2006 Subaru Outback rough idle and stalling

I have a manual 2006 subaru outback 2.5L XT that idles rough pretty much all the time and the rpm gauge fluctuates at stop lights between 500-1000rpm. The car also stalls at low speeds and almost always about 1/2 a mile from my house when the engine is cool and at a stop sign or light. It seems to not do it as much when the engine is warm. No mechanic can figure out what it is yet. Gas mileage is getting bad because of it, check engine light is on. One mechanic thought then an engine coil because there was a misfire in cylinder 4. He switched that one to 3 to see if it would throw the check engine again and it did not. Then they thought it was the fuel injector because they were switched and now they are back to normal. Now they think it is a head gasket. Any help?

Have they checked the valve lash? Valves out of adjustment can cause symptoms like this.

No no one has mentioned this yet. I finally to it to the dealer to have it checked. I have read about several things that can cause this. Some include: ECM/ECU Reflash, MAF sensor, Idle Air Control valve, Vacuum leak, Fuel Pressure problem. Do any of these sound like they are of valuable consideration? We will see what the dealer finds tomorrow. I will post with his diagnostic findings.

Since this problem could be major or minor, the first thing I would suggest is putting a vacuum gauge on it. This is easy to do, will reveal much, and unfortunately is a tool that many mechanics do not use.

If the manifold vacuum is not correct then a compression test should be run. If the compression is low on any cylinder then the valve lash should be inspected. If the valve lash is tight (especially with an exhaust valve) adjusting the lash may be a short term fix because a tight valve is often damaged. It will only get worse over time.

A vacuum gauge could also reveal whether or not there’s a vacuum leak. A vac. leak can cause this kind of symptom and an Idle Air Control valve problem can do the same.
From the sound of things these mechanics are skipping basic diagnostics and throwing parts at it.

So what I am hearing from the mechanic right now for an initial diagnostic that the issue has to do with the turbo and a bad o-ring. They have had a couple turbos in this week that have had the same issue with the same symptoms: running rough when cool, stalling, and running a bit smoother when the engine is warm. I made mention of the vacuum gauge to the person I talked to, but he seemed the think the mechanic was certain that this was the problem. Any thoughts?

just to clarify-they think it is a bad intake o-ring…

Slightly skeptical. The turbocharger is pretty much irrelevant at idle speeds and the only way an O-ring would cause an idle problem would be if it allowed an air leak that was below the throttle plate, not in the intake tract between the throttle plate and turbocharger.

Any kind of minor air leak above the throttle plate should not affect the idle but could possibly cause a hesitation or bucking problem.

It would be interesting to know which O-ring he’s talking about and where it’s located.

So the mechanic at the Subaru dealer came in with a final diagnostic. I am told there is a service bulletin out on this issue from Subaru how to fix this if there are symptoms similar to mine. The seals @ the intake manifold need to be replaced. The mechanic did a compression test on the cylinders that were throwing the check engine light and those valves checked out fine so he was able to eliminate that. I guess there have been a couple cars in this week at that shop with this problem, 3 out of the 4 were turbos. I am also told that the issues arise the colder the weather gets. He said he has seen quite a few 2005’s do this and one 2008. Mine is the first 2006 he has seen. They say that they stand by their diagnostic, so we will see if that solves the problem. Any final thoughts? I will let everyone know how it turns out.

An intake manifold leak below the throttle plate will cause a rough idle on any car made and will often be more noticeable when cold.
Again, spending 10 seconds connnecting a vacuum gauge would show this instantly, just as it would any problem involving tight valves and compression.

Glad the compression test came out fine although the numbers are not given here. Ideally, you’re looking for 180-190 PSI on each cylinder.

I will find out the numbers. So what is the difference or how to you tell if it is the seals at the intake manifold or the head gasket. Is there a difference?

Any leak at the intake manifold will not show up on a compression test. If there is a head gasket problem that caused a breach into one or more of the combustion chambers then yes, that will show up.

While it varies a bit, the numbers you’re looking for on a compression check should be in the 175, 180, 190 PSI range and all numbers should be fairly close to each other.

We will see if this work will solve the problem. The final diagnostic from the tech was “leaking intake manifold to tumble generator seals”…

So I had a new intake manifold set installed (gasket) per the dealer’s diagnostic. This did not solve the problem. It stalled as soon as a pulled out of the parking spot at the mechanic. They took it for another test drive and thought it had to do with the car not being programed to run in idle without the heat or AC one. They took it for a drive and it was no longer stalling. Although the new intake manifold seal did solve the car’s rough idle problem, now it just stalls without warning. Again it ONLY stalls when the car is cold and within the first 1/2 mile from my house when either stopping, slowing down, when the car is in neutral, and when slowing down when going up hill. Once it is stalled it now has a hard time restarting, as though it is not getting enough fuel or something.

On a side note, Subaru said that if it were the head gasket that was an issue, with these cars it would be visible once underneath the car. I am still suspicious…it is a phantom.

Please Help. Anyone?

A head gasket problem is not always evident to the eyeballs. Some head gaskets weep coolant or engine oil externally and this can be seen. This does NOT affect the way the engine runs.
Some head gaskets leak internally and may allow coolant to mix with engine oil. This also does NOT affect the way the engine performs. (although it can be damaging over the long run)
The only way a head gasket fault affects engine performance is if there is a breach into a combustion chamber. In this case, it will run poorly all of the time.

Just offhand, this sounds like a vacuum leak (symptoms are worse on a cold engine) or a problem with the Idle Air Control valve. An IAC fault does not always set a code. I can’t say as I was enamored of this diagnosis from the get-go as any vacuum leak is easily verifiable with a vacuum gauge and I’m not buying into the programming angle either.

Has there no one simply connected a vacuum gauge to see what’s going on? This tool usually takes a whopping 10 seconds to attach and reveals things instantly.
While the reading will vary based on a number of factors, what one generally looks for is a gauge reading of 17-20" inches of vacuum with the needle being rock steady.

i have a 2005 Subaru Outback 2.5 XT with this EXACT same issue. Has been happening on and off for over a year now. I was not able to replicate it at a garage until just yesterday. I would describe it as a reduction in power or long hesitation while attempting to accelerate, most noticeable within the first few minutes from first starting the vehicle. I never get on the throttle enough to engage the turbo so I assumed the turbo was not the issue (although there seem to be PLENTY of other issues with Subaru Turbos…).

Finally yesterday, I had it into my mechanic. He said Cylinder #2 is mis-firing. Went on to replace the coil and spark plug. This did not fix the issue. This morning he did a leak down test and found that 2 exhaust valves are “burned out”. I’m not very car tech-y (much better with bicycles and things that are man-powered), but it sounds like 2 of exhaust valves related to Cyclinder #2 are burned out and need replaced. In addition, ALL OTHER valves need an “adjustment” (which should have been done as part of routine maintenance at 105,000 when the previous owner had the car, but did not happen…).
We’re talking ~$3,000 to get 2 burned out valves replaced, 14 remaining valves “adjusted” which means head gaskets, shims, valve “seats”, and the heads have to be machined. The labor will be half or more of the total bill.

As of this moment I am exploring other options – trade, new engine, etc… The vehicle has 150,000 miles and has been perfect up until this point.

Anyone have any suggestions? Does this sound reasonable? I consulted with 4 other garages, one of which specializes in Subaru Repairs, and he says he sees this “all the time” as a result of skipping the recommended valve adjustment at 105,000 miles.

I’m worried that spending $3k now will only allow the engine to last long enough for the Turbo to Blow, which it sounds like is a good possibility within the next 15 to 150,000 miles…

I am having similar problem with my 05 XT, although I did have a turbo failure last month, and the rough running occured after the turbo was replaced, $2700 in Rochester, MN. I have 80k on mine, and was told that I need to change oil every 3k mi to keep bearings in turbo lubed.
I too am not ready to dump another 3k into this car, but would really like it in good mech shape.
BTW did you or previous owner change water pump timing belt? If so do you know what mileage it was changed?

I had the same problem. Idles fine when cold, rough and stalls after the engine is hot. No CEL light.
After cleaning the MAF sensor, MAP sensor, EGR valve and throttle body (twice), replacing the upstream O₂ sensor (downstream previously replaced), then decarbonizing the intake manifold (twice, using CRC intake valve and turbo cleaner), sprayed carb cleaner on all vacuum hoses and their connections, I gave up and took it to my mechanic.
The mechanic diagnosed it as a bad throttle, while also stating that “throttles rarely go bad”. If my other car was working, I would have paid a visit to a local scrap yard to pull 2 used throttles (2006-08, most Subaru models, so they should be abundant) but I didn’t. They installed a new $750 throttle.
It DIDN’T fix it!!!
I took it back to them and they replaced the ignition coil - the one (NGK) I installed to replace the one ("economy favorite at RockAuto) I installed to replace the original coil, which had a cracked housing on one of the wire connections. They installed a Beck Arnley coil (the only aftermarket brand they trust) and it seems to have fixed it. No trouble after two 1/2 hour drives and two 2-hour drives. Idles correctly now.

Good investigation there @mauser . Thanks for update. I wouldn’t have guessed the throttle as the problem myself. Poor engine performance only when hot, me, I’d suspect a problem w/ ignition system electronics, coils, or crank & cam position sensors as the first guess. Lots of other possibilites of course, heat can adversely affect most engine parts, fuel pump, other parts of fuel system, any of the various circuit boards in the car could have broken traces, etc. Fuel trim test would be worth doing in any event. For problems with idling when cold, then I’d be looking more at the throttle.

For some reason I’ve never had a problem w/heat affecting engine performance on any of my cars. I did have a difficult to diagnose rain-related problem with an ignition coil on my VW Rabbit. Only figured it out after I removed coil and inspected it with magnifying glass. Problem was an almost imperceptible crack in the casing.

The awful thing is that it wasn’t the throttle. I think my mechanic dismissed the possibility of the ignition coil because it had already been replaced a couple of years ago.
I guess the approach should be: replace the cheap parts first…

IMHO the best approach is to to complete the diagnostic process before replacing any parts at all. The manufacturer usually provides a sequence of steps in their shop manual procedures that defines the diagnostic process. That’s the better approach, unless there’s a part that is obviously not working just by visual inspection. In some cases there’s so many interconnecting links in the chain of possibilities that there’s just no practical alternative but to make a guess and replace a part, measure the effect of the replaced part. But that should be considered part of the diagnostic process. If there are multiple and equally-likely suspects for which part to replace, replacing the less expensive one then makes sense.