has a short timer to give the engine a chance to stablize if the ECT is already > 51F on startup.

fueltrimupdatepoints [ECMTuning - wiki]

www.ecmtuning.com

You'll see the ECU switch to closed loop in the log and watch the STFT start to adjust while the ECU waits for the O2 to switch. If it takes too long or doesn't before the trim reaches the max it gives up for now.

Have one more tid-bit to throw at the closed loop question. In the 1997 FSM there is a chart on page 13A-113 that seems to show that the "Normal Condition" is OPEN-LOOP when coolant temperature is Less than 86 deg F.

That's the chart for a non-turbo 2G.

 

[Kryndon]

Alright, it APPEARS it fixed itself on its own, however I have slight suspicions it was due to too rich idle cells between 1000-1500 RPM at idle. I went and changed the 2 main ones from [60][65] to [58.5][66.5] and dropped the surrounding cells by -1.

Today I fired her up, it was by far the coldest of the days, 30*F, it died as it cranked because my battery isn't the strongest and I didn't keep the key to crank long enough. But it started to idle great, entered closed loop even quicker, at 55 seconds and at 59*F. Then I watched closely, and it just idled great, no stalling, didn't have to give it any gas at all. I then took off and cruised around the city. Once fully warmed up, and once the Combined Fuel Trims stabilized, I was getting a great idle, to the book. 0.63 TPS voltage, 0.25g/sec, 800RPM, 33 ISC pos, all the good stuff. It didn't try to stall once even with full accessories etc.

I did notice, however, that at the end of my trip the STFT wanted -4% correction so my suspicions would be correct that it may have been experiencing too much fuel at idle. For now I will assume the issue was solved. Too much fuel and too little fuel can have similar effects!

 

[We're on Boost]

That's the chart for a non-turbo 2G.

This is specified at 2,500 rpm though, and the chart is for the Non-Turbo 2 Liter.
I can't find a chart like this for the Turbo 2 Liter, or for the 1g.

 

[steve]

Sorry my old eyes missed the disclaimer and the brain stuck on the fact that there is nothing in common between the two ECUs software including their behavior in closed loop for the chart to seem useful to the discussion