LS3/E38 crank priming tabel / how to control fueling during cranking

[hjtrbo]

That's not good either. It's a pain in the ass that it's there and you need to tune for it, or at the very least make sure you're rich when heat soaked as opposed to lean. Would hate for you to do a lean WOT run with a hot engine and knock knock. It's generally tuned at the end once everything else is locked away.

The MAF tuning should only take half a dozen good driving methods logs. If you're always adjusting fuel and then reflashing then adjusting again you'll get no where. Which is what you're finding out I'm picking from your recent findings.

Set up a filter on your graphs so injector tip temp >65 degC is ignored and combine that with ignoring throttle transients.

Here is an example of the filter I use, it is very conservative and only stable throttle gets through 98% of the time.

------------------------------------------------
Closed loop
ITT < 60 deg C [] ignore 0.5s either side of a shift [] smooth stable throttle [] ignore idle, coast, PE & DFCO
[6215.241]<60 and [2517.161.avg(500)]=0 and [2517.161.avg(-500)]=0 and (abs([50090.156.slope(1500)])+abs([50090.156.slope(-500)]))<2 and ([6310]=7 OR [6310]=9 OR [6310]=14 OR [6310]=15)=0
------------------------------------------------

------------------------------------------------
Open loop
ITT < 60 deg C [] ignore 0.5s either side of a shift [] smooth stable throttle [] throttle > 75% & in PE (ignores transition into & out of PE)
[6215.241]<60 and [2517.161.avg(500)]=0 and [2517.161.avg(-500)]=0 and (abs([50090.156.slope(1500)])+abs([50090.156.slope(-500)]))<2 and ([50030.91]>75 and [6310.avg(250)]=9)
------------------------------------------------

[TRStech]

First of all, thanks for sharing this knowledge! This is way beyond my current skills, but I will take this as a starting point to dive into that.

Concerning the Offset, true, I had this thoughts to. But since Im currently just sitting in idle to get airflow corrected at the very low end of the curve I expect this to not cause serious trouble. Once I get back to driving, I will bring the the values as found back in. Do you think this is a resonable strategy for now?

This topic came out while I was in the process of correcting the very low end of the curve manually, since I wanted to smooth it a little over the point where I wasnt able to get the engine to in the drive logs, means lower then 550rpm or with other words below ~1850Hz MAF. I did this by copying the lower end of the curve to Excel to get a resolution of this section and then I flantened out the bump which was in there at the low end. I did this by correcting the area which was not reached during driving. I left the other values alone.
Later, when I tested those settings I started to take notice of the fuel trims in the area slightly above, so lets say up to 3000Hz, whats also an area the car will run in for a good amount of time during the test. After several driving logs and corrections I still had a pretty constant -4.7 LTFT in this area and I started an attempt to correct them, not necessarily expecting to get it done without driving. It was just a try. But then I stumbled over the wired LTFT behavior I described. Today I will see if the "rich after reflash" is actualy the root cause by reflashing several times between two calibrations with zeroed offset and whatching the changes in LTFT for both.

My issue is, that very long sections of the driving cycle are taking place in the low airflow region. There are also some idle sections and coast down when deaccelerating is done without gear/in idle. This means, a pretty good percentage of the time the car runs in idle and therefore I want to have that perfect. Also, from my understanding, each transission means, that the ECU is leaning more and more towards base mapping and away from the trims. So each section of acceleration and deacceleration are critical, since an imperfect mapping in this region will cause to the mixture to drift away from ideal until a phase of constant speed is entered and the trim learn takes over again. I hope this understanding of the ECUs behavior is matching the reality.

[hjtrbo]

But since Im currently just sitting in idle to get airflow corrected at the very low end of the curve I expect this to not cause serious trouble. Once I get back to driving, I will bring the the values as found back in. Do you think this is a resonable strategy for now?

You may be able to get away with. Only way to tell will be if you are still going around in circles after a few goes or it converges to good.

Personally, in the 1 drive cycle, first start of the day (cold engine), I'd do a long idle up to operating temp, then go for a long drive at various loads observing good throttle behaviour. That will give you an excellent set of data. Wash the car and put it away. Have a beer at night time, look over your log with good filtering, make adjustments in the tune using 'multiply percent half', then rinse and repeat the next day. Do this over a few weekends and you will be dialled in. There will always be some variation day to day. Always want to hit a little on the richer side so +0% to -5% trims. This way during different atmospheric conditions, you will be safe when in open loop PE.

There will always be some discrepancy in the trims when idling in gear vs out of gear. If you're with 5% in both those situations I would call it good.

My issue is, that very long sections of the driving cycle are taking place in the low airflow region. There are also some idle sections and coast down when deaccelerating is done without gear/in idle. This means, a pretty good percentage of the time the car runs in idle and therefore I want to have that perfect. Also, from my understanding, each transission means, that the ECU is leaning more and more towards base mapping and away from the trims. So each section of acceleration and deacceleration are critical, since an imperfect mapping in this region will cause to the mixture to drift away from ideal until a phase of constant speed is entered and the trim learn takes over again. I hope this understanding of the ECUs behavior is matching the reality.

Yes, as the engine transitions away from a stable air flow situation, there is some biasing towards the GMVE air flow. If you log maf airflow and gmve airflow and dynamic airflow you will see when this is happening. When the air model biases towards the GMVE table, it will also contribute towards the injector pulse. Hence why we always tune MAF and the GMVE.

This might interest you if you're good with formulas...
https://forum.hptuners.com/showthrea...eous+VVE%2FMAF

[TRStech]

Hi all,

I found another culprit during my measurements. Ive got issues with a rich condition, as it seems during rpm falls down to idle from a higher rpm or directly after. It seems that CL is struggeling to deal with that. both (pre and post converter) O2 sensors show high voltage starting at the middle of coast down from about 2000rpm. This condition persists for arround 2-5s. After that, a lean condition occures for roughly the same amount of time before CL oszilation starts again as usual.

This seems not to happen when rpm is not going completely down to idle. I have to do further measurements to confirm this, but it seems that the rich condition does not occure as long rpm drop is stopped above idle rpm by the accelerator paddle or the idle control rpm function of the scanner. Therefore I expect some kind of idle transition which causes this issue.

The consequence of this effect is a portion of 5s in which HC goes up from ~5 to ~500 ppm, CO from ~0,01 to ~2,5 Vol%. Since the city cycle for my emissions test contains 9 points were engine is brought to idle while highway part has only one, this also mirrors my issue with way to high values in city cycle but acceptable in highway.

Does anyone have any suggestions which values or tables i can look on to tweak this?

Thanks,

Soeren!

[07GTS]

possibly transient fueling there is the evap factor and impact factor, they control the estimated loss due to fuel on the port walls and evaporation from sitting in the hot ports ect plus all there variables, if durin a decel the evap is to high it might be adding too much in so i could try that first see if u notice anything or u could increase the map/tps values higher so it dosnt count a small change as a transient event, the are quite a few things to do on that tab

[TRStech]

Want to let you know what I found out concerning this specific issue: It was the base airflow beeing high.

I corrected the MAF airflow for idle conditions pretty much to optimum and it turned out beeing pretty low. (5.5-6.0g/s @550rpm) The base airflow for idle was higher then that by arround 15%. I tried to get data for a new base airflow table, but later I decided to just multiply the factory table by 80% and this solved my basic issue with the rich conditions after dropping back to idle. Idle drops, especially whith cold engine, are a little less stable but still completely acceptable and the downstream O2 sensor spikes as well as the emission readings for those transients were massively reduced.

I did some road testing and also there the post O2 rich condition is gone. Only thing Im still fighting is the effect, that there is still a short rich condition when the engine was in Coast, then clutch is opened, rpm drops and short before idle (when the ECU is adding fuel to stop the rpm drop) the clutch is closed and rpm go up again without any load requested by the driver.

I expect it would be hard to get also this specific transition to perfect, but furtunately I dont think this will be necessary, since such a condition should not be seen in the defined driving cycle.