Jump to content
LegacyGT.com

Fine Knock Learning super high ... super randomly


AutoXCivic

Recommended Posts

All right Subagurus ... I'm at a loss.

 

Occasionally when driving my AP will register really high Fine Knock Learning and Feedback Knock ... like up to -12. There seems to be no rhyme or reason for when these readings show up or stop. Sometimes the high numbers will hang around for 1/4 of a tank, and sometimes just for the short drive to work.

 

The only common thing between these events is that the readings only show up between 1800 - 3000 rpm and 0 - 30% throttle input. Once I get over 3000 RPM or 30% throttle input these readings drop back to zero. The speed at which I adjust the throttle doesn't matter, I can roll onto it or punch it and the readings will immediately drop to 0 or thereabouts.

 

I've replaced both O2 sensors, all of the top side grounds, all four coil packs (used but lower mileage than mine). The other odd thing is this happens no matter what tune I use. I usually switch between the stock tune from Cobb and the Stage 2 91 octane tune. I used to use the stage 1 but that was where I first started experiencing weird FNL and FK issues. The guys I got to looked at everything and suggested trying the stage 2 tune because they liked the timing maps better (initial issue was likely due to bad gas).

 

Engine-wise the only performance modifications I have are an aftermarket downpipe of some sort (was on the car when I bought it) and a Grimspeed up-pipe. Engine was out of the car and partially apart recently and I was told everything looked good.

 

I've logged several of these events and even my tuners are baffled. Everything that's getting logged seems to indicate a healthy engine. No issues at high RPM or WOT.

 

Any ideas what to try next?

Link to comment
Share on other sites

If you've reset the ECU, then your fine-learn numbers should all be at 0.0. Reflashing different maps will reset the ECU, so changing tunes will automatically reset your fine-learning table.

 

Fine learn knock will only appear within certain ranges. Do some reading on how the "Subaru Knock Control Strategy" works and it might clear up some questions. Namely if you see some FLKC for a couple seconds, and then it disappears as you change throttle position or as the engine speed changes, it's because you've changed what FLKC cell the car is currently in. It's based on engine load (named "Calculated Load" on an AP) and RPM. If there's any FLKC learned into your ECU's non-volatile memory, it will stay there until the ECU is either reset, or until the ECU "unlearns" it, which only happens when you're driving in that particular cell. For example, if my ECU "learned" -4° of timing in the FLKC cell between 1.3-1.5 g/rev load and 0-3200RPM engine speed, then the only time I'll see that FLKC value is if I'm driving in those ranges (i.e., I'm under 3200RPM and I've applied enough throttle to put a 1.3-1.5 g/rev load on the engine). As soon as I exit those conditions, either by letting off on the throttle or adding more throttle, or if the RPM goes above 3200RPM, then the FLKC value will change to whatever is stored in the next corresponding cell. If I stay in those conditions for extended periods of time, then the ECU will start adding timing back as long as it doesn't hear any knock (by default at 0.35° at a time). If the ECU hears any knock at all, it'll pull more timing (by default, 2° per knock event).

 

If you're getting knock continually (especially at higher loads/RPMs), it's likely your tune could need work. A lot of times at low RPM/loads, the ECU will "hear" knock that's caused by loose heatshields or other metallic rattles in the engine bay.

Link to comment
Share on other sites

I've reset the ECU a few different times, and it always comes back eventually (sometimes pretty quickly, sometimes it takes a while). Always in the same RPM range and throttle position.

 

Resets have been performed by using the APs "ECU reset" function, and many of the map changes are made by updating the "Real Time" map and not a full reflash. Should I try a full reflash?

 

I've done quite a bit of reading on the knock strategy, and it kind of makes sense.

 

FKL will randomly jump to -8 - 12 and hang out there for a while. Then after a while (usually within the same tank of gas) will stop doing it. There doesn't seem to be any particular trigger for it "learning" or "unlearning" this.

 

I have 0 knock under high load/high RPM conditions (it's perfectly happy to do WOT pulls). It's just this weird cruising area that's causing problems.

Link to comment
Share on other sites

Are you sure you're looking at Fine Learning knock? That will not randomly jump up, it will happen sequentially. Feedback knock is the same thing, but it won't be stored.

 

Log your Knock Sum as well and see if that's increasing with your FLKC. If you see your knock sum increase while in an FLKC cell, then you should see your FLKC bump down 2° every time (until it hits the maximum correction, which should be around 12° IIRC). Honestly, if it's knocking that much, I'd expect the DAM to drop below 1.0.

Link to comment
Share on other sites

Yes I'm sure. I have gauges for both feedback and fine learning knock on the AP. I'll see if I can find some of my logs when I get home and post pieces of them. The logs are all pretty long since there's no sure fire way to recreate the symptoms. Or maybe I'll make one on the way home today.

 

I will pop a gauge for the DAM up on the AP for the drive home and see what that's doing.

Link to comment
Share on other sites

  • 2 weeks later...
  • 3 weeks later...

LV - learning view.

 

I'm having the same issue as well. Higher rpms and loads look good. It is random. Sometimes my IAM is dropping below 1 (today it was down to .125 IIRC). It will quickly jump back up to 1.

I'm wondering if I have a vacuum leak. I have checked the usual places and zip-tied everything I can. My air bypass pipe (inlet to IC) is pretty chewed up where it clamps to the IC air control valve. I have ordered a replacement. Seems like a vacuum leak would be more constant.

I also have some suspension klunking going on in the front (from a shitty AAA alignment). I'm wondering if that is causing false knock. I think I need to replace an endlink. Already tried to re-torque the suspension bolts on the ground.

Link to comment
Share on other sites

No clue on the AP and LV. I use a vag-com cable and rom raider. If you can't get LV on AP, then I'd recommend just getting a vag-com cable. Its only $20. I just got mine a few months ago from here - http://www.ebay.com/itm/VAG-COM-VCDS-Cable-USB-Scanner-Tool-OBD-2-FTDI-409-1-VW-Audi-Ross-Tech-INPA-BMW-/291675299286?hash=item43e933ddd6:g:FboAAOSw1vlUyU7M

Works great. Just follow the instructions on this site for install of driver and software. Really easy.

LV is a free download, so you'll only be $20 in and have access to more tools. If your phone supports OTG, then another $15 gets you the BtSSM app for your phone (well plus $5 otg cable).

 

[url=http://www.ebay.com/itm/VAG-COM-VCDS-Cable-USB-Scanner-Tool-OBD-2-FTDI-409-1-VW-Audi-Ross-Tech-INPA-BMW-/291675299286?hash=item43e933ddd6:g:FboAAOSw1vlUyU7M][/url]

Link to comment
Share on other sites

Hmmm. Okay. I'll look into getting either access tuner (paid software?), or that vag-com cable. Thanks.

 

ATR used to be free, but now I think Cobb requires people to take a $150 tuning class before they're allowed a license. One more reason to just learn how to tune OpenSource.

Link to comment
Share on other sites

Archived

This topic is now archived and is closed to further replies.



×
×
  • Create New...

Important Information

Terms of Use