Please help - '03 LLB OBW turns over but won't fire

[live2fish]

I love my Subaru - 2003 LL Bean OBW H6-3.0L - even though I had to have it towed home Thursday night.

I'm new here so if my answers are spelled out elsewhere, please refer me to them - TIA.

I drove about 2 miles, came to a full stop, and started off again. Then the engine felt like it wasn't getting enough gas and it started to "sputter". I looked at the gas gauge, it was about 1/8 tank; I pumped the gas pedal and the engine took off again. I stopped to fill up with gas and it didn't start back up again.

I checked the battery - 12.5 V with ignition off, 11.0 V when cranking, and two weeks ago I measured when running at 13.8 V (I thoought the alternator was squeaking). Tried a jump-start anyways, didn't start.

I checked the fuel filter - gas pumping in and out. Changed it anyways as they were on sale.

I checked the fuse panel and all the fuses are okay - did not check the relays but read here that there are no relays for the ignition. In this forum I read that there are coils for each plug and the power to the coils is regulated by the on board computer?

I disconnected and reconnected the negative battery clamp thinking it mght "reset" something and tried to start the engine. Still only cranks but won't fire for more than a couple seconds - or not at all.

My CEL comes on intermittently, had it checked once and was told it was an O2 sensor. I had this happen on a 1993 Subaru and drove it w/o issues for years so I have ignored it with this car too (I feel your eyes rolling). I don't know which one it was.

If I understand correctly, there is no Haynes Manual for the H6-3.0L? I could check ALLDATA at my library for info - fixed my A/C with info from that site.

It doesn't seem likely that all 6 coils/plugs would fail at once? My fear is that it could be the timing chain (not a belt if I understand correctly). My hope is that I am overlooking something simple.

I will have it towed in for repairs if I can't get it running, but I have more time than money right now!

Thanks for any help to figure this out. What is the next step?

 

live2fish

)><(())*>

 

P.S. 150,000 miles, bought with 135,000 about a year ago.

Edited March 27, 2010 by live2fish
add miles

[Gloyale]

Map sensor?

[live2fish]

Map sensor?

 

Gloyale,

 

Thanks for the reply. I am a newbie, so please forgive all the questions.

This map sensor (intake manifold air pressure sensor?) is mounted on the throttle body, correct? How does the MAP sensor [or it's malfunction] prevent the engine from starting? What is the test for it's correct operation?

I didn't see or feel any cracked/brittle hoses. Would a vaccuum leak by itself prevent engine start-up?

I removed two spark plugs and they looked pretty new - NGK PLFR6A-11. Spark plugs were not easily accessible!

 

Thanks Again,

live2fish

)><(())*>

[Fairtax4me]

MAP sensor should not prevent the engine from starting. A crank or cam angle sensor would do that. a bad coolant temp sensor could also flood the engine and keep it from starting.

Have you tried putting the throttle pedal on the floor while cranking? Crank it for 10-15 seconds first, then floor the throttle, then crank again for 10-15 seconds. Setting the throttle at WOT (wide open throttle) will tell the ECU to cut the fuel injectors while cranking which will clear unburned fuel from the combustion chambers.

 

Another thought is, could be a massive vacuum leak. Check the brake booster hose for cracks/splits and proper attachment at the booster and throttle body/ intake manifold. Check the PCV hoses for the same at the manifold and at the block.

Check the large breather hoses on the intake box and make sure they're all still attached.

[bheinen74]

unplugged coil pack?

unplugged ignitor

vacuum hose off.

[live2fish]

Thanks for the pointers! I tried unplugging the coolant temp sensor and starting it up before this (I read about it somewhere on this board so I gave it a try) and it made no difference.

I'll recheck the vaccuum hoses and I'll check the PCV valve hoses. Coils are all plugged in.

 

I got my hands on a scanner and the only trouble code I could read from the ECM was P0118 - Engine Coolant Temp Circuit high input.

 

The Freeze Data Frame function read:

DTC that caused freeze frame

P0118

Engine RPMs

158 RPM

Load Value

1.9%

MAP Sensor

97KPA

Coolant Temp

-40ºF

Short Term Fuel Trim1

0.0%

Long Term Fuel Trim1

0.0%

Short Term Fuel Trim2

0.0%

Long Term Fuel Trim2

0.0%

Short Term Fuel Trim3

-100.6%

Long Term Fuel Trim3

-100.6%

Short Term Fuel Trim4

-100.6%

Long Term Fuel Trim4

-100.6%

Vehicle Speed

0 MPH

Fuel System 1

Open

Fuel System 2

Open

 

Newbie questions again. Do I replace the engine Coolant Temp Sensor or is there a wiring/contact problem between the ECM and the sensor? What do "Fuel System 1 Open" and "Fuel System 2 Open" mean - open circuits, bad fuel injectors, or...?

 

Thanks Again,

 

live2fish

)><(())*>

[CNY_Dave]

Fuel pump could be giving flow with no real pressure, or pressure with no real flow.

 

Really need a gauge on the rail while cranking.

 

I thought on some subies, if the MAP/MAF went bad it might not run at all, but would run in limp mode if it was disconnected.

 

Also, is tapping the cam (crank?) sensor one of the tricks?

 

Dave

[OB99W]

[...] Do I replace the engine Coolant Temp Sensor or is there a wiring/contact problem between the ECM and the sensor? [...]

ECTS failure is common enough, but wiring/contact problems are also a possibility. There are a few places that contacts could be corroded, although there are two primary ones. First look at the ECTS connector itself. If that's okay, check the large multi-pin connectors that couple the engine harness to the bulkhead harness at the rear passenger side of the engine. If those seem good, the odds are better that the ECTS itself is bad rather than there being a wiring problem.

 

If you decide to change the ECTS, disconnect the battery negative cable before pulling the sensor's connector. Don't overly tighten the new sensor, or you could cause damage -- the factory manual says it's supposed to be torqued to 16 N-m (which converts to 12 ft lbs ... although the manual says 1.2 ft-lbs, which might be a bit loose ).

 

Don't be concerned about the fuel trim numbers, etc., for now. If the ECU is getting bad coolant temp info, that will be reflected in the fuel delivery.

[oldgray1]

I may be way off base, but my 2001 LLB had the same (or similar) symptoms, but I remember it threw a code..forget the code number..that pointed to the Idle Air Control (IAC) valve. Ordered a new valve and gasket; installed and started right up. Cleared the code with a scanner, no problem after that . Also not a bad idea to shoot some Seafoam or other throttle body cleaner into the throttle body and get that whole intake cleaned out. Mileage was about 155000 and 6 cylinder, auto, etc.

[CNY_Dave]

I may be way off base, but my 2001 LLB had the same (or similar) symptoms, but I remember it threw a code..forget the code number..that pointed to the Idle Air Control (IAC) valve. Ordered a new valve and gasket; installed and started right up. Cleared the code with a scanner, no problem after that . Also not a bad idea to shoot some Seafoam or other throttle body cleaner into the throttle body and get that whole intake cleaned out. Mileage was about 155000 and 6 cylinder, auto, etc.

 

Would ir start if you gave it a little throttle?

 

Dave

[oldgray1]

To Dave,

I seem to recall giving more throttle did not help.

[live2fish]

Thanks again for all the advice. I followed the testing procedures from ALLDATA.com for the ECTS. The first step required an engine start and a read of the ECTS data with a scan tool. I couldn't do that as the car doesn't start and the scan tool I borrowed doesn't read "live data". The rest of the steps led me to the last step and instructed to replace ECTS: the resistance between connnector E8 pin # 2 and engine ground was less than 5 ohms - 4.5 ohms. Of course the meter could also be off +/- 0.5 ohms.

 

If a bad temp sensor was preventing start-up, would the engine start simply by plugging into a new ECTS (not threaded into the engine) that would be detecting ambient temp - 60º? I guessed that it would, but it did not start, either because the ECTS isn't the problem or the battery was weak from previous testing and I forgot to recharge it.

 

I was told a "quick and dirty" method of testing for fuel pressure was to try to squeeze the fuel line closed downstream of the filter while cranking the engine with the gas pedal to the floor....if you could squeeze it closed when the fuel pump was pumping gas, there was a good chance that the fuel pump was bad. Is that a good indicator or not?

 

When I was going to give it another shot, I spotted a cracked plastic hose bracket/holder. It was molded to the hard plastic PCV tube that threads between the throttle body and the "air chamber" (newbie name for the big black plastic thing that sits on top of the throttle body). Upon further investigation I found a hairline crack right along the hose production "seam" and it split wide open when I pulled the PCV hose off the connector at the top of the engine. I missed this twice while checking the vaccuum and PCV hoses before - maybe I need glasses!

 

Would a cracked PCV hose be enough to prevent the engine from firing?

 

As I plod along with this project I am rumbling down the road in my trusty rusty '86 Chevy 3/4 ton plow truck. I just hope it holds together long enough to get my Soob running! LOL! Most of my mechanical experience has been from working on the old Chevy and replacing Subaru front axles. I think I am finally getting my head wrapped around the "ECM" system running my vehicle.

 

I'll replace the hose and fire it up again - the battery was on trickle charge all night and is charged back up - might get a new battery too.

 

Thanks again for all the pointers!

 

live2fish

)><(())*>

[live2fish]

I'm back after working on the Subaru. I tried most all the suggestions and then some. Here's what I've done so far:

-Replaced ECTS and that took care of the error code.

-Replaced the battery (it was old and drawing down fast with all the cranking)

-Replaced the plugs (tips were rounded, but no unusual wear).

-Checked for spark on all of them; they all have spark.

-Replaced the fuel filter.

-Replaced the fuel pump; pressure is better.

-Tapped on the camshaft and crankshaft sensors.

-Checked the vaccuum hoses, PCV hoses & tubes; replaced cracked PCV tube; didn't find any disconnected or other cracked hoses.

-Started engine twice by using some starter fluid in top intake PCV tube (I saw that some use Sea Foam in the intake to clean it ).

 

The engine started and ran very rough.

-While running checked for injector "noise" with screwdriver and detected slight vibrations.

-While running checked for fuel pressure by trying to squeeze fuel line closed at the output side of fuel filter - I could squeeze down on it some, but not all the way closed as before I replaced the fuel pump (pressure tester I borrowed wasn't working, needle didn't move).

-When I gave it some throttle, the engine stalled and wouldn't restart.

 

I scanned it again and got "cylinder misfire" codes for each and every cylinder 1 through 6. I started the code diagnosis and got to the resistance test between the injector ECM connectors and chassis ground and got "no reading" or zero on the two I tested before I hung it up for the night. I know the electric meter works.

 

Is the problem a missing ground in the injector engine harness and if so, what is the repair?

 

Thanks again for the help and suggestions. I've learned a lot from this forum.

 

live2fish

)><(())*>

[WoodsWagon]

Stop using ether to run the engine, that will do permanent damage to the engine. Bad ************ to have anywhere near a gas engine. The only thing you can use ether on is a direct-injection diesel, nothing else should ever have ether put through it. I have personally shattered a piston with ether.

You'll be replacing an engine rather than diagnosing a no-start before too long if you keep using ether. Use propane or gasoline in a squirt bottle.

[Fairtax4me]

Ever had the timing chain replaced?

I wonder if it would be worth having a look at the timing marks to verify that the chain didn't jump.

[CNY_Dave]

I'm back after working on the Subaru. I tried most all the suggestions and then some. Here's what I've done so far:

 

-Replaced the fuel pump; pressure is better.

 

The engine started and ran very rough.

 

-While running checked for injector "noise" with screwdriver and detected slight vibrations.

 

-While running checked for fuel pressure by trying to squeeze fuel line closed at the output side of fuel filter - I could squeeze down on it some, but not all the way closed as before I replaced the fuel pump (pressure tester I borrowed wasn't working, needle didn't move).

 

-When I gave it some throttle, the engine stalled and wouldn't restart.

 

I scanned it again and got "cylinder misfire" codes for each and every cylinder 1 through 6. I started the code diagnosis and got to the resistance test between the injector ECM connectors and chassis ground and got "no reading" or zero on the two I tested before I hung it up for the night. I know the electric meter works.

 

Is the problem a missing ground in the injector engine harness and if so, what is the repair?

 

Thanks again for the help and suggestions. I've learned a lot from this forum.

 

live2fish

)><(())*>

 

I would say the global misfires and the lack of good ground are consistent with it being able to idle but not take any throttle.

 

Tracking down a bad/missing ground connection without a wiring diagram will be tough- you are going to have to check all the grounding points that are anywhere near the injector loom, and work backwards into the loom until you find the bad spot.

 

The problem with the ground (if this is the problem) is likely right at the grounding point, very close to it, or where the harness gets chafed/pinched by something.

 

And, you really need to get a working fuel pressure gauge on it- there could still be a problem there.

 

Dave

[OB99W]

Ever had the timing chain replaced?

I wonder if it would be worth having a look at the timing marks to verify that the chain didn't jump.

See my posts in http://www.ultimatesubaru.org/forum/showthread.php?t=104712 . (Actually, read the entire thread. ) Compression testing of all cylinders would be very useful info, especially since the work that's been done hasn't lead to a solution yet.

[oldgray1]

Here's another wild one. Had a Dodge Dakota with a "turns over, but no fire" symptoms recently. Went to the forums for information... Heard everything..plugs, wires, coil/ coil pack, cam sensor, temp sensor, fuel pump , etc. Turned out to be the crank sensor. Don't know if our Subarus work the same way, but the computer on the Dakota won't send a signal to fire the coil and get the whole thing going without that signal input from the crank sensor. You could crank the engine all day long and it would never start. I'm certainly not much more than the average shade tree guy on this stuff, so thanks for reading. By the way, the actual crank sensor itself was not bad. turned out that a mouse, or some other marsh rodent, had chewed through the wire and made the sensor useless. So, in support of all others who suggested checking wires, connections, etc., it's all good . It makes you feel pretty good to solve some of the real tough ones without just throwing parts (and money$$) at it. Just my 2.5 cents!!

[Fairtax4me]

Yeah the cranks sensor idea isn't so "wild" on these. Those things go bad all the time. And yes, it does work the same way on Subarus, as well as on pretty much every other engine that has computer controls.

 

You'll want to find a FSM for the car for wiring diagrams and ground locations if you think it's a wiring issue.

[live2fish]

Thanks again for all the pointers. I will not use starting fluid in my Subaru again. I don’t think it did permanent damage.

 

I do have a manual and wiring schematics, and I have some limited experience with diagnosing electrical problems. Step 2 of the diagnostic procedure asks if signal voltage between ECM connector and chassis ground on faulty cylinders exceeds 10 V with key “on”. I get no voltage with red lead on ECM connector B137 - #1, B136 - #6, #5, #4, #3, or #1 and black lead to ground. That doesn’t seem right, but it’s less than 10 V, so step #3.

 

I brought home the “Fluke” meter from work to double check the resistance test between the fuel injector connectors and engine ground – still nothing - tested all 6. Step 3 of the diagnostic procedure says that if measured value is less than 10 Ω to “repair ground short circuit in harness between fuel injector and ECM connector.”

 

Even with the diagrams it doesn’t seem like it will be easy to find the bad ground. Injectors connect to the ECM through connector B22 – the 6 wires tested for voltage in step 2. The injector yellow/blue wires are all connected to the camshaft position sensor connector which is connected to the ECM by two wires through connector B21 and also to four solenoids through connector B20. From there the yellow/blue wire makes its way back to the ECM and the main relay which are grounded through connector B47 - #1. All of these are part of the engine harness and grounded by either GB-3, GB-5, or GB-6, I think.

 

The closest grounds are two 4-wire engine grounds bolted onto the top of the intake that come right out of the wire loom under the intake. The loom is kind of pinched there with a tight zip-tie.

 

I don't have my hands on a compression tester or a working fuel pressure tester yet. That crank postion sensor sure is tough to get at!

 

I'd really like to get this car running, tired of driving the plow truck around.

 

Wish me luck.

 

live2fish

)><(())*>

[OB99W]

Thanks for posting all the readings, etc., and for making it obvious that you have "experience with diagnosing electrical problems". Based on that, I think we can talk you through this.

 

The yellow/blue wires are power supply for the injectors and solenoids that you mentioned. There should be battery voltage (~12 volts) on them when the key is "ON". If not (which is what seems to be the case, but please verify that and let us know), then either power isn't getting to the main relay contacts, or the relay coil isn't being energized. (It's also possible that the main relay has failed, but that's less likely.)

 

I know that you checked fuses already, but please verify again that SBF-5 is good (feeds power to main relay contacts). Then, with key "ON", check voltage at ECU connector B134 pin #2 (power to main relay coil, according to the 2001 H6 diagram I have on hand).

 

Get back to us with your findings.

[OB99W]

I forgot to mention -- also check the B22/E3 pin #10 connection (should be where the yellow/blue wire power is fed).

[live2fish]

I am grateful for the help and advice I am receiving here from experienced Subaru technicians, mechanics, and others. Thanks for the help OB99W. Measure twice, cut once, I always say – good advice.

 

Answers to your questions:

- KEY ON – 11.8 V at each injector connector yellow/blue wire.

- Checked SBF-5 and it is good - has continuity.

 

I have to say I didn’t like where you were going with this next because it was “Step 5 Engine Starting Failure Diagnostics – ECM Power Supply & Ground” and my probes had been too worn to go through the back of the ECM connector do it before. I have the good meter now so I crossed my fingers and ….

 

- …KEY ON – 10.9 V between B134 pin #2 and ground! Really, exceeds 10 V, I tested it three times using different chassis grounds, cycling power each time to make sure. Whew!! How much does an ECM run these days anyways? Go to step #6 (down there ↓).

 

- KEY ON – 11.8 V B22 #10 – GE.

 

I have done the previous steps 1–4 and gone through steps 6-9 also:

 

Step 1 – resistance between B47 main relay terminals 3 – 5 & 4 – 6 less than 10 Ω when relay terminals 1 & 2 are powered up with 12 V. Go to step #2.

 

Step 2 – resistance between B134 #22; B136 #8, #17, #18; B137 #8, #9, #21, #31 and ground all less than 5 Ω. All zeros. Go to step #3.

 

Step 3 – KEY ON: 11.9 V B137 #10 – ground, and then it got weird.

Found 0.0 V B134 #14 – ground; B134 #14 wire is supposed to be green/red and it is orange. B134 #14 is just one up and one over from B134 #22 black/white wire (terminal engine ground, I think). Be nice if an ECM connector diagram was pasted on the inside of the protective cover plate.

Found a green/red wire at B134 #3 and it was 11.6 V B134 #3 – ground. If that’s the right wire for the ECM input voltage, we’re golden, otherwise??

B134 #3 is supposed to be violet/green and goes to the rear defogger according to “E/G(LH6)-03”, pp WI-179. Go to step #4?

 

Step 4 – resistance between B134 #2 – ground > 1Ω. Go to step #5 (up there ↑).

 

Step 6 – I’m not sure if this test is KEY ON or OFF; I think ON because the previous step #5 was KEY ON and it doesn’t say to turn ignition off until step #7. I did it both ways with two different outcomes.

KEY ON: 10.9 V B47 main relay terminal #2 – ground > 10 V so go to step #7.

KEY OFF: 0.0 V B47 main relay terminal #2 – ground < 10 V; it directs to repair open circuit between ECM connector and main relay.

 

Step 7 - KEY OFF: resistance between B47 main relay – ground ≈ 0.3 < 5 Ω. Go to step #8.

 

Step 8 – KEY OFF: 12.2 V B47 #5 – ground; 12.2 V B47 #6 – ground > 10 V. Go to step #9.

 

Step 9 – KEY ON: 0.0 V B137 #2 – GB & 0.0 V B137 #3 – GB < 10 V.

(The yellow/blue power wires are too fat and I measured voltage with the connector unplugged - which I don’t think would change the input voltage result?).

Diagnostics direct to “repair open or ground short circuit in harness between ECM connector and main relay connector.”

 

It appears I have an open circuit between B47 #4 and B137 #2 & #3. The diagram shows that the yellow/blue wire that runs from B47 #4 to B137 #3 and the yellow/blue wire that runs from B137 #2 to the solenoids and the injectors intersect and connect somewhere in the loom between the ECM connector and the main relay. The two wires must still be connected in the loom which might explain why I get power to the injectors, but no start because there is no ECM signal to the injectors and who knows what else? Maybe the connections in the ECM socket at B137 #2 & #3 are bad.

 

Should I pierce the wire insulation just before the socket on #2 & #3 and check for the voltage? Should I go buy a box of electrical tape? Do they make a pin extractor that small? I suppose.

 

I have this [false] sense of being closer to finding the problem and the solution. Thanks again for all the help!

 

live2fish

)><(())*>

[OB99W]

[...]

Step 9 – KEY ON: 0.0 V B137 #2 – GB & 0.0 V B137 #3 – GB < 10 V.

(The yellow/blue power wires are too fat and I measured voltage with the connector unplugged - which I don’t think would change the input voltage result?).

Diagnostics direct to “repair open or ground short circuit in harness between ECM connector and main relay connector.”

 

It appears I have an open circuit between B47 #4 and B137 #2 & #3. The diagram shows that the yellow/blue wire that runs from B47 #4 to B137 #3 and the yellow/blue wire that runs from B137 #2 to the solenoids and the injectors intersect and connect somewhere in the loom between the ECM connector and the main relay. The two wires must still be connected in the loom which might explain why I get power to the injectors, but no start because there is no ECM signal to the injectors and who knows what else? Maybe the connections in the ECM socket at B137 #2 & #3 are bad. [...]

You should be seeing ~12 volts (assuming the battery is sufficiently charged) on all the interconnected yellow/blue wires with the key "ON". Since the voltage is there at the injectors, the main relay is closing as it should. Not seeing it at ECU B137 #2 and #3 certainly indicates something's amiss. If the meter probes are too thick to backprobe the connector, a piece of thin bare wire, a straightened paper clip, or even a pin can be inserted in the back of the connector to make contact. (I'm reluctant to pierce insulation unless absolutely necessary.) In the spirit of "measure twice", please recheck B137 #2 and #3, and use a known-good ground reference (just in case there's a grounding problem).

 

If you still determine that those two ECU pins aren't getting power, then I'd agree that there's a break between them and B47 #4 (main relay).

 

If you can get to the yellow/blue wires on any of the solenoids powered that way (EGR, Idle Speed Control, etc.), see if they're getting ~12 volts with the key "ON" as well.

[Fairtax4me]

Wow! You're having a lot of fun! A little trick I learned (kinda shade tree but it works as long as you're careful) when you have wire that is suspected to be bad between two connectors and you don't want to tear into the harness, splice a new wire of equal gauge in place of the "bad" wire between the two connectors and see if the problem continues. You can run the wire wherever you want, and it's easy to put everything back the way it's supposed to be if it doesn't fix it. I've done this a few times, and I think it's best to cut the old wire for testing, then strip it and solder it back together, and heat shrink wrap it if the "new" wire doesn't cure the problem.