|
|
|
|
Thread Tools | Display Modes |
09-26-2010, 05:06 PM | #1 |
Scooby Specialist
Member#: 245218
Join Date: Apr 2010
Chapter/Region:
South East
Location: FL
Vehicle:2013 STI Limited |
Fine Knock Learning in ATR
While on the dyno and even on the street I can see the "Fine Knock Learning" column start pulling timing. It is usually about -1.4 but i have seen it go to -2.8 and starts at 5100 RPMs.
Now the weird thing is that when we flashed in Cobb Stage 2 boost levels, then reset the ECU and made a pull, the FKL was still showing it was pulling timing above 5100 RPM but there was still no knock activity. The ECU was not picking up knock, but it was pulling timing in the FKL table. I am confused. Wouldn't the FKL table only pull timing based on what it learned from the knock correction tables? If the ECU did not pick up any knock, how would the FKL table have a learned amount of timing to pull right after an ECU reset? Anyone have any input on the Fine Knock Learning tables and how they work exactly? I dont understand how the Cobb Stage 2 timing and boost levels were casing this table to pull up to 3 degrees of timing in the FKL table every time we made a pull on the dyno and then I was able to replicate it on the street. Thanks,
* Registered users of the site do not see these ads.
|
09-27-2010, 09:33 AM | #2 |
NASIOC Vendor
Member#: 4803
Join Date: Mar 2001
Location: Austin, TX
|
Fine knock learning is the ECU's means of fine tuning timing at the current DAM. When the DAM changes, the fine knock learning table is cleared. If the fine learning drops in a particular cell, that, in and of itself, is indication that there was a knock event. If the DAM drops, that is an indication of a knock event. If the feedback correction drops, that is also an indication of knock. These are three responses to knock for which only one can change at any given time in response to a knock event.
Logging "knock activity", unfortunately, is not an accurate means of determining all knock events. This is because the time scale of a knock event is extremely small and it is difficult to catch it when logging. It is better to look at the 3 knock responses described above. The DAM and fine learning are learned values, so you can see it drop. Feedback correction is held for a period of time well after the knock event (relatively speaking), so you are able to "catch" it in the logs. That said, you will have some knock response at times, regardless of the tune. The Subaru knock detection is less accurate at lower RPM/load where you are more likely to get false knock. The higher the load where the knock event occurs, the more importance you should place on the knock event. Bill Subaru ECU Specialist |
09-27-2010, 01:22 PM | #3 |
Scooby Specialist
Member#: 245218
Join Date: Apr 2010
Chapter/Region:
South East
Location: FL
Vehicle:2013 STI Limited |
Makes sense.
I will just add more fuel up top and see how it reacts. |
09-28-2010, 05:00 PM | #4 |
Scooby Specialist
Member#: 245218
Join Date: Apr 2010
Chapter/Region:
South East
Location: FL
Vehicle:2013 STI Limited |
Ok so I took away 1 degree of timing from the dynamic advance and adding a touch of fuel. Now it is happy with no more Fine Knock learning.
|
10-01-2010, 09:08 AM | #5 |
Scooby Specialist
Member#: 245218
Join Date: Apr 2010
Chapter/Region:
South East
Location: FL
Vehicle:2013 STI Limited |
I take that back. It is still pulling 1 degree of timing in FKL. Lame.
|
10-01-2010, 09:31 AM | #6 | |
NASIOC Vendor
Member#: 4803
Join Date: Mar 2001
Location: Austin, TX
|
Quote:
Bill |
|
Thread Tools | |
Display Modes | |
|
|
Similar Threads | ||||
Thread | Thread Starter | Forum | Replies | Last Post |
Knock Learning sustained before boost hits, sluggish feel. | GSter | Engine Management & Tuning | 2 | 08-08-2010 11:07 AM |
HUGE knock learning but DAM stays at 1? | erikm | Engine Management & Tuning | 18 | 05-22-2010 06:59 PM |
-11 knock learning? | erikm | AccessPort | 0 | 05-05-2010 01:47 PM |
Data logging newb: Knock learning vs. Correction | Liquidr1 | AccessPort | 13 | 03-10-2010 01:42 AM |
increasing load values in ATR | Bpipp | AccessPort | 1 | 09-20-2009 06:39 PM |