I found this on another website. It clearly explains the operation of SFTF and LFTF. very informative ...


If you have the means to log over the OBD-II port you can watch what the short term and long-term fuel trims (STFT and TLFT) are doing. These are the corrections that the PCM is applying to add or remove fuel based on its O2 feedback. There are different cells (spots or ranges of correction) based on what MAP voltage level you're at and if it's above or below ~2k rpm's (as well as separate corrections for idle/decel). Short term is on the spot temporary changes of up to 25% correction to add or remove fuel. If the PCM continually has to add 10% more fuel under certain operating conditions/ranges (above 16psi boost and above 2k rpm's, for example), the long term would get adjusted so the short term can go back towards zero correction. The PCM always likes to keep short term as close to zero correction as possible. Long term can also provide up to 25% correction, and both together can make up to a total of 50% correction. The PCM makes these corrections by altering the injector pulse width which is the time the fuel injectors are "on" and flowing fuel.

Both the short and long term fuel being cell based means they use a matrix based on RPM and load (determined from the MAP sensor) to determine which cell the engine is operating in. Each cell can store its own fuel trim correction. Under normal driving with open throttle the engine management uses 12 cells, referencing MAP voltage against RPM's. It basically uses low RPM (below around 2k rpm's) and high (above 2k), referenced against 6 different MAP voltage set points. There's also a deceleration and idle fuel trim cells for above and below 2k rpm's.

So as an example, one cell might have a range of 3.6V to 3.9V on the MAP sensor with the engine operating above 2k rpm's. So if you had a LTFT correction of 10% it would be adding 10% fuel when the boost is between ~10-12 psi (3.6-3.9V). As soon as the MAP voltage goes over 3.9V you'd enter the next cell which may have a different LTFT correction.

When you're looking at fuel trims remember that there are only two different reference points for rpm's in regards to the fuel trims. There's one set of fuel trims for above ~2k rpm's and one set of fuel trims for below. For above and below that 2k rpm point there are six different fuel trim corrections stored (called cells) that is MAP voltage based.

So if you're at 4k rpm's and 3.5V on the MAP that would fall in a certain cell. At 5k rpm's and 3.5V it's still in the same cell (because the MAP voltage didn't change and it's still above 2k rpm's). Once the MAP voltage changes a certain amount or the engine drops below 2k rpm's you may drop into a different cell that may or may not have different fuel trim corrections. A lot of the older PCM's used many more tables and cells to store different fuel trims. My educated guess is that because the NGC (Next Generation Controller) used in the SRT-4 is model-based and a lot more powerful than older PCM's they can rely on it being able to use O2 and other feedback as well as calculations to make faster, more accurate "live" decisions on changes to the fuel trims without having to rely on extensive cells for different rpm and MAP voltage ranges like the old PCM's.

These short and long term fuels trims can be viewed on most scan tools or data loggers that can monitor generic OBD-II information and will be represented as a percentage. A positive number (such as 15%) means the PCM is adding fuel and a negative number (-12% for example) means its removing fuel. Some older pre-OBD-II cars like GM use a different system based off a block learn value between 0 and 255. In those types of systems 128 would be the middle or average value with zero correction and a higher number would mean the PCM is adding fuel and a lower number meaning that it's removing fuel.[/quote]

Quote Originally Posted by Blackbird R/T
The long-term is adjusted by the PCM to keep short-term closer to zero/no correction. The PCM adjusts long-term based on the short-term adjustments, and the short-term is based on the O2 feedback. And just because you have +5% fuel in short-term doesn't mean it will automatically change long-term to +5% either and short-term back to zero every time. You can use short -term for drivability issues and light acceleration/boost to see what your fuel system is doing. And if you wanted to you could take a look at WOT in open loop to see if the long-term values are ever changing based on adjustments made by the PCM in closed loop to the long-term.

The short-term and long-term can both add and remove up to a total of 25% of the pulse-width. That's what the numbers you're seeing on a scan tool mean if it is displaying a percentage value. A positive number is the PCM adding extra pulse-width (add fuel/richen) and a negative number is removing pulse-width (remove fuel/lean out). Pulse-width is the time in milliseconds the injector is on and can also be converted to duty cycle.

After startup the PCM uses short-term correction (which means O2 feedback) in open loop warm-up when the engine coolant reaches ~30-35°F. When the engine goes into a closed loop mode and is above ~170-190°F it will adjust or allow the long-term adaptive correction values to be updated based on the short-term direction of movement/correction. It's important to note again that the long-term adaptive memory and cells are used in all operating conditions (including open and closed loop modes) but only change when above that temp, in closed loop, and when the engine has been running for over two minutes.

The long-term values are adjusted by the PCM in order to keep the short-term as close to zero (no correction) as possible. It probably has an algorithm in the programming that alters the long-term if it continually sees short-term in a certain cell having to add/remove fuel, how much it add/removes, and how long the short-term has been making corrections. The long-term can add/remove up to 25% pulse-width, so if the LTFT max's out, the short-term can still alter an additional 25% when in closed loop (O2 feedback) for a total of 50%. But when you got to WOT (open loop/no O2 feedback) it will only be using LTFT and subsequently a max of 25% change in pulse-width.

During acceleration the car is still in closed loop mode with the O2 feedback being used (which means short-term values are adjusted). This might occur when you roll onto the throttle while cruising and start building boost but aren't at WOT. That's a likely scenario of how the short-term and eventually the long-term get adjusted under boost. When you go to WOT it reverts to open loop and only reads (not modifies) the LTFT. Since the last cell range is for 3.9 volts and above, that correlates to about 12 psi and above on a stock/S1 car.

Also the short-term memory is lost when you shut the car off. The long-term adaptive memory is held as long as you don't disconnect power NGC.
[quote="Blackbird R/T"]Long-term isn't just some some average, but then again it is, and it's only adjusted based on the short-term movements. How's this for an example. The short-term goes to +5% in a certain cell to increase the injector pulse-width to richen the mixture. The long-term is at 0% (no) correction but after the NGC sees it keeps having to add 5% in that cell it eventually adjusts the long-term to 5% and the short-term back to zero. You're still getting the same 5% extra fuel.

Now the car starts to run a little lean again when operating in the same cell. This time it has to add 3% pulse-width. So now you have 3% in short-term, plus 5% of long-term, for a total of 8% extra pulse-width being added. If it continues to have to add that 3% it may bump the long-term to 8% and short-term back to zero. So short-term may never get to as big of a correction as long-term. It could also go from long-term adding fuel in a cell and suddenly the short term goes negative in that cell trying to remove some of that fuel. In that case both numbers might converge towards the baseline of none/zero correction (the PCM decreases LTFT, the mixture leans, then the O2 sensor tells the PCM it can increase the STFT from a negative value back towards zero).