Help Troubleshooting Subaru 3.6L
 

Hi All, figured I'd ask the Subaru junkies-ahem-enthusiasts for advice.

My wife's car is a 08 Subaru Tribeca with a 3.6 H6 (EZ36D). Driving down the road, it started lurching as though it were misfiring. Surely enough CEL showed up. Stopped it, checked to make sure there was nothing visibly wrong, got it to a gas station to fill up because it was pretty low, then got it home. Pulled codes and sure enough, two codes for misfiring on cylinders 5 and 6. I believe these are the back cylinder on both banks.

So some history. This car doesn't get driven much, just had an oil change and has a new catback because the old one had a hole that failed inspection. I didn't have time to do it myself, so I let the inspecting shop do the recent work. I had driven the car 2 days before and there were no problems. It had about 1/8 tank of gas when this happened. The Tribeca is stored outside if that matters. I doubt any ignition components have ever been replaced.

An interesting note is that above ~2500-3000 rpm it drives fine. No lurching at all. It's a sudden change like hitting that rpm (don't know which rpm value precisely) flips a switch.

My troubleshooting so far:
I didn't see any leaks, oil level looks fine, and I filled the gas tank with 93 (it'll take 87 but I wanted to see if bad gas was the problem) and ran the engine for a bit in neutral to see if it got better. I do not think it's a matter of resetting codes because the physical symptoms are still there. Easy next step is spark plugs/coil packs, but it's a bit odd that two different cylinders would go out at exactly the same time? Another possibility a friend pointed out is fuel pressure problems.

Any thoughts?

Fuel filter.
Lower rpms not getting enough flow starving cylinders and getting misfires. Higher rpms create enough flow to run fine.
Sounds counter intuitive but I have had it with other cars.

Yeah, that was a possibility I talked about with my friend. I suppose I should start there. In the fuel tank... that's now full of fuel... :bonk:

Not sure where the filter is on those. Maybe get lucky and it is an in-line?

The internet says any Subaru after 2004 is in the tank.

Well that sucks.

Is there an access panel in the trunk or passenger area to do this through without having to drop the tank?

i had an 06 but it was just the 3.0 with variable lift and timing on the intake. no issues ever besides speed sensor wires to 140k.



move the coils around and see if the codes move to those cylinders. a ratcheting gear wrench or a 1/4" drive fits.

There is theoretically an access panel under the 2nd row seats. But it seems that I am going to have to drain the full 17 gallon tank first or capillary action is going to spray gasoline all over my interior... 17 gallons is a lot of gas to store... I don't have any tanks that large, and I probably shouldn't be driving it around to burn it.

Apparently the pump is inserted into the filter, so I have to disassemble that assembly then reassemble it with the new filter so it's not exactly a quick pull and replace. I suppose I can probably just buy the whole assembly.

That sucks!

i have the PDFs of the 06 service manual if you need a section.

I think I will be alright. Youtube is a wonderful place.

Were you speaking off the cuff, or did you really mean capillary action because, unless I am misunderstanding the situation, I am failing to see how that phenomenon applies in your situation. If you are referring to the tank being filled and overflowing when you access the port then that will happen. I'm pretty sure all you have to do is syphon off a quarter tank, so the fuel is at or below the height of the access port. There should be no need to drain the tank completely. If you are referring to fuel vapor pressure causing fuel to spray everywhere then that seems less likely if you have a functioning EVAP system. Regardless, siphoning the tank should purge the vapor.

If you disconnect the fuel lines before removing the pump assembly from the tank, the fuel will run up the narrow pipes connected to the pump assembly. I don't think you can remove the pump assembly before removing the fuel lines, but I could be mistaken.

Also, I'm pretty sure I saw this happen before, but I could be mistaken as to the physical cause if it.

Get a couple cans, some fuel hose the same size as the one feeding the rail, one of those barbed tube things to join the hoses, blah blah... Jumper at the fuel pump relay. Profit.


Might take a while. ... clogged filter and all.

You can purge the lines by disconnecting the fuel pump relay and turning the car over. If you don't then the line will be pressurized, so yes, when you pop off the lines you will spray fuel. A terrycloth towel around the connector should limit spray, but purging the line should limit the volume of fuel leakage to something fairly minimal.

That's not the problem. Watch this. I basically need a hose and stopper anyway, but I have to drain the tank so it comes out slowly enough to put a hose on before it gets everywhere.

https://youtube.com/watch?v=kSXQJja8AVI

New dilemma. Is the risk of the pump actually being messed up worth the ~$300 difference for just replacing the whole unit... and it would save me time... :mad0259:

(42072 vs 42021 or 42022)

https://parts.subaru.com/a/Subaru_20...10-421-03.html

You would find out if you try using it to drain the tank. ...maybe. If it drains slowly and the filter is dirty, then... else if filter is clean and it drains slowly, then..

It's a good thought but...

The filter looks to be sealed in a plastic case, you buy the case with the filtration element inside. I'm saying this because I don't think I can check the filter to see how dirty it is. Plus it would mean opening and resealing the tank twice because I have to wait for parts.

I searched the usual places like Alldata and a few technician sites but only found one reference to anything like your problem and the cause was a dying battery.

It makes sense in a way. Most alternators reach max output somewhere above 2K rpm so if your main battery is shot, it might take that many rpm to feed enough current demands to make the engine run normally.

An electrical load test might help and imagine a DIY oriented auto parts store has a free tester.

If you do go with a fuel system repair, definitely do it outdoors and not in the garage.

Good luck.

:(

Most motorcycles with an in-tank fuel pump publish Gallons-Per-Minute rate in the repair manual.

Disconnect the fuel line and route it to a holding can.
Apply a known voltage to the fuel pump and it should move X amount of fuel in Y minutes.
Then you know if it's the pump and filter or something else.

Also gets fuel out of the tank in the mean time and don't have to break into the tank.
There should be specs for this available somewhere.

=========================

Looks like the 86 manual only wants you to use an in-line pressure gauge and Techstream.
Bleah.

Good thoughts. Brought the battery to AutoZone to get tested. Read as having a low charge and they charged and tested it. It "tested bad" as in it probably wouldn't hold a full charge. It was 5 years old in a car that doesn't get driven much so I got a new one.

New battery installed, took like 10 seconds to start the car which I found a little odd. It cranked fine. High idle on startup was fine, but when it warmed up a bit and lowered the idle the same issue started happening. :cry:

I'm thinking taking a while to start might further indicate fuel flow problems? Alternatively, the alternator could have an issue resulting in a similar situation to a bad battery causing low fuel pressure...

Disconnect and plug the vacuum to the pressure regulator. See if it changes how it idles. If Tcoat's theory holds, high manifold vacuum at idle increases return flow above the pump's ability to maintain pressure. It makes sense. Best thing would be to actually measure rail pressure.

So, let's pretend I don't know what vaccum to what regulator you are talking about (it's the internet therefore I really know everything right? ;) )

Are you saying that the fuel pressure regulator is controlled by some air vaccum line to the manifold? Not familiar with how fueling systems work so sorry if I misunderstood.

Assuming that is true, if I disconnect the regulator from the line, I need to plug the end of the line to maintain correct manifold pressure, and leave the regulator open to the atmosphere to see what that does to idle. What would you expect it to do if Tcoat is correct, versus some other problem?

Also, is it safe to start and run the engine this way? I would think there is a reason they need the fuel pressure regulator in the first place?

Yup, but I was wrong. Just looked it up. They went to a returnless system before 2008. Oooooops. :(

I'm pretty sure it also blows holes in Tcoat's theory. The only way to tell for sure is to test fuel pressure real time.

To answer your last question, disabling the pressure regulator would raise rail pressure but not to some crazy unbounded level. It would be the same as when manifold vacuum is lowest - at initial startup and what would be delivered near to full throttle. It would run rich at the low end. Not a showstopper, and would confirm the issue.

I don't see how it invalidates the theory that a clogged fuel filter is preventing flow at lower pump settings.

Possible culprits:
1. Fuel filter is clogged
2. Fuel pump is dieing
3. Some fuel pressure regulation device is malfunctioning.
4. Coils/plugs
5. Alternator (possibly also the reason the battery was shot?)

I can probably check the the Alternator voltage output pretty easily, (obd readout or multimeter) to make sure it's at 14V+ right?

That just leaves 1-4. I think it's going to be more cost/time effective to just replace the filter and pump together than try to rig something together to figure out how to measure fuel pressure in a running engine and fi more diagnosis. If that costs me $300 and an afternoon to learn that wasn't it, well I learn better the hard way anyway lol. Plus, the filter at least is overdue for replacement anyway.

If that doesn't solve the problem, the plugs are due for replacement, and then coil packs. If I am still in the dark after that, well, it might be time to pay somebody to fix it...

Edit, first, I'm going to inspect the fuel lines under the car. I wonder if they may have been damaged when I got the cat inspected...

Fuel delivery issues are still in play but I can't draw a straight line to the counter-intuitive poor performance at idle while improving under power. That we are dealing with a returnless system invalidates only my specific case in support of Tcoat's observation.

Nothing wrong with replacing the pump assembly. It would eliminate 1-3 in one fell swoop. You still have to drain the tank to replace the pump. 300 bucks would buy lots of pressure gauges. ;)

Definitely check the charging system. Do the headlights dim at idle? Does the engine stumble when everything's turned on? Cabin fan, headlights, brake lights all at once?

It occurs to me that we haven't discussed relocating coil packs. If the misfire is consistently on 5 and 6, swap a couple. They don't last forever. Worn (or conventional) plugs are hell on coil packs.

Thanks for the ideas.

Fuel pump pressure is controllable electronically, so the best guess is it lowers the target pressure at lower rpm to put less strain on the pump, then steps it up as more fuel is needed. At some point that increase results in enough pressure to keep all 6 injectors going. It may not have a feedback loop in the control of the pump, so it is not adapting to increased pressure demand.

I think I might have a fuel pressure test kit from my grandfather, probably need a bunch of modern connectors though lol. What should the pressure be? I cannot seem find specs using my apparently weak google-fu. I suppose if it's lower at idle than it is primed but no flow that is evidence that the filter is the problem?

Everything electronic I can control works fine all together, and ecu voltage is still reading 14.2ish so I think charging is fine? Engine stumbles at low rpm as described, and is remarkably consistent.

Yeah I hear you on coil packs, but the chances of two different packs on different banks going at the exact same time are pretty low. I cleared the codes and checked again, still 5 and 6. I suppose it's cheaper to try to swap coils around before getting that in-tank assembly though...

reset the codes, swap the coil packs to 1 and 3, drive and then check codes.



if it stays on 5 and 6, check the plugs. if the plug holes are swimming in oil, get the valve cover gasket kit for both sides.

Is that a thing for these engines? I can picture it.

Spent a few hours this morning shuffling plugs and coil packs around, using every single socket, wrench, and adaptor in my arsenal of course (ok just the 10mm and 12mm sockets and wrenches). Reset codes, still reports misfiring on 5 and 6. No visible damage to coils, plugs were dry and looked fine.

Sooo, fuel supply next?

At least I can say I've done spark plug stuff on a Subaru lol.

Another long shot: The misfires are isolated to #5-6, it isn't coil packs so yeah, something's off with fuel delivery.

I recall that D4-S mapping is compound DI/Port at lower rpm, exclusively DI at midrange, then both again at the high end. So, could the #5-6 port injectors be clogged? Maybe bench testing them would help.


https://www.ft86club.com/forums/atta...1&d=1338416188

see if the intake manifold bolts are tight on the drivers side. (it's an o-ring type gasket.)

Why would a Subaru car without Toyota's D4S have this problem?

Cylinders 5 and 6 are on opposite banks...

That's putting it nicely!

yes, i finally found the diagram in the manual. a google fail on my part.





the 06 manual shows a drain plug on the tank.

coil packs
spark plugs
injectors
valve springs

My FR-S was checked in that order for the exact same problem. It was valve springs.