LS1Formulation

iTrader: (13)

The purge valve is basically a solenoid that opens and shuts the line off. If the canister is full, be it with vapor or gas, it's going to pull it in regardless. All intakes end up with some sort of oily condensation in the ports, if it's not gas, I wouldn't worry about it. Have you tried spraying brake clean around the intake ports when the engine warms up and it runs rough? If there is a gasket problem or a crack, that will likely show it.

SnowMan01

I guess I've done basically the same thing by using a propane torch. I used it when the engine was cold and found nothing. I could try it again when cold and see if something shows up. I am a bit leery of using anything flammable when the engine is up to operating temperature. My research leads me to believe a propane torch is the safest thing to use when the engine is hot. I've also used a water bottle mistifyer with no luck.

mrvedit

SnowMan, so you find it "safer" to flood your engine area with propane gas while the engine is running, rather than brake cleaner? Hey, each their own. I have used propane too, but brake clean as the advantage of also cleaning parts.

You have tried so many combinations, that its sometimes hard for us to keep track of them all. I personally thought we had ruled out the EVAP system, but e.g. @gametech still thinks that is the source. Not being there myself, I can't rule it out.

In any case, I have to commend you for your determination at figuring this out yourself. Compared to your 82, I'm "only" 69 and probably like you have rarely needed the "checkbook" to fix things. (In contrast a checkbook is the only tool my neighbor has. Sad.)
I suspect this is getting frustrating, but we are happy and determined to keep giving you help.
If you add your City/State to your profile, someone might offer to come and help you in person.

SnowMan01

Completed some more testing, this time on the fuel injector pulse width. The results are as follows:

O2 Sensor disconnected and engine in open loop status
Injector PWM was cycling between 6 & 7 ms at start up - coolant temp 18 degrees C
rough idle started when coolant temp reached 64 degrees C - Injector PWM variation increased to cycle between 6 & 10 ms (it only occasionally reached 10), most of the time it was 8 to 9.5 ms. In some cases the PWM changed up or down 3 to 3.5 millisecs within 4 seconds.

I waited for the engine to cool down and then I repeated this test again. The results were essentially the same.

Next I waited for the engine to cool down again and repeated the test however this time I hooked up the O2 sensor so all sensors were hooked up
The results are as follows:
Injector PWM started at about 6 millisecs and slowly decreased down to 2 millisecs. The loop closed at a coolant temp of 60 degrees and the engine started to rough idle. Injector PWM was cycling between 2 and 3.5 millisecs at this point. I next disconnected the evap hose at the purge valve and the idle smoothed out. Injector PWM varied between 2.5 to 3.5 millisecs even though the idle seemed to be smooth.

Does this make any sense to anyone. Should PWM's vary so much in so little time (3 or 4 secs)?

SnowMan01

Using anything flammable could be dangerous. The way I look at it is that using flammable liquids such as quick start or carburetor cleaner would result in coating the engine with a flammable fluid. If anything went wrong you could possibly have a raging fire on your hands. If a gas vapour such as propane was ignited by a spark or a hot engine it would all be over in a flash. Unlike flammable liquids, a gaseous vapour such as propane would also have a tendency to dissipate in short order so you might end up singeing your eyebrows or the hair on your hands. At least that is what I think. Besides I use a propane torch to do soldering so it is very handy for me to use. I rarely use brake cleaner anymore since I purchased two new vehicles in 2009. They have both required very little mechanical maintenance over the years so I don't have any brake cleaner around.


Thanks for the suggestion that I might get an in person offer to help but I live near Toronto so I doubt there is anyone in this area that belongs to the forum.

LS1Formulation

iTrader: (13)

Have you tried capping the intake port as previously discussed? I'd like to see the EVAP disconnected, but the intake port plugged, so we can definitively rule out the EVAP ( or look more into it, if the rough idle issue goes away.) Injector pulse width will vary a lot in a short period of time, especially in closed loop. The ECM is making adjustments constantly. One thing you haven't done yet, I don't believe, is a compression test. We should be verifying that the engine is completely healthy as well. Do it cold, then hot.

SnowMan01

LS1Formulation

Yes I reported on disconnecting the EVAP hose from the purge valve and then plugging the end that is connected to the intake. The engine was idling smoothly when the hose was disconnected but when I plugged it the engine started to idle very rough and looked like it would stall.
I have just come in from performing some more tests on the injectors with the instrument you recommended. The first time I used it I subjected each injector to just one long pulse (250 ms) two times and the fuel rail pressure dropped the same amount each time so I just concluded that the injectors must be OK. After thinking more about the tests and reading up on how injectors work I decided I should maybe do the mode 3 test which would be closer to real life operation of the injectors, 100 pulses of 3.5 ms each. Again I did the test on each injector two times. Increasing the fuel rail pressure back to 50 or so psi after each test. Three of the injectors dropped the pressure about 13 psi each time they were pulsed. Very close to the 12 psi drop when I did the test last week. For discussion sake I will refer to these 3 injectors as the "good" injectors.

One injector caused the pressure to drop 17 psi the first time I did the test on it. When I repeated the test the pressure dropped 20 psi. I thought I was doing something wrong so I checked out my connections and repeated the test. The pressure again dropped 20 psi. I was really scratching my head and decided to repeat the test on one of the three "good" injectors. I did the test twice and each time I repeated the first results, i.e. the pressure dropped 13 psi. I went back to the "bad" injector and did the test there again. As soon as I connected the pulser to the battery I heard a noise from the injector and the fuel rail pressure was dropping quickly on it's own BEFORE I had a chance to select the mode. In a few seconds the pressure dropped to zero. This drove me crazy so I decided to step back for dinner and do the test after having something to eat.

Well I returned after a snack and did the injector test again on the "bad" sensor and this time the result was the same as the other three injectors. What is going on?

I next decided to repeat the fuel rail pressure test so I could compare it to what I measured last week. After 15 minutes the pressure had dropped from 51 psi to 43 - a 8 psi drop. Last week the pressure dropped 5 psi in 15 minutes. Today it is dropping that amount in 10 minutes.

Any suggestions? I am at a complete loss!!

SnowMan01

Could this "bad" injector be leaking/sticking and would adding injector cleaner help. Is it worth a try to change this one only? The car only has 70,000 km on it.

LS1Formulation

iTrader: (13)

That injector definitely could be sticking, but pouring injector cleaner in the tank is not likely to have much of an effect. You can change one injector if needed, it shouldn't be a problem. Since you like to tinker, you may want to use the injector pulse tool to attempt to clean the injectors. I figured out a method for doing so that works really well. You'll need a hand operated vacuum pump ( Mity-Vac or the like ), and some Berryman's ChemTool injector cleaner. And you have to remove them from the engine, of course. Get a stainless bowl and fill it partly with the cleaner, enough to submerge the outlet end of the injector. You'll still have to figure out a good way to keep the leads on the injector - I'd order a injector connector pigtail off of Amazon to make it much simpler. Once you have the leads connected to the injector, find one of the rubber adapters that come with the vacuum pump that fits the inlet side of the injector. I find that a tapered one fits into the end nicely. With the injector outlet end submerged, use the vacuum pump to pull about 25" of vacuum on the injector. Once that is done, push the continuous button on the injector tester and the vacuum will backflow the injector with the cleaner, and flush any particles backward out of the inlet. I did a set of injectors awhile back where at least 3 had been sticking and the amount of crud that came out was really surprising. The outlet holes in an injector are very, very small, and it doesn't take a lot for them to get clogged and cause issues. You may have to pump vacuum back up a couple of times on each injector to fully backflush them.

mrvedit

Many (most?) injectors also have a filter on the inlet side which can be pulled out with a sheet metal screw and either replaced or at least cleaned. No idea if this is feasible with your particular injectors.
Did you ever tell us what year and model car you are working on?

SnowMan01

In my first post I stated it was a 2009 Pontiac G5. Added info would be it is an SE with a 2.2 L engine.

A distressing turn of events has occurred. After performing all the injection pulsing yesterday I reconnected everything (I think) and now she won't start. It turns over but I see no evidence of any combustion. I took all the plugs out to see if I had saturated them with gas yesterday because I injected a lot the gas with the pulser. I thought I might have flooded the thing and would find the plugs very wet with gas however the plugs were completely dry.

Any thoughts by anyone of how I can find out what might be wrong. I am thinking of taking out a plug to see if it sparks while it is turning over.

LS1Formulation

iTrader: (13)

Hmm... I'm not sure there. Definitely should check for spark, but there should be no reason it won't start simply by reconnecting the injectors. Might want to also check fuses, just in case you touched the wrong thing and fried a fuse.

SnowMan01

Great analysis LS1. That was the first thing I did this morning. Found the injector fuse and it was fried alright. It started right up with the new fuse.

However all that pulsing of the injectors changed nothing. It still has rough idling when it goes into closed loop

I decided to check out the EVAP canister so after driving up the ramps I crawled under and disconnected all the hoses and electrical connectors. I was fully expecting to find a very dirty cannister with lots of gas in it but found none of this. The thing was very clean and absolutely NO signs of gasoline. Put everything back together and started her up again.

Once it went into closed loop and idled roughly I again disconnected the evap hose and it ran smoothly. This time I watched the live O2 sensor 1 and short term fuel trim. They both looked good. Short term fuel trim averaged around 0% (from about + 3% to -3%). The O2 sensor varied in a beautiful sine wave pattern between about .1 mv to .9 mv. The engine idle RPM was 1500. I then reconnected the EVAP hose and the RPM dropped to 900. The STFT dropped dramatically to the -20% area and stayed down there until the engine finally stalled. Next I tried to duplicate the extra air coming in through the hose by loosening off a hose clamp on the air intake hose downstream from the MAF. The engine ran very rough so I loosened the clamp even more and slowly pulled the hose off the intake duct (I was trying to let the same amount of air into the intake as the disconnected EVAP hose - albeit on the other side of the throttle body). This did nothing and the engine finally stalled. The STFT stayed in the -20% region. Seems to me this would this rule out a vacuum leak.

Does any of this give anyone any new ideas?

I am not yet all that familiar with the EVAP system and will have to do some reading to find out how it all works. There is a fuel tank pressure sensor on the canister which has 3 electrical connections, one says 5v, another gnd and the third I could not make out. It is the only component I could not check out electrically. I did not try because I didn't want to take a chance on damaging it. Does anyone think this little sensor could be malfunctioning and causing me this headache?

mrvedit

Thank you for repeating what car and engine you have; sorry I forgot, but I monitor posts and stay active in three sections.
Your last post is an good summary of what is going on.
Please confirm that you disconnect the evap hose right at the intake manifold? Assuming true, then this is a huge vacuum leak which makes your engine run properly with a STFT of only +- 3%. What RPM is it running then?
A STFT of -20% is probably the limit of how much the ECU can lean the mixture.
This should indicate the ECU is not injecting the extra fuel (unless its defective).
Your info about the injector pulse widths in post #44 confused me more than it helped.
Your experiment about bypassing the MAF is also more confusing than it helps, me at least.

One hypothesis is that an injector is leaking the exact amount of fuel to compensate for the disconnected EVAP hose. Connected, the mixture is too rich and the ECU closed loop cannot compensate for it. However it doesn't explain why it runs fine when cold, unless the injector needs to warm up to leak. I suspect that members who know this stuff better than I are also stumped.

LS1Formulation

iTrader: (13)

I definitely am, I can't figure this out. Here's what I would do, even though it's "throwing parts at it": Buy one single injector, and replace the one that caused you an issue when you hooked up the tester. Since it's the only one that didn't work correctly, it's suspect. An injector shouldn't be more than $50. If the problem continues, swap the new injector in place of each injector and repeat testing. It's tedious and it's going to get gasoline everywhere ( buy some shop rags, and do it in your driveway on a nice day ). You don't have the higher end type of scan tool that can help diagnose issues like these, so you have to do work arounds. One other thing I'm wondering is, with everything hooked up and the car running rough, if you push down the accelerator slightly, do the fuel trims go back to normal and the car runs smoother? It doesn't seem like an airflow issue, but you never know. Also, check your air filter if you haven't already.

SnowMan01

I did some research on how the EVAP system works and I am left scratching my head. Here are some of the conflicting descriptions I have found regarding the vent solenoid.

"It's primary function is to close off the vapor canister during self-diagnostic tests, effectively sealing the system. When the system is not undergoing testing, the vent solenoid remains open, allowing fresh air to enter the charcoal canister."


A youtube video I watched said the exact opposite. It said the vent solenoid is normally open and fuel vapors will enter the charcoal filter from the gas tank. The vapors then pass through the charcoal filter where the bad hydrocarbons get trapped. The remaining clean air then exits the canister to the outside atmosphere through an unenergized solenoid or "open" vent valve.

Here is another description.
"It is the job of the EVAP vent solenoid to allow air into the charcoal canister so that the EVAP system can pressurize and function properly. When the solenoid is opened, air is allowed into the EVAP system to relieve any pressure that may be present as a result of purging the EVAP system."


So what is it. Does air enter or exit the vent when it is in the open position? Or does it do BOTH but at different times.

I was wondering if once the purge valve is activated then the vent valve remains open so fresh air can be drawn into the charcoal filter and "purge" it of hydrocarbons. Otherwise I would think the charcoal filter would become plugged with the hydrocarbons. That is what makes sense to me. If the ECU commands a leak test when the car is running then it would also make sense to me that both the purge valve and the vent valve would be closed at the same time in order to check for leaks. Does that make sense? Does anyone know how often and when does the ECU check for leaks in the EVAP system?

If what I described above is correct then maybe for some unknown reason the vent valve is stuck in the closed position and that would not allow air and hydrocarbons to enter the intake manifold when the purge valve opens. That might explain why when the purge hose to the intake manifold is disconnected then the engine idles smoothly. The purge valve is operating properly because I put my hand on it and I can feel it opening and closing many times in less than a minute. Also I checked it out previously by applying voltage to it and it opened because I could blow air through it with my mouth. I also previously checked out the vent solenoid as well and could see that it was functioning properly. It did not appear to be sticking at all so I don't know why it would stick in the closed position when I reinstalled it, if that is what is happening.

Anyone have any thoughts on what I've written here? I need an EVAP education!!!

Tomorrow I will try to look at the fuel trims as suggested. Thunderstorms and showers are predicted so I don't know. I will also try the injector swap as suggested as soon as possible.

mrvedit

I suspect that different EVAP systems work a bit differently, meaning "conflicting" descriptions can all be right.
This is why the first suggestion was to complete disconnect the EVAP system and plug where its hose enters the intake manifold.
It seems you have done that, several times, and it made no/little difference. Assuming true, lets ignore the EVAP system.
Since we don't know this engine, we have to rely on you to know what every hose going to the intake manifold is - presumably a large hose for the power braking booster, the previously mentioned EVAP hose, likely a hose which is part of the PCV system (which you could also remove and plug), possibly a small hose going to the fuel pressure regular. (Hmmm, if there is a hose to the fuel pressure regulator, a torn membrane could suck fuel in.) There may be other emission system hose, but like very small ones which won't matter.

SnowMan01

Well I have spent a lot of time educating myself about EVAP systems and I am convinced I know how mine functions. The vent valve always remains in the open position until the ecu gives instructions that a leak test is to be done. I thought the vent valve might be sticking closed and therefore when it went into closed loop and the purge valve was opened that no air could be drawn through the canister and therefore the engine was not getting enough air to run properly. In my mind this was borne out when I disconnected from the purge valve and the engine ran smoothly since it was getting all the air it needed through the EVAP hose. I ran more tests today and found that the vent valve did not have any voltage applied to it for the twenty or so minutes I was doing the testing so the air vent was always open during this time. I had the vent valve in the house last night and cycled it back and forth numerous times, not once did it stick.

With everything hooked up properly again the engine ran smoothly until it went into closed loop. I checked the voltage to the canister vent valve and the voltage was zero so the ecu was not calling on the vent valve to close. At this point I could feel the purge valve opening and closing very quickly. Once it went into closed loop the STFT started dropping down to over -20%, sometimes even hitting -27%. By this time she was on the verge of stalling so I stepped on the accelerator pedal and watched the STFT. The STFT went up to between -7% and 0% at about 2,000 RPM. I pushed it up to 3,000 RPM and the trim rose further to between about 0% and +10%. When I took my foot off the accelerator the STFT dropped like a rock, again down to -20% and of course it stalled.

The problem definitely seems to be getting worse. Dies very quickly now when it warms up whereas the last time I had it on the road it only died once while waiting for a green light.

Tomorrow I will be getting new injectors. Sure hope that solves my problem.

I have tried looking at the G5 forum but those guys don't get into the nitty gritty detail. Compared to you guys, who I consider professionals, they are amateurs.

BTW my air filter is very clean and was replaced around Sept of last year. As you can tell from one of my first posts it only has 70,000 KM after 15 years so I don't put many miles on it.

LS1Formulation

iTrader: (13)

Your TB is drive by wire, and there's no idle air control as far as I know. I wonder if your throttle body is screwed up and simply not allowing enough air in at idle. I believe there is a TB relearn procedure you could try, but I'm not 100% positive. My 2013 Camaro has that capability.

SnowMan01

Yesterday was a very good news day for me and a busy one at that. Tons of gardening work and still much to be done. By the end of the day I was too worn out to share the good news of my G5. LS1formulation I should have heeded your advice of posts 15 and 19 in which you suspected the MAF sensor . It would have saved me a lot of grief and some money. The only thing is that I would not have learned so much from you guys.

I got side tracked with the EVAP system so I did not look into the MAF sensor very much because when I disconnected it, whether the engine was cold or in closed loop, it died instantly.

I got to thinking the other night that maybe as a last resort I should pursue the MAF sensor a little bit more before getting into a messy injector swap.

The last time I had a MAF sensor problem (a year ago) the engine light came on and a scan displayed a code for a faulty MAF sensor. The sensor looked clean to me but I purchased some MAF sensor cleaner anyway and gave it a try. I erased the code and of course the engine light disappeared. In a few days the engine light came on again and once again a scan displayed a code for a faulty MAF sensor. The thing that seemed odd to me was that the engine seemed to be operating just fine. I purchased a new Amazon sensor and installed it. Everything was OK with the engine until the rough idling which reared it's ugly head just a month ago. I NEVER suspected that the new MAF sensor could be a problem. I remembered what LS1formulation said about Amazon sensors so I decided to try something. I am a pack rat. I cannot seem to throw things away. I searched my "junk" and low and behold I found my old MAF sensor. I replaced the Amazon sensor with what I expected would be a faulty sensor. I wanted to see if ANYTHING would change with a different sensor. Boy was I in for a pleasant surprise, the engine ran beautifully even in closed loop.

The whole time I displayed the injector PWM, H2O S1, the MAF and STFT live during the closed loop idle of 700 RPM.

The PWM was very steady in closed loop. The average value was in the range of 3.0 and varied only +/- .2 ms.
H2O S1 varied between 60 and 800 mv in a sine wave fashion.
MAF was also very steady between 3.5 +/- .3 g/s
STFT - quite steady with an average of about 0 +/- 2%

All of these attributes except of course for the H2O S1 were much more steady than the previously reported erratic values with the Amazon MAF sensor. The engine coolant temp reached 83 degrees C in the 15 or so minutes the car was running.

I shut the engine off and immediately reinstalled the Amazon MAF sensor while the engine was hot and started her up again. I just had time to read the MAF because the engine started up running very rough and then stalled in no time. The MAF was 10 g/s, a far cry from the 3.5 for the original sensor. I then took out the Amazon sensor and installed the original 15 year old sensor again. I did not bother to watch all the live data this time because the engine returned to a smooth idle. I only watched the MAF live which showed again that it was 3.5g/s.

So it seems the problem is solved. I took her out for a road test a couple of times and all was fine.

The tell tale sign for me that the Amazon MAF sensor was faulty was the high MAF reading. I guess it was over reporting the amount of intake air so the ECU was flooding the engine with fuel. Is this a correct analysis?

I would like to give you guys a BIG thank you for helping me solve the problem. And also a HUGE thank you for bringing me out of the carburetor/distributor stone age. I now have a much better understanding of the modern day workings of the internal combustion engine thanks to you. Thanks again.