I am just now seeing your fuel trim discussion. I am going back and reviewing your posts now to get caught up.

George

 

Forgive me if some of these comments cover well understood ground but lets see where we get:

There is the potential that some of this is PID selection related, IOW is LTFT really the proper LTFT? I don't know these software packages well enough to help you there but we just have to make sure that we don't chase a non existent problem due to parameter selection. That said:

LTFT is the computers attempt at establishing an integrated (over a long time period) compensated fuel delivery to maintain the desired 14:1 air fuel ration (minimum emissions). That's all that closed loop (fuel controlled by the O2 sensors) cares about. If performance sucks while compensating, that's not considered, just minimum emissions. If a bad fuel injector is dumping raw fuel into one cylinder the computer is leaning that bank and the other cylinders will suffer. As the engine ages (compression, etc) , a fault (vacuum leak, bad valve, etc) or a sensor drift becomes persistent, then the computer sees this constant STFT trend and bakes that into LTFT so that STFT spends most of its time closer to zero. If something changes like fixing a vacuum leak then as that unmetered air goes away, STFT would immediately go negative to remove all that extra fuel that was being added to match the extra air but over time the computer would see this persistent change and LTFT would start drifting more negative and STFT would drift in unison back towards zero. This happens slowly. In the OP's graphs LTFT changes many percent in a few seconds which is not normal, at least not at a constant load. If the fuel values were recently erased then some cells may jump around under load changes (not sure on our computers) Did those LTFT rapid changes occur while driving? Same question for Les' more stable LTFT numbers? The question is whether this is the right PID or is LTFT moving around that fast as driven by the computer? Les' LTFT's are very stable as you would expect over a time frame of a minute or two.

STFT's are pretty stable in a steady state engine condition with no change in the load (idle or cruising on flat ground). STFT changes more rapidly but under a constant load they usually more around +/-5% or less. When driving around at different loads the STFT graph will shoot around a lot. Unless a very specific condition is being diagnosed or your doing a vacuum leak/propane test I wouldn't put too much stock is STFT's when driving around.

Fuel trim at any given time is the addition of STFT and LTFT (mindful of sign) and should be less than +/-10, even lower on a newer engine. Theoretically perfect is all zeros. When the engine is shut off the computer remembers the LTFT and starts back up using those values. All of this only matters in closed loop (O2 sensor control). In open loop the computer uses look up tables based on non-O2 sensor values like TPS, MAP, ECT, etc.

The first thing to solve is the wildly fluctuating LTFT numbers.

The only way to really assess the O2 sensors is with graphing. Once they heat up they should switch between 0.15 and 0.85 volts, passing through the 0.45V bias voltage (most of the time) at around once per second at idle with about a 100ms rise and fall time. If some of the edges (but not most) go back without crossing 0.45v that's normal. O2 sensors can get sluggish as they age causing performance issues. A common test for this is force lean (vacuum leak) or force rich (propane) and watch the reaction time and ramp of the O2 sensors.

When you floor the accelerator the computer goes into a forced rich mode, over riding the O2 sensors and the O2 sensors will peg rich until you let off the gas. Its the only time other than open loop where the O2 sensors are ignored.


George

 

Thanks George for the second opinion.

 

I have not used fuel trims in the way that you guys may be using it here, graphing LTFT under possibly changing load conditions. This would cause the computer to change fuel cell control locations (the individual look up table locations) on the fly. I have read that in some vehicle types, LTFT can change abruptly when the fuel cell changes under differing load conditions, particularly if a LTFT value has not been computed for that cell yet (cleared memory) but I have not seen this in our vehicles. This morning I went back into the shop manual for some clues. This is not discussed but there are two different LTFT PIDS referenced, one called average and one not called average. I did not know that there were two choices for LTFT until now. When I get a chance I am going to get out my scanners and take a look, particularly on my Snap On which might shed some light on this. I think I understand fuel trims pretty well but you guys present an interesting situation with all of this graphing under possibly varying load conditions.

If the OP's LTFT's are changing dramatically as graphed under a static load then I think there is a problem.

George

 

Yeah, I thought you have a very good handle on the fuel trims. Why I think it is interesting to get your opinion. And here we are trying to help someone who will actually listen, comprehend, and diagnose!! Doesn't happen very often.

 

Ya this OP has taken direction well and learned a lot. He will go far with that attitude.

George

 

Thanks George and Les! So to answer you Les, I have thought about the poly mounts vs misfire scenario quite a bit and can say for almost certain there's something wrong with the engine still, the poly mounts may have a different feel but going off what we have found with the fuel trims I think something is definitely wrong. I can almost hear it missing at idle I swear, maybe I will get a video later tonight so you guys can hear. When I'm idling at red lights my heart sometimes starts beating fast just hoping it doesn't stall out LOL(it never has), it's definitely running rough though... About the starting issue, I said before I need to prime it every time before starting it, well I have been testing this more lately and it seems more hit and miss if it will start or not. I have come up with this, it will occasionally give me a long crank time, maybe 5% of the time and I haven't been able to narrow it down into a pattern. This is what usually happens, I will get in my truck to start it, prime it on ACC, then it will just not start, I will prime it again and let it crank for a second (long crank) then it will start. OR I get in my truck to start it, try and start it with no prime and it doesnt start, then turn it off and back on and crank for a second (long crank) it will start. And I can't narrow it down to anything particular happening, it could happen hot or cold, and also wet or dry outside. The next time it does it I am going to try and not cycle the key back and just keep cranking on the first time to see if it will eventually catch or not. I usually like to prime it before starting anyway as I thought this was giving me a higher chance of it starting but I'm not so sure anymore so for the next few days I will try just always cranking the engine right away without priming at all and see what I come up with.

Now thanks for chipping in here George alot of good information there, so the average fuel trims is what I was having trouble with in dashcommand where they were reading 0 all the time. I believe now I am reading the true fuel trim values, I may even go check my wired scanner vs torque while keeping the engine off so they dont change and see how they line up. I have a video for you here with some comentary so it's a little easier to see whats going on... But yes, the long term trims are changing durastically when I change the load on the engine. It seems it will happen mostly while I'm driving only, in the video I was holding a constant rpm going a constant speed and you can see the long term trims pegged way positive and the second I let off the throttle they go back down to near 0. Then in park holding constant rpm this doesn't happen as much. Be advised in the video where I was holding "2500 rpm" in park for a few seconds it was really not a steady 2500 rpm, the 1500 and 3000 rpm measurements were the most accurate. So I'm thinking about what you said about the different fuel cell controls and each having there own long term trim value, I think this answers me in the video where I'm saying "it seems like the long term trims are changing with the throttle". I am thinking the computer is changing fuel cell control locations when I am driving under different conditions and each has its own long term trim value, I am just saying this based off what you said George so please correct me if I'm wrong. And at last guys please call me Zach no need for OP.

 

I'll try to look at this tomorrow to further understand this LTFT behavior under varying loads and the issue of the two different LTFT parameters. Zach: what are your choices for fuel trims called in your scanner?


George

 

So with my wired scanner it only shows long and short term bank 1 and 2, on torque it shows the same with the addition of "fuel trim bank 1 sensor 1" and its not on the picture but "fuel trim sesnor5" with the latest value of -3.12%, I'm not too sure what these last two values mean, maybe you guys do. On dash command it has the GM pid long and short term fuel trim average for bank 1 and 2 and also the SAE long and short term trims for bank 1 and 2. The picture in dash command is with the SAE fuel trims and they seem to be recording the same as everything else.

 

Zach, I think I can see what is being talked about at about 5:00 in the video. LTFT went to completely different values in park depending on the RPM. IIRC I think mine usually went back to near zero at steady but different RPM. Also the adjustment (cycling) is kind of lazy in comparison / not as fast at higher RPM. I wonder if a sensor input like MAF or something else could be causing a problem. Maybe we should go through and check the function of each input sensor and rule them out as possibilities. Pretty easy to do with a scanner. Also I still wonder about restricted exhaust, but that will be evaluated with the snap (running) compression testing you are doing this weekend. Seems the starting problem is independent of the roughness.

I have heard about some of these V6 engines being a "shaker" because of the weird balance required on these even fire V-6 engines. Also why the engine mounts get broken so often (IIRC both of yours was broken). As I understand it they custom tailored balanced these particular engines at factory by adding weights in the front balancer if they shook. I just want to make sure that shaking roughness is not being interpreted as a misfire. That's where the cylinder balance test would be very handy to confirm if a cylinder is actually doing it's job or not. Does it idle bad all the time, and does it idle the same each time at same RPM? If so, we may be able to figure out a way to do the cylinder balance test by disconnecting the IAC and then shorting each spark plug wire one at a time and evaluating the RPM drop. Trying to envision a good way to short out the wire while running without getting the P*S shocked out of you.

George, please continue your Fuel Trim studies!