*** Knock, do you have any?

[05pearl]

I cannot find IAM, Knock Sum and FLKC in my ECU parameter list. I believe I downloaded the latest definitions (to my desktop) and pointed romraider to this location.

 

Any advice as to what I need to do to find these "missing" parameters?

 

Thx

[tommypenguin]

Many of those parameters for me only appear when the cable is connected to the car and I have rom raider up. Try that out and see if you then have knock sum, flkc, etc. available to select from the list of parameters, if not double check and make sure you are pointing to the most current definition list inside of romraider.

[05pearl]

Many of those parameters for me only appear when the cable is connected to the car and I have rom raider up. Try that out and see if you then have knock sum, flkc, etc. available to select from the list of parameters, if not double check and make sure you are pointing to the most current definition list inside of romraider.

 

I'll try that, thanks.

[LittleBlueGT]

I suppose I should add something else:

 

Many many tunes I have seen on our cars do not have extended load ranges for things like FLKC or for rough correction range.

 

What this means on a car with a slightly larger turbo is this:

 

Car hits say 3.5 load. The unmodded code in the ROM stops FLKC at anything over about 2.1 load. Now the ECU will only react to knock via FBKC, it will not learn any timing in the upper load ranges, because that parameter has not been extended.

 

I have also seen tons of logs where there was a bit of knock, and the ECU reacted with FBKC. FBKC is only designed to be used if there is a significant change in load, in a transitory situation. But once the turbo is spooled, if things are fairly steady FBKC should not be the result of knock, FLKC, or a change to the IAM should be.

 

The problem gets compounded when ones look at there LV and say all is well. Remember, if the ECU reacts to knock via FBKC (or is forced to cause FLKC is already turned off) it is not stored or remembered in any way. You could be getting it on many WOT runs, but then see that your IAM and LV are good, and go off on your blissfully unaware merrily way.

[weitek]

Great thread. I think I picked up a few tips. Thanks

 

Edit: What is the general consensus about knock during very low load/rpm conditions? I have my own theories about this and have taken some unusual measures against it but I'm curious what you guys think?

 

Example: Crusing on the highway at 70MPH with constant throttle input, manifold pressure at -8PSI, ecu reports feedback/fine learning knock and pulls from 2 to 8 degrees.

 

I've seen it on cars ranging from stock to stage 2/350whp level. No amount of reduced timing in the tune seems to have any effect on this.

Edited July 6, 2009 by weitek

[rtbrjason]

Great thread. I think I picked up a few tips. Thanks

 

Edit: What is the general consensus about knock during very low load/rpm conditions? I have my own theories about this and have taken some unusual measures against it but I'm curious what you guys think?

 

Example: Crusing on the highway at 70MPH with constant throttle input, manifold pressure at -8PSI, ecu reports feedback/fine learning knock and pulls from 2 to 8 degrees.

 

I've seen it on cars ranging from stock to stage 2/350whp level. No amount of reduced timing in the tune seems to have any effect on this.

 

I've learned to live with it.. I've spent a fair amount experimenting with reduced timing and adding fuel etc only to go back to my original settings in the end. No matter what I did, it would show knock at loads between .3 and .9 or so at 2400-3000 rpm. I could see a knock sum count of 3 or 4 just backing out of my garage/driveway. A typical 20-30 min trip could show a knock sum in the 30s.. I can almost make it do it at will in that range by applying and releasing the gas pedal. My WOT logs in every gear are always completely clean. I've been debating raising the threshold to somewhere in the .8 to 1 load range.

[weitek]

I've learned to live with it.. I've spent a fair amount experimenting with reduced timing and adding fuel etc only to go back to my original settings in the end. No matter what I did, it would show knock at loads between .3 and .9 or so at 2400-3000 rpm. I could see a knock sum count of 3 or 4 just backing out of my garage/driveway. A typical 20-30 min trip could show a knock sum in the 30s.. I can almost make it do it at will in that range by applying and releasing the gas pedal. My WOT logs in every gear are always completely clean. I've been debating raising the threshold to somewhere in the .8 to 1 load range.

 

That's exactly what I've seen. It seems to be really bad on the LGT when starting off - Franz and I are pretty sure the clutch/throw out bearing is noisy. For my STI it usually only did it cruising in residential areas or on the highway. I also spent lots and lots of time experimenting, any it always happened no matter what I did. I ended up moving the knock threshhold to 1.05 load and above and have been super happy with it. It drives a lot better now since it's not pulling timing all over the place. I was just wondering if anyone else had found a better solution but it is nice to know I am not alone.

[LittleBlueGT]

My knock thresholds are even higher, around 1.2 load IIRC. I had the exact same problems, now I don't worry about it anymore.

[weitek]

My knock thresholds are even higher, around 1.2 load IIRC. I had the exact same problems, now I don't worry about it anymore.

 

Wow awesome. You just made my day.

[NSFW]

Great thread. I think I picked up a few tips. Thanks

 

Edit: What is the general consensus about knock during very low load/rpm conditions? I have my own theories about this and have taken some unusual measures against it but I'm curious what you guys think?

 

I saw a fair amount of knock below 2500 rpm and 1 g/rev when I was on the stock tune, so I've pretty much just ignored it when I started tuning. Usually when the knock sum increments in my logs, it's from lifting my foot off the throttle.

[weitek]

I saw a fair amount of knock below 2500 rpm and 1 g/rev when I was on the stock tune, so I've pretty much just ignored it when I started tuning. Usually when the knock sum increments in my logs, it's from lifting my foot off the throttle.

 

I've seen that one too. Do we know why that happens? Like do you think it's just noise or a tuning issue?

Edited July 7, 2009 by weitek

[NSFW]

I don't really know, but I assume there's probably detonation occasionally as the ECU cuts fuel as you lift your foot. Hopefully there's so little fuel (and air) when that happens it's not destructive. I don't think it's possible to tune that out because the ECU has to cut fuel when you lift your foot enough.... but I don't lose sleep over it.

[fishbone]

^It's called shift shock from what I remember, pretty predominant for MT guys and I had it happen on my 5EAT as well. I'd be in manual mode, 3rd gear and at the end of the logging pull I'd let off without upshifting near redline and it would ALWAYS induce knock. It has since been addressed and does not seem to do it anymore.

[rtbrjason]

I get most of this knock while in my normal cruising range while in 4th gear. Shift knock as I understand it is more prevalent in aggressive driving where you're rapidly transitioning from a high load/rpm area of the table to an area with considerably different fueling/timing.

[Ridgeracer]

I never see any FLKC with this. Only FBKC.

[weitek]

I've seen a mix of both.

[dease42]

I don't want to hijack the thread, and I'm just getting started learning about how to tune, but I've done a couple data-logging sessions and I'm trying...

 

My question: I did a 1-3rd gear WOT log this morning, and I'm seeing the parameter "Knock Correction Advance (degrees)" get as high as 5.5 in 1st gear, 4 in 2nd, and 3 in 3rd.

 

I don't have good logs to post (will in the future), my only concern is the high numbers... So, should I lay off the throttle until I can learn/diagnose/log more? Am I doing damage by being frisky with the pedal?

 

I'm starting to understand more of this, but it's slow going and I'd hate to screw something up before I even get started.

 

If the log would help, I can post it.. the parameters I logged (and yes, I know these aren't optimum) are:

Knock Correction Advance (degrees)

Manifold Relative Pressure (psi)

Mass Airflow (g/s)

THrottle Open Angle (%)

 

.. Please forgive my ignorance .. and Thanks!

[weitek]

I don't want to hijack the thread, and I'm just getting started learning about how to tune, but I've done a couple data-logging sessions and I'm trying...

 

My question: I did a 1-3rd gear WOT log this morning, and I'm seeing the parameter "Knock Correction Advance (degrees)" get as high as 5.5 in 1st gear, 4 in 2nd, and 3 in 3rd.

 

I don't have good logs to post (will in the future), my only concern is the high numbers... So, should I lay off the throttle until I can learn/diagnose/log more? Am I doing damage by being frisky with the pedal?

 

I'm starting to understand more of this, but it's slow going and I'd hate to screw something up before I even get started.

 

If the log would help, I can post it.. the parameters I logged (and yes, I know these aren't optimum) are:

Knock Correction Advance (degrees)

Manifold Relative Pressure (psi)

Mass Airflow (g/s)

THrottle Open Angle (%)

 

.. Please forgive my ignorance .. and Thanks!

 

Knock Correction Advance is the ecu adding extra timing. It is not an indication of danger or a problem.

 

At the very minimum you should be logging Feedback Knock Correction + Fine Learning Knock Correction. Then if you see those values go negative that means the ecu is pullling timing because it senses knock. That is when you should be concerned.

 

I suggest logging the following parameters until you have a better understanding of what your car is doing.

 

A/F Learning #1 (%)

A/F Sensor #1 (AFR)

Engine Load (Calculated) (g/rev)

Engine Speed (rpm)

Feedback Knock Correction (degrees)

Fine Learning Knock Correction (degrees)

Ignition Timing Advance (degrees)

Ignition Total Timing (degrees)

Manifold Relative Pressure (Corrected) (psi)

Mass Airflow Sensor Voltage (V)

Primary Wastegate Duty Cycle (%)

Throttle Opening Angle (%)

[dease42]

Thanks weitek, that's a relief. I looked at the rest of the logs from this morning (general commuting), and I noticed some larger numbers there too, at like 30% throttle, so I kind of assumed it was ok after the post, but it's nice to have an expert confirmation

 

I'll try your suggestion and see where that gets me.. thanks again!

[weitek]

Thanks weitek, that's a relief. I looked at the rest of the logs from this morning (general commuting), and I noticed some larger numbers there too, at like 30% throttle, so I kind of assumed it was ok after the post, but it's nice to have an expert confirmation

 

I'll try your suggestion and see where that gets me.. thanks again!

 

Careful who you call an expert!

 

Welcome.

[iNVAR]

ugh just got my vagcom cable today and started logging in RR. very confused even from reading just this thread. i plan on reading more, but in the meantime, is this log cause for concern? outside air is about 82F i believe, in nyc, so altitude is just a tiny bit above sea level. this is a WOT 3rd gear pull from about 1700RPM up to 6800. the only thing done to the car is a Cobb AP stage1 93 octane map.

 

it looks like there's knock...? also, sorry if i'm missing some parameters. i already ran out to the car, but i think i got most of the important ones. the "roughness" ones look useless... so i guess i'll remove that?

 

thanks in advance.

romraiderlog_20090715_120602.csv

[LittleBlueGT]

Looks like it might be a version #1 AP??????

 

That log is of concern, notice at time 6659 there is some FLKC? That is learned knock from before, not necessarily a big deal, but keep an eye on it w/ learning view.

 

Then on lines 10425 and 11076 there are two knock events in the log, that is from actual perceived knock during that WOT pull. Perhaps you could try the 91 octane map out?

 

If it is version #2 AP then you could maybe get someone (like Infamous) to tweak the tune a bit for you with ATR.

 

Two other things of note:

 

1) you don't really need these items in the log:

-roughness monitor

-only log 1 FBKC and 1 FLKC (I just do the 1 byte ones)

-fine learning table offset

-MAP (seeing as you already are logging MRP)

-coolant temp

-knock correction advance (if you know IAM is at 1.0, then you can figure out from the other parameters)

 

-EGTs are not really needed, but could have useful info

-IAT is a good idea IMO

-load is not needed, as it can be calculated (really easily using airboy's sheet) (yes there are load comps, but they don't come into play very much)

-total timing is a good one to log

-notice how the knock sum incremented at a knock event, then you can see that the ECU used FBKC as the strategy to counter it? (more on that in 2)

 

 

2) notice how the ECU decided to use FBKC for the high rpm knock events? That to me says a few things are likely the case. First off the ST maps I have from COBB for the AP vs #1 cap off FLKC at 2.2 load. That is BAD! In your log those two knock events happen above 2.2 load. When the ECU uses FBKC there is no learning happening. Dumbing it down a bit (as I don't totally understand it) the ECU is designed to use FBKC when the load is rapidly changing, like fast throttle transitions, but should use either FLKC or rough correction (changing IAM) to respond to knock events that are not happening in a fast transition. What should happen in those high load areas is your car should store a value in the fine learning knock table, then the next time your car sees the same situation (same load/rpm region) it will have learned to pull a bit of timing, thus making it safer.

 

Right now the way it stands, your car could get that knock every single time you are at about 6000 rpm, but the ECU will never learn anything. A simple tweak to the tune would solve this.

 

If you do have AP vs#1 then a simple RR/ECUflash can tweak the base map for you to solve this.

 

A lot to think about, but I hope that gives you something to mentally chew on for a while.

Edited July 16, 2009 by LittleBlueGT

[iNVAR]

Damn. Thanks very much for your input so far. Looks like I have a lot to digest!

 

Okay, to provide a bit more detail: It's a an AP v2, with the latest v1.18 map. Also, forgot to mention it's a 5MT. Also included another WOT 3rd gear pull I did, but this one only goes to about 6000 RPM. This run was about 3 minutes before I did the last datalog.

 

I'm not going nuts trying to fix this knock unless I really, really have to. In about another month, I'm going Stage 2 (invidia catted DP, spt catback, avo tmic, catless invidia UP) and I'm going to have it dyno tuned.

 

So I guess my question is if this is something I should be really, really worried about and absolutely have to fix now. My main reason for datalogging now before I go stage 2 is to make sure the engine is (for lack of a better description) "healthy."

 

I will definitely try sticking the 91 octane map on, and also a "stock" type map and do some more logs and see what happens.

 

Again, thanks!

romraiderlog_20090715_120424.csv

[LittleBlueGT]

Damn. Thanks very much for your input so far. Looks like I have a lot to digest!

 

Okay, to provide a bit more detail: It's a an AP v2, with the latest v1.18 map. Also, forgot to mention it's a 5MT. Also included another WOT 3rd gear pull I did, but this one only goes to about 6000 RPM. This run was about 3 minutes before I did the last datalog.

 

I'm not going nuts trying to fix this knock unless I really, really have to. In about another month, I'm going Stage 2 (invidia catted DP, spt catback, avo tmic, catless invidia UP) and I'm going to have it dyno tuned.

 

So I guess my question is if this is something I should be really, really worried about and absolutely have to fix now. My main reason for datalogging now before I go stage 2 is to make sure the engine is (for lack of a better description) "healthy."

 

I will definitely try sticking the 91 octane map on, and also a "stock" type map and do some more logs and see what happens.

 

Again, thanks!

 

I think you can answer your own question.

 

Let's start with:

 

1) is the knock consistent, or a one time thing?

[iNVAR]

Yeah, the logs make it look consistent. Man this is irritating. Aren't the Cobb OTS maps supposed to be pretty conservative, argh. Guess I'll toss on another map and see what happens, that ought to be the first step. Thanks. Have two more logs from that same day... One of a very quick pull through 1-2, and the knock that's consistent occurs between the 4100-5000 RPM range. Edited July 16, 2009 by iNVAR