Fuel trim refers to adjustments being made dynamically to the base fuel table to get the proper ratio of fuel to air. Short term fuel trim (STFT) refers to adjustments being made in response to temporary conditions. Long term fuel trim (LTFT) is used to compensate for issues that seem to be present over a much longer period. Fuel trims are expressed in percentages; positive values indicate lean (add fuel) and negative values indicate rich (subtract fuel).
Fuel trim is generally calculated by using a wide set of data values, including O2 sensors, intake air temperature/pressure (or the more elegent MAF sensor), engine (coolant) temp, knock sensors, engine load, throttle position (and change in throttle position), and even battery voltage can effect fuel trim. Long term fuel trim generally should not exceed +- 8%.
My knowledge of the inner workings of the Honda EFI system is very limited (on the scale of how much there is to know, it's a beautifully intricate system). But in my experience, the ECU uses the values calculated via the LTFT to clean up the fuel under open loop. So even if your fuel tables aren't perfect, the ECU will fix it for you. This is why you'll notice that after flashing the car, it'll begin to feel more smooth and WOT runs will feel more responsive after a few days of driving. Also this is where calibrating the MAF properly becomes extremely important. If the MAF is too far off, the car will constantly be adjusting the LTFT on a dramatic scale and you'll experience erratic performance and fuel economy.
Comments
My experience with MAF calibration
As per Mucter's insight, I have been calibrating my MAF for about 3 weeks and his comments are spot on! I did about 36 datalogs (used the for MAF calibration process @ http://www.accesstuners.com/tech-resources/maf-calibration-spreadsheet-and-demo-video). I basically didn't add or pull any fuel from the fuel tables in ATR but rather tweaked the MAF sensor table to adjust the LTFT. I noticed that once the MAF was calibrated the subsequent datalogs showed my AFRs @ WOT (wide open throttle usually 78.04-78.43) were in the range of 13.5 - 13.9, all this while my LTFTs were in the range of -.8 to +2.9. By the time I got to reviewing my AFRs @ WOT for the 9th datalog I saw that they were about 13.1 - 13.3. This is a direct correlation to the insights shared by Mucter. Of note is that the ATR Helpfile mentions this but it is sort of vague if you have never done this before. Now it makes sense!
I notice the same thing , it
I notice the same thing , it just takes some time