New logs, confusion regarding FLKC

[bnguyenbb6]

I just sorted out some problems with my car. I replaced an injector and logged my TA maps and a Cobb Stage 2 map. What exactly is FLKC? I thought it was just a type of KC used to refine a car while cruising but my TA maps show changes in FLKC at WOT. And is my Cobb map knocking at higher RPM's and boost? I'm pretty new to this and any help/feedback is greatly appreciated.

Brian

 

Maps:

Cobb Stage 2 93oct

TA Stage 2 93oct

TA "STU" 93oct pull 1

TA "STU" 93oct pull 2

cobb stg2.csv

Stage2.csv

STU1.csv

STU2.csv

[m sprank]

I'm no authority , but it looks like the Cobb map was knocking in the lower RPM range at WOT and the TA maps are knocking in the high RPM range. Maybe send a PM to Infamous1 and ask for advice.

[bnguyenbb6]

Thanks, I've already emailed these logs to Mike @ TA. Anyone else have any feed back though?

[Ridgeracer]

I'll give you some feedback.....like the knock that you have..

 

First off, when your car detects knock, depending on how severe it is, the ECU will do one of three things to correct for the knock; Apply Feedback Knock Correction, Fine Learning Knock Correction, or reduce Dynamic Advance (same as IAM). As a side note, if dynamic advance drops, you have very severe knock in multiple areas of your map. Since all three respond to the same signal, its best to just log Knock Sum if you have RomRaider as your logger. As for your logs, I think the Cobb Map is the worst as far as knock is concerned. The other three aren't that great either since its a lot of high load correction.

[bnguyenbb6]

^^lol this guy

 

Thanks for the reply though. Yeah that Cobb one was a real eye opener. I thought it was supposed to be a safe map. Do you think the knock is just from the tune of could be be something else?

[m sprank]

Its from the tune. I had lots of knock issues running the OTS Cobb maps. Ended up going opensource I was so frustrated by it.

[LittleBlueGT]

Looks to me like another case of COBB stage2 (or many other tuners) not raising the Fine Learning threshold. Anything above 2.2 load and the ECU will just respond with FBKC, and never FLKC. That way it can knock every single WOT run you do, but never drop IAM (DAM) and never be remembered in the Fine Learning table.

 

In most cases the knock you got on the COBBstage2 log would have been remembered in FLKC (if the load limits where changed).

 

It looks like TA has that already done.

[bnguyenbb6]

What is the different between FLKC and FBKC? Sorry I'm being a big noob, I still don't quite understand the differences between the two.

[LittleBlueGT]

You got to read the stickies on Romraider

 

Cliff notes:

 

Fine Learning Knock Correction is just that, it learns to change ignition timing based on previous knock. COBB stage2 keeps the OEM specs of turning it off at about 2.2 load, but on a stage2 tune you see about 2.8 or higher load.

 

Feed Back Knock Correction is an instantaneous reaction to knock, on many tunes it will pull 2 degrees of ignition timing per knock even that it hears. FBKC will not result in any learned timing.

 

FBKC is usually used (over-simplifying as I don't completely understand the whole process) when load is changing fast, ie: turbo spooling, quick change in throttle

 

FLKC is usually used in a steady state environment, like cruise, or after the turbo has spooled and things aren't rapidly changing.

[merchgod]

IAM and FLKC are learned corrections. The conditions to allow for changes to IAM/FLKC are much more restrictive than FBKC. The idea is to potentially make changes to FLKC or the IAM when conditions are such that the accuracy for determining knock is higher. One of the key conditions is the change in load. The change in load has to be relatively low (along with a number of other conditions) before a change to FLKC/IAM can occur.

 

The IAM and FLKC are linked. The IAM gets timing roughly where it needs to be and FLKC fine tunes that timing at that particular IAM (FLKC table is cleared whenever the IAM changes). Again, changes do not occur to these unless knock accuracy is relatively high (at least that is how I see it, not knowing the engineer's motives).

 

FBKC is not learned correction. It doesn't matter if FBKC is -12, it doesn't impact IAM or FLKC or anything else. Basically, it is a catch-all correction for conditions when there might be false knock (but not overly so), where you still want to correct for a knock event but you don't want it to be learned. FBKC has very few restrictions. A number of conditions must ALL be met (including the load change threshold) in order to DISABLE FBKC and even potentially allow for a change to FLKC/IAM. If even one of those conditions are not met, then FBKC will be the active knock response (within its own RPM/load ranges). That means that FBKC can occur regardless of the change in load.

[bnguyenbb6]

I have a new log of a revised TA Stage 2 map.Logged knock sum and killed off some other parameters.

 

So let me get this right, FLKC learns to pull timing from previous knock occurrences. Does this mean that even if FLKC pulls timing(as seen on the log), my map/tune is still safe? I have knock sum did not move, meaning no knock(yes it starts at 16, it goes off when you hit a bump in the road)?

 

I did not look at the Learning View before hand.

 

Thank yall in advance, I know I'm being slow . This is the first car I've dealt with any sort of datalogging/tuning. I'm learning...slowly.

TA s2 v3.csv

[LittleBlueGT]

I have a new log of a revised TA Stage 2 map.Logged knock sum and killed off some other parameters.

 

So let me get this right, FLKC learns to pull timing from previous knock occurrences. Does this mean that even if FLKC pulls timing(as seen on the log), my map/tune is still safe? I have knock sum did not move, meaning no knock(yes it starts at 16, it goes off when you hit a bump in the road)?

 

I did not look at the Learning View before hand.

 

Thank yall in advance, I know I'm being slow . This is the first car I've dealt with any sort of datalogging/tuning. I'm learning...slowly.

 

You had two counts of knock in that log, they were both when you let off the throttle, don't worry about those.

 

In the WOT part of the log you had no knock, good.

 

Beginning at line 10566 you had 2 degrees of learned knock correction, it didn't knock this time, but it had in the past, and that is why the ECU "learned" to reduce timing there. The ECU is doing its job.

 

Post up a learning view, do you know what that is?

 

The learned knock may have come from a run where the TMIC was heatsoaked, may have come from a number of different things. It is not really horrible, but worth keeping an eye on it.

 

The ECU can conversely learn to add back that bit of timing that it pulled. Everytimg you go over that same load/rpm spot during driving, for a predetermined length of time, with no knock, it will advance it a bit.

 

If you notice that you continue to get knock in the same spot continually, I would get TA to tweak the map slightly. If it goes away after a few OWT runs, and you don't see it back, then don't worry about it.

 

DO learn how to do a learning view!

[bnguyenbb6]

Thanks for the quick reply and patience! After reading many threads and posts, I just still wasn't sure whether I should be worried about FLKC pulling timing. Thanks for sorting it out. I'll pull up a learning view later tonight.

[monsterface]

Looks to me like another case of COBB stage2 (or many other tuners) not raising the Fine Learning threshold. Anything above 2.2 load and the ECU will just respond with FBKC, and never FLKC. That way it can knock every single WOT run you do, but never drop IAM (DAM) and never be remembered in the Fine Learning table.

 

In most cases the knock you got on the COBBstage2 log would have been remembered in FLKC (if the load limits where changed).

 

It looks like TA has that already done.

 

I know I'm bumping an old thread, but just curious what is the solution for this? I'm building my map off a Cobb stage 2 OTS map and this worries me... what is the ideal fine learn threshold you should set to?

[serx7]

Looks to me like another case of COBB stage2 (or many other tuners) not raising the Fine Learning threshold. Anything above 2.2 load and the ECU will just respond with FBKC, and never FLKC. That way it can knock every single WOT run you do, but never drop IAM (DAM) and never be remembered in the Fine Learning table.

 

In most cases the knock you got on the COBBstage2 log would have been remembered in FLKC (if the load limits where changed).

 

It looks like TA has that already done.

 

I know I'm bumping an old thread, but just curious what is the solution for this? I'm building my map off a Cobb stage 2 OTS map and this worries me... what is the ideal fine learn threshold you should set to?

 

You should extend the FLKC load range to values that your particular car will not see.

 

In ATR, under Ignition Tables -> Knock Control -> Fine Knock Learning -> Fine Knock Learning Modify (Load Range), you should see 4 values. The 05 LGT/OBXT Cobb Stage 2 OTS have these set like this:

 

0.70 | 0.75 || 2.10 | 2.20.

 

IIRC this basically means FLKC is enabled when Load > 0.75 g/rev, and is disabled when Load > 2.20 g/rev.

 

W/ my upgrades, I can pretty easily see loads > 3.50 g/rev when ambient temps are cool. In my case, I raised the 3rd and 4th values in the Fine Knock Learning Modify (Load Range) to 3.55 | 3.60. This means that my map keeps FLKC enabled until Load > 3.60, which is a value I haven't yet seen. I've come very very close to hitting 3.60 though (3.58 a couple of times in 25F temps), so I'll probably bump that out to 3.70 or so before winter arrives.

[monsterface]

Thanks for reply serx7.

 

I'll definitely be changing these table values, I'm seeing about 3.0 g/rev right now. I expect to see up to 3.2 g/rev after some more tuning.