Knock Correction vs Knock Correction Learn values?
 

Just curious what the difference between the two are when looking through an EcuTek log?

Pretty sure the actual correction value of non zero means there was knock. Someone else may chime in the learn value is adjustments to the timing so it doesn't happen again. I think the learn value goes back to its default (5?) at each startup.

I don't see any actual correction but in random parts of the log I find tons of learn correction. They don't seem consistent either.

Here's the example I was talking about:
http://www.datazap.me/u/swift996/ver...0&data=5-13-14

I am no expert but have you had a look at RomRaider forums here

http://www.romraider.com/RomRaider/H...ndKnockControl

and here

http://www.romraider.com/forum/topic5371.html

Your IAM is 0.5 which would indicate there is enough knock being detected to pull timing , I believe IAM should be around 1 if no significant knock is detected.

The "Knock correction" parameter is strange maybe its not being monitored/read correctly by logging software.


"KC learned" seems to indicate knock at times. Even stock 86 seem to indicate knock around 1500-3000 rpm low load.

My stock BRZ can have 'KC Learned values" as high as 6 on pump fuel low load 1500-3000 rpm,

With a stage-1 vishnu tune this dropped to 3 or 4 in same range same fuel.

best to get an expert to look at it

knock correction is FBKC (feedback knock correction), i.e. instantaneous timing correction based on knock sensor feedback. this is the 'default' subaru knock control strategy. KCLV is the currently applied advance value, which can be calculated as IAM * (current cell in the advance map). KCLV should always be positive in a properly tuned car, whereas FBKC can only ever be negative.

IAM of 0.5 means your car is pulling half of the timing out of the advance map in coarse correction mode. it only enters this mode and drops the iam after seeing what it perceives as significant knock. this needs to be looked into. your IAM should always be 1. if it's not, it requires timing adjustments to ensure that it always is. in many cases, on these cars, this involves lowering the advance value in the knock-proned areas of the map (0.6-1.0 load, 2800-4000rpm) to less than 3deg so that coarse correction mode won't be active in this range, thus preventing the IAM from dropping and robbing gobs of power across the board. fbkc will register in this area no matter what you do, at least in my experience.

Wondering then what you would say to someone experiencing the following:

Tune flashed and started at 1.0 IAM - take off for around 20 minute drive, IAM drops to around .5 before building back up to 1.0. Car shut off and left for a few hours to cool off, come back do the same run (no reflash) with about the same inputs - no movement of the IAM at all. I've seen this pattern from several different e-tunes on my own car, same parts and fuel otherwise. It seems that IAM on an initial learn cycle can and will adjust - once the learning is complete the car seems much happier.... or am I just missing something? Based on this I'm wondering if a lot of people are getting in the car after the flash and imediately logging and seeing this initial dip, and wondering? (KC through this dips into -1 to -2 territory a few times, KCLV ranges depending on the e-tune from 0 to around 7 - 9. Will try to attach my last initial log with the tune in the car - which daily for me now stays at 1.0.

My point in saying this is I rarely see someone post a log relative to what type of learning the car is currently using - I would understand adjusting if the tune is in fine correction mode and constantly getting kicked out due to knock, but on initial run? Wondering how many are paranoid when they see the IAM start to fall. (Granted, if the car continuously has a knock value/event in the negatives this completely doesn't apply I would think)

there is no startup phase with regard to knock control as far as i can tell. is the car warmed up completely before you take off for that initial drive?

What about fueling ?? does car remember fuel trims ?? if its "corrected" a lean mixture maybe that would cause the reduction in knock on subsequent drives ??

That was my first thought. I've found that the trims can go a bit wide in either direction during warmup, sometimes causing a larger than normal ltft for the first few minutes of driving. I'm not sure if this is unique to my car as I haven't logged it on any but mine. Better coolant temp fueling comp tables would certainly help.

Also, if you have miscalibrated aftermarket injectors this will happen, often far more pronounced as the car runs in port only mode during part of the warmup phase.

I think I read somewhere that their is some sort of catalytic converter warm-up phase that results in higher idle and more fuel for a while after startup, not sure if Shiv disables this in sage 2 catless header tunes, seem to remember some talk on it ??

Yup car is/was warmed up before driving off. And as I said, if I go log a normal drive now with the tune in place, very little knock and a happy IAM. Not too worried about it because it only happens to me on initial tune switch/flash, just an oddity I figured I'd mention to see if anyone else has seen similar behavior.

Sent from my SGH-M919 using Tapatalk

ya i believe you want you iam to stay 1.0 and just control your ignition thru the advance tables to avoid the ecu from cutting the multiplier.

Couple things.
1. I believe the knock learning / IAM system is much more "sensitive" right after a reflash. I recall reading about that on romraider at some point. So, if after the first drive cycle the reading are more stable then take it easy for the first drive cycle.

2. Your motor could be more noise prone when it's cold, assuming that fits your scenario and your not letting it cool over night between your drive cycles.