The Stumble/Stutter Thread

[eightballrj]

Lets see if we can get a handle on this issue and get a repository going of what mods people have, what they have tried to do to fix the stutter, what tests they have run with either good results or no results.

 

Let's define stumble/stutter as any form of hesitation at part throttle at either steady throttle input or increasing throttle input.

 

TYPE OF STUTTER/STUMBLE

Steady throttle/Increasing throttle/Both:

 

MODS

Intake:

Exhaust:

Fuel:

Turbo:

Engine Management:

Tuner and revision of tune(if you don't mind):

 

MAINTENANCE

Sensors changed:

Plugs changed:

Coil Packs changed:

 

OTHER SYMPTOMS

Check Engine Light:

Etc:

 

RESULTS

What has and hasn't worked for you:

If something worked, how long has it been working for:

Edited March 25, 2012 by eightballrj

[eightballrj]

I'll start

 

TYPE OF STUTTER/STUMBLE

Steady throttle/Increasing throttle/Both: Increasing throttle (often), steady throttle (very rarely)

 

MODS

Intake: Perrin Turbo Inlet, FMS TMIC

Exhaust: OBX catback, CNT Catted DP

Fuel: Stock

Turbo: BNR 16g

Engine Management: Accessport

Tuner and revision of tune(if you don't mind): Infamous1, Rev 1 (on a retune, changed mods)

 

MAINTENANCE

Sensors changed: Cleaned MAF, no change.

Plugs changed: none

Coil Packs changed: none

 

OTHER SYMPTOMS

Check Engine Light: Just got a PO302 the other day and reset ECU and it's gone and didn't come back

Etc.

Had boost leak a while ago and fixed it. It did help a little.

 

RESULTS

What has and hasn't worked for you: New tune and/or resetting ECU mostly "fixes" the stutter.

If something worked, how long has it been working for: Only lasts for a few weeks and I have to reset again.

Edited March 26, 2012 by eightballrj

[GearOnePerformance]

Clearing a CEL is not a fix. If there is ever a CEL regarding a driveability concern like a misfire or lean/rich condition it NEEDS to be fixed prior to diagnosing a stumble/stutter. In most cases the CEL is the cause. Also, to set a check engine light, a number of different parameters need to be met. For example, sometimes a problem like a misfire may need to occur more than once in a single drive cycle to turn the light on. Just an FYI from a subie tech! Ill be glad to help out with this topic as well!

[m sprank]

We have seen many different versions of the "stumble". We have come across many different fixes.

 

For some it is a slight vacuum leak. Post turbo.

 

For some it is a dirty MAF.

 

For some it is a dirty or faulty TPS. In older cable TB cars it was the idle air solenoid.

 

For some it is a front (A/F) O2 sensor that is reading off.

 

For some it is a leaking BPV (leaking to atmosphere).

 

It can be a total mystery when combined with 5EAT's too.

[eightballrj]

You guys made me remember some things I have checked with varying success. Right now it seems like what caused my misfire code it a tune or sensor issue. The longer I run any one tune it seems like the stutter gets worse. Until the day I got the misfire code, I had a very rough idle. When I reset the ECU the idle and drivability got better. So either the tune is funny(which I doubt because even at Rev 1, infamous is pretty good) or a sensor is giving funny readings and messing with my fuel trims, etc. I do, for sure, have some bad endlinks. I have heard people say that they got flash knock readings from the click/thump they make. Those will get fixed ASAP. So maybe the computer is pulling too much timing in some places in closed loop and causing a misfire. But I don't know how much timing is able to be removed by tuner ecu. So maybe this isn't a problem.

[eightballrj]

False readings not flash readings*

[m sprank]

I have personally tested a set of Perrin endlinks that caused false knock. Installed them in three different LGT's. The knock went with them to all three cars and went away when I removed them. Switched to Kartboy.

 

Stumble is not in the tune. It is the ECU reacting to something. Trick is finding it. Sometimes like a needle in a haystack. Sometimes it is never found.

[eightballrj]

Gotcha. Thanks Mike. I'll start digging for the needle.

[m sprank]

Good luck to you. Did the car ever drive without a "stumble". If not, it might never go away. For sh#$ and giggles clean the ground under the TMIC. Next to the starter. That is the cars main ground and most important to ECU communication. It could actually be a slow communicating sensor causing it. You would be amazed at what can effect CANBUS communication and the steps engineers took to attempt to avoid interference.

[NSFW]

TYPE OF STUTTER/STUMBLE

Steady throttle/Increasing throttle/Both: 2000-2250 RPM, light throttle. Most pronounced after a cold start, but often still noticeable when warm. Without even thinking about it, I just got into the habit of driving at 2400+ to avoid it, and I've been doing that for so long that I'm not sure when the stumble started. I think it may have been there since the car was stock.

 

MODS

Intake: KSTech 73mm CAI

Exhaust: Invidia catted DP, Perrin Mid/Y, stock cans

Fuel: AVO 245, then later Aeromotive 340

Turbo: ATP 3076

Engine Management: Stock ECU

Tuner and revision of tune(if you don't mind): Me, about 175 'official' revisions plus a couple dozen one-offs here and there. Only a few of those flashes were related to the stumble, mostly the revisions have been due to changing mods, adding power, tuning tip-in, changing boost levels, etc, etc... lots of little tweaks here and there.

 

MAINTENANCE

Sensors changed: I've got two MAF sensors, didn't notice a difference between them.

Plugs changed: Last summer, didn't notice a difference.

Coil Packs changed: Never tried that.

 

OTHER SYMPTOMS

Check Engine Light: I've had a few, but never anything related.

Etc: Car runs great otherwise.

 

RESULTS

What has and hasn't worked for you:

I've tried forcing the car to stay in open loop, to see if the was related to the O2 sensor or closed loop - made no difference.

Tried adding and removing ~5 degrees of ignition timing in the affected area - no difference.

Tried adding and removing ~10 degrees of AVCS in the affected area - no difference.

Changed fuel lines, changed fuel pressure regulator to an Aeromotive 1000 - no difference.

 

If something worked, how long has it been working for:

Nothing has really worked. It's pretty mild after the engine has warmed up, but it doesn't go away completely.

 

Going to try speed-density as soon as I have the time and energy. I made a speed-density ROM for 2005 LGTs but haven't had time to really tune it, so the ROM is still in MAF mode. If the problem is MAF-related, that should fix it.

Edited March 27, 2012 by NSFW

[No DRFT]

TYPE OF STUTTER/STUMBLE

Steady throttle/Increasing throttle/Both:

Stumple at idle. first start up and after warm up. Intermittent problem.

 

MODS

Intake: AVO air filter

Exhaust: Borla cat back, PDE catless up pipe. Flow catless shorty down pipe

Fuel: 93 octane

Turbo: VF40

Engine Management: Cobb AP

Tuner and revision of tune(if you don't mind):

TDC protune stage 2 map (old)

Cobb stage 2 OTS map (old)

Infamous stage 2 OTS map (current)

 

MAINTENANCE

Sensors changed: PCV valve

Plugs changed: Yes.

Coil Packs changed: No

 

OTHER SYMPTOMS

Check Engine Light: No

Etc:

 

RESULTS

What has and hasn't worked for you: I had the V.1 AP with TDC protune stage 2 map. It did a lot of stumbling at idle. I've cleaned the MAF sensor and upgraded to V.2 AP stage 2 OTS map and the stumbling went away. I'm not sure if it's a result of a bad tune or dirty MAF sensor, or both.

Revision: 1

Installed Grimmspeed EBCS and Infamous Stage 2 tune. The stumbling at idle in drive mode is still occurring intermittently.

If something worked, how long has it been working for:

Only have less than 1,000 miles since the upgrades. Will revise if conditions change.

Revision: 1

Stumbling came back, now intermittently does it at idle while in drive mode.

Edited March 31, 2012 by No DRFT

[Scooby2.5]

TYPE OF STUTTER/STUMBLE

Steady throttle/Increasing throttle/Both:

 

MODS

Intake: AVO

Exhaust: Autospeed

Fuel: 93 and E85

Turbo: AVO380

Engine Management: Cobb AP

Tuner and revision of tune(if you don't mind): Calvin Dotson Cobb Plano Protuned

 

MAINTENANCE

Sensors changed: MAF no difference

Plugs changed: 25K on plugs

Coil Packs changed:

 

OTHER SYMPTOMS

Check Engine Light: NO

Etc:

 

RESULTS

What has and hasn't worked for you: Nothings worked

If something worked, how long has it been working for:

 

My car is doing this as well. I noticed that if I ever so slightly push on the pedal and watch the Throttle position on the AP it actually goes down for a split second before it increases.

 

For example at idle with no input it is sitting at say 6. As soon as I just barely press on the pedal it goes to like a 3-4 before going up and the AFR goes a bit rich and then it hunts for a second.

 

Myself and KCwagon have the same exact issue. Once it goes rich like that as the throttle position number goes down then it is out of whack for a bit until it settles in.

Both of us were tuned at Cobb Plano, but I would think this portion of the tune was not touched.

 

Sometimes I wonder if its an issue at the throttle body or the sensor on the pedel causing this, but its weird that both of us are having the same issue.

 

If I hold the throttle at the point when the TP number goes down a bit the car really starts idling rough and will almost die.

 

Its like it does not like this partial throttle and who knows what the number going down does.

 

Anyone else have this or can test it?

 

I just went on the AP monitor section and monitored throttle position while pushing the pedal ever so slightly with my hand.

Edited June 27, 2012 by Scooby2.5

[NSFW]

I noticed that if I ever so slightly push on the pedal and watch the Throttle position on the AP it actually goes down for a split second before it increases.

 

This might be due to the transition between the idle DBW mode and regular driving DBW mode. The ECU learns to idle at whatever throttle opening angle is needed, but when you touch the pedal it switches to the DBW table. With Subaru's factory tune, the throttle plate angle table has 0% in no-throttle / 750 RPM corner. What you're seeing might be due to the ECU trying to use that value. It would be easy enough to put 3%-5% in that corner to test this theory.

[Scooby2.5]

Hey NSFW, I also notice it is worse on really HOT days.

 

Last couple of days its been in the high 90's here already. This morning though it was in the 70s and the problem wasn't near as bad.

 

Its also not bad at all hot or not with the AC turned off.

 

Thing is the idle on mine is set to 850 both AC on or Off.

 

Mine has something to do with a blip on the throttle or a slight opening of the throttle.

 

Its really starting to irritate me.

 

Here is some info from the service manual

 

1. Engine stalls during idling.

1) Manifold absolute pressure sensor

2) Mass air flow and intake air temperature sensor 3) Ignition parts (*1)

4) Engine coolant temperature sensor (*2)

5) Crankshaft position sensor (*3)

6) Camshaft position sensor (*3)

7) Fuel injection parts (*4)

 

2. Rough idling

1) Manifold absolute pressure sensor

2) Mass air flow and intake air temperature sensor 3) Engine coolant temperature sensor (*2)

4) Ignition parts (*1)

5) Air intake system (*5)

6) Fuel injection parts (*4)

7) Electronic throttle control

8) Crankshaft position sensor (*3)

9) Camshaft position sensor (*3)

10) Oxygen sensor

11) Fuel pump and fuel pump relay

 

3. Engine does not return to idle.

1) Engine coolant temperature sensor

2) Electronic throttle control

3) Manifold absolute pressure sensor

4) Mass air flow and intake air temperature sensor 5) Accelerator pedal position sensor

6) Oil temperature sensor

 

4. Poor acceleration

1) Manifold absolute pressure sensor

2) Mass air flow and intake air temperature sensor 3) Electronic throttle control

4) Fuel injection parts (*4)

5) Fuel pump and fuel pump relay

6) Engine coolant temperature sensor (*2)

7) Crankshaft position sensor (*3)

8) Camshaft position sensor (*3)

9) A/C switch and A/C cut relay

10) Engine torque control signal circuit

11) Ignition parts (*1)

12) Accelerator pedal position sensor

13) Oil temperature sensor

5. Engine stalls or engine sags or hesitates at acceleration.

1) Manifold absolute pressure sensor

2) Mass air flow and intake air temperature sensor 3) Engine coolant temperature sensor (*2)

4) Crankshaft position sensor (*3)

5) Camshaft position sensor (*3)

6) Purge control solenoid valve

7) Fuel injection parts (*4)

8) Electronic throttle control

9) Fuel pump and fuel pump relay

[Scooby2.5]

"Before drive by wire technology was introduced, if a throttle stuck open a driver could generally put a toe under the accelerator and lift up. Occasionally after servicing or repair, the wire or cable between the accelerator and throttle would not be correctly reinstalled causing sudden acceleration. However, with the ETC, the movement is all done by electronic controls moving an electric motor. But just moving the throttle by sending a signal to the motor is an open loop condition and leads to poor control. Most if not all current ETC systems have a closed loop system whereby the ECU tells the throttle to open a certain amount according to an algorithm based on the geometry of the throttle. Then, if due to dirt build up in the throttle bore or a damaged TPS a signal is sent from the TPS to the ECU, the ECU can make appropriate adjustments to compensate, though it might result in surging, hesitation or uneven idle."

[NSFW]

TYPE OF STUTTER/STUMBLE

Steady throttle/Increasing throttle/Both: 2000-2250 RPM, light throttle. Most pronounced after a cold start, but often still noticeable when warm. Without even thinking about it, I just got into the habit of driving at 2400+ to avoid it, and I've been doing that for so long that I'm not sure when the stumble started. I think it may have been there since the car was stock.

 

[...]

 

RESULTS

Tried adding and removing ~5 degrees of ignition timing in the affected area - no difference.

 

If something worked, how long has it been working for:

 

Last weekend I revisited the ignition timing idea. I noticed that my stock 05 LGT ROM had as much as 15 degrees less ignition advance than a stock 09 LGT ROM, and I started to wonder if maybe Subaru learned something in the 4 years after my car rolled off the assembly line...

 

For example, my ROM has 32 degrees at 2000 RPM and 0.5 g/rev, whereas the 09 ROM has 45 degrees.

 

At 1600 RPM and 0.5 g/rev, I have 24 degrees and the 09 ROM has 41.

 

At 2000 RPM and 0.25 g/rev, my stock ROM has 26 degrees, and the 09 ROM has 45 degrees.

 

So, I made a couple of new ROMs, with timing values from the 09 LGT in the 800-2800 RPM and 0.5-1.0 g/rev area. One ROM uses the values from the 09's Cruise timing table, the other uses the values from the 09's Non-Cruise table. The non-cruise table runs the most timing in this area (the comparisons in above this paragraph came from the non-cruise table) and it seems a little smoother. I need to do more A/B testing to be sure though.

 

The result has been significantly smoother driving in the sub-2500 RPM / sub-1.0 g/rev area.

 

I didn't make any AVCS changes because this problem is worst when the engine is cold and AVCS doesn't start operating until the engine warms up. I didn't make any AFR changes because this corner of the fuel table is all 14.7 anyway.

 

I've attached a screenshot of the timing table that I've been running since Saturday, using values from the 09 LGT's non-cruise timing tables. The highlighted cells are the only ones that I have tweaked for this experiment. The corresponding dynamic advance table has no advance in the 0.25 and 0.5 columsn, 1.05 degrees in the 0.75 g/rev column, and 1.4 degrees in the 1.0 g/rev column.

 

I still have more experimenting to do, but I was so impressed by the results of this change that I figure it's worth sharing right away.

[Pako]

TYPE OF STUTTER/STUMBLE

Steady throttle/Increasing throttle/Both: Increasing Throttle from 65% to WOT.

 

MODS

Intake: Stock

Exhaust: Cobb DP/Magnaflow CBE

Fuel: 91 Stock

Turbo: Stock

Engine Management: Open Source (Tactrix+RR+LV)

Tuner and revision of tune(if you don't mind): Stage 2 Infamous1 Tune, Revision 3 complete.

 

MAINTENANCE

Sensors changed: Original

Plugs changed: Original

Coil Packs changed: Original

 

OTHER SYMPTOMS

Check Engine Light: N/A

Etc: N/A

 

RESULTS

What has and hasn't worked for you: Going to the dealer and complaining about it. They said it might be the MAF sensor going out. It has been like this since I bought it new in 2007. 2008 LGT MT

If something worked, how long has it been working for: Getting e-tuned by Shamar has eleminated any previous stutter that I experienced. That, in of itself, was worth the tune in my opinion.

Edited August 9, 2012 by Pako

[bmx045]

Last weekend I revisited the ignition timing idea. I noticed that my stock 05 LGT ROM had as much as 15 degrees less ignition advance than a stock 09 LGT ROM, and I started to wonder if maybe Subaru learned something in the 4 years after my car rolled off the assembly line...

 

For example, my ROM has 32 degrees at 2000 RPM and 0.5 g/rev, whereas the 09 ROM has 45 degrees.

 

At 1600 RPM and 0.5 g/rev, I have 24 degrees and the 09 ROM has 41.

 

At 2000 RPM and 0.25 g/rev, my stock ROM has 26 degrees, and the 09 ROM has 45 degrees.

 

So, I made a couple of new ROMs, with timing values from the 09 LGT in the 800-2800 RPM and 0.5-1.0 g/rev area. One ROM uses the values from the 09's Cruise timing table, the other uses the values from the 09's Non-Cruise table. The non-cruise table runs the most timing in this area (the comparisons in above this paragraph came from the non-cruise table) and it seems a little smoother. I need to do more A/B testing to be sure though.

 

The result has been significantly smoother driving in the sub-2500 RPM / sub-1.0 g/rev area.

 

I didn't make any AVCS changes because this problem is worst when the engine is cold and AVCS doesn't start operating until the engine warms up. I didn't make any AFR changes because this corner of the fuel table is all 14.7 anyway.

 

I've attached a screenshot of the timing table that I've been running since Saturday, using values from the 09 LGT's non-cruise timing tables. The highlighted cells are the only ones that I have tweaked for this experiment. The corresponding dynamic advance table has no advance in the 0.25 and 0.5 columsn, 1.05 degrees in the 0.75 g/rev column, and 1.4 degrees in the 1.0 g/rev column.

 

I still have more experimenting to do, but I was so impressed by the results of this change that I figure it's worth sharing right away.

 

 

The heads flow differently in 05 and 06 vs 07+.. perhaps they realized less timing (on a stock setup) could be achieved with the higher flow of the B25 heads...?

[NSFW]

Beats me. All I know is that adding timing made it run smoother. It's not 100% smooth, but it's much better.

[bmx045]

why not just use timing editor and smooth the stock timing tables?

[Scooby2.5]

Well I totally removed and properly installed my fuel pump. Fixed the issue with not holding fuel pressure, check valve in the pump.

 

While doing this I had the battery disconnected for a few days. Although i have tried to do a reset for this problem to no avail.

 

Now the weather suddenly has cooled off from in the low 100's to 75-90 degrees and my car is running perfect.

 

A/C on or off no matter. Car idles rock solid and no "hunting"

 

As I posted earlier, mine was only doing this during A/C on.

 

Maybe somehow it is heat related.

[700watts]

I hate to post this, but I have noticed that my stutter/stumble is directly related to the weather.... the hotter it is, worst it gets … I also didn't have or didn't notice the problem before I started modding.... I also noticed that when the stumble/stutter happens my AFR bounces off 15.8ish. Just throwing it out there….could the problem be that to achieve a stoich ratio of 14.7 ….that the computer is “swinging” to far rich then lien? I am not a tuner nor claim to be, hence my lack of vocabulary.

[NSFW]

why not just use timing editor and smooth the stock timing tables?

 

You know, that's a really good idea... I'll have to try that next. Where can I get a copy?

 

The change that I made last week wasn't about the smoothness of the tables though, it was about the smoothness of the engine itself, around 2000 RPM. I added timing (a lot, up to 20 degrees) in the low-RPM, low-load corner. The engine now stumbles a lot less that it used to in the 2000 RPM area.

[seabass07]

I get something similar, but I've always assumed it was heat soak. Might be worth some experimenting with the higher timing values.

[ssbtech]

Is this "stumble/stutter" the same as the turbo surging issue?

 

I've noticed mild acceleration in 5th gear, as the boost increases there's a stutter feeling, almost as if the clutch is grabby/slipping a bit (it's not slipping)