The narrowband trimming is primitive in its approach. The narrowband sensor, to put it simply, has three states: Rich, Good, Lean. The ECU doesn't know the magnitude of the rich/lean state. Say the sensor output sends 'rich'. The ECU has a look-up table, and trims by a preset value. The sensor sends 'rich' again, so the ECU trims again. Then the signal may be 'good', so the nothing more happens. At another point, it may send 'lean', so the ECU adds a bit. It's an iterative approach, and the throttle must be in a steady state for a long enough period for the ECU to trim to the 'good' point.
Unfortunately, the 'good' point is not really that good. It's usually a small range with the stoichometric point of 14.7 smack in the middle. Great for meeting emissions standards, not so good for lean best torque.
Truth be told, using a wideband device, whether it be a fixed sensor or in a dyno, is really the only way to get your AFR right. Second best is looking at the plugs. Reading the plugs is still a valid way to do it, but you end up with a compromise.
The narrowband log will just be a number that has no meaning. It may as well be counting house numbers.
If you don't want to invest in a wideband setup or some dyno time, may I suggest turning off the lambda in the map and read the plugs and use seat of pants.
--- --- --- --- --- --- --- --- --- --- --- --- --- --- --- --- --- --- --- --- --- --- --- ---
..[You must be registered and logged in to see this image.].In GRiSO we trust!
.