I'm getting a weird hiccup in 2nd gear between 5k-5400 rpms and i thought when it first happened it was a boost leak until i logged it, the car is pulling -9.1 on fine knock learning in that rpm range. My car is tuned by Ron at RSMotors, i sent him an email but being as its sema week and he's normally a busy guy i don't know how quick he will be able to respond. Just wondering what could be causing this and should i stay out of boost until i figure this out? thanks guys ill attach all 3 of my logs i took tonight.
upload them to datazap.me please and post a link. Awesome website for easily uploading logs and viewing them. Looking at your spreadsheet though that looks like real knock and I wouldn't make it do that again without your tuner looking at it. Large timing correction at full load like that is baaad. What fuel is this on? Pump gas? 91 or 93? Could be bad fuel or maybe the tune is a little too aggressive. 13.5 degrees of advance around 5k rpm might be too much for the octane you have.
It's tuned on 91 but I always put in BP 93. Thanks guys! I contacted Ron and he told me to reset my ecu and do a couple more logs for him and then we will see from there.
Could be a loose knock sensor, or heatshield/exhaust contacting something and setting off the sensor, but for this kind of thing it's "guilty until proven innocent". Ron will get you sorted
So I reflashed my map as Ron told me to do so, and it doesn't seem to be there anymore. I did one 2nd to 3rd gear pull while logging. Now just waiting to get home from work to check it out
The ECU learns certain parameters over time, so if it's not there now, it is still very possible that it could come back. I'm by no means the expert here, but I would highly recommend posting this up over at iwsti (they have a specific forum/thread for posting logs and soliciting feedback). Most guys over there are fairly experienced. Another side note, you will commonly see posters request that you drive the car a few days after resetting the ECU to start to "learn" the same old issue from before the reset. And no offense @EricS but I don't buy the "your knock sensor might be loose" excuse. I'm willing to bet that is rarely the root of the issue. I will say that Subaru' ave some pretty ridiculous Phantom knocks that show up. But at full load and knocking at the same place.... I'd say it is definitely an issue you will want to look into.
No worries TMF, I should have been more clear - I wasn't meaning to say that what Jake described is not a real issue/knock, I was more intending to throw out there the general info that knock detection is an indirect measurement and that it is not infallible. When I said it 'could be...', I should have worded what I said more carefully to convey that it's not very likely what's happening here, just that it is _possible_ in the general sense for the sensor/ECU to declare things other than knock to be knock. Man I can't write very well tonight must be tired, good night!
Will do, if it comes back I'll look into it more. Really hoping it doesn't. If it does i'll probably bring it to Ron personally and see what he thinks. I'll post up the logs tomorrow on iwsti. Thanks for the help guys!
@Matt the logs are in my second post. the ones i took last night all look fin, not notice of feedback knock or fine knock learning.
I thought you were going to post more up lol But if they look good all the better. For so long I never used the AP, unless I needed to change a map but now I monitor, which I should've been doing.
If you notice it comes back after the ECU relearns from the reset please post the logs here too, I am curious like a cat...thats why my friends call me whiskers. lol The one piece of missing info from ECU logs in general is a reliable AFR reading in boost. The stock O2 sensor is technically wideband but its location makes it extremely non-linear when under boost and basically useless for this kind of diagnosis. Real AFR would be nice to see right around the hiccup but I am sure Ron will have a wideband if/when it comes back and he looks into it.
Post up your "good" logs. Also, is the column "I" supposed to be total timing?? If so, that is your problem. 5 degrees under full load at that rpm is bad, real bad. Really, your whole timing curve seems pretty low actually. I would ask you what your egt's are running, but MNSubaru doesn't believe in any extra gauges. Russ
Yeah column "I" is total timing and you can see it was supposed to be 13.5 and then the ecu saw some knock and pulled 9.1 degrees at 5100 rpm resulting in the 4.5 - 5 degrees of timing for a bit. What is weird is that there isn't any "feedback knock correction" (which indicates a transient problem that just happened) but there is 9.1 degrees of "fine knock learning" from 5100-5500 rpm which means there was REPEATED knock seen in this rpm range. Obviously when you reset the ecu it wipes out this learning value but if the problem is still there OP will see "feedback knock correction" and will eventually turn into "fine knock learn" correction.
not to get way to off topic or anything, but what gauges do you guys recommend getting? Boost, wide band, oil pressure are the main 3 I've been told to have, any other ones you recommend. I'm planning on doing some stuff over the winter.
my whole thing is that every where i've read(IWSTi/Nasioc) is that if there is anything that happens in the Fine Knock learning category that it would be learned away in increments of -1.4 but it just disappears out of no where once i hit 5600 rpms.
Yeah those are the most important. EGT also if you want to add another. I think you are getting the Fine knock learning and feedback knock mixed up. When the ECU sees knock it pulls however much timing it decides using feedback knock correction and then gradually adds the timing back in small increments as long as more knock isn't seen. The fine knock learning gets stored longer term if the ECU sees feedback knock in the same load/rpm range a few times. The ecu must have not seen repeated knock after 5500 rpm so that is why it never stored any fine knock learning there. If you hadn't reset the ECU those -9.1 cells would have been reduced slowly over time as long as it keeps not seeing any knock. If they come back then you know the problem is still there.
To be clear, in your "bad log" at the top the ecu never actually saw any knock. It preemptively pulled 9.1 degrees because it was anticipating the knock from past experience. That big of a drop in timing at peak torque is why it felt like a stutter or like you hit a wall for a second.
Even looking at the Cobb tutorial on YouTube for data logging it says that large knock in the fine knock learning will be slowly taken away in increments of -1.4
Yeah increments of 1.4 will be added to the cells that are -9.1. If the cell for a specific load/rpm is already 0 then there is nothing to be learned away... The fact that it goes from -9.1 immediately to 0 in your log is not significant. You have to think about your ignition timing as a 3d graph with Load and rpm as x and y and timing is the z axis. There are multiple ways to get to certain load/rpm values Have you done logs in higher gears? A 4th gear pull instead of your 2nd gear pull might make it show up again. 2nd gear logs on the STI 6 speed are not really long enough for some problems to show up.
yup i've done logs in 4th, thats the gear its recommended in. I haven't had any issues in any other gear, only 2nd.
Could have been bad gas maybe? I know you said you put 93 in on a 91 tune, but could have been a bad tank of gas.
