LNF STFT and LTFT Tuning

[Google]

stft - ltft = % of error in fueling. if your trying to log maf error by stft+ltft then set your histogram up with Table - MAF Airflow vs. Output Frequency and then load your custom PID ltft+stft, not a sensor

[Ch1ck3n]

Okay, so this morning I decided to open the log file again and give this another shot.. I'm currently not even in the same time zone as my car so all I'm working with is this one test log I have.

I open VCM scanner and decide to start totally anew.
Reset to all the factory imperial defaults of the program.

I then go in to table view and load all the PID's I think i need to scan for MAF tuning, plus some others (STFT + LTFT ([PID.6]+[PID.7]), the PID for Mass Airflow (HZ), Mass Airflow (LB/MIN), air load, ve airflow, etc) .



After I set up my table view I went into my histogram display and started to set these up. I deleted all of the histograms out of the display so only the histograms I want for MAF tuning are there.

The first histogram I set up is STFT + LTFT vs Maf Correction Table.

Set my custom PID as the plot value and then in the right tab, set the table to have the PID plotted against as the MAF Correction table.



Then I go to my next histogram and set it up for my LTFT+STFT PID to be plotted against the MAF Airflow VS Output Frequency table (I left all the values for the columns in one histo cause I'm too lazy to change it all until I at least figure out what I'm doing wrong ).



So as that's all I should need to get started, I proceed to load the log that I've been working with and get this.





And my table view has for some reason dropped all of the added pids and sensors i added.



Any ideas of what I might be doing wrong?

I had considered it might have something to do with the fact that the cfg Ihave now wasn't the one loaded when I ran the scanner on the car for the log. Is that a possibility?

So this is a big huge writeup on what I'm doing and I'm sure I'll get a reply that will tell me I'm an idiot and point out the exact problem almost instantly.

Sorry for the huge rant and long, fruitless post, I'm just getting kind of frustrated by this, not really sure why I can understand all the tuning aspects and the editor itself but the scanner is kicking my ass.

[otherguy84]

The fact that you are now using a different config file sounds like your issue. When you create a log it records the "table" entries and plays them back to you based upon your histogram setup etc. Unless you log the car AFTER you add the entries in the table the log has no record of those values to manipulate and display.

[Ch1ck3n]

Thats kind of what I thought too.... I guess I'll just have to wait till next week to give it another shot then

[tazz]

Both. When you unplug the solenoid that makes sure the LT freezes, so if you do it at 0 after a flash... well you get it I hope

Need to mention I have 07 Auto Pontiac Solstice GXP.
Ive noticed this.
After a write entire calibration both my trims will set at 0 at idle for about 2min until it gets out of open loop. In closed loop the STFT's really go high +20% but adjust down over the next 2 min to -.08+.08 at idle. My LTFT's will still be 0 until I drive for a while then settle in over the next 50 miles or so to there resting place. Once my LTFT's settle in the STFT's bounce around to maintain 14.7 with the settled LTFT's.
When I do a write entire and unplug the sloenoid right after the flash it seems like the unplugged solenoid freezes the table used for the open loop trims since my STFT's stay at 20% when at idle no matter how long or many miles driven they never come down.
?
Should I wait to unplug the solenoid until the open loop tables kick in and then freeze the tables and trims then since at that time the LTFT's are still 0? Since i usually drive my vehcile when it's open loop this way should be more accurate? If I don't wait won't my trims be skewed higher than usual and throw off any logged STFT's that I will use to figure the needed adjustments?
Thoughts comments.

[tazz]

Anyone
I just don't want to go chasing my tail on this since Im not even sure I can dial the trims in as perfect as some have stated. Ive seen 5% swings in my LTFT's on a day to day, week to week basis while everything else has stayed the same including commute and driving habits. For my vehicle it seems to all depend on temp and weather as to where my LTFT's will end up. Some days they can be almost 0 or -.08% while other days it can be at 5% and my MAF Frequency and correction have not changed.

[Ch1ck3n]

It's always based on temp and weather, the ecm is just compensating. Your best bet is to just work out an average over the course of time. It will never be perfect 0, you can only get so close.

Got my logging problems all sorted out and im sitting within 3% on trims.

[tazz]

Thanks
So going back to my original ? on if I should unplug the solenoid after the it goes into closed loop and the LTFT is still sitting at 0 after the STFT have settled down and log or leave if locked in open loop with the STFT way high and log? I would think unplug after the STFT have settled down in closed loop since that what my vehicle operates in 98% of the time.
Your saying just say F unplugging the solenoid and just log STFT's + LTFT's and average them out over time correct since they will never be perfect.

[Iam Broke]

Like was posted before, the LTFT's will compensate for temps & humidity changes. I have found that if I keep the LTFT's slightly positive, (+.8 to +1.6) the trend for the STFT's has always been to run slightly negative (-.8 to -2.1) resulting in LTFT+STFT being +/- 0 to 1% on a 75*f day with avg humidity. It's taken me months to get the MAF correction to this point, but I'm pretty anal.

It's a slow learning process to watch the ECU learn & hunt on the fuel trims. They seem to settle in after 4 fairly long key cycles so I try to make a minor change only once or twice a week and log a few hours. I drive 2 hrs a day so it's pretty easy to get a good log.

Keep in mind with my MAF relocated to the cold side CP my ECU pulls fuel on accel, the stock location usually tends to add fuel. (positive STFT's) YMMV.

[Ch1ck3n]

Thanks
So going back to my original ? on if I should unplug the solenoid after the it goes into closed loop and the LTFT is still sitting at 0 after the STFT have settled down and log or leave if locked in open loop with the STFT way high and log? I would think unplug after the STFT have settled down in closed loop since that what my vehicle operates in 98% of the time.
Your saying just say F unplugging the solenoid and just log STFT's + LTFT's and average them out over time correct since they will never be perfect.

IIRC (please correct me if I am wrong), the LTFT's tend to look for trends in the STFT's. Unplugging the purge solenoid effectively removes LTFT's from the equation so that you can concentrate solely on dialing in your STFT's. Once the short terms are pretty close, the long terms will follow suit when you plug the solenoid back in.

I unplugged my purge solenoid and tuned and now have STFT's within +-2% and my LTFT'S within +-.07%.

[Iam Broke]

Yes, you can do it that way if you wish. I choose to leave the Purge sol connected so I use the LTFT's to tune the entire MAF freq across the board, then work on the MAF correction for STFT's from the LTFT+STFT histo.

I don't like making the MAF freq table lumpy with individual cell changes. I tried that when I started tuning the 'balt. I offset the entire curve by a given percentage to null the LTFT's out. Just works better for me in my mind I guess.

[Ch1ck3n]

Do you leave the MAF correction table alone and just use the freq table as its own correction factor then?

I tried adjusting my trims without unplugging the solenoid and I just seemed to be chasing my tail.

[Iam Broke]

Do you leave the MAF correction table alone and just use the freq table as its own correction factor then?

I tried adjusting my trims without unplugging the solenoid and I just seemed to be chasing my tail.

No, If you want to start over.... Set the MAF correction table to all 1's. Leave the purge sol corrected & drive for several miles/long trips. See where the LTFT's settle.

Multiply the ENTIRE MAF freq table by the % needed to trim LTFT close to zero. I keep a stock copy to base changes off of instead of compounding changes. In other words, if I shift it 3% positive and find I need 4%, I'll go back to the stock one and add 4%. I try and keep the shape of the curve as close to original as possible just change the offset so to speak.

If LTFT's go to +3.1, multiply the entire MAF freq by 1.03 to bring them down, if -3.1, multiply by 0.97. Rinse & repeat a few times over long trips.

Once you are pleased with the LTFT's, move on to the MAF correction table using LTFT+STFT histo to dial in the errant cells. (Average readings of all cell hits) Ignore cells with less than 50 hits or so and try and keep the changes between adjacent cells as smooth as practical. It takes many long trips to see the trends. Keep records and look for where you may be chasing your tail on short trips ect. I use a 50 mile commute to tune mine so I get a lot of data.

My LTFT's only move ~.8 now during a long drive and the LTFT+STFT's are +/-1% depending on weather & humidity.

It takes several long trips before the ECU settles in, you can't do this well in a day IMO. You can get pretty close within a few logs, but if you have major bolt ons & an intake it takes a while.

Practice makes perfect.

Edit... The reason I use LTFT+STFT for the MAF correction table is in case the LTFT's drift while logging I still have a good idea which cells in the correction table need work.

[Ch1ck3n]

Would you say its more of a personal preference kind of thing or is there a difference in methods performance-wise?

I've been using the technique as prescribed by BYT and have my fuel trims right where I want them, but if it will improve response and power I'm down for anything.

[Iam Broke]

Nope, just different ways of getting the trims perfect. If yours are dialed in no need to start over. You made it sound like you were having issues.

[Ch1ck3n]

I was at first without unplugging the solenoid, but after doing it byt's way it made things pretty much idiot proof.

I am a fan of smooth things myself, so I was just curious about your method. Whatever gets the job done though eh?

[-MD- HHR]

Thanks for posting this!

[DoMiStIc_RuSh_06]

Is there anyone who can post up a copy of the histograms that we are suppose to be using. I tried to make them but i dunno if i did it right.
Attachment 25825

[Iam Broke]

Use Shabby's LNF config from the repository.

[DoMiStIc_RuSh_06]

ok....got it. Thanks

Still trying to figure out how to modify the trims so i dont mess them up.....Im still learning the features of hpt too, so that doesnt help.