I don't watch the phone while driving so it wouldn't matter but, if I record using a combined graph can I go back and view the log as seperate graphs?
Does it have to learn to idle?
- Thread starter Sqrly
- Start date
I don't watch the phone while driving so it wouldn't matter but, if I record using a combined graph can I go back and view the log as seperate graphs?
And if you want to get
really silly you can enable GPS recording, then replay selected data points with a map, like a video! Here is one playing at 5x speed...
I believe tjbaker57 posted a way to test the map using a vacuum pump (hand) and connecting it to the input "nipple"(removed from the intake manifold) while leaving the map connected electrically to the vehicle. Turn the key to on. Apply some form a vacuum via the pump and read the output on the obd app. The reading of the pumps meter and the app should match. I found at least one MAP that failed with high vaccum readings after a few cycles of vacuum / release / vacuum.... almost like the sensor got stuck.
And it's stalling again.
This is why I wanted to change the title to "Learning to stall" because that's what it seems to be doing.
It idles fine for a few days after doing the TB relearn and then goes back to the same stalling when the ac compressor turns off after warming up for about 5 minutes idling in the driveway while I clean the windows with the hvac on defrost.
This is why I wanted to change the title to "Learning to stall" because that's what it seems to be doing.
It idles fine for a few days after doing the TB relearn and then goes back to the same stalling when the ac compressor turns off after warming up for about 5 minutes idling in the driveway while I clean the windows with the hvac on defrost.
Just for curiosity sake,,, what is the outside temperature during this ??
And is your HVAC system RPO CJ2 or CJ3?
I thought about that. This morning it was 46f when I was warming it up (the overnight low was 36-38). On Oct 31st when this issue first appeared it was 50f (overnight low of 48). It also stalls once it's completely warmed up and 60f outside.
FYI. I removed the AC relay and it idles smoothly while the knob is in the defrost postion or when it's in the vent position with the snowflake button on. Whatever confusion this causes the ECM it appears that it never tries to turn the compressor off because ...
When I move the knob off defrost or manually turn off the snowflake button, it tries to stall even though the AC relay (Engine Compartment #44) is removed.
How do I know which HVAC system I have?
Last edited:
I ran a short log of the RPM, Throttle Position, and MAP pressure while parked this morning after it was warmed up and the AC relay was removed. I let it idle for a bit, then I revved it up once and then held it at 1500 rpm for a bit. It's not easy to hold it at a specific rpm with drive by wire 
2024-11-18 10-04-41.brc
2024-11-18 10-04-41.brc
If you adjust your HVAC temperature with vertical sliders you have the CJ3 (manual) system.
The CJ2 (automatic) system has knobs for adjusting HVAC temperature plus it has temperature sensors both inside and outside the vehicle.
OK, so not near the 34° F compressor cutoff point for a CJ2 system.
I was just reviewing some A/C & PCM HVAC modules data and had a thought..... We have been observing the compressor shutoff and engine stalling. What if the sequence is reversed? The PCM will disable the compressor upon detection of an unstable idle. Maybe there is an engine idle stability issue and the compressor shutoff is a result, not the cause?
It's the CJ3 then as I have the manual system.
I can cause the idle severe idle drop that sometimes results in an engine stall by turning off the snowflake switch when the clutch is engaged.
I can duplicate this even when the ac relay is removed. This leads me to think that with the relay removed the system thinks the compressor is running but never gets what ever signal it's waiting for to turn it back off.
When it's doing it by itself, the idle drops more each time until finally it stalls. This doesn't start until the engine is not 'high idling' due to low temperatures.
In fact, in the summer, I don't recall it ever doing a 'high idle' at start up. The idle would settle down to normal in like 5 -10 seconds after starting the engine in the morning and it never stalled.
Currently it idles at about 1500 when first started cold and it slowly drops to normal over a few minutes.
What gets me is it seems to be "learning to stall".
I cleaned the TB on the 2nd and did the TB relearn and it idled fine for 2 weeks.
I did the TB re learn again on the 16th and it idled fine for 2 days.
This all started about the same time the over night temperature outside started dropping close to 40.
Maybe it's just a coincidence but the colder it gets, the worse the problem is getting.
Last edited:
Hmmmmm...
Well, there is a compressor feedback signal to the HVAC module. So that module should surely know if the compressor clutch is powered or not.
If the relay is out then the clutch cannot be powered and thus the feedback signal to the HVAC module (connector 1, terminal B12) must be absent.
So at least the HVAC module knows the compressor is not running.
While the PCM does control the actual enabling of the compressor it does so at the direction/request of the HVAC module. The PCM does not decide on its' own to engage the compressor.
With the relay in place it goes like this:
A press of the "snowflake" button while the compressor is enabled causes the HVAC module to send a message for the compressor to be disabled (8815981001 for anyone interested).
The PCM then reports the current status of the compressor (enabled) and then acknowledges the previous message from the HVAC module.
Now the HVAC module acknowledges the status report from the PCM.
Then the PCM reports the current engine torque and the disables the compressor and again reports the status, now disabled.
The HVAC module acknowledges the PCMs status report of the now disabled compressor.
The PCM again reports the current engine torque.
All this happens in maybe 200 milliseconds or thereabouts.
This is typical of the goings on of the serial data bus. Almost nothing works like the old days of control loops and active switches anymore. It's all programming and sensor inputs.
We know the PCM adapts the idle to compensate for the load of the compressor.
So when the compressor is stopped what adjustments are removed?
I am curious about what happens with the throttle angle, injector pulse width, and other parameters the PCM may adjust in anticipation of the release of the compressor load.
It's too cold here for me to attempt any tests with my own 2002 4.2. Mid-teens °F each night and tomorrow's high is forecast at 29°F.
I agree.... during the AC "mode", the system usually increases RPM to handle the load. Upon AC "off", the system drops "stuff".... whatever reduces "fueling". However, if the system is already running in a "leaning" mode (ie.negative trims... more than just just a few %), then it may be likely further fuel reduction is enough for a rpm stumble / stall.
Just let me know what you'd like me to log and when, like 1st thing in the morning when it's warming up or on the 2 mile drive to the kids school and back in the morning, or in the afternoon and I'll do it even for curiosities sake. Note: I usually warm it half way (160) up then shut it off for about 20 minutes before driving to the school. Now that's it's freezing overnight I'll warm it up for another 5 minutes just before leaving.
I just finished cleaning the windows this morning. It's 34f outside. The rear window was iced and the front was not because the vehicle is parked facing east. The dash gauge read 160 after about 10 minutes, the time it takes me to pull it in and clean 4 windows and the heater was blowing hot air.
The AC relay is still out. It idled fine and defrosted the windshield fine. The engine stumbled only once when I moved the rotary knob off defrost to floor only.
Last edited:
Mooseman
Moderator
Having the relay out makes the PCM "think" the compressor is running when commanded. If you disconnect the low pressure switch on the dryer/receiver instead, this will completely disable the compressor and the PCM will not command it on. Although this won't fix your issue when you need the A/C, it should prevent the idle stalling or may unmask another issue if it's still stalling.
I agree since the trims are not "normal", this could be the root cause of the stalling. There's something causing the rich condition that the PCM is trying to correct. Did you check your fuel pressure regulator to see if it's leaking into the intake? That's a possible culprit as it sometimes fails and leaks fuel into the intake resonator it's connected to.
The goal is to capture data leading up to, and during a stall event.
If a stall can be captured in a data recording there is a chance that some condition leading up to the stall may be evident.
Naturally there is no guarantee such data will reveal a cause but not looking is a guarantee of not finding
As to what to monitor/capture that is at best an educated guess. I am thinking towards what causes a stall.... fuel mixture issues (too rich, too lean) so grab long and short fuel trims. One would think fuel trims would act upon the injector pulse timing as that is the primary means of controlling fuel delivery. Cylinder 1 is usually used as the representative value for such purposes. Throttle angle might indicate if the PCM is trying to raise the rpms so grab that.
An issue with data recording is how much to record. Select too many sensors and you end up with too slow a refresh rate between data points. Too few sensors in the record and you risk missing important data.
sidebar: while writing this my woodstove stalled! Too little air
This observation of a relationship of the stall with the HVAC system really is a peculiar thing. Difficult to see a logical path between the two given the presence of the trouble while the A/C relay is removed.
Keep in mind that the low pressure cycle switch is not wired to the PCM at all. It is only a sensory input to the HVAC module where the HVAC modules' programming responds to this input by granting or withholding the "A/C Request Withheld" parameter which can be viewed with scanners.
So disconnecting the low pressure cycle switch indirectly stops the compressor in the very same fashion as any of the other parameters listed above. That is by causing the HVAC module to not request compressor engagement.
Now the A/C High Pressure sensor IS wired to the PCM. And the PCM reports this high pressure value via the serial data bus upon changes in the value. In this way the HVAC module has access to the A/C high pressure value. So depending on PCM programming it does at least technically have the ability to know whether or not the compressor is running regardless of the A/C relay.
Actually. no, I did not check the vacuum line on the pressure regulator ... until just now. I pulled it off and smelled it, blew through it, no gas smell at all.
I ran a log a bit ago while sitting idling in the school parking lot, engine fully warmed up, and pressing the snowflake button, turning it on and off. It doesn't seem to have saved the entire log for some reason.
On = LTFT goes up. Off=LTFT drops. 2024-11-19 12-21-16.brc
I ran a log a bit ago while sitting idling in the school parking lot, engine fully warmed up, and pressing the snowflake button, turning it on and off. It doesn't seem to have saved the entire log for some reason.
On = LTFT goes up. Off=LTFT drops. 2024-11-19 12-21-16.brc
Ok, tomorrow morning l'll put the ac relay back in and I'll start the logging before I start the engine and just do my normal routine up to the stall and restart.
I'll log rpm, STFT, LTFT, and what, throttle position? I don't see any "throttle angle or injector pulse width" in my app.
And while we're at it, is this log normal and/or accurate? This is the timing advance while idling at normal temperature.
Reminder: Whoever worked in this vehicle before it came into my possession was an idiot and preferred duct tape and silicone over oem quality parts.
Last edited:
Injector pulse width for each cylinder should have been imported with the 2nd set of sensors I linked for you. There should be a bunch of sensors that all begin with "[GM]" for the sensor name.
Edit : see 3rd private message on Nov 2.
Last edited:
Here is a comparison recorded several years ago. I see similar variance at idle though your looks tomsrop a little lower than mine did.
I can't say that I have ever graphed my LTFT but do gage monitor it alot / "always". It never bounces around all over the place. My "reads" on the net indicate that the LTFT should be "stable" depending on some factors like throttling and load. Further, the fact that your trims "tally" so much negativeness is not good.
Hopefully, something more certain will show up.
In a truck with no MAF, the only governing sensor inputs are the MAP and intake air sensor for a "speed density system". Since your problems are in cooler times, maybe monitoring the intake air temp might help.
My Yukon sure does change. Remember there are a bunch of different "fuel trim cells" and each of them hold values for a specific range of conditions. A cell for idle fuel trims another cell for decelerations, yet another for wide open throttle fuel trims and a handfull for the in-between conditions.
Here is a 5.3 liter LM7 going over Wolf Creek Pass here in Colorado.
Ok, got it. I only imported the misfire and tcc pids from that set. I'll import the pulse width ones now.
Well that looks interesting... I wonder why they would call it LONG TERM FUEL TRIM if it is looking like there is no long term.
Anyway, when I monitor my trims via gage, the value is pretty static. I would think that there is no "special method" used by the app to "flatten" the response. I also mention the "stability" of the throttle can "influence" the results.
ADDED: this morning, I did a short idling / graphing to watch the stft and ltft. I haven't uploaded anything BUT my ltft at idle are flat. I can cause a "bump" by light touching the peddle, but the flatness returns there after once the system settles.
This is on the torque app.... 2008 4.2 with 240kkm.
ADDED: we are probably talking the "same" just a different time frame... you are showing a time axis of minute / pedal moving, while I am talking "idle" or constant peddle "cruising".
Last edited:
When I do an import I usually import all then remove the ones that don't work or I don't want.
On the log you posted above from your Yukon, I see the LTFT mostly above 0. Mine never goes above 0, it only gets lower.
Where's this difference coming from?
Where's this difference coming from?
The actual above or below running isn't that important as the system is "engineered" to be "at zero" in "perfect operation". However, "perfect" never happens especially with age. The important thing is the size of the value. Small increments above or below likely indicate some adjustments the system has determine are needed to deal with age to meet 14.7:1 fueling mixtures. It is the large adjustments that are at issue. 25% (with time) will cause a coding for either "running too lean" or "running too rich" as that is out side the limits that the system can compensate for. Running too rich uncompensated for will result in destroying the cat while running too leaning may cause overheating / misfires and such.
The values that your graphs show seem to point at the system trying to lean out the fuel because it is detecting a richness in the exhaust.
I think this log shows more than just turning on and off the AC... as there seem to be some pedal action (more than just system response).... is that true?
hmmm.... the system responded with something to go to over 1000rpm.... that seems pretty drastic for an idling vehicle.
This morning, temp gauge on mirror said 48, I started the log, started the engine, turned on the hvac fan, turned the hvac knob to defrost+floor. I let it idle until it stalled then restarted it and let it idle until it stalled again. I never touched any of the pedals or anything else for that matter.
Note: While stalled (not running) I had a "Check Gauges" light on and the DIC said "stop engine" and "low oil pressure".
Q: How do you get the app to display more then 3 seperate charts at once? I had to edit these captures together.
2024-11-20 07-59-37.brc
Note: While stalled (not running) I had a "Check Gauges" light on and the DIC said "stop engine" and "low oil pressure".
Q: How do you get the app to display more then 3 seperate charts at once? I had to edit these captures together.
2024-11-20 07-59-37.brc
wow that is some "ugly" system. I don't believe that the LTFT should be bouncing around like that in a "idle" situation.... the vehicle speed is zero so there is nothing happening other than system response. I would say you engine temp is low and seems to "flatten" possibly indicating a stuck open thermostat. To me, the fact that it is running "rich", might be an indications the system still has not reached "normal operating temperature" and thus "fueling" / "choking" to ensure operation. My truck reaches "nomal" in about 4-5 minutes in our current 30f ish temps.
Ugly = IKR
The LTFT is responding to the AC compressor turning on and off. If the HVAC system wasn't in defrost or AC it would stay flat.
The coolant temp does reach 190 and stay there but it takes a few more minutes then that log was run for.
Note: It doesn't stall if it's in drive and not moving because my foot is on the brake. I only tested this for a minute or so sitting in line at the school so I'm going to run another log with he same settings but with it in drive and the parking brake on after it sits for a few hours.
Ok, this is interesting and even more confusing.
I just ran another log with the same sensors and the same conditions as in post #71 above except 2.
1. It's warmer outside, 64f instead of 48f.
2. I put it in drive (with the parking brake holding it still) right after I started, before I turned on the HVAC, it and left it in drive until about 12:30 into the log.
You can see it never stalled though the rpm did change a little when the compressor went on and off while in drive.
Now the interesting part.
After I put it in park you can see where the rpm dropped a lot as the compressor turned off but, it is not doing it consistently. Sometimes it's smooth, other times it nearly stalls and now that it's warmer outside it only stumbles and doesn't actually stall.
@budwich You may be right about the thermostat. It doesn't seem to hit 190 until I start driving it. Those are cheap so I think I'll replace it.
EDIT: FFS, those aren't cheap either, I have to replace the entire housing
2024-11-20 12-21-41.brc
I just ran another log with the same sensors and the same conditions as in post #71 above except 2.
1. It's warmer outside, 64f instead of 48f.
2. I put it in drive (with the parking brake holding it still) right after I started, before I turned on the HVAC, it and left it in drive until about 12:30 into the log.
You can see it never stalled though the rpm did change a little when the compressor went on and off while in drive.
Now the interesting part.
After I put it in park you can see where the rpm dropped a lot as the compressor turned off but, it is not doing it consistently. Sometimes it's smooth, other times it nearly stalls and now that it's warmer outside it only stumbles and doesn't actually stall.
@budwich You may be right about the thermostat. It doesn't seem to hit 190 until I start driving it. Those are cheap so I think I'll replace it.
EDIT: FFS, those aren't cheap either, I have to replace the entire housing
2024-11-20 12-21-41.brc
Last edited:
Mooseman
Moderator
The changing throttle position, despite not touching the pedal, is the PCM adjusting it to maintain idle speed. This would be its response to adjust idle when the A/C compressor kicks on and off. It wouldn't use fuel trims to adjust it. Think like the old days of carbs when there was a solenoid to push the throttle open a bit with the A/C on.
Fuel trims is a response to the O2 sensor telling the PCM it's running rich or lean. If it were ME, I would replace the upstream O2 sensor. I know it's a shotgun approach but at the mileage it has, it could be going bad. Something is making the PCM think it's running rich and, according to this latest log, it's intermittent, going between 0 and -15 on LTFT.
It's normal for the cluster to tell you to shut off the engine with no oil pressure. It's an automatic response when the ignition is on but there is no oil pressure despite no RPMs.
Fuel trims is a response to the O2 sensor telling the PCM it's running rich or lean. If it were ME, I would replace the upstream O2 sensor. I know it's a shotgun approach but at the mileage it has, it could be going bad. Something is making the PCM think it's running rich and, according to this latest log, it's intermittent, going between 0 and -15 on LTFT.
It's normal for the cluster to tell you to shut off the engine with no oil pressure. It's an automatic response when the ignition is on but there is no oil pressure despite no RPMs.
The 0-15 LTFT change is exactly when the compressor goes on and off.
When the compressor turns off the LTFT goes from 0 to -15, shortly after the STFT jumps up.
I think you posted right before my last log + pictures?
I'm fairly sure that one or more sensors is malfunctioning, I'm just trying to be sure which ones before I go buying them.
Don't forget, it idles smooth as a babies bottom when the compressor is not turning on and off.
Please consider, this vehicle was driven for many miles with K&N air filter that had never been cleaned. It had so much dirt crusted on it I had to throw it away. I'm suprised a vehicle would run with that much air flow restriction.
Don't forget, it idles smooth as a babies bottom when the compressor is not turning on and off.
Please consider, this vehicle was driven for many miles with K&N air filter that had never been cleaned. It had so much dirt crusted on it I had to throw it away. I'm suprised a vehicle would run with that much air flow restriction.
Last edited:
Here's another capture. Sitting idling fully warmed up.
I was watching the following sensors.
O2 Sensor V
O2 Sensor STFT
STFT
LTFT = Never changed, sat at -18
Throttle Position = fluttered about .4%.
Is this what it should look like?
I was watching the following sensors.
O2 Sensor V
O2 Sensor STFT
STFT
LTFT = Never changed, sat at -18
Throttle Position = fluttered about .4%.
Is this what it should look like?
The capture is good and as expected for a hot idle vehicle. Your "observed" LTFT is the problem.... no way it should be sitting in that far negative. In fact, when the STFT swings negative, the sum is almost at coding (-25%)... running too rich.
This is what I see too but I don't have the knowledge to know what it means.
I also see that just before it does stumble, there's 3 small changes in the STFT graph, and when it does not stumble, there are only 2. I've no clue what this means but it is there.
