Cause of not starting when gas low, then starting after 2 pumps of gas pedal?

[mountainwalker]

This is for our 2003 Outback VDC 3L H6 which is in very good condition. Never happened before, but happened twice in the last few weeks. We usually fill up way before the gas tank empties, usually at ½ or 1/3 tank, but sometimes get close to empty. When we have gotten close to the empty line or light and parked the car for the night, we’ve never had a problem starting the car on the first try the next day. Recently twice when we let the gas go very close to the empty line or light, parked it and returned to start it the next day, the car didn’t start right away. It clearly wasn’t an electrical problem so I pumped the gas pedal twice and the car started fine. After driving the car for 15 minutes still without filling and parking, returned 45 minutes later and car started fine.

 

What could be the cause of the car not starting right away when parked for the night when close to empty? Dirty injectors?

 

We do fill up from time to time on Chevron Techron before a longer drive to keep the fuel system clean (supposed to be same as using a fuel system cleaner).

[Fairtax4me]

Pumping the gas pedal has no effect on fuel injected cars. You probably have an issue with fuel pressure, not an injector issue. There are a few threads recently discussing a fixture on the fuel pump assembly that causes fuel pressure to bleed off. Seems to be an increasingly common issue on the 2000 and later models, especially with the 6 cylinder engine.

 

Here's one thread about it: http://www.ultimatesubaru.org/forum/showthread.php?t=135288

 

Waiting to hear if the proposed fix in that thread will work. The only current way to fix the problem is to buy a used part, which will likely fail again in the not too distant future, or buy an entire fuel pump assembly from Subaru to the tune of around $400.

Edited August 13, 2012 by Fairtax4me

[mountainwalker]

Thanks, similar problem written about at Subaru Outback Forums:

 

http://www.subaruoutback.org/forums/110-gen-2-2000-2004/45780-gen-2-h6-fuel-pump-o-ring-failure-ethanol.html

[mountainwalker]

Asked the better of the two local dealerships if they're familiar with this problem - they said they just replaced the entire fuel assembly at a cost of about $785 in parts and labor. Ouch. Our indie garage is great and always much more reasonable, but if they can't source the parts they'd still have to get a whole assembly.

 

Does anyone have a good source for the cap and o-ring parts aside from junk yards (and if the original Subaru parts are designed or made badly to begin with won't they just fail again)?

Edited August 13, 2012 by mountainwalker

[CNY_Dave]

Someone is testing out the 2004 replacement fuel-filter that comes with a cap. The cap may or may not fit.

 

Just that cap may be available at the dealer.

[idosubaru]

the H6's around that year have weak throttle position sensors. i've never heard of one not allowing it to start like you suggest and that certainly wouldn't be fuel level dependent....but if we assume the fuel level was coincidental and i have seen TPS's cause no starts in other Subaru's (just not EZ30's) then maybe it's possible.

 

one of mine also has a funny characteristic, and oddly enough it only does it on long extended drives - it will "run out of gas" right around the time the low fuel light comes on. done it multiple times, i just don't let it get down to empty and all is good

 

if the original Subaru parts are designed or made badly to begin with won't they just fail again?

that's what i keep wondering. what is *causing* it to fail?

 

maybe replacing the fuel filter is a good idea...if it gets clogged could that increase pressure on the pump components? doesn't seem likely to me but something must be causing this?

Edited August 14, 2012 by grossgary

[mountainwalker]

grossgary, if it was the throttle position sensors, why would I be able to start after 2 pumps of the gas pedal?

 

Don't know if related, but a few years back we had an issue with the throttle dying when coming down off of speed, say coming off a highway, and then stopping at a red light. The idle speed would fall low enough that it would just stall. Started right back up. Only happened about 3 times, and then we had a major servicing, and then it never happened again.

[1 Lucky Texan]

would depressing the pedal to the floor put the system into 'flood clear' mode?

 

wondering if an injector is leaking fuel into a cylinder - just doesn't seem likely to prevent starting with a 6 cyl. car, I dunno.

 

mountain - do you ever smell fuel during the failed start events?

 

Fair, gg, cny - mountain once said he sometimes hears a whooshing noise when removing the gas cap, is there a tank vent or other vapor recovery/vent system failure that could relate?

[mountainwalker]

The whooshing we sometimes (not always) hear when removing the gas cap is a vacuum inside the tank - meaning air rushes inside the tank, not out. Usually hear that after driving for a while at highway speed and then stopping to fill up.

 

Remember what I've reported about the failed start has only happened twice ever and in the last few weeks, and only after parking the car overnight when the car was at or right near empty the day before. Two presses of the pedal, and it started. No smell of fuel.

 

Have heard people say you shouldn't let the gas get down too low for fear of gunk floating in the tank clogging things up, but isn't that what a filter is for?

 

We only fill premium/super gas per Subaru instructions for our car.

[1 Lucky Texan]

The whooshing we sometimes (not always) hear when removing the gas cap is a vacuum inside the tank - meaning air rushes inside the tank, not out. Usually hear that after driving for a while at highway speed and then stopping to fill up.

 

Remember what I've reported about the failed start has only happened twice ever and in the last few weeks, and only after parking the car overnight when the car was at or right near empty the day before. Two presses of the pedal, and it started. No smell of fuel.

 

Have heard people say you shouldn't let the gas get down too low for fear of gunk floating in the tank clogging things up, but isn't that what a filter is for?

 

We only fill premium/super gas per Subaru instructions for our car.

 

 

I feel your car's symptoms seem more like an EVAP system problem(charcoal canister, purge solenoid or ???) , but there's no reason not to be prepared for the fuel pump 'cap' issue.

[CNY_Dave]

grossgary, if it was the throttle position sensors, why would I be able to start after 2 pumps of the gas pedal?

 

Don't know if related, but a few years back we had an issue with the throttle dying when coming down off of speed, say coming off a highway, and then stopping at a red light. The idle speed would fall low enough that it would just stall. Started right back up. Only happened about 3 times, and then we had a major servicing, and then it never happened again.

 

The TPS is a potentiometer where a wiper rubs along a resistive strip. Exercising the wiper can clear crud out of the way or get the wiper to 'wipe' along a slightly different path.

 

So pumping the gas could possibly affect how a bad TPS works.

[mountainwalker]

1 Lucky Texan, thinking you might be right about a leaky injector.

 

OK, so brought it in to our good indie garage, asked them to check for fuel filter cap and 0-ring deformation - none found. Only error code thrown was P0183 code (Fuel Temp Sensor A, High Input Circuit). No Check Engine Light, just the error code.

 

Mechanic couldn't replicate the no-start problem I was having. After leaving, I filled up a full tank of gas. Everything was fine until two days later. Went to start car in the afternoon after not driving it all day - no start - and this time it wouldn't start 3x until it started on the 4th try. I drove some blocks and then parked to see what would happen - car wouldn't start on first and second try again.

 

I'm thinking maybe a leaky injector that's flooding the cylinder? Any suggestions?

 

Found info related to the P0183 Error Code here:

 

http://www.ultimatesubaru.org/forum/showthread.php?t=77153

 

http://www.ultimatesubaru.org/forum/showthread.php?t=7578&highlight=P0183

 

I had copied and pasted summary notes from what I found in these threads:

 

Code:

P0183 Engine Fuel Temperature Sensor A Circuit High Input

 

Description:

The comprehensive component monitor (CCM) monitors the EFT sensor circuit to the PCM for high voltage. If voltage were to exceed a calibrated limit and a calibrated amount of time during testing, the test will fail.

 

Possible Causes:

*Open or Short to power in harness*

*Damaged EFT Sensor*

*Improper harnes connection*

*Damaged PCM*

 

Diagnostic Aides

Verify EFT-PID value to determine open or short

 

-It uses fuel temp to know when to run the integrity test. It basically makes sure you have no leaks in the fuel system, which would allow raw fuel in the atmosphere. In an endwrench article for enhanced evap soobs it says:

 

To test the integrity of the [fuel] system, the system shuts off all access of the tank to the outside atmosphere. [then a bunch of other steps].

 

The testing of this system is only conducted once per drive cycle and only under very specific conditions:

• The fuel tank must contain less than 9.6 gallons of fuel.

• The fuel temperature must be less than 113°F.

• Engine speed must be over 1500 RPM.

• Vehicle speed over 28 MPH.

• Throttle position must be mid-range.

• Intake manifold vacuum must be equivalent to cruising vacuum.

• 455 seconds must have elapsed since engine start.

 

-the ECU uses the fuel temp sensor input to know when it can do the evap fuel system integrity test.

 

-Typically when the CEL illuminates with a fuel temp sensor code, either the fuel filter is partially plugged, the pump is defective, or the sensor is bad. The fuel continuously circulates from the tank to the injection unit and back to the tank again whenever the pump is running. Any restriction (plugged fuel filter or pinched line) will cause the fuel to heat up and trigger a code.

BTW: in tank pumps are good because they are submerged in fuel which acts as a heat sink to keep the pump cool. Running the car on empty a lot can lead to premature failures of the fuel pump. Ideally you should always refill BEFORE the low fuel light comes on.

 

Also in winter you should keep the tank above 1/2 full all the time as doing so will reduce moisture buildup in the tank which will reduce the chance of a frozen fuel line, and reduces corrosion in the tank. It also means that you will have "extra" fuel in an emergency.

 

In addition, apparently having the wrong ECU for your model Subaru can also throw the P0183 code. Doubtful in my case as I'm sure I have the OEM ECU.

Edited August 24, 2012 by mountainwalker
added info I had to look up

[uniberp]

So it was the fuel filter?

[mountainwalker]

Thinking it could be a leaky injector. Also will have them look at fuel filter.

 

I have depressed the pedal 2-3 times after not starting the first time - it has started for me on the 3rd or 4th try after pumping the pedal 2-3 times.

 

Someone suggested that starting on throttle full open is a good way to test flooding - if it starts on full throttle, indicates flooding.

[uniberp]

or air filter? Can no longer actually cause a rich condition, but can starve it altogether.

[nickb21]

To expand on some of the gas pedal 'pumping' thoughts... I could have sworn I read on this board (or another) that holding the gas pedal down when cranking on many (EFI) cars (including subaru), will tell the ECU not to pump in additional fuel during cranking.

 

By cranking a few times before starting in combination with maybe the ECU detecting WOT when you crank - perhaps that is enough to clear out any fuel from a leaky injector? Maybe just holding down the pedal the 1st time while cranking would lead to a faster result? That is of course if the ECU does cut fuel.

Edited August 25, 2012 by nickb21
.

[1 Lucky Texan]

suppose one cylinder were flodded - seems likely the car would start anyway.

 

if ALL were flooded, then ,from what I understand, wot will help clear the xtra fuel and start the car - not sure how, likely cut back or kill the injector flow seems like the only option.

 

 

from alldatadiy;

 

*******************

How to Start a Flooded Car Engine

Q: Would you explain what could be reasons for flooding of an engine and how to start the car if flooding has occurred?...Lal Singh

A: So we’re all on the same page; a flooded engine is an engine that has received too much gas for the amount of oxygen available for combustion. When this condition occurs the inside of the cylinders, as well as the spark plugs, will become wet with gasoline. A cylinder wet with gas will result in a reduction of cylinder compression, loss of lubrication to the piston and rings and cause spark to the spark plug to be misdirected. Because the large majority of engines on the road today are fuel injected we’ll cover how to start them if flooded. The computer has a program in it that is called the Clear Flood Mode. For the computer to activate the Clear Flood Mode it needs to see information that is out of the norm. This abnormal information is the sensor on the throttle reporting to the computer that the throttle is being held wide open, along with a crank signal to the starter. This combo of information causes the computer to respond by reducing, and in some cases completely shutting off, the fuel injectors. This process allows air only to enter the cylinder and dry out the abundance of gas. As the correct balance of air and fuel return you will feel the engine start to fire at which point the throttle can be released and the computer will return to normal fuel delivery. **********************************************************************

[Fairtax4me]

A leaking injector should be evident if the fuel system is primed with the spark plugs removed.

 

You can also check for that with a fuel pressure test.

[mountainwalker]

So the problem went away on its own - no issue for the last 6 weeks. At least nothing other than removing the fuel pump to inspect it and putting it back.

 

The progression was: Several times the car didn't start on the first turn of the key when the gas was very low close to empty. One pump of the gas pedal and it started. After pulling out the pump to inspect it, nothing was found wrong. Then right after filling up a full tank of gas, it wouldn't start right at the gas station on the first try as well, until after a single pump of the gas pedal. Then the problem disappeared, though I've been staying at my usual above 1/4 tank full and haven't been testing things by running very low.

 

What I can't explain is why: A) Pumping the gas pedal once was all that was needed to start the car after it wouldn't start up on the first try. Why the issue went away.

 

I should mention that the Crank Shaft Pulley (Harmonic Balancer) just came apart on me yesterday, and I've just ordered a replacement, though don't see how it's connected to this.

 

Two thoughts:

 

1) My indie mechanic suggested that perhaps the fuel we Californians are limited to by state law may mean a higher chance of impurities getting into your fuel pump, and that whatever was in there, had a chance to get out after running on a new tank of fuel.

 

2) Because I had driven the car twice within the space of a very busy two weeks close down to empty, that perhaps some debris got into the filter, which when it was removed to look at it, or when it had a chance to run more, got out. I typically always fill up before 1/4 tank - that's generally a good habit to keep considering most modern fuel pumps are cooled by the gas they are sitting in, and that gunk floating at the top has more of a chance to get in when you're running on low gas.

[CNY_Dave]

I'm skeptical about the pump cooling thing as the pump sits, what, about halfway up in the space? So the gas could be just under a half and the pump would be out of the gas?

[mountainwalker]

That's what's so frustrating about a problem that isn't consistent. If it were consistently causing a problem, it's much easier to diagnose. Our mechanic couldn't duplicate the issue, and even for us, after it happened several times, it didn't happen again (though we've been sticking to our usual refilling before the tank is below 1/4 full these days - it was unusual for us to let the tank get down low close to empty twice in two weeks).

 

Anyone have a guess as to what would cause this a few times over 2 weeks after letting the gas tank get down close to empty twice, and cause this both when the tank was near empty but not quite empty, and again when the tank was full, and then not cause this for 6 weeks?

 

And could this possibly have any connection to the Crankshaft Pulley coming apart now? Don't think so, just checking.

 

What systems would you check now if any? Like to nip any issues before they become major problems, and before they hit you when it's least convenient.

Edited October 18, 2012 by mountainwalker

[mountainwalker]

So yesterday with a full tank of gas, when I tried to start the car, the dashboard lit up and I heard a clicking sound coming from behind the dash, but the starter didn't turn over or catch. Same thing on the second try. On the third try it started up. That happened each of the two times I used the car afterwards. I had just replaced the Harmonic Balancer (Crankshaft Pulley) a few days before, after the rubber/plastic on the pulley had separated.

 

Perhaps all along it's been the starter motor or gearing around the starter motor going bad. But can't understand why the problem hasn't been consistent.

[Fairtax4me]

Clicking could be the starter interlock relay, or could have been the solenoid on the starter. I would not think either of these relate to the previous issues.

 

The contacts in the starter solenoid are a common weak point and deteriorate heavily with use. They are easily repairable if you can find a supplier for the contacts.

 

Relays can go bad on occasion. There could also be a wiring problem such as a poor ground connection.