The famous stumble...fixed?

I've always been surprised/impressed that MY stumble has never moved from it's rpm spots, or at least seen from a fuel correction graphing standpoint. my biggest spot being ~1700 rpm.

Dont need to. Just relocate the vacuum reference. You want the plenum, not a single runner close to the valves.

 

I relocated the reference and have the older simpler FPR, still stumbling, but not always (seems to happen more during cold starts). I actually think it's not fuel related, but perhaps timing related, but haven't had a chance to catch it while datalogging.

Could be timing.

 

 

Just figured it would be an easy thing to try to add a little damper to the FPR signal line, if anyone is still having these issues. I'd try it myself, but i sold my car a while ago, and it no longer stumbled anyway

I have relocated the reference, switched to a simpler STi FPR, and still have some stumbling in the 2400-2800 RPM range. Infamous tuned out some of it, but it's still there; typically on lighter throttle.

Could be timing.

 

 

Just figured it would be an easy thing to try to add a little damper to the FPR signal line, if anyone is still having these issues. I'd try it myself, but i sold my car a while ago, and it no longer stumbled anyway

 

It might, but I'm not that hopeful, especially since I haven a fuel filter on the feedline, which kind of works like a damper and reservoir (similar to running couple feet extra line).

It is a combination of things. As the ECM learns the stumble gets worse. Once you have your fuel lines and FPR reference straightened out you can eliminate the learning in the rpm and load range of the stumble. I have never had a customer complain after these steps, but I have heard complaints of "its only 99% gone" online.

 

 

 

After that its a nuance of a fuel injected and computer controlled engine.

It is a combination of things. As the ECM learns the stumble gets worse. Once you have your fuel lines and FPR reference straightened out you can eliminate the learning in the rpm and load range of the stumble. I have never had a customer complain after these steps, but I have heard complaints of "its only 99% gone" online.

 

 

 

After that its a nuance of a fuel injected and computer controlled engine.

 

just for clarity, are you saying that adding the 3 foot section of fuel line and changing the FPR reference from the OEM location to the main plenum are the two things that need to be done?

 

I mean, i think we all know (and has been shown) to change the FPR reference for less system noise, but what about the fuel lines need sorting, from professional experience?

Extending them 12-18" helps to reduce the "water hammer" effect.

It does make alot of sense. I work with devices that move liquids, and as things get smaller and smaller, we have more and more issues with building in enough compliance to keep things in check, because working around it with software is near impossible due to the countless parameters that can change between any two events. easier to eliminate the problem than to tune a system around it, at least in this case.

^ that's a good way to explain the problem.

In which case heiche's suggestion for a damper would help. If my parts WRX has a damper (that's in one piece) I'll try to install it and see if it helps.

I’m not sure if the theory of this being caused 100% by resonance holds true.

 

I plumbed the fuel pressure regulator reference to the BOV line, studded was still there.

 

I replaced my fuel pressure regulator and dampers with a Radium setup, stutter was still there.

 

I replaced the fuel rails with Radium ones along with ID1050X’s, two radium pulse dampers, and all rubber hose and it’s still there.

 

At least on my car it seems to corrotlate with short term AFR correction overshooting. I think this is becaus the per cylinder compensations no longer match, but I don’t have a 4-channel EGT to try and fix that.

I tried a T fitting and a couple feet of hose, and it didn't help much.

I tried a damper on the inlet side of the fuel rails, it didn't help much.

Tried connecting the FPR to the BOV hose, and it helped a lot.

Tried adding more AVCS advance in the cruise area, and it helped a lot.

 

At this point the stumble is almost completely gone, so I'm almost satisfied. Almost.

At least on my car it seems to corrotlate with short term AFR correction overshooting. I think this is becaus the per cylinder compensations no longer match, but I don’t have a 4-channel EGT to try and fix that.

 

I thought there was a table to set the AF Correction min/max limits, but I don't see it now. But if you can reflash, you can disable closed-loop fueling, which means no AF Correction or AF Learning. I posted a thread about that on a RomRaider a while back. I do it sometimes to work on MAF scaling:

 

http://www.romraider.com/forum/viewtopic.php?f=33&t=7897

 

If your AF correction theory is right, you should see an improvement. If not, it's one more thing to cross off the list, and it doesn't take much to try and set it back to normal afterward.

I’m not sure if the theory of this being caused 100% by resonance holds true.

 

 

No one said it is 100% resonance. It is a combination of things. Resonance is one. Flutter caused by valve operation is another. ECM corrections is another. In the end there may be unidentified gremlins contributing as well.

 

 

 

Honestly, I feel far too much time and energy has been spent on an "issue" that... I personally never experienced driving any of my 4 LGT's, that Subaru never felt was a large enough "'issue" to warrant even looking into, that most of my customers never experienced and that in the end a little throttle overcomes.

 

 

 

But, for those that enjoy the "challenge", have at it.

I thought there was a table to set the AF Correction min/max limits, but I don't see it now. But if you can reflash, you can disable closed-loop fueling, which means no AF Correction or AF Learning. I posted a thread about that on a RomRaider a while back. I do it sometimes to work on MAF scaling:

 

http://www.romraider.com/forum/viewtopic.php?f=33&t=7897

 

If your AF correction theory is right, you should see an improvement. If not, it's one more thing to cross off the list, and it doesn't take much to try and set it back to normal afterward.

Yup in all Denso-era and many Hitachi-era Subaru ECU ROMs there are not just AF Learn A-D-F range limit tables, but also an upper and lower bound values that define from what g/s and to what g/sec AF Learning #1/2 and 3/4 (only #1/3 on H4s) can register and be recorded in memory. If defined, you can tweak both and they do take effect.

 

Its just a question of establishing or expanding the ROMs definition to get at them and then either disable AF Learning say below 3g/sec or above AF Learn C (to avoid D bleed into OL).

 

The latest 'rage' on Hitachi ECU Subarus is running full-time Closed Loop, as opposed to our good ol' full-time Open loop. I guess in theory, it should be possible to enable full-time CL on any OE semi-wideband A/Lambda Sensor Subaru... knowing it wont read any richer than 11.x:1 (some can be recalibrated to be fairly accurate down to 10.5:1 but no more).

 

I wonder what full-time CL would do on an old 4/5th gen LGT if it could and would be configured... maybe it would help here or be beneficial in general?

Edited September 11, 2018 by Perscitus

I’m not sure if the theory of this being caused 100% by resonance holds true.

 

But it sure seems like a red X (Shanin problem solving term for major contributor to an issue), and it's been documented (by COBB) and makes sense that moving the reference from right before a single cylinder/valve to the main plenum is improved as far as system noise is concerned. sure, it's not the whole pie, but it's a clear contribution to the issue.

 

 

At least on my car it seems to correlate with short term AFR correction overshooting. I think this is because the per cylinder compensations no longer match, but I don’t have a 4-channel EGT to try and fix that.

 

Are you still using OEM cast manifolds? may i be so bold as to ask why you haven't tried some 4 channel EGT testing? you seem like the prime candidate for it. i'm sure some development parts and/or funding could be wrangled up...

Edited September 11, 2018 by Flinkly

Are you still using OEM cast manifolds? may i be so bold as to ask why you haven't tried some 4 channel EGT testing? you seem like the prime candidate for it. i'm sure some development parts and/or funding could be wrangled up...

 

I’m on Tomei ELH, so even easier to add the EGT to. I considered it when i put everything on, but it was SUPPOSED to be a simple rebuild with a new turbo and ELH added while everything was out. Not this never ending money pit science project.

 

So why haven’t I done it so far? Lack of a garage or private driveway to drop the headers, lack of a welder to put in the bungs, and lack of disposable income for the 4-Chanel thermocouple logger. If I could find someone local willing to donate a driveway and welder for a day I could probably scrounge up for the logging hardware.

Edited September 11, 2018 by utc_pyro

well, you could also drill and tap your OEM and swap back and not have to mod your tomei. I've been eyeing this place for a 4 channel converter that i can read from an arduino:

 

https://www.playingwithfusion.com/productview.php?pdid=87&catid=1004

 

 

 

EDIT: i guess per cylinder stuff would (of course) be different between ELH and OEM...

Honestly, I feel far too much time and energy has been spent on an "issue" that... I personally never experienced driving any of my 4 LGT's, that Subaru never felt was a large enough "'issue" to warrant even looking into, that most of my customers never experienced and that in the end a little throttle overcomes.

 

I've got 3 Subarus and two of them are stumble free.

 

05 LGT, 5MT/6MT - used to stumble a lot, now stumbles very little

05 OBXT, 5EAT - no issues

09 LGT, 5MT, no stumble at all

 

To be fair, I think some of the stumble in my 05 LGT was self-inflicted, but it's also something that I noticed during the first test drive.

 

And I wonder if the torque converter in my OBXT helps to damp out a little bit of stumble. Or maybe there really is something different, but I can't figure out what it could be.

 

I saw a list of ROM revisions for my 05 LGT and Subaru said that one or two of them were intended to address stumble. I think "roughness" might be the word they used. I looked for it just now, but I can't find it.

Trying to get rid of my "stumble".

- I have moved the reference vacuum to the main line off the center of the intake (teed with the BPV)

- I have a longer 2 ft fuel line

- I have replaced the original FPR assembly with a STi FPR

- The MAF was replaced with OEM last year

 

I still stumble.

 

I took this log on my way home from Louisville this evening. I was going up a steady incline holding essentially 63 mph in 5th gear. I'm seeing a lot of variation that I wasn't expecting to see, but I'm not sure what the source is. Is it that the TPS is reporting a bunch of variation in the throttle plate that's causing everything else to try to adjust? Or is it something else? Thoughts?

 

BtSsm_20181007_184954.csv

Edited October 8, 2018 by Infosecdad

Throttle changes don't throw any red flags for me, it's only moving about 2%, which is not too bad.

 

Where you on cruise control? Also how steep of a hill was it? Reason I ask is, you're sitting at 5-6psi the whole time. I live in the mountains, my typical highway destination involves climbing a mountain from 600ft to 2,300ft. At 70mph (2900rpm), I'll sit at -3inHg pretty consistently, infact it never hits boost in cruise control by itself, it will sit at 0psi at worst.

 

I'm on a big16G with 8cm housing and running wastegate only, but I did see similar trends on my VF46 too (after wastegate table re-tune).

Throttle changes don't throw any red flags for me, it's only moving about 2%, which is not too bad.

 

Where you on cruise control? Also how steep of a hill was it? Reason I ask is, you're sitting at 5-6psi the whole time. I live in the mountains, my typical highway destination involves climbing a mountain from 600ft to 2,300ft. At 70mph (2900rpm), I'll sit at -3inHg pretty consistently, infact it never hits boost in cruise control by itself, it will sit at 0psi at worst.

 

I'm on a big16G with 8cm housing and running wastegate only, but I did see similar trends on my VF46 too (after wastegate table re-tune).

 

I was footing it, no cruise control and it was a 5-6% grade. My most consistent stumbles are going up a decent grade (I live on a ridge) or getting up to speed with a light foot.

That's still odd, especially for it not to gain speed, you might have some mechanical drag (brake dragging, tires, etc.).

 

Alternatively your timing could be very low thus car is not generating enough torque. I just plotted you're timing and you are running about 16* at 1.40-1.60 g/rev and 2.4-2.8k rpm. For perspective, at 1.40g/rev I'm running 24 & 29 in that rpm range, and at 1.60g/rev I'm running 18 & 22. I could probably smooth mine out, but it hasn't knocked so I didn't bother .

 

Why does it matter? Running that much less then MBT, will result in an inefficient burn which will cause the car to stumble.

Interesting, that's good to know.

Here is my base timing table.

 

 

What things should I be looking at related to this?