Rich running/missing forester (codes): CPS?

[ldeikis]

Hello all.

 

Backstory: Wife's 98 forester had lit the CEL a couple weeks ago with a P0170 (fuel trim malfunction)--I dumped in a can of sea foam, cleared it, and it hasn't come back in a few hundred miles. I picked up fuel/air/oil filters because the car was due, but hadn't yet installed them.

 

Today the car started on only 2 cylinders, running horridly, and smelling VERY rich. No CEL yet. Put in the new fuel and air filters because why not, started it up, same symptoms, and it lit the CEL with a P0341--cam position sensor. Popped the CPS out and tested it with a good digital meter and it's reading infinity across the terminals... I found it hard to believe the car could even RUN that way, so I dug out my old analog meter, which is reading .2 ohms (almost total continuity). BUT, I can't get the old analog meter to zero properly, so it might be FUBAR. The digi is a Fluke that I use for work regularly and literally trust with my life.

 

Questions: Does the CPS care if it's tested with a digital (low voltage) meter or an analog? Do I need to buy OEM or brand-name on this (prices seem to range from ~$20 to over $100)? What else should I be checking?

 

Car has a *very* slight exhaust leak, mostly muffled by an infuriating loose heat shield... can't imagine it's nearly enough to contribute. We've only had it since late last winter, a little less than a year... PO is "pretty sure" the timing belt got done 20k ago, but I planned on doing it all this winter to be safe = no real knowledge on when belt/tensioner/etc were done.

 

Help.

 

Thanks.

 

Luke

 

EDITED: I had the P0170 typed as 0172.

Edited December 8, 2011 by ldeikis

[john40iowa]

I am not versed enough to contribute to your main dilemma, but the rattle with heat shields I attest for. All old Foresters have them. Simple fix, buy large C clamps and tie em up. My car has three and no rattles.

[naru]

Somethings wonky.

Cam sensor spec is 1-4k ohms.

Try your meters on the crank sensor.It is also 1-4k.

 

The sensors are simple coils so they don`t care what type of meter you use.

[ldeikis]

Went and picked up a new analog meter to double check. Both crank and cam position sensor are coming in at 2k Ohms, which I gather from this forum is correct. Can't explain why it's throwing a code unless the circuitry is wonky somewhere. Connections seem clean, tried a dab of dialectric on the terminals with no change.

 

I don't understand why it would spit a cam sensor code, nor why that would act like this and I have to wonder if it's not a red herring one off. I tend to think the P0170 is more helpful even though it's not throwing that one anymore--the nose confirms it's running *really* rich, like raw-gas rich.

 

I don't want to run it this gassy longer than necessary but I think I'm going to button it up and idle it until it spits a code again. Does 0170 imply a front oxygen sensor on these cars?

 

Luke

Edited December 8, 2011 by ldeikis

[porcupine73]

Can you see if it is giving pulses on your analog meter to show it is actually working properly? It might read in spec impedance wise but not be outputting a nice waveform.

 

It seems someone had mentioned it can still start with the cam position sensor faulty, that it can interpret enough from the crank position sensor, but it has to play around to figure out if the cylinder is on the intake stroke or the power stroke.

 

On the flip side, the soobs, esp of that era, will run very rich when it is cold out, at least until they warm up. Subaru had said something like they did that to help make sure it would start even with gasoline that had a low vapor pressure at low temps (probably i.e. not winter blend gasoline).

 

Your waveforms might look different, esp for the cam position sensor, as they played around with the divots on the back of the cam sprocket between some years/models.

[ldeikis]

Unfortunately the meter is just a rat shack multimeter and doesn't have a scope, so I don't have access to a way to see the sine.

 

EDIT:

 

Reinstalled sensors and took it for a ride to try to throw a code again. Car can barely get out of it's own way--I needed a rolling start to pull up the gentle bump into my garage. Shut car off, turned it back on to hook up scanner and it finally threw a CPS code again. Freeze data says:

 

Code P0341

RPM: 888

calc load 5.9%

MAP 17.4 Hg

Coolant temp 169*F

ST FTRM1 0.0%

LT FTRM1 0.0%

veh speed 0

fuel sys 1 open

fuel sys 2 n/a

 

I guess I ought to call the stealer and go pick up a CPS huh? Still doesn't make any sense to me, but oh well.

 

 

All help welcome.

 

Luke

Edited December 5, 2011 by ldeikis

[porcupine73]

I think Subaru mentioned using an analog meter set to a fairly low voltage AC range to check the output of the crank and cam position sensors while cranking the engine. You can't read the waveform per se but you can see if you are getting the 'heartbeat'.

 

It's not super uncommon for soobs of that era for the cam position sensor to go faulty. Walking into a dealership of course to get the part means you will pay top dollar for it at most dealers.

 

I think P0341 was what my '96 thew when it died at a stop light one time. Then a little while later it would not even start and kept dumping that code, and it was fine after I replaced the sensor (was just gamling that it was the sensor that was bad).

[naru]

So the Fluke flunked?

Hope you don`t really trust your life to it.

 

I would check for AC voltage generated by the sensor while cranking.

Should be 0.1 volts or more.IIRC,I`ve used a digital meter for this measurement w/o problems.

 

Maybe it is intermittent.I would abuse it a little w/the ohmmeter hooked up.

[ldeikis]

Had the wife turn the car over while I monitored the CPS with the meter. It definitely is pulsing, but the voltage it's putting out is way too low for me to read it on the meter... but that's what I'd expect from a hall sensor like that.

 

My intuition still tells me this isn't the problem...

 

Is it possible that a fueling issue is causing the engine to idle so low--nearing a stall, or about to stall--that the time between pulses on the CPS is long enough to trigger the ECU thinking its got no signal? I only ask b/c the freeze data shows the code was thrown at 888rpm, which is way below the car's normal idle I believe... and because I'm having trouble believing this is my issue. Also, it's taking a while for it to throw the code after it's been wiped off, whereas I'd think an actual faulty CPS would toss immediately.

 

FYI, the fuel pressure regulator is all happily hooked up with no signs of leaking; engine doesn't respond to tweaking any of the vacuum lines, and none are disconnected or loose that I can see.

 

Local dealer doesn't have it in stock anyway, so I'll have to order it online. I'm going to keep thinking on it, but I have to commit by this evening as I'm going out of town for 3 days working and need this part waiting when I get home.

 

Anything else to check???

 

 

EDIT--Yea, funny on the Fluke, I wouldn't think a sensor like that would need an analog meter... and I do trust my life to that thing somewhat regularly, though not usually reading ohms. The wife already took off for work tonight, so I can't try testing the CPS while turning over with the digi--I wouldn't think it would respond in a readable way, though... new analog metr which is correctly zeroing trembled appropriately, I think.

Edited December 5, 2011 by ldeikis

[porcupine73]

My intuition still tells me this isn't the problem...

 

Anything else to check???

 

You could look for the Subaru service manual for that model&year and see what troubleshooting procedure it has for the code you are getting. The Subaru procedures are pretty thorough; has you check to make sure connectors and wiring is good, goes through pretty much all possible causes of the problem. Usually the very last step if everything else checks out is to replace a part.

[ldeikis]

You could look for the Subaru service manual for that model&year and see what troubleshooting procedure it has for the code you are getting.

 

Yea, I've been looking. I tried to download a copy from an eBay seller and it won't uncompress correctly. I'd be happy to pay for a paper copy used or new--I'm a big fan of having the right info--but I can't even find one.

 

I'm leaning towards just ordering the part, since that's what the car says it wants...

 

Luke

[jarl]

By all means, don't feed the eBay manual sellers. All the information they have to offer -save very few of them- is all available through the web.

 

The Subaru manuals are available through the "My Subaru" page:

http://www.subaru.com/my-subaru/index.html

 

You have access to 50 docs an hour for three days for $34 or so. Or just ask the right questions and you may get the right answers

[naru]

Free 97 manual here

http://www.finleyweb.net/JonsStuff/SubaruDocumentation.aspx

 

FSM doesn`t have much to say anyhow.

Merely check continuity of the wires to the ECU and thier isolation from ground.

If good,replace sensor.

[naru]

888 rpm is well above cranking speed,so no problem there.

I would inspect the protusions on the cam sprocket if possible.

 

I see the sensor wires are enclosed in a grounded sheath.

If this was comprimised,stray emf may induce enough of a voltage in the sensor wires to overwhelm the miniscule signal voltage.

 

I would do a quick check for alternator AC output.

(high elec. load,2500 rpm-measure at the alt.-should be under 0.5V)

 

Check the fuse in your Fluke.

When I blow the fuse in my meter it will still read voltage but not ohms or amps.

[jarl]

FSM doesn`t have much to say anyhow

 

I'm glad to say you are somewhat off... Personally, I'm of the opinion the FSM is worth every penny of it's price. The "diagnostics" section of the FSM has plenty of information, it's just not included with most of the copies of the FSM you find around.

 

For people checking their CPS, the FSM for my '99 OBW mentions between 1 and 4 kOhms between terminals with the sensor off the car. This value may change if you measure the sensor without removing it, though...

[naru]

Actually,I find the FSM dignostics wanting.

For me,in most cases,a quick look at a wiring diagram suggests a more logical diagnostic path.

 

They are useful,of course,but,the diagnostics suck sometimes.

[ldeikis]

So I got my hands on the FSM and ran through the relevant tests. Not a lot to the CPS--either the sensor's bad or the wiring's bad. I *still* have trouble believing that's the issue.

 

While I was at it, I went through the diagnostic for the 0172. Diagnostic says with the car at operating temp, the MAF should be reading 2.2-4.2 g/s at idle, and 8.6-14.5 g/sec at 2500rpm. My scanner only reads lb/min, but the internet converts those to .29-.56 lb/min and 1.14-1.92 lbs/min respectively (calculator here). My little scanner shows ~.95 lb/min at 1000rpm and ~2.00 lb/min at 2500 rpm, which is about twice what the FSM says... which means the MAF thinks the car is breathing a lot more air, so the ECU would dump a lot more fuel, creating a super rich environment, no?

 

I was off to the store to get some spray can MAF cleaner, but my VW decided now was the perfect time to toss it's alt belt, and it's a nightmare to access so that pretty much tied me up.

 

Can anyone weigh in on whether that g/sec > lb/min conversion is fair? Are the ranges my scanner picked up far enough out to blame? Can I trust this Actron scanner to accurately read the MAF via the OBDII port? Do I need OEM for the MAF--the only junkyard anywhere near here is closed for the winter (or anyone got a good one they wanna sell?). Is it reasonable that the MAF could be that far out of wack and the ECU not know it?

 

Also: In the diagnostic, I had to skip the bit about testing the fuel rail pressure b/c I don't have a gauge for that. Is that a pretty standard universal thing I can pick up at NAPA or autozone? I'm happy to own it if it's not super expensive and has other applications; if it's a zillion dollars and suby specific maybe I don't need one.

 

Also: While trouble shooting, I started the car with the MAF disconnected to see if it would limp along. Ran like crap, then stalled. It also tripped the AT fluid temp light, which is now flashing incessantly. WTF?

 

Thanks. I won't be able to work on this thing til Friday night or Saturday at best; wish I could have sprayed down the MAF and tried that, but I'll have to do that before I spend $ on a new one.

 

Luke in NY

 

EDIT--for clarity, the car hasn't thrown 0172 in a couple hundred miles, but is registering the CPS code pretty much immediately.

Edited December 6, 2011 by ldeikis

[naru]

2 lb/min= 15 g/sec so that is only slightly over spec of 14.5

 

0.95 lb/min= 7.2 g/sec so that is a fair bit higher than the 4.2 g/sec spec.

1000 rpm seems too high for idle speed so not totally unexpected

Still seems on the high side though.

 

Good idea to clean the MAF.

I think the MAF needs to be waay out before a code is thrown.

 

Fuel pressure is usually the first thing to check on an injected engine.

Sometimes you can borrow or rent a gauge locally.

You can spend a zillion dollars on one if you are a sucker,but,there is no need to.

 

Here is a inexpensive one.

http://www.ebay.com/itm/Fuel-Injection-Pump-Pressure-Gauge-Tester-Gasoline-Test-Tools-Tuner-Cars-Trucks-/260908497390?pt=Motors_Automotive_Tools&vxp=mtr&hash=item3cbf5bb9ee

 

I bought the same one years ago.Works fine.Mine didn`t come w/the fancy case.

[ldeikis]

I'm working on getting my hands on a fuel pressure tester.

 

 

 

In the meantime, I ordered the CPS as well as a front oxygen sensor--kind of a lot of money, but I need to have this car back on the road, and if it doesn't need it now, it likely will some day.

 

CPS made no difference to the car at all. Big surprise.

 

I also sprayed off the MAF sensor with a can of MAF cleaner fluid, which made no difference.

 

Went to install oxygen sensor and Liberty Subaru sent me the wrong part--the one they sent has a square connector and the car has a triangular one. I think they sent the rear sensor instead of the front? Waiting on a call back from their guy to make that right.

 

While I was warming it up a little to ease removing the old sensor, I was watching the data on the scanner and noticed the car starts out in open loop mode, then switches to closed as it warms up.... but it's running sloppy rich and missing right from the get-go. Does this mean the 02 sensor definitely isn't responsible, or does the car base its open-loop settings off previous 02 sensor input? I cleared the codes on the ECU prior to observing this, but I don't know if that wipes off everything or not.

 

Thanks. I'm off after a pressure tester. This is getting old.

 

Luke

 

EDIT: Got a fuel pressure gauge on it. Reads 30psi at idle with everything hooked up (FSM spec is 23-30psi), 39psi at idle if I disconnect the regulator vacuum hose (FSM spec is 33-40).

 

Happy for opinions on where to poke next, as well as whether open-loop mode runs off factory defaults or off learned trim settings (even with cleared codes). I think next I'll try disconnecting injectors to see if I can isolate a particular cylinder missing, and then start swapping around injectors. Wondering if maybe the seafoam could have loosed some crap resulting in a blocked and dripping injector or something.

Edited December 12, 2011 by ldeikis

[ldeikis]

Help. I'm running out of ideas over here.

No surprise, new CPS changed nothing.

New OEM front 02 sensor changed nothing.

Fuel pressure tests ok.

Checked harness to CPS--which is still the only code reliably being thrown--and it checks out ok.

New fuel and air filters.

Cleaned MAF--which looked spotless--with no change.

Code reader says ECU is accurately understanding coolant temp.

No detectable exhaust or intake leaks, no loose hoses.

Car continues to reek of raw gas and miss horridly.

 

It won't run well enough to even drive to a shop with my tail between my legs--I'd have to get it towed, and we live in the sticks. Really don't wanna do that. Intuition is still saying it's related to the seafoam a tank ago, maybe loosened some crap up that's clogged or blocking something... maybe causing an injector to not close all the way or I don't know what. Am hesitant to feed it more cleaner though until it's happy... a rough running engine is a lot easier to diagnose than a non-running one.

 

All suggestions welcome.

 

Thanks.

 

Luke

Edited December 21, 2011 by ldeikis

[1 Lucky Texan]

Can't you get a crank or cam angle sensor code if the timing belt slips?

[jarl]

Is it possible to test the flow from the injectors on these cars? Maybe an exhaust leak?

[ldeikis]

Can't you get a crank or cam angle sensor code if the timing belt slips?

 

If I can't isolate the problem to an injector or wiring/plug combo, I think a jumped timing belt will be my last hope before giving up. I'd planned to do the belt later this winter because the PO couldn't be 100% sure it'd been done. Is there a way to visually confirm the cams and crank are still lined up right without removing all the stuff you have to yank to actually change the thing?

 

Luke

[smiffy6four]

If you want to change the belt anyway, why not do that now and eliminate the possibility of jumped timing? Has the car back-fired at all? That can make the belt jump.

[davebugs]

Take the ends off the timing belt cover and look at the cam sprocket TDC marked against the top/back of the timing cover on both sides. If they align (both TDC at the same time) odds are the belt is good (could still have nubbies missing though).

 

But that's an easy first step that just takes a 10mm deepwell for me.