What is a good amount of time for LTFTs to settle

[David-L]

I am wondering what length of time or how far would you typically recommend to drive for LTFTs to become steady and change very little? How long are LTFT averages stored? I know they can be reset by command or by disconnecting the battery or a few other ways, but assuming nothing happens to reset them, how long do the averages go back? Typically/Generally speaking of course, I realize every ECM and situation is different. A week, a month, 50 miles, 200 miles? It has to be some finite amount of time or the values would never change based on the STFTs.

[blindsquirrel]

LTFT is based on the running average of the STFT. LT just adjusts slower. If ST is positive then LT will drift positive, no matter how long it's been since trims were reset.

You can experiment with this, you know, to get a feel for how it works. Create a vacuum leak and watch how ST and LT react, fix vacuum leak and watch how they return to normal.

[Matt Vardaman]

You can pump the brake pedal rapidly with the engine running and it will make enough of a vacuum leak to drive the trims 15% positive or so. Under Fuel>Oxygen Sensors>Ltft you can check out all of the settings the pcm uses to drive the long term trims. After tuning, my ltft's settle out after about one day. I do drive a lot though.

[David-L]

Thanks for the replies blindsquirrel and Matt! Okay, so it sounds like its a relatively short term average on the order of possibly hours instead of weeks or thousands of miles. I knew it is a moving average, I was just unsure of how far back in time the average goes.

So what would be a reasonable amount of driving time to assume the trims in and near the cruise areas are at a good stable average? a few hours?

[blindsquirrel]

It depends. Why are you asking? Just general curiosity or are you thinking there is some way to use the single, lumped trim average number for something - something other than an overall health check? Because that's all it's good for. Spotting problems and general trends.

If you want to use trims for tuning and want a longer-term average, in your graph set the required cell hits to a higher number, like 100 instead of 25.

[5FDP]

I computer could learn new LTFT's in 10 miles pretty easy if you asked me. I've repaired plenty of cars with problems from massive intake leaks, injector problems or low fuel pressure problems. They all ran like crap and once the problem was correct it really didn't take long at all for the computer to see that everything is back in line and start fixing the fuel trims.

[David-L]

I'm kind of asking for several reasons. General curiosity and checks/validation are the primary reasons. I did use the sum of the ST and LTFT with a bunch of filters for steady state data and 150 cell hits to tune my MAF in the idle to light load range and it seems to have worked incredibly well in just one iteration. I stayed well away from heavy load/WOT/PE areas.

[blindsquirrel]

So create a small vacuum leak to see how long it takes for trims to move from short term into long term, and how long it takes to correct after the leak is fixed. You can command the purge valve open if you want to see it correct for a rich condition.

The brake pedal thing is good for a quick check to see that the trims and O2s can react as expected, not so good for watching the system work over a longer time and seeing where/how the long terms settle out.

[retlocer]

i am no expert but i searched the same info on google & learned it takes 50 minutes or 100 miles. & i find it believable as trims tend to settle in around those parameters at least with a p01 running 3 bar sd

[blindsquirrel]

Oh, well, found it on Google then it must be true!

It's not that simple. It depends on the conditions. How quickly values transfer from short term over into long term depends on how much of a correction it needs to make. It depends which FTC it's in. And more.

Just create a vacuum leak, watch the trims and how long they take to stabilize. This does not require anything but an EFI engine, a scan tool, and a mildly resourceful brain.