OB99W

EBSCO hosts several databases, the Auto Repair Reference Center being one of them. Apparently these databases are available by subscription, usually to businesses and libraries. Where I live, just having a library card provides access to the EBSCO ARRC, via the library's website and "navigateway". Instead of a username and password, you enter your library card number. Anyone interested could check with their public library, and see if they have similar online access. The ARRC has useful information. Some of it is generic (not specific to model or even brand), but if you have free access, it seems worthwhile to look.

Since the blower works at both low and medium-high speeds, the problem area is very limited. There may still be a problem with the switch, or just wiring as Tiny Clark suggested. I can offer electrical troubleshooting suggestions, provided that: 1) You verify that you have access to a voltmeter and know how to use it 2) You tell me what the colors of the wires are (including any color stripes, such as "red/yellow") at the switch connector, and their positions (terminal numbering); I'll need that to verify the wiring diagram I have is correct for your '98.

DOHC: http://www.soko.com/engines/ej25%20dohc.jpg SOHC: http://www.subaru-faq.ru/faq/gfx/inline/ej25-new.jpg

Sorry, I didn't realize that the last year Subaru used narrow-band sensors was model year 1999. Wideband sensors don't generate voltage, unlike the earlier design; they rapidly change in resistance (indeed responding to A/F ratio directly), and that resistance change is used to cause a varying voltage. Leaning and enriching the mixture to check the sensor's response as I suggested previously should still work, but the result of that would probably be most easily observed on 2000+ models via a scanner. I do agree that a borderline sensor might be harder to detect.

Yes, monitoring O2 sensor voltage can reveal quite a bit. Since a failing sensor can still have "correct" average voltage but be sluggish in operation, an oscilloscope is a better choice than a voltmeter to see what's happening. However, an O2 sensor that checks bad on a voltmeter is bad. When working correctly (once things are sufficiently warm), the voltage fluctuates around 0.45 volts or so. You can make the mixture leaner by pulling a vacuum hose, or richer by temporarily partially blocking the air intake. The O2 sensor voltage should change very rapidly after you do either, but in opposite directions. Rather than explain the theory and what to expect, I found a Web page that says it well: http://www.aa1car.com/library/o2sensor.htm

You should be able to get a vacuum gauge meant for engine diagnosis at a decent automotive supply shop for under $20. Here are a couple of Web pages that explain how to interpret the readings: http://www.classictruckshop.com/clubs/earlyburbs/projects/vac/uum.htm http://www.obd-codes.com/howto/diagnoseengine.php

You say "the other settings work properly". If that means the wipers work at both low and high speeds, and that they park correctly at the bottom of the windshield (no matter where in the cycle you shut them off), then the wiper motor is okay. If so, then it would seem that the problem could be with either the switch or the interval/intermittent wiper module, or a connection between them; the switch, of course, makes the connection necessary for the intermittent function, and the module "pulses" the motor to initiate the periodic sweeps. On my OBW, I can hear a relay "click" each time a sweep is initiated in the intermittent mode. Sorry, but I don't know where the module is located on the Forester.

That's not surprising. The parts are typically aluminized or galvanized. Welding vaporizes or otherwise removes that plating, which typically isn't particularly thick anyway. Not only that, but the temper of the metal is changed by the heat, and the weld material obviously doesn't have the flexibility/ductility of the pieces being joined. Some welding fluxes are themselves corrosive. If you could clean the weld area and coat it with an anti-corrosive from the inside, that might make enough difference to be significant. Unfortunately, the welded joints often can't be reached after assembly.

This kind of problem can be hard to diagnose. Looseness tends to exagerate the effect of imbalance, and vice versa, etc. Any front end looseness needs to be dealt with first. Besides imbalance, other tire-related possibilities are out-of-round, or runout (due to bent rim). However, since new tires and rotations didn't help, those seem to have been eliminated. I'm also assuming that the problem existed before the rotors were replaced, because imbalance in a rotor can sometimes cause problems that balancing the wheel and tire off the car can't resolve. Speaking of things being bent, is there any possibility of a damaged half-shaft? You could jack up the OB and try turning each wheel by hand, verifying that there's minimal runout (wobble). While rough roads can cause front end problems to make a vibration worse, it's also possible that a bad engine/trans mount could be confusing things. Have you checked them? It's sometimes useful to determine if the problem is road-speed specific or not. If the shake starts at 60MPH in top gear, does it also begin at the same speed if you're in one gear lower?

It probably would reduce exterior rust. However, it's unlikely to make much of a difference in exhaust system life, since most of the rusting occurs from the inside out, due to the corrosive nature of exhaust gases. If you do protect the exterior, don't let appearance provide a false sense of security. It's a good idea every so often to tap on the pipes, etc., to see if things still seem solid (or collapse ).

The change from 25 MPG to 20 MPG would indicate that the mixture is richer now than it was before. The replacement sensor could be miswired, or just not compatible. The ECU may be running "open loop"; is the CEL lit? As Nipper mentioned, the ECU may have to be reset. If the original O2 sensor is still available and undamaged you could try remounting it, or consider getting a new OEM one. By the way, were the 3 connections just plug-in, or did you have to do any splicing, etc. to wire the replacement? It's also possible that there's another cause of the gas mileage change. What exactly was done to repair the "blown" intake manifold gasket? Also, an intake leak will lean the mixture; sometimes the ECU won't compensate fully, and if the leak isn't too severe, the result can be improved MPG.

From your reports, the engine runs acceptably when cold, which is when the ECU is in "open loop" mode. You've said the rough running begins after warmup, which is when the ECU should already be in "closed loop". Disconnecting the O2 sensor normally forces the ECU into "open loop" operation. You could try unplugging the O2 sensor, and see if the engine performance when warm is better that way. If so, the O2 sensor might be bad, but running okay with the O2 sensor disconnected doesn't prove it's the culprit, because when the ECU runs in "open loop" it can mask other problems. By the way, for those of you with some test gear, it's not difficult to check O2 sensors. An oscilloscope makes it easy, but even a voltmeter with a reasonable input impedance is often sufficient.

Not quite. The procedure calls for "grounding terminal #5 of connector B82 a 6 pole black, right side of steering column.", according to a not-so-well-written part of an End Wrench article. I had to find and use a 6-pin black connector designated B82 when I diagnosed the ABS on my '99 OB. Terminal pins 4 and 5 don't exist in B82 on the '99 (they're just empty spaces, so you can't do AT diagnosis that way on a '99). B82 might be the same connector on the '97 as the '99, but may include something at terminal location #5. The B82 connector on the '99 was taped up to the wiring harness, near the green connectors. For info about the location and terminal pin numbering, see my post (#14) in this thread: http://www.ultimatesubaru.org/forum/showthread.php?t=46918&page=2 Note that I am not certain that B82 is the same connector on the '97 and '99, so if you are unable to verify that it is the correct location, or have any other doubts, please don't make the diagnostic connection.

The ECU runs "open loop" until the engine is warm. That is, it ignores things like the O2 sensor data (no voltage when cold) and uses a rich-running default. Since your engine runs okay until warm, it's a clue that the ECU is probably getting bad sensor data after warm-up, and misadjusting the mixture. OBD-II monitors how well the cat works; sensor 1 is before the cat, sensor 2 is after, and comparing the readings allows determination of cat efficiency. With all the trouble codes you've been able to retrieve, I'm surprised that none of them have caused the CEL to light. Does it come on as it should when the engine is first being started?

I assume that you're wiping the dipstick before reinserting to take a reading (if not, it's difficult to get an accurate reading). If so, use a clean, white, lint-free cloth, paper towel, or napkin. You can then get a much better idea of the color of the ATF. That's a different situation. The trans can certainly seem to be problematic when the engine is misbehaving that badly. Misfire can have several causes. Remember that besides spark, air and fuel have to be there in the right amount and at the right time. Can you tell if the misfire seems to be lean or rich? Something as simple as a bad batch of gas, especially if it partially plugs the fuel filter, could be the culprit. Or, possibly something getting sucked into the air intake. Even exhaust system problems can cause misfire. If ignition cables are bad enough, leakage (especially if they get wet) can certainly cause misfire. You can sometimes see the problem if you look under the hood with the OB in a darkened area, especially at night. Since the problem came on suddenly, I'd suggest thinking back to anything that happened very recently, like a gas tank refill immediately preceding it. If something obvious doesn't show up, I'm still wondering if this could be sensor-related. Best of luck.

Nipper, wouldn't an open circuit connection that prevented the transfer duty solenoid (Duty Solenoid C) from being powered have the same effect? I'm mentioning that because Aksnokat reported that the flashing sometimes wasn't happening, which could indicate an intermittent connection. For the curious, or those comfortable enough with their skills to get more deeply into the problem, there's info available including how to read trans trouble codes for certain model years via the ATF OIL TEMP light: http://endwrench.com/pdf/feb2004pdf/4EAT.pdf http://endwrench.com/images/pdfs/4EAatPh2Win04.pdf http://endwrench.com/images/pdfs/4EAT.pdf http://www.endwrench.com/pdf/drivetrain/02Winter4EATDiagServ.pdf http://www.endwrench.com/pdf/drivetrain/02Winter4EATTransOp.pdf Code "24" indicates a "C" solenoid-related problem.

A "vacuum leak" is an unintended opening at the intake side that "leaks" vacuum; in other words, it's a low-pressure area when the engine is running, and will suck in things that are at higher pressure, like atmosphere or a test fluid. So spraying the "fake smoke" into the intake therefore wouldn't be useful with the engine running, since it won't exit at the area of the leak. It possibly won't be evident with the engine off, either, since not much would be propelling it towards the opening. I'd also be concerned about how that gas mixture might affect a running engine. Why not try the other suggestions first?

Okay, I've done a bit more research for you . On the '99 OB, and hopefully on yours as well, the latch is about 4 inches to the passenger side of midline. I intentionally stuffed the strap of mine inside, and was able to find the latch by sliding my left hand from the passenger side towards that position; I did it that way since the plastic storage area for the seat belts makes it more difficult to get in from the other direction. It's in fairly deep, I'd say about 5 inches or so; unless you have a large hand, much of it will be out of view. The spring on the mechanism isn't very strong, so as long as it isn't jammed, pulling upward with one finger is enough to release the latch. I hope that helps.

The constant 2K tach reading accompanied by the engine misbehaving leads me to believe that you're experiencing an intermittent sensor problem, possibly due to a bad electrical connection. While the ATF condition you're describing doesn't sound great, it doesn't seem as bad as your initial description. It should probably be changed eventually, but may not relate to the current problem. I'm assuming that you're not driving the OB, so the "warm up" was just from idling. It takes about 20-30 minutes of idling to get the trans near operating temp if it's cold, so the fluid may not have expanded enough to get an accurate level reading. However, assuming that you're parked on level ground, a reading after warmup from even just idling that's "on the low side of the cold marks" would seem to be somewhat low. You might try adding no more than 1/2 pint of ATF (Dexron III type), and recheck the level after a 10 minute drive. Add more if needed, but don't overfill. The bad O2 sensor may or may not be contributing to the problem; it certainly can affect fuel/air mixture and emissions. In my opinion, what the tach is indicating is more likely to be related to other things, perhaps speed sensors. Speed sensor data is used by both the engine and transmission, and neither will run correctly if that data is untrustworthy. By the way, although the trans holds about 10 quarts in total, draining via the plug on the pan will only allow about 4 quarts to be changed. If you want to replace all of the ATF, it may be best to have a shop do a "flush". That will usually run about $100. Without knowing exactly what's wrong, I really don't want to advise you on the safety of moving the OB. Sorry. Perhaps someone else can offer some further advice.

To start with, as you might expect, "brown" ATF isn't a good sign. However, "frothy" has me more concerned. If we're talking air bubbles, the trans really isn't going to like that. Proper operation is dependent to a great degree on the inability of ATF to be compressed much, and air of course can be. So pressure isn't applied correctly when ATF is aerated. It also doesn't convey heat as well when airy. Etc. There's "enough fluid"; are you sure? Too much can be as bad as too little. If you're not exactly sure how to check the level, see this: http://www.endwrench.com/pdf/drivetrain/02Winter4EATDiagServ.pdf If the froth is more than just air, that's another concern. There may be a small leak in the transmission cooler, which is inside the radiator. That can lead to coolant getting into the ATF, with drastic results. If there is any milky tan or pink emulsion on the trans dipstick, or in the radiator reserve ("overflow") tank, it could indicate a trans cooler leak. If you can be a bit more descriptive of the "froth", etc., it might help with the diagnosis. A complete fluid change is probably in order, but there's not much sense in doing that if there are other issues that need to be seen to first. By the way, are you saying that with all that was going on, the tach never seemed to indicate revs over or under what would be "normal"?

I believe that indicates a general electrical problem. The first thing I'd suggest is looking for loose/corroded connections, especially grounds. If nothing is obvious, have someone check for trouble codes.

I agree with the possibility of hydraulic lock, and also if it happened that an oil change is a good idea. For future reference: http://endwrench.com/images/pdfs/ProperInsideEWFall05.pdf

NOMAD327 is mostly correct, but the strap is about 3" to the passenger side of center (at least on the '99). On my OB it aligns with a seam in the seat fabric, or roughly with the edge of the driveshaft tunnel. You can definitely retrieve it with your fingers. When my OB was delivered, even after reading the manual I couldn't figure out how to get the seatback down flat. The manual didn't make it clear that the lifting of the seat bottom was a standard feature, so at first I thought it was an option that I didn't have. However, after slipping my hand in the space between seat bottom and back cushion, I was able to find and snag the strap. Apparently the dealer had latched the seat with the strap inside during prep.

Hey, this is a "no genuflecting" zone! All you needed was the correct length 3/8" extension (2", or 3"?), the correct socket (mid-depth?), possibly a universal joint, a strong rod-shaped magnet or magnetic retrieval tool, and a 6" (approximately, use what length works for you) piece of rubber tubing that fits snugly (but not too snugly ) over the plug insulator. (I'm sure others will add to the list.) Oh, and lots of patience and some spare skin for grafts. ;)

Okay, but it's sometimes like this... http://www.netapp.com/news/report/feb-mar03/images/balancing-act.gif