[StL_Stadtroller]

Hello Subaru folks,

first time posting here, please be gentle with me!

I've got - what we'll call for now - an '02 Outback with about 40,000 miles on the clock. Plugs recently replaced, new s-belt and idlers and tensioner, fitted with a K&N Apollo intake system, and ceramic coated headers into a high-flow CAT and muffler.

It's been in the vehicle and running for about 1,000 trouble-free miles, but suddenly has developed a *very* subtle rough running / vibration issue.

At idle I can ever-so-slightly perceive a bit more vibration than previously, and at highway speed I can feel it as an ever-so-subtle low harmonic... even my rear-view mirror is vibrating a bit from it.

I have no CEL, OBDII tests through the Torque Android app come back clean. No change in available power that I can perceive.
No odd noises, no leaks, etc that I can detect.

I know it's the engine and not suspension/wheels because I can de-clutch and the vibration goes away. Also I can feel it very subtly at idle.
I've checked all fasteners, mounts, etc to rule out something just being loose.

In my old-school carburetor mind I would take a wild guess of a dead cylinder, but this is not indicated in the OBD tests as I stated above.

If anyone has any ideas or tips with where to start and what to check (either physically or with Torque/OBD) let me know!!!

now, here's the kicker... this engine is actually in a 1987 Volkswagen Vanagon Syncro... but the entire engine and wire harness and ECU and all are just as they were in the original Outback...

 

[CNY_Dave]

You are a very very bad man...

Is this hooked to a subaru trans? If so are all 4 wheels powered?

Trans or engine mounts can do what you say, maybe one is overstressed and has bottomed out.

And, where the heck are the pics?!!! We need pics!

 

[StL_Stadtroller]

You are a very very bad man...

Is this hooked to a subaru trans? If so are all 4 wheels powered?

Trans or engine mounts can do what you say, maybe one is overstressed and has bottomed out.

And, where the heck are the pics?!!! We need pics!

LOL!
Hey at least I didn't hack up an SVX as was my original plan!

But no, the engine is mated to the stock Vanagon Syncro transmission through an adapter bellhousing. I'm running a manual trans flywheel and clutch, and Subaru starter. There are solutions to run both the engine and trans in a 2wd Vanagon, but none yet for a Syncro.
All 4 wheels are powered - through a viscous coupler in the front diff. It doesn't use a center differential like the Subaru AWD system.
But regardless, my driveshaft to the front isn't even installed yet, so that takes the AWD system out of the equation.

Like I said, I can feel this vibration / roughness even sitting still at idle, so I'm not thinking it's anything suspension or rotational.. it's just more pronounced at speed.

I'll be re-checking the mounting bolts this evening anyway. (stock OEM Subaru engine mounts, brand new).

Pictures of my entire, insane conversion process can be found here:
https://picasaweb.google.com/108389533870702883054/SyncroSubaruConversion?authuser=0&feat=directlink

 

[CNY_Dave]

Hey no FAIR the plugs are really easy to get to!!!

 

[1 Lucky Texan]

you might try a vacuum gauge test.

what's the idle rpms?

Ours is glass smooth so, I suspect you really have 'some' kinda problem.

TPS is a sorta weak spot.

dirty IACV/throttle body

high mileage engines can have burned exhaust valves. shouldn't be a problem for you.

maybe the actual tensioner is failing?

maybe the crank pulley is loose/failing?

 

[StL_Stadtroller]

you might try a vacuum gauge test.

what's the idle rpms?

Ours is glass smooth so, I suspect you really have 'some' kinda problem.

TPS is a sorta weak spot.

dirty IACV/throttle body

high mileage engines can have burned exhaust valves. shouldn't be a problem for you.

maybe the actual tensioner is failing?

maybe the crank pulley is loose/failing?

Warm idle reading is about 650-690rpm

Yeah, if the motor had run this way from the very beginning I would never have suspected a thing. The rumble/vibration only caught my attention at all because you could have balanced an egg on the intake manifold before.

Didn't have any wrench time last night but I did stop at the FLAPS and got a bottle of Seafoam for the tank, and a spray can of it for the throttle body. sometimes you get lucky, right?
This motor sat in storage for at least a year, now that it's running again who knows what may have worked loose in the fuel system, eh?


I did pull this from TorquePro yesterday, my diagnostic test results (still no actual engine error codes thrown).
I'm not sure what "TIDS" are, but here you go:

Mode $06 report generated by Torque for Android
================================================

Vehicle VIN: Not present
Vehicle Manufacturer: Unknown
Vehicle Calibration ID: D6HE000B

Unit and scaling information are not supplied with the data from the ECU for this type of vehicle. Consulting the manufacturers service book for this information is recommended.

Test report:
------------------
TID:$01 CID:$01
- Rich to Lean sensor threshold voltage(constant)
Min: 0
Test result value: 65,535
PASS
----
TID:$02 CID:$01
- Lean to Rich sensor threshold voltage(constant)
Min: 8,483
Test result value: 65,535
PASS
----
TID:$06 CID:$01
- Lean to Rich sensor switch time(calculated)
Min: 102
Test result value: 102
PASS
----
TID:$06 CID:$02
- Lean to Rich sensor switch time(calculated)
Max: 41
Test result value: 40
PASS
----
TID:$07 CID:$01
- Minimum sensor Voltage for test cycle(calculated)
Max: 61
Test result value: 11
PASS
----
TID:$0c CID:$01
- Misfire counts for last/current driving cycles(calculated)
Max: 55,303
Test result value: 44,674
PASS
----
TID:$0d CID:$01
-
Max: 52
Test result value: 28
PASS
----
TID:$0e CID:$01
-
Max: 52
Test result value: 27
PASS
----
TID:$0f CID:$02
-
Min: 0
Test result value: 65,535
PASS
----


End of report.

If I can get the time tonight, I may do something like pull one injector plug at a time, or coil at a time looking to find one that doesn't make a difference in idle. That would tell me if I have a faulty injector / plug / coil... right? It's what I'd do with one of my air-cooled VW's anyway..

 

[StL_Stadtroller]

*maybe* have this figured out?

Tonight I went through each cylinder, disconnecting 1 injector at a time, then 1 coil at a time.
Every one made a very distinctive change in idle, a clear "WhumpWhumpWhumpWhumpWhumpWhumpWhumpWhumpWhumpWhump"

The roughness I have been experiencing is FAR more subtle than even that.

Along with this wrenching, I also did a straight compression test, and got a range from 185-200psi.

I tried to do a leakdown test as well, but got fed up with burning the @#%!!$ out of myself trying to get each cyl at TDC and keep it there... gave up after 2 cylinders.
I may re-visit that, if someone can give me a trick to finding TDC on these engines? the pulley doesn't seem to be marked..

Anyway, I don't think it's that anyway.

I then blew a can of SeaFoam throttle body cleaner through the intake, revving the snot out of it.
It just may have done the trick... or at least reduced it enough to where "I'm not sure" any more.

Still no engine related error codes, though I do have persistent codes for the Torque Cut Module which is not in use (high/low).. no one in the Vanagon conversion world has yet figured out a way to eliminate those codes. Those have been there from the start however.

I logged some data on my drive home, with "Long term fuel trim" and attached that tab-deliminated .TXT file if anyone can interpret that info.

Also a few screen shots of the Torque app when I could stop long enough...

 

[1 Lucky Texan]

LTFT all look like single digits so, I'd say that's not bad. (from my very limited reading)

um - I guess coolant temp is Celsius?

Torque looks interesting. I wonder it would run under Bluestacks on my laptop?

 

[StL_Stadtroller]

LTFT all look like single digits so, I'd say that's not bad. (from my very limited reading)

um - I guess coolant temp is Celsius?

Torque looks interesting. I wonder it would run under Bluestacks on my laptop?

uh... hmm.
My dash display screen caps are English units, but the data in the tab-separated TXT file appears to be Metric.
so yeah, that must be Celsius.

 

[cardoc]

Okay,

Mode 6 is the monitor system for emission controls and the values in the list are only a set calculation for the monitor setting to take place and explaining them would take more than typing it here. Its passing the testing. I did notice that the Catalyst monitor isn't set. Subaru PCM's will set the Cat monitor in short term unless it is watching something to try and determine where the problem lies so it can let you know through the MIL and code setting.

This engine needs to run average temperature of 200F. The STFT is the amount the computer is adding or subtracting fuel to balance the AF ratio to 14.7:1. Yours is high. -25% means the computer is withdrawing the fuel, reducing the injector cycle because somewhere the PCM is getting that the engine is running too much fuel. I take it that the altitude reading is correct? If so, the BARO should be between 26 and 29 depending on local weather when the log was recorded, high or low pressure system. With that, you don't have a vacuum leak because your engine vacuum is where it should be. MAP sensor looks good also. So, anytime the trim is off by +/- 7% at temperature, you have to look into the sensor feedback to the computer or a mechanical issue effecting the ignition system.

Type of fuel - Ethanol sucks. Especially in cooler weather.
Correct spark plugs - H6 use specific plugs.
Coil burn time correct? Full output?
Timing. Did you have the engine cover off for any timing chain work? If not, look into lazy cam or crank sensor. Debris on the magnets or gear reluctor ring (AKA sludge).
MAF sensor - At idle, the MAF should be showing 4.2 to 4.7 grams per second. I didn't see a MAF reading. If its high, the computer adds fuel, low it reduces, then adjust again based on AF sensor feedback.
Can your software log AF Correction? Knock Correction? Injector Pulse?

You may want to look into Tactrix hardware and RomRaider software. It is Subaru specific.

Looking at your Alt reading, you are at the same altitude as I am. Ambient temp and intake temp variations will be different but base learning curve should be the same. Idle at avg 650, MAP absolute should run avg 5.5, Relative MAP will be determined by the barometric pressure at the time the log is taken and will be the vacuum minus actual BARO with an avg of 20 inHg or a -7 to -9 relative. Knock correction should be zero, timing avg 14-15 BTDC. Fuel injector pulse about 2.0.

If all the Ignition system checks out, its a balance issue with the engine. The negative STFT is where you need to start. Pull a plug from both sides of the engine and look at how they appear. High amount of carbon is heavy fuel, whitish is lack of fuel. Check the AF sensors, MAF output, grounds, then all else.

If you can log the other parameters listed above it could help.

 

[StL_Stadtroller]

uh, holy crap dude, is this English?
Lots of this was over my head but I will address what I can...

Okay,
Its passing the testing. I did notice that the Catalyst monitor isn't set. Subaru PCM's will set the Cat monitor in short term unless it is watching something to try and determine where the problem lies so it can let you know through the MIL and code setting.

Not sure about this.. I'm not running a Subaru OEM Cat Converter but do have the OEM O2 sensors plugged in. Is this what you're talking about?

This engine needs to run average temperature of 200F. The STFT is the amount the computer is adding or subtracting fuel to balance the AF ratio to 14.7:1. Yours is high. -25% means the computer is withdrawing the fuel, reducing the injector cycle because somewhere the PCM is getting that the engine is running too much fuel. I take it that the altitude reading is correct? If so, the BARO should be between 26 and 29 depending on local weather when the log was recorded, high or low pressure system. With that, you don't have a vacuum leak because your engine vacuum is where it should be. MAP sensor looks good also. So, anytime the trim is off by +/- 7% at temperature, you have to look into the sensor feedback to the computer or a mechanical issue effecting the ignition system.

Plugs look great to me, nice coffee-stain color.

the only think off the top of my head I can think of that would effect fuel ratio may be the K&N Apollo intake system (less restrictive than OEM?) and my DRASTICALLY shorter exhaust run than when this engine is in a Subaru car. CAT and muffler are also high-flowing varieties by Vanagon standards but I don't know how they stack up to a complete Subie exhaust run.

Type of fuel - Ethanol sucks. Especially in cooler weather.
Correct spark plugs - H6 use specific plugs.
Coil burn time correct? Full output?
Timing. Did you have the engine cover off for any timing chain work? If not, look into lazy cam or crank sensor. Debris on the magnets or gear reluctor ring (AKA sludge).

Nothing but Ethanol here, but I do run Premium (91 octane)
plugs are correct and new
Coil burn? Don't know how to check
Engine only has 40,xxx miles so I elected not to remove and mess around with the timing chain chest.

MAF sensor - At idle, the MAF should be showing 4.2 to 4.7 grams per second. I didn't see a MAF reading. If its high, the computer adds fuel, low it reduces, then adjust again based on AF sensor feedback.
Can your software log AF Correction? Knock Correction? Injector Pulse?

If it does, I haven't found the PID or know how. This engine has a MAF? where?
in a VW the MAF is in the intake runner but I have no such component, or apparent connector for one in the wire harness.

Looking at your Alt reading, you are at the same altitude as I am. Ambient temp and intake temp variations will be different but base learning curve should be the same. Idle at avg 650, MAP absolute should run avg 5.5, Relative MAP will be determined by the barometric pressure at the time the log is taken and will be the vacuum minus actual BARO with an avg of 20 inHg or a -7 to -9 relative. Knock correction should be zero, timing avg 14-15 BTDC. Fuel injector pulse about 2.0.

If all the Ignition system checks out, its a balance issue with the engine. The negative STFT is where you need to start. Pull a plug from both sides of the engine and look at how they appear. High amount of carbon is heavy fuel, whitish is lack of fuel. Check the AF sensors, MAF output, grounds, then all else.

If you can log the other parameters listed above it could help.

This is largely over my head.
But I will add that I re-used the O2 sensors that came with the engine and parts, which jangled around in boxes and bins for many moons before I re-installed them. I recently learned that this may not be good for them (hey, be easy on me, I'm used to a single carb and only 2 wires going to my engines!)

I also learned that, strangely, a faulty main o2 sensor may be contributing to my repeated VSS failures I posted out elsewhere.

Could the trim and stumble issues all be O2 sensor related??


Thanks for the thorough response!!!

 

[cardoc]

No MAF. It runs off the MAP sensor, so the PCM needs air temperature, pressure (BARO) and Vacuum readings to help with fuel trim.

It hasn't set the CAT monitor for 1 of 3 possibilities that I can see from the previous data and your last set of pics: you haven't driven it enough, the temperature isn't sustaining proper operating conditions or that CAT is too small. Actually, that CAT really looks too small anyway. Do you know what its rated for?

The length of the exhaust is not an issue. Neither is the air intake system as long as you have the air temp sensor somewhere before the throttle body. I can't see it in the photo, but I take it the catalyst monitoring O2 is mounted in the center of the CAT?

The front O2 sensors are actually A/F sensors that monitor the oxygen content of the exhaust for the computer to utilize in adjusting the fuel trim to the proper mix for combustion, 14.7:1. Either your engine is not burning all the fuel in combustion or the computer is getting a false reading from the front O2 sensors. Were the plugs on both sides of the engine equally colored? If not, the side with the darkest blend is the side with the problem, either mechanical, ignition or with the front O2. NOTE: It would benefit you and your engine both that in the event you replace an O2, use Denso. They react faster.

Coil burn time is sometimes available in scanner programming. Not always. The coil burn time is the amount of time the computer utilizes the coil for ignition per cycle. Don't worry about this if its not available on your logging program. If a lazy coil is the issue, it will rear its head soon enough.

Altitude makes a huge difference in the equation of fuel trim. If the PCM thinks its at sea level and its actually at 2,000 feet, it will run leaner fuel; and vice versa. Your logger showed altitude at 485 to around 600 feet I believe? If this is correct for your actual altitude above sea level, then the barometer on the engine is feeding the correct information and this also should be dismissed. If not, you need a new one, and start over.

Also, electrical flow is a big issue, especially with computer systems. With that, check to make sure your battery is good and supplying the amperage its rated at. Make sure you have an excellent ground to the engine, frame, and body of the van. It is important to have grounds from the battery feed every element of the total system. If you have the computer grounded to the body, you need a large supply from the battery to the body. A lot of people think backward and put large positive cables all over the place and neglect the grounds. Electrical flow is from negative to positive, thus, large number of ground connections helps the electrons flow faster and easier. 4 gauge cable is sufficient for battery to block. 4-6 gauge for feed to the frame and body. Also, either add in a couple 6 gauge cables or braded ground straps from the engine to the body. I have repaired numerous issues just by insuring that the amperage at the battery is within 100 amps at the block and body. When its not, I check for bad grounds and replace accordingly. On a side note, I've had quite a few Land Rovers that display errors with the ABS, Traction and Stability programming. Dealers and other shops have tried to sell the owner a $5k repair. I spend 2 hours cleaning and replacing and often times adding ground connections and the problem is solved.

Think of it this way: If your computer is not getting enough amp feed to run all those sensors and actuators, its like running a laptop with a low battery; REALLY SLOW processing. Which effects the outcome of what its trying to accomplish.

Take one item from the list and check it. If its good, move on. If not, repair/replace and start over. A record of the data stream at cruising would be good to compare with one idling. It would show the changes the computer is making in trim, timing, etc.

 

[StL_Stadtroller]

Cardoc, you are a gentleman and a scholar.

I will try to address as much as I can:

1) altitude - this is correct. I am in St. Louis Mizzery and we're anywhere from exactly 400-650 feet above sea level here.

2) CAT is Magnaflow brand, don't know what it's rated at but I can probably find out. The rear O2 sensor bung is in the bevel edge of the exit end. As you can see space is a challenge with the Vanagon so this is a typical part used in the conversions.

3) Plugs looked consistent across the board. Also newly replaced. If there is a means to read the coil burn through the TorquePro app, I have not yet found it.

4) Electrical... this is a fun one! In the Vanagon, the battery (grp 41) is way the **** up under the passenger seat! I made a up a new double-aught gauge wire (11 feet long!!!) to run from a brand new battery back to the starter B+ post. From there a 10ga wire runs to the junction box in the left rear of the engine bay, where the ECU (you call it a PCM though, I'm not caught on to modern terminology yet!) wiring harness taps in for power. The EC... er.. PCM itself is under the back seat on the left side of the van where the old VW computer lived.
I have the OEM Subaru wire strap from the left side fuel rail cover connected to the chassis, as well as a large ground strap from the transmission to chassis.
There is another 14ga auxillary ground wire from the PCM harness to the firewall near the PCM itself.
If I remember right I see 13.4v at the PCM connector.

I've been doing some reading about Fuel Trim the last few days, learning a lot. General (ie non-Subaru tech) mechanics I've talked to seem to think my readings are not of any concern.

All I can say is that it's all very VERY subtle and I don't know what to make of it yet. I may just ride it out until something becomes more apparent.

We're leaving tomorrow on a 5,000 mile trip from here to New England and a few Maritime provinces so if anything is going to happen it's going to happen I betcha.

At least on this trip I will be able to source parts and take it to a shop if I have to!

Cheers,
Brian

 

[cardoc]

Enjoy the trip. Its a great drive.

Just a note on your last post:

The large negative amount of fuel trim is an indicator of an issue.
The large B+ is good, but again, electron flow is from negative to positive and the grounding is most important. If you had run a large ground instead, it would have been more beneficial.
I am also inclined to believe it is the ethanol blend gasoline. The higher the blend, the worse the car runs. Try using the major oil company stations on your fill ups and see what happens. It may work itself out once the ECM learns the trims on the drive. Subaru programming is stand alone in the way it "learns".

PCM is the "old school". In the older computer controlled vehicles, other than European, they had a singular control module that handled all the systems, thus PCM = powertrain control module, whereas now, with multiple computer systems in vehicles, its referred to as ECM = engine control module.

EBCM = electronic brake control module
TCM = transmission control module
BCM = body control module
and the list goes on. Chevrolet trucks and many of the sedans have as many as 23 computers tied in together and operating everything bumper to bumper.

 

[lockmedic]

Or more. . . And not a single one of them are plug and play. Out of the box they won't work (or will but it'll set a bunch of CEL's or won't work properly) without programming them with TIS.

We have a Tech 2 and a TIS2web subscription and do a lot of work for the area independents and rebuilders. One outfit has bought a couple of flood trucks and dang near went broke buying modules and paying me to program them.

 

[cardoc]

Or more. . . And not a single one of them are plug and play. Out of the box they won't work (or will but it'll set a bunch of CEL's or won't work properly) without programming them with TIS.

We have a Tech 2 and a TIS2web subscription and do a lot of work for the area independents and rebuilders. One outfit has bought a couple of flood trucks and dang near went broke buying modules and paying me to program them.

GRAVY work though.

 

[jones]

an alternative mechanical cause ?

StL_Stadtroller,

I have been a lurker on this site for the ten years I have owned a 2003 H6 Outback. During that time ( and 203,000 Km) I have done all my own maintenance and have picked up a few tricks along the way.

If, at some time in the past, the harmonic balancer has been removed without making a special jig to lock it; or the crank has been jammed in place by wedging a screwdriver or the like between the Balancer and the cast plate behind, damage to the rubber inserts at the back of the harmonic balancer can result. Such damage leaves an out of balance Harmonic Balancer which can produce the subtle vibration to which you refer. Left unchecked the vibration slowly causes increased vibration as the rubber is squeezed out and the subtle and occasional vibration becomes far less subtle and constant until a perceptible "wobble" in the pulley/harmonic balancer can be easily observed at idle.

I hope this helps. The local dealer commented when I bought a new Pulley/Harmonic Balancer that it was the first that he had sold where there had been no accident damage to the front of the subject vehicle.

Good Luck.

regards

Jones