2003 Pontiac Montana
Automatic FWD
EVAP small leak code
180,000 kms
After the engine reaches full temperature, the van stalls when braking to a stop while steering, or while turning at idle, or when rolling in reverse at idle. Shifting back into park, the engine starts up again fine. I cleaned the IAC and the throttle body butterfly valve area. The van has never stalled while just sitting at idle, but once at full temperature, and with A/C off and no steering, the idle feels a bit rough, hovering between 650-700 RPM.
Well if you do have a vacuum leak that will suck air and make the car stall out. So try to find where the leak is on the EVAP system. But then again it is a Pontiac Montana and they were one of the worst cars the GM ever made. Lots of things can make them stall out and sometimes the evap code is just a anti-pollution code that doesn't really affect the running of the vehicle. I would look for a vacuum lake first though
@scottykilmer the EVAP code pre-exists this issue by years. Does that change your assessment?
the EVAP code pre-exists this issue by years. Does that change your assessment?
Scotty does not usually revisit topics he has already answered.
It's quite possible the leak has gotten worse over the last year. What do your fuel trims look like?
@chucktobias oh okay thanks. I don't know how to check fuel trims. I'll look into how to check for vacuum leaks and how to check fuel trims
@jwestern It's very simple to check fuel trims, even most low-cost scan tools can read live data. You should also be looking at other data such as sensor outputs. Even though it's old that's a computer-controlled engine and there is a lot of information in the data stream. Details on interpreting live data are freely available online, particularly on youtube, that can be found by searching.
@chucktobias I guess my scan tool is a super low-cost one because it has no live data capability. I ordered one that does have it. I see that one can bolster a case for vacuum leaks with the fuel trim readings. Will report back my findings
@jwestern Yeah, there are still some super-basic ones out there that can only read codes. The trouble is that contrary to popular belief looking for codes is just the bare beginning of the diagnostic process. There's a wealth of information in the data. My $50 scan tool can read and even graph live data, it doesn't take a very expensive scanner to do that kind of thing these days.
@chucktobias I'll have it tomorrow. In the meantime I pumped cigar smoke into the intake and didn't notice smoke coming out of anywhere. Not sure a MacGyver smoke test like this is conclusive though.
New symptom: when the engine is at full temperature, after stalling, I cannot start again until the engine cools down.
That could be a sign that your fuel injectors are leaky. Hot engines require less gas than cold engines (mechanics and engineers use the term running rich, if the car is using more than 14.7 parts air to 1 part fuel, such as using 13.1 parts air to 1 part fuel. Your engine runs lean, if the mixture is greater than 14.7 parts air to 1 part fuel). 14.7 parts air to 1 part fuel is stoichiometrically perfect. Too rich on a hot engine means the car won't start.
Press the gas pedal to the floor and crank it over. You clear the too-rich mixture, and it starts the car after a couple seconds.
Long-term trim sits around -11%, seldom moving more than 1% from there.
Though not super high that does indicate that overall the computer is subtracting fuel, so you would otherwise be running too rich. If it was vacuum leak I'd expect to see the opposite. You could have drippy injectors, as @justin-shepherd mentioned. Do your spark plugs show signs of carbon fouling? Try his suggestion when it won't start.
Short-term fuel trim usually oscillates between +-4%, excursions to +-6%, occasionally farther. When I press the throttle, this reading plunges to -15% then climbs back towards 0% over ~five seconds Long-term trim sits around -11%, seldom moving more than 1% from there.
@jwestern here's a short article explaining fuel trims which you might find useful: https://trade.mechanic.com.au/news/understanding-short-term-and-long-term-fuel-trims1
Note the section on diagnosing high negative fuel trims and possible causes.
New symptom: when the engine is at full temperature, after stalling, I cannot start again until the engine cools down.
Short-term fuel trim usually oscillates between +-4%, excursions to +-6%, occasionally farther. When I press the throttle, this reading plunges to -15% then climbs back towards 0% over ~five seconds
Long-term trim sits around -11%, seldom moving more than 1% from there.
Ok, I did a proper smoke test and don't see leaks. I also pulled 3 spark plugs, pictures are in this google drive folder. They don't look fouled by a rich mix to me -- what do you think? The other 3 plugs are really difficult to get at, so I didn't pull them.
Is this looking like a faulty O2 sensor?
I haven't been able to try the starting trick yet. Waiting on a wheel bearing before I can drive again. Thanks @Hixster for the link.
@jwestern If anything those plugs look like they may be running a little lean. Your scan tool should be able to monitor your upstream O2 sensor output. That vehicle is old enough that it probably has a narrow-band sensor which should continuously vary between around 0 to 1 volt as it adjusts mixture on the fly.
You're right about the O2 sensor readings. I've uploaded videos of the live graphs for O2 upstream and downstream, under idle conditions and ramping up to 2000 RPM. I also included pictures of the MAF graph at idle and at 2000 RPM. All of these readings are with the engine at temperature. Same link: https://drive.google.com/drive/folders/1HOvowSE_IEfBn-4NVhxhRsqGZRNBphC9?usp=drive_link
@jwestern Upstream O2 sensor output looks OK. Downstream looks pretty wonky at first but then evens out with minor fluctuations around .5 volt, which is normal. (That sensor only monitors catalyst efficiency and does not affect engine operation.) That V6 engine uses the same O2 sensors for both banks?
MAF sensor data looks kind of raggedy but I think that's due to the auto-scaling of the scan tool's graph, the actual fluctuation isn't that much. At idle the g/s is fairly close to the engine capacity so that's OK. As you rev up the engine the MAF sensor reading should increase steadily along with the throttle position.
I'm not sure whether it's the same O2 sensors for each bank.
So, aside from trying the starting trick (which I will be able to try in a few days when I get a new wheel bearing in), does the evidence point to a next step? Could I be missing important evidence because I didn't look at the other 3 spark plugs?
@jwestern It's possible you might find something different on the other 3 plugs. If you go to the trouble of reaching those I'd also do a compression or leakdown test. (That thing is over 20 years old.) If everything else checks out I'd concentrate on the idle control system. (I don't know if 650-700 RPM is normal for that engine, many idle at more like 800-900 RPM.)
I got the other 3 plugs out and performed a compression test (cold engine):
BACK LEFT: 195 PSI
BACK MID: 197 PSI
BACK RIGHT: 201 PSI
FRONT LEFT: 199 PSI
MID FRONT: 201 PSI
RIGHT FRONT: 193 PSI
I also took borescope videos inside the cylinders. There are differing amounts of carbon buildup on the cylinder heads. Note some of the observed wetness I think is likely unburned fuel which was residually present in the fuel lines during the compression test, because I didn't see it beforehand.
I also took new pictures of the plugs and labelled them according to which cylinder they were in. All this material is uploaded here:
https://drive.google.com/drive/folders/1HOvowSE_IEfBn-4NVhxhRsqGZRNBphC9?usp=drive_link
I replaced the plugs and spark plug wires and then drove around. The van stalled once. I didn't push it much after that. So there may have been an improvement, but the issue was not eliminated. The idle RPMs or degree of roughness did not change. The long-term fuel trim started at -25% and then crept up to -15% or so over time.
Next I will inspect the fuse box and ECU pins for burning or corrosion. There's a pending O2 High Voltage code -- we'll see if that gets stored -- maybe it's a transient from post-compression test conditions?
The earlier EVAP code was probably just the gas cap, as its ring was cracked. I replaced the cap -- we'll see if it returns.
Any additional guidance is appreciated.
@jwestern You've pretty much eliminated internal engine problems so the trouble. That long term fuel trim however indicates the computer is compensating for an overly rich condition. You might want to record live data when the problem occurs to check afterwards for clues as to what's causing the problem.
@jwestern Another thing you could do since internal engine problems have been ruled out would be to try the ATS engine and fuel system cleaner products that Scotty recommends.
I found a likely culprit: chaffed, exposed, and corroded battery cable going from the positive terminal to the under-hood fuse box:
https://drive.google.com/drive/folders/16gytrhhM0dFf7-_GvoX97VNfeHEjxRUc?usp=drive_link
Thanks for all your help. I'll repair this and let you know how it goes.
@jwestern Great if that's what it is. Bad battery connections can cause a host of oddball problems hard to track down.
It would explain the dependence on temperature and moisture. Will find out tomorrow when the parts come in.
Did a 3 hr drive yesterday with no problems. Thanks for all your help, I learned a lot.
How long does it take long-term fuel trim to settle after a battery disconnect?
@jwestern Glad you found the problem, but this is an example of why it's always best to check the simple things first! How long it takes for long term fuel trims to settle down depends on the ECU firmware and how much of a correction needs to be made.
Ok, the start-up trick (flooring the gas and cranking over for several seconds) doesn't work. Just have to wait 20+ minutes before trying again.
I replaced the IAC and MAF -- still stalling. It got worse after a day of rain (stalling began earlier, and no-restart was immediate). Makes me suspect an electrical issue. Grounding fault?
Began trying to get at the 3 hard-to-access spark plugs. Nightmare. Got one out, and it looks similar to the first three.
Some other possibly relevant history: the very first time the van stalled was 5 minutes after getting my rear tires changed. This was ~3 hrs after one of the rear tires blew out. I had a really hard time removing the spare tire as the metal components on the cable thru the center of the tire were seized together, and didn't put the spare back because the cable was stretched after that. The timing of this first stall is rather conspicuous, although I'm not sure what connection there could be -- screwed up a ground while wrestling with the spare? But why the dependence on engine temperature?