Knock fix with Australian Fuel 98/100 RON

[steve99]

If your in Australia your BRZ/86 will knock on 98 or even 100 RON fuel. Even on stock Tune. And your IAM will likely be dropping to less than 1.

After seeing the knock on 98/100 I would only run 95 in an absolute emergency.


While its less on 100 RON (United) their is still some knock evident in 2000-3200 RPM range. Much less severity and less often but very occasionally you can get IAM to drop in city driving.

Note it knocks on STOCK tune more than on Open Flash Tablet Tunes in this area. It will generally result in your IAM dropping to 0.7 or below in city driving, it will recover with light pulls to over 4500 RPM
100 RON seems much better at high RPM/LOAD 5000-7500 RPM and does not need timing pulled in that area (on my car anyway), but 98 knocks in this area as well.

The ROM's below should vastly reduce knock on 98 and 100 RON fuels, they don't completely eliminate the 2000-3200 high load tip in knock as this seems to be a trait of these motors, but its very much reduced and IAM no longer drops below 1.

OFT logging now monitor's FLKC and Knock correction (FBKC)

FBKC is most noticeable in 2000-3200 RPM high load
FLKC is most noticeable in 5000-7500 RPM high load


KCA = (values in "Knock Advance Max A table" * IAM ) + FLKC(usually negitive)

Total_Timing = Base_timing + KCA + other offsets due IAT/coolant temp ect

FBKC is what will cause your IAM to drop if its severe enough and often enough. Not sure of the exact parameters but appears. Something like events over -1 ((FBKC) and say 4 in a minute or so , just guessing here. Conversely if no FBKC knock events are detected over a period of time (seems to be a longer period) your IAM will start to rise again towards 1.
FBKC used to alter IAM (coarse correction) does not retard timing instantly.

If you get enough FLKC it can also cause the IAM to drop

FLCK retards the timing instantly by decrementing the KCA value at that point instantly.

If your IAM drops below 1 then it decreases the timing added by KCA, effectively retarding timing over entire rpm/load range.

KCA = ((Values in Knock Control max A Table) * IAM) + FLKC (usually negative)


IAM= Ignition advance multiplier (coarse correction)
FBKC= Feed back knock control (coarse correction)
FLKC = Fine learning knock control
KCA = Knock control Advance
IAT = Intake Air temperature.
KC Learned = Knock correction Learned - Should match values in "Knock Correction Advance Max A " for each RPM/Engine Load point table if you have no knock.

If you have an OFT it now monitor's FLKC directly. So You don't have to derive it from KC Learned (KCA) value. Which would mean looking for a drop in the KC Learned value when compared to the "Knock correction advance max A table" that was not due to IAM . You can usually pick it as a sharp chunk out of the KC learned value as opposed to a fairly smooth curve. So you need to look at the KC Learned value for an RPM vs Engine Load point and compare it to values in "Knock Correction Max A" table in RomRaider for your ROM.

FLKC = KC Learned - (Knock Correction Advance Max A Table value * IAM)

If is less (by an amount say more than 0.5) then you likely have knock or your IAM =1 than it likely knock (FLKC). The CPU Averages values between the RPM/LOAD cells so hardly ever exact. If you IAM is less than KC learned values are always less.

Eg you log shows at RPM=6000 Engine Load=1.1 IAM=1 KC Learned = 3.6 you look in "Knock Correction Advance Max A' table and find value is 2.1 then you have knock (FLCK) around that point.


I used Tactrix logging to find FLKC as it was just easier, no reason you cannot use OFT it is now able to log FLCK directly.

Note the OFT logging has now been upgraded to log FLCK (although it not quite working properly yet)

On an Open Flash Tablet you need to log

RPM
Engine Load
Knock Correction (FBKC)
IAM
KC Learned
FLCK

Correcting persistent knock

If it's FBKC (Knock correction) on tip in then a small amount of knock seems inevitable and if it's just pulling less than -0.7 or so deg then I wouldn't be concerned. If it monitoring over -1.0 then you could pull 0.35 of a degree or so in the area to reduce it. If, however, it's showing any FLKC (Knock Learning or Fine Correction) over -0.7 then I would definitely pull some more timing. note that FLKC can sometimes be positive as ECU returns timing it pulled previously.

FBKC is active during rapid changes of load, ie transitioning to WOT.
FLKC is when load is more constant, ie during WOT or near steady throttle states.




Their is more to it that what is written here, this is just the stuff I think I understand but no guarantees



The only table changed in each ROM is the "Base Timing B" the changes are fairly minor timing retard on 0.35 to 1 degrees.

Proceedure to correct Knock:

Always try different Brands of Petrol first its amazing how different blends from different companies with the same Octane rating perform sometimes it worth 2 or 3 degrees less knock correction if you find the right brand of fuel.

Make sure your fuel trims are good LTFT less than 7% better if its less than 4% if it over 10% you rearly need to scale your MAF first. also check your AFR is following the values in the OL FUEL Tables for your tune. Running lean will cause knock.
Log your tune and find out where your getting Feedback Knock correction (FBKC) or FLKC (fine learning Knock correction) note the RPM and Engine load where you consistently get knock.

Say your getting -3 degrees of either FBKC or FLKC at 4000 RPM engine load 1.2 then open the BASE_TIMING_B table in Romraider or similar and remove about half of that from the value at that point so if the value was 25 make it 23 ie about 2 degrees less, sometimes you have to take a bit from the adjacent cells as well but maybe a bit less maybe 1 degree from the cells adjacent. Try to keep your adjustment smooth and avoid digging a big hole in the timing table better to smooth the entry and exit.

sometimes it takes a couple of revisions, just take small steps and be patient.

You could make similar changes to stage 2 tune, I have not run it so just check for Knock and apply similar changes to Base timing Table.

RomRaider has some good info on Subaru Knock correction but it for WRX ECU, however FA20 is most likely similar.

http://www.romraider.com/forum/topic1840.html

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


Good explanation of knock by @arghx7 who I suspect is an Engineer in the automotive field.

http://www.ft86club.com/forums/showp...69&postcount=1

and why knock sensors don't always get if right or detect false knock

http://www.ft86club.com/forums/showp...47&postcount=4

Kodename47 knock correction tool
http://www.ft86club.com/forums/showthread.php?t=74890

Kodename47 Tuning Tools
http://www.ft86club.com/forums/showthread.php?t=75710

Knock Correction for 91 Octane USA fuels
http://www.ft86club.com/forums/showt...47#post2216047



Intake Air Temperature (IAT) vs Ignition Timing Advance tables (correction for high ambient air temps)

If your Knock is Temperature related i.e , all good in winter or cooler months but knock evident when it hotter you may consider making the Intake air Temperature (IAT) Ignition timing Retard tables a bit more aggressive see below

Standard IAT vs Ignition timing retard




I made the Temp at which the compensation starts a bit lower




No Temperature compensation changes have been done to the attached ROM's

Disclaimer : Use the attached ROM's at your own risk.

Note these are G series Manual Transmission ROM's just copy changes to your ROM if you have Auto. Its only the "Base Timing B" table that's changed

ROM's below are based on Open Flash Tablet Tunes V1.5x A01G series ROM Manual Transmission . I tried to pull the least amount of timing possible so you may have to pull a bit more than I did especially if you have mods to exhaust of intake.

Changes look something like this for BASE_TIMING_B table.

I don't like removing timing from the Knock_Correction_Advance_max_A table as it will limit the amount of timing the ECU has available to it to correct knock, even though it would have the same effect on the overall timing applied by ECU. It better to leave that safety net and pull the timing from the BASE_TIMING_B table.

note BASE_TIMING_A is for limp home mode so leave that one alone :-)








@Wayno has kindly applied these changes to the "I" series automatic ROM's as well for G series manual 700 A00 and A01 ROMs see third attached zip file

Disclaimer : Use the attached ROM's and information in this post at your own risk.

OFT V2 Stg1 A01G T K S 98 Rom has been adjusted for 98 ron fuel, has some extra temperature compensation for high Intake air temps, has added ingnition retard should coolant reach or exceed 100C (212F) and has mods to quieten down high rev cold start. Based on OFT V2xx tune


NOTE: The tunes below are quite old now I posted this info a log time ago. Their is nothing wrong with them but further progress has been made. This was just a simple way to get the OFT tunes running safely on fuel that was not as good as the 93 AKI USA fuels the OFT tunes were calibrated for. We have since discovered more ways to combat knock and make the OFT tunes better, smoother, increased fuel economy ect. @Wayno has a couple of very good posts on how to go further with OFT tunes. the info posted here is still relavant for knock correction though :-)
http://www.ft86club.com/forums/showthread.php?t=83945
http://www.ft86club.com/forums/showthread.php?t=94822

[Nardi330]

Awesome Steve. thanks for your help.

[Turdinator]

Sweet, I haven't bought an OFT yet but I plan to soon and i will def give these a go. Are these the corrections you made after or before rescaling your MAF?

[steve99]

Originally Posted by Turdinator

Sweet, I haven't bought an OFT yet but I plan to soon and i will def give these a go. Are these the corrections you made after or before rescaling your MAF?

No scaling done, only table changed was Base Timing B

ROM is based on Open Flash Tablet tune 1.5x , Thanks also to Shiv.

[akyp]

Here I'm also on 98 RON fuel and I saw some knocks in what little logging I've done. Will give your maps a try.

[Burn]

I'll compare against my latest copy too

[steve99]

Originally Posted by Burn

I'll compare against my latest copy too

From memory your running headers on sage 2 ?? if so stage 2 map runs a bit more base timing than stage 1, also runs a bit more fuel

[Burn]

Originally Posted by steve99

From memory your running headers on sage 2 ?? if so stage 2 map runs a bit more base timing than stage 1, also runs a bit more fuel

Yup, Stage 2 catless headers

[steve99]

Updated 98 RON Map, pulled a bit more timing 5000-7400.

[Victor Draken]

Why not pulling times only from the "max advance table A"?
So you retard timing only when at WOT and not on partial throttle etc.
If i remember correctly.

[steve99]

Originally Posted by Victor Draken

Why not pulling times only from the "max advance table A"?
So you retard timing only when at WOT and not on partial throttle etc.
If i remember correctly.

Total timing applied = Base Timing B + Advance Max A

This applies through rpm/load range not just WOT

I originally pulled from Advance Max A but this then limits timing that can be pulled due knock control (ie IAM ) so thought it safer to leave Advance Max A alone and pulled timing from Base Timing B table instead. Overall timing applied would be the same though


Base Timing A does not appear to be used possibly limp home mode or AVCS fail or something.

[Victor Draken]

gotcha.

[AL008]

I installed my overpipe and new Grimmspeed Headers gaskets because there were little leaks. This is my new datalog in 3rd gear:

http://www.datazap.me/u/toy-sooby/3r...e?log=0&data=1

[chickencornsoup]

well done steve,

dumb question: will it cause any engine damage for the level of knock you have seen on OFT tune 1.5x ?