Ali,

this is the last time I will try to explain this to you. Feel free to argue against it as you have with the same arguments instead of looking at actual data and trying to understand the system. I won't argue with you anymore on this subject however.

You seem to be missing the main point about knock correction that I have made over and over and you even stated that you knew before I explained it.
Knock correction is the combination of values taken from the ignition correction map which are then adjusted by multiplying them by (IAM/16) combined with values taken out of the fine learning table. This combination is what you see as "knock correction" in deltadash. So as you can see, the values in the fine learning table are a direct part of the knock correction values reported by DD. When there is knock outside of the range of IAM learning, the fine learning table is adjusted by the ECU to pull enough timing to stop the knock. This adjustment is also reflected in the knock correction values. For example, if the fine learning table was full of zero's and the IAM was at 16 (no knock), the knock correction values in Keegan's logs would stay at around 7-9 degrees on the top end. Now in reality, Keegan's ECU raised the IAM to 16 while he was driving within the learning range of the IAM and then the ECU sensed knock on the top end where it is outside of the learning range of the IAM. Since its outside this range, the ECU does not adjust the IAM, but rather has to use the fine learning table to pull timing on the top end. It did so, pulling up to 9.5 degrees of timing on the top end in response to knock and as you can see in his logs, these negative values from the fine learning table are reflected in the knock correction values as they dip into the negatives on the top end.
Remember that the fine learning table values are a direct part of the knock correction values in this way:
knock correction = (IAM/16) * Ignition Correction Map(load, rpm) + Fine Learning Table(load, RPM).
So when the knock correction values go negative on the top end while the IAM is still at 16, you can see that the only way that can happen is because the Fine Learning Table part of the knock correction equation takes on a negative value with a higher magnitude than the (IAM/16)*Ignition Correction Map part of the equation.
On top of that, when the IAM is 16, it imposes a maximum on the knock correction values. This means that the Fine Learning table cannot advance timing beyond this level so it cannot contain any positive values. Again... this is only the case when the IAM is 16. It can and does, however, correct for knock with negative values in this case. When the IAM is lower than 16, the fine learning map can sometimes have positive values in it, trying to advance timing in areas of the map where it will not cause knock since the entire map has been shifted down by the drop in IAM.

There is a lot of information here, so please read it carefully as I've posted the same explanation several times and you seem to ignore it. Feel free to talk to Nate or any other tuner about this and I'm sure they will confirm this functionality of the active ignition system.

Thanks