RomRaider

The knock has me concerned, partly because of the -2.8 in the 4800-5800 row (which means that it pulled 1.4 degrees of timing, and still knocked), and partly because there's just so many negative numbers the max load column. That problem might go away when the boost is brought under control, but it seems like timing and/or AFR are too aggressive for a setup with unproven boost control.

An easy question for you guys:

When active, FLKC is applied everytime you enter a load/rev range where a value is stored. The only way these stored values will be decremented is if you enter that particular cell and there is no knock, right? My car picked up some noise in the @1.6 g/rev and it's been taking a while for this to come down. It started at -1.4 and eventually made it's way to -0.7 and I can only assume it is taking so long because I've been simply putting around town and not really hitting that specific load/rpm area. Therefore the ECU hasn't spent enough time in that particular cell to trigger a decrease. Does that make sense?

Thanks.

Yes, you would need to be in that cell's load/rpm range (twice consecutively as far as that routine is concerned) along with all the other conditions listed that would also have to be met.

Thank you.

(This might be another dumb or obvious question)

I noticed something while looking through the stock 2008 STI map. In the Fine Correction Rows (RPM) and Fine Correction Columns (Load) tables, the stock map defines:

Rows: 2400, 3400, 5400, 10000, 11000, 12000
Columns: 0.90, 1.30, 1.60, 2.30

Should these be scaled on a modified map to account for the increased load and to utilize the seemingly unused 10000, 11000, 12000 columns?

Another thing I'm confused about is that FLKC is enabled between 1600-6300 rpm and between 1.0-4.8 g/rev, but the row/column tables seem under utilized. Would I be correct in assuming that the ECU looks at the nearest row/column when incrementing/decrementing a stored FLKC value? For example, if I experience knock 2.4 g/rev at 4000rpm, I should see -1.40 populated at 3400-by-2.30. So, with regards to the stock map, any FLKC values at loads above 2.30 g/rev and RPM above 5400 get lumped in the last cell. Is this correct?

Thank you.

Look at the learning view load/rpm axes - that is how FLKC is stored and applied. The "Fine Correction Ranges (RPM)".."(Load)" determine when FLKC can change. However, FLKC is always applied outside of idle, so a change within that range can be applied across a broader range as dictated by the FLKC table. Extreme OEM values will give you corrections over a broader load and rpm range. That is the more conservative approach.

That makes sense, but what about the 10000-11000, 11000-12000, 12000+ RPM rows? Those will never be hit, so shouldn't the rows be rescaled so that the ECU can make use of the extra cells? I think Cobb does this.

Well, that is up to you. By doing so, you would be splitting up the ranges into smaller chunks in the upper RPM area. The advantage of this is more precise control by the ECU of this learned correction. The disadvantage is if the knock is present across multiple rpm ranges (of those "added" by rescaling), then the ECU will have to learn a correction for each relevant "additional" cell.

Could you guys offer a little advice on this situation. I noticed my car didn't seem as fast a normal recently and decided to check out what the knock levels were at. To my horror I discovered this:



I'm running the XPT Stage 2 map, tweaked by Gabedude, with an APS 65mm intake. The ecu was reset about 500 miles ago. I always run a 94 octane 10% ethanol blended fuel and I shocked by the IAM value. The temperature has been a consistent 70ish for the past few weeks.

Interesting. Bump for any feedback. Have you posted this anywhere else? Let me know because I'd like to follow it. I have a very similar problem and to this day no one has been able to figure it out.

I haven't posted anywhere else. I did not do a vishnu-type reset after the ECU reset in this case. After I posted this is occurred to me that I had to run 1 tank of 91 octane because I was out of town. I then drove about 200 miles back home on the highway and I'm thinking that crappy gas was probably the culprit. I have recently recent the ECU and will periodically log to see whats up and post back.

Yea, the gas could have been a factor... but your IAM is awfully low too. Keep an eye on it and get another Learning View and post it up here soon. I'm running an XPT Stage 2 map as well, and my car is running insanely lean. I've done everything in the world except scale the MAF (which I'm thinking I may have to do).

Hi:
I have a question regarding the Feedback Correction Ranges in the 32 bit ECU.
I have the following values set.
In this log
I get FBKC at times when the RPM is below the "Enable Above" threshold of 1800, but the Load "Enable Above" is met.
So does this mean that the RPM and Load thresholds are evaluated independently as opposed to evaluated together?
Thanks,
Dale

FBKC is disabled at startup because you're below the 1000 RPM limit. If you rev above 1800, it becomes enabled. It will not be disabled until RPM drops below 1000. In the meantime you'll see FBKC at 1001 RPM through 7000 RPM.

Ah right, I forgot about the hysteresis and that both the Load and RPM are still above the disable threshold.
Thanks.