RomRaider

Newbie that is getting into tuning and would like input on my approach. I want to install an aftermarket catted 3-in ID downpipe so I can put in a wideband and properly tune. Car is currently completely stock. Steps I'm going to take:
1. Install downpipe/wideband
2. reduce WGDC by 15% to prevent overboost due to the reduced backpressure
3. Run idle and 2500 rpm cruise and log AFR, fuel trims and actual/target boost
4. compare and adjust WGDC if actual boost isn't stable, adjust MAF scaling if fuel trim is greater than 5%
5. Run 2500-4500 rpm pull and log AFR, fuel trims, actual/target boost, knock correct advance, ignition timing
6. compare and adjust WGDC or MAF scaling as needed.
7. Run a WOT pull from 2500 to redline and log AFR, fuel trims, actual/target boost, knock correction advance and ignition timing
8. compare and adjust target boost so it's smooth across rpms, pull timing if there is high (2.5 degrees or more) knock correction

Am I missing something?

for just starting out i think you should aim for zero knock on your WOT pulls, once you've got some experience with your car and had a listen with det cans then you can decide if you know better than the ECU for listening to knock events.

Appreciate the advice. So is there a low possibility of knock occurring while in close loop and that's why I should just focus on WOT pulls? And if knocking does occur, my assumption would be AFR is the culprit (too lean) since backpressure is reduced with a wider downpipe, volumetric efficiency of each cylinder would increase and therefore more fresh air in the cylinder leading to a lean AFR. My thought process in the right direction?

Also before putting the downpipe in I plan to do a few logs at idle, partial throttle and WOT to get a baseline and ensure there aren't any issues with the stock tune.

but if your MAF based for fuelling then if the cylinders are pulling more air due to higher VE then the MAF sensor simply reads a higher voltage because more air is getting sucked in. If your Speed Density tuning then yes need to worry about VE but MAF bases is slightly different.

Be aware of knock all the time, but if your doing WOT pulls aim for zero knock rather than "some knock is good". You said "pull timing if there is high (2.5 degrees or more) knock correction" but you mean pull timing until fine and feedback knock correction are zero and IAM is 1.00.

You're right, I'm on MAF so it should compensate for higher air flow. Correct, I will pull timing until fine and feedback is zero and IAM is 1.00. Thanks for the info.

I did a 2nd gear partial throttle pull up from 2k-6k rpm and had knock at about 4200 rpm (log attached). This is the second log I ran under the same conditions, knock occurred right around the transition from CL to OL.
Map: Stock
Mods: Completely stock except for a Cobb AOS (haven't installed the downpipe yet)
Vehicle: 2009 WRX
Fuel: 91 Octane
Conditions: 45F, 5000ft elevation (Denver area)

I'm thinking AFR is too lean (around 14) for that engine load and rpm, which is why the AFR Correction spikes to 12% before knock occurs and I should shorten the CL/OL delay. The stock map has AFR target for open loop at that engine load/rpm at 11.

Thoughts?

having a look at your log, you didnt log manifold pressure - i would normally log manifold pressure and target boost rather than boost error.

also a/f sensor#1 and your wideband should be logged as well to get an idea on what actual AFR is vs the target value.

and yes a lot of stock maps stay in closed loop for a LONG time when coming onto boost and this can generate some knock, start off with a very very low cl/ol transition delay and also make sure your AFR is actually tracking fairly close to your target.

I changed the CL to OL delay to 150 (stock ROM value is 750) and attached the ROM. I added MAP and target boost to my logs. Still no wideband, probably will be able to get around to it in the next few weeks. I selected the Manifold Absolute Pressure (4 byte) parameter but I'm pretty sure what is getting read is relative pressure because the numbers are too low.

Ran a few 2nd gear partial pulls but they all were roughly the same so I only attached two logs. First log (..._072913.csv) got feedback knock kicking in at 4000-4500rpm, MAP hit 8.5psi and MAF hit 134 g/s . But my second log (..._073218.csv) I didn't get any knock but MAP only hit 5psi and MAF hit 109 g/s. I also pulled RR Learning Table Values. AF Learning values look good except for cell #1 but that's not the cause of knock. I've driven about ~100 miles since flashing the ecu with the shorter CL to OL delay so the AF learning should have settled by now.

My initial thought is AFR is too lean at the onset from CL to OL for the that high of a MAF; even though the final base fueling is the same, since there is more air flow that means more fuel which means higher cylinder temperatures and knock. To fix this issue:

a) Enrich the AFR in that area of the Primary Open Loop Table so it's not as lean when transitioning from CL to OL but it's already set to around 11:1
or
b) If I understand the Minimum Primary Open Loop Enrichment (Throttle) table correctly, I enrich these values so this is the leanest the AFR will be at those throttle positions.
or
c) shorten the CL to OL delay even more to 50. I'm just not sure on how it'll affect drivability.

I'm leaning more towards option b but want to make sure I understand the fueling logic correctly.

your fuel trims are looking pretty strong, get your wideband on and see what your fuelling is actually doing - it may be you need to scale the maf more before playing with the fuel targets.

You could even log A/F Sensor #1 and compare it to final fueling base to see how far off they are to get a vague idea how much fuel you're actually getting. If your AFR is miles and miles out compared to target then most likely there's a boost/vac leak somewhere you need to find before you go round in circles for days trying to tune around it.

I totally forgot to add A/F Sensor #1....thanks for the reminder, I'll get some more logs and and see what it reads.

I plan on cleaning the MAF and checking vacuum lines, air box, etc. I'm thinking the MAF may have drifted since it's the one from the factory.

When the ECU changes from CL to OL, should Final Fueling Base immediately go to the values in the Primary Open Loop Fueling Tables plus whatever the applicable compensations are? Or does is it a slower transition?

I'm trying to understand why my Final Fueling Base is still above 12 after changing to Open Loop when, based on my Primary Open Loop Fueling, it should be around the lower 11's. I looked through all my compensations and it isn't enough to put my FFB in the 12s. Even the AF Learning isn't enough to put it into the 12s. Looked through forums and couldn't find an answer if the FFB immediately transitions (for example ~14.7 in close loop, then once open loop starts FFB goes to ~11) or if the FFB slowly transitions using intermediate values.

if your cl/ol delay is 0 then it will usually drop instantly to the open loop target. Is it possible your engine load is still low enough that your in a 12ish area of the fuel map?

I put a picture of my Primary Open Loop Fueling Table. It is the stock table. In my log my engine load is 1.7 g/rev after going into open loop and continue to go up. This should put me in the mid 11's once I'm in open loop and get richer form there. But my Final Fueling base is in the 13s and high 12s between the point I transition to open loop and Feedback Knock Correction kicking in.

From the CL/OL transition point (@13.9 seconds of the log) and Feedback Knock kicking in (@14.4 seconds) it's about half a second. You're probably right I need to set CL/OL delay to 0 to give the ECU more time to enrich the AF mixture.

I had a peek at your rom and you are right you definitely don't want any transition delay, lean under boost is bad news. Some people fully delete closed loop, but I like it for fuel economy. I zero'd out the delay and smoothed out your fuel table. Feel free to change/richen it up but my cars seemed to like mid-low 11's however everybody is different. Bludgod is right you need to get your maf scaled before you do anything else

withgusto22, I appreciate it! I plan to remove the transition delay and run some more logs. I'll take a peak at the rom you attached when I get a chance. Still gotta throw the wideband on so I can verify MAF scaling is good....been busy with kids and school.

Finally got the downpipe and wideband installed....found out the O2 sensor that came with the AEM wideband was a dud, it constantly reads lean....

Anyways I found out the knock I've been seeing (look at previous logs that I've posted on this thread) is actually "preemptive feedback knock correction". The way I tested it was I changed the Feedback Correction Retard Value from -1.4 to -1.01, did a partial 2nd gear pull like before, and feedback knock correction kicked in around 4200rpm pulling timing by -1.4 when it should have pulled -1.01 The thread I linked below goes into it but basically the ECU pulls timing using value different than FBKC if a couple conditions that may lead to knock are met . I've also attached my log and rom for anyone that's interested (don't mind the target boost and MAP, I'm still working on refining the WGDC to compensate for the larger diameter downpipe).

viewtopic.php?f=40&t=8509