New AF/Front O2 Sensor - Super Rough Idle

Preface: Did Stg3 installs last Friday with GS TMIC, AEM CAI, ID1000s, DW65c, Tomei ELH + UP, and VF52. AF Sensor apparently crapped out during extraction from stock headers, so Mikey/EFI was not able to do a "complete" tune until the O2 sensor has been replaced. Car has been regularly maintained, with 90k + timing belt done at 93k.

 

Okay, here we are:

I swapped out the AF sensor with a new Denso (#234-9035 Upstream), reset the ECU, and let the car relearn, but idle remains rough as hell. The car did stall out the first couple of times (turned over, tach gradually just went down, and car died). Managed to do a 5 minute idle + 20 minute drive datalog for Mikey. When stopped, tach bounces around 850-950rpm, and I have to give a little gas to keep it from dropping too low and risking a stall out.

 

I'm taking a crash-course on what these datalog parameters mean, since I'm not even sure if the new AF sensor is actually working. Anyone willing to take a look?

 

Too big to upload, so here's the Google Drive link. May have to download it, since .csv might be not a recognized file type.

 

Any help interpreting these results would be insanely helpful.

 

Thanks

I am no tuner. But I did a quick plot of your AFR over time with rpm overlaid. I could see that for the first 40 sec of the log, it appears that you are idling. But during that lapse of time, your AFR hovers around 20. So it is like super lean. Should be around 14.2-14.7 during idle.

I have another new Denso AF/O2 sensor coming in from Amazon today, and also have an OEM AF sensor on order at a local Subaru dealership (but it'll arrive on Friday at the earliest). I know that these sensors are incredibly finicky, so my fingers are crossed that it was just a defective sensor, and that the new one arriving today will resolve this idling issue/AFR being all over the place. I'll also clean the MAF sensor..

 

Can having the CAI and ID1000s in combination throw the AFR off this much?

Can having the CAI and ID1000s in combination throw the AFR off this much?

 

sorry. Can't answer that question. Don't know much about these ID1000s. But I am just wondering if these injectors are just putting so much fuel that the ecu is trying to compensate by cutting them off or something, thereby running lean and like crap. Could be completely wrong there though..

 

Doesn't your tuner know all this stuff? He should be able to help troubleshoot.

 

And make sure you do not have a boost/vacuum leak either. That could make the car run like crap too, if it is an obvious one.

No worries - I appreciate your input, regardless. The car was tuned without issues and driven down from EFI back to NJ without issues, but the tuner did say explicitly that I needed to replace the front O2 sensor. The car hasn't seen any pulls since its time on the dyno, and all the other parameters were fine then, so I don't see how there'd be a boost/vacuum leak that just sprung up. I'll have to check all the hoses later tonight, then..

I see. What is your target boost at WOT? Can you post an LV?

Target boost is 19 +/- 0.5. Sorry, I automatically interpret LV as "left ventricle," which is obviously not applicable here.

 

Edit: Will have to download AccessTuner Race from Cobb when I have the chance to pull a LV, apparently.

OK. yeah, pulling an LV (learning view) can help troubleshoot as well.

 

Are you still using the stock throttle body hose (between intercooler and throttle body)? It is notorious to leak at high boost. Ask me how I know..

No, I swapped it out for the TB hose that came with the GS TMIC kit (also installed at EFI).

OK. all vaccum hoses ziptied? (especially the hoses on the blue T underneath the intercooler).

Not zip-tied, but with those little Corbin clamps. They were completely fine prior to my installs last Friday and during the tune, and I certainly hope EFI did a good job putting everything back together.. I may have to remove the TMIC tonight and double-check the hoses, then.

One more thing, I would log 'roughness misfire' per cylinder and have a look at idle.

Okay, I'll add that to my datalog parameters then. Kind of lame that after all this, I can't even enjoy the car, but hopefully it's something stupid like a popped off hose.

Yeah. I hope so..

Just reposting this from the other thread for future use. Maybe it will come in handy for someone else? In regards to the log you posted above:

Your "A/F Learning 1" is much too high. It should be less than +/- 1. I think you have a vacuum leak that is causing unmetered air (post MAF) to enter the intake and the car is trying to compensate. It looks like the excess air and is causing your AFR (A/F Sens 1 Ratio) to be lean at idle. It looked like it was around 16-17 AFR at idle when it should have been 14.7 AFR.

 

Similarly, your fuel trims (A/F Learning 1B, 1C) were above the value considered normal (i.e. 8).

 

Is there a mechanical issue, or is it just an incomplete tune? I don't know.

I would think that larger injectors would naturally cause the car to run rich. The same pulse duty cycle to the larger flow injectors will dump more fuel.

 

Definitely yank off the TMIC and check for vacuum leaks. There are several places to pay attention to. Follow the hoses from the EBCS to the wastegate and turbo cold side. Check the other little hoses from the inlet. Check the BPV vac hose.

 

Whatever you do, don't go WOT until you can get the fueling issue sorted out. You don't want to be running that lean under boost. You'll risk the engine.

try replacing rear 02 sensor as well, it's likely tuned out but still can cause problems

^ I love your signature

Update: Replaced the AF sensor again - nope, still rough idle. Took off the TMIC, and voila -- the damn BPV vacuum hose had popped off the intake manifold side, which explained the lean conditions. Popped it back in, corbin clamped it, reset the ECU, let the car relearn idle, went for a short drive to datalog for Mikey, and everything seems to be in good shape. Fingers crossed.

Told ya . Yeah, so make sure you do zip tie all these vacuum hoses. Or else, they WILL pop off once again due to your high boost level.

Yeah, seriously good call. They've now all been clamped and/or zip-tied. Fortunately, the hose was fairly accessible and easy to pop back on.

Good to hear. I don't recall if you are using an AP at all times when you drive. If not, I would seriously invest in this:

http://legacygt.com/forums/showthread.php/btssm-bluetooth-adapter-app-229709.html

 

This will let you view critical live parameters (knock, AFR, fuel trims which will give you an indication of boost/vacuum leaks, misfires per cylinder, etc...). You can even log and pull LVs whenever you want, with this thing too. Awesome device. And it runs on your freakin phone or android tablet! Anyway, fyi..

I am using an AP at all times. After overboosting once last year and having to drive to PHL airport and back in pure limp mode, I make sure to always keep my AP in the car while driving now. AF Correction 1 still hits -25/+25, but it looks like my AF Learning 1 usually stays between 0 and 5. Mikey said my fueling looks good though.

OK. That's good.

 

Yeah, so if your AF Learning parameter stays between +/-5%, then you should be free of vacuum/boost leaks. Whenever the learned values stays consistently above +/-5%, then you should check for leaks, or revise the tune. Pull LVs on a weekly basis too.