OB99W

Being biased up like that, one might think there could be leakage from the sensor's heater, or that the ECU had a problem. However, it's been shown a replacement ECU (admittedly, a used one) didn't change the voltage, and that with a new sensor, things worked okay even though the voltage remained high. See http://www.ultimatesubaru.org/forum/showthread.php?t=66744 (and then maybe you can explain it to me ). Actually, the consistency of the reading, in conjunction with the code occasionally thrown, is why I've mentioned that the front sensor might need to be changed. Given the various conditions under which data was read, I'd have expected to see more swing. The O2 sensor isn't wired via the connectors that I've asked 2000 Legacy to check/clean, but some grounds are. Once that's been done, we can evaluate what else to consider.

You're welcome, although love is a bit stronger emotion than perhaps warranted. I'm not sure that you've correctly identified the proper connectors. The ones we want are not typically booted. One is grey, the other brown, and they should be about as far back as the throttle body, but down on the passenger side near the rear of the engine, on a bracket roughly behind cylinder #3. See the MY02 wiring folder, and open the "Bulkhead Wiring Harness (In Engine Room)" PDF. Page WI-168 lists the connectors, and WI-169 shows a (difficult to use ) location diagram. The "Engine Wiring Harness and Transmission Cord" PDF shows the mating connectors on page WI-178. If they look particularly bad, use water first. If not, rubbing alcohol alone should be adequate. Yes, it's near the ECU. It should be a 12-pin, inline, blue connector. See "Bulkhead Wiring Harness (In Compartment)", page WI-174. It's shown on the diagram right near B134, B135 & B136 (ECU connectors). Because it's an interior connector, it's less prone to corrosion, but worth looking at just to be sure. You can leave the ECU connectors alone (at least for now ). There's no need to apologize. You've had so many ideas thrown at you that I'm sure it became overwhelming. Those connectors are how all the engine sensors connect to the ECU, and they're vulnerable to road-salt-contaminated water exposure. That makes them worth checking, especially when there are diagnostic readings that don't clearly point to other causes. The misfire codes you're getting aren't specific to a cause, although the code for the front A/F (O2) sensor that popped up a couple of times still might mean it should be replaced. Okay. I wish I could assure you that cleaning those connectors will resolve everything, but at least the job is neither difficult to do nor expensive, and there's a history of them causing problems.http://set2music.files.wordpress.com/2009/10/fingers_crossed.jpg

While DOT 4 has a higher boiling point than DOT3 when uncontaminated, water more quickly reduces the boiling point of DOT 4. If you're good about changing brake fluid on a regular basis, DOT 4 is probably a better choice. However, if you only change it when there's a brake problem, DOT 3 could make more sense for you.

In that case, you were right to note that it was unexpected for the system to be "CL". The ECU's determination of whether to be "OL" or go "CL" is based primarily (although not exclusively) on the temperature it sees from the ECTS. After four hours of not running (and assuming an under-hood ambient temp of no higher than 96F, what the IAT indicated), it strains credibility to believe that the engine coolant was still at 149F (what the ECT reading was). Other readings have also been somewhat erratic, or at least not easily explained, and I know that you've replaced the ECTS. That leaves either an intermittent problem with the ECU itself, or connection problems as possible causes. In my post #38 of this thread I mentioned the problem with corrosion and contamination of the multi-pin engine-to-bulkhead harness connectors. The 20-pin E2/B21 mating connectors are where the MAP, TPS and ECT sensors are routed. Poor contact and leakage conduction due to salt contamination in that and E3/B22 have caused all kinds of weird problems that made diagnosis difficult. If you haven't already done so, I would suggest that you disconnect the E/B connectors and inspect/clean the contacts. Water is best for salt removal, alcohol works for drying afterward. I use 70% isopropyl alcohol (drugstore "rubbing alcohol") in a spray bottle. It works well because the 30% that's water can dissolve salts, while the 70% alcohol is a drying agent. Spray them, wait 1/2 a minute, spray again, repeat several times. If you see anything greenish coming off, keep spraying. Once the connectors are clean, they need to be thoroughly dried. A hair dryer or other source of warm (but temperature-limited) air works well. If that's not available, let things air dry in the sun. You also might want to check the B83 ground/shield junction connector. If you don't mind reading another longish (only 7-page ) thread concerning a 2000 Legacy with strange symptoms that eventually had a simple resolution, see http://www.ultimatesubaru.org/forum/showthread.php?t=100611 . I think it's worth reading to the end, but I might have a slight bias . Although the problem presented itself differently than yours, it may be that the cause is similar. (Perhaps more than coincidentally, note the last paragraph of post #38 in that thread as well.) The OP, rvac99, was in Narragansett, probably no more than 50 miles from Colchester.

I had written a detailed reply to that latest post (#112), when I remembered something that I previously meant to follow up on. Your post #107 began: About four hours later I reread your reply and wasn't sure that what I did earlier was correct so I went out and plugged in the 3130 and started the car and took the following readings in 5 minute increments (1:13 PM, 1:18 PM & 1:23 PM): Fuel Sys 1: CL (I was surprised, I thought this would be OL when I first started the car) Calc Load (%): 7.4, 4.8 (9.8 at 2000 RPM's) ECT (F): 149, 186 (194 at 2000 RPM's) This could provide an important clue: Were the first readings (at 1:13) taken immediately upon starting the car? (Or had the engine already been running for 5 minutes at 1:13?)

Have you tried removing all the spark plugs (just in case one of the cylinders is hydrolocked)? EDIT: Apparently 987687 posted while I was writing mine. At least we both had the same idea.

Impreza 2001 airbag info: http://www.subaru-car.ru/imp1/imp01_body_3.pdf http://www.subaru-car.ru/imp1/imp01_body_4.pdf http://www.subaru-car.ru/imp1/imp01_wir_1.pdf

You've mentioned both the cable and your computer, but not the software. What program are you running? Cable info, includes links to software: http://allobd.com/proddetail.asp?prod=TAC-OpenPort1.3U&cat=45 EvoScan: http://www.evoscan.com

True, but hydrometer readings can tell you whether there's a chance that the battery can be salvaged by proper charging, or if one or more cells are likely beyond salvaging. Using both a load tester and hydrometer provides more insight than either alone.

By the way, the real test for CCA is done at 0 deg F, and measures how much current the battery can deliver for 30 seconds while still being able to output (for a 12 volt battery) a minimum of 7.2 volts. That corresponds to 1.2 volts per cell for the typical six-cell battery. To draw 100 amps from a "12 volt" source would imply a resistance of 0.12 ohms. With a fixed resistance, as soon as the battery voltage drops, the curent draw does as well. At 9.6 volts, a 0.12 ohm load is only going to draw 80 amps. So, of course, the tester gives an estimate of CCA, which in itself may or may not be very accurate.

You might be able to determine the actual age from a date code that's usually stamped into the case. Lead-acid batteries get sulfated as they age, especially if they've ever been fully discharged or left in a state of partial charge for any significant length of time. That can impact their capacity quite a bit. Certain charging techniques can reverse the sulfation to an extent, and recover some of the capacity. However, if there's a substantial difference in output from one cell to the next, it's unlikely that the battery can be sufficiently rehabilitated so that it will be reliable on some frigid mid-winter morning in Buffalo. Assuming that it's not a sealed type, and you can get your hands on a battery hydrometer, compare the cells and see what the balance is. If it's not to bad, let us know and I'll post links to charging info that might improve things... ...or just buy a new one (or take your chances ). I nursed a battery through a couple of rough winters, but it "got" me one day when I really needed it. I had enough, charged it one last time, and got a replacement.

Thanks for retesting and posting the results. You didn't quite do what I asked (I wanted some of the data at both 1500 and 2500 rpm, not at the median value of the two [2000 rpm] ), but for now we've got enough. Most of the readings fall within acceptable ranges. Things that don't: 1) Fuel trims should normally be no more than about 5% (positive or negative) if all's right -- the higher (over 10%) negative readings you've posted indicate that the ECU sees the mixture as rich, and is cutting back on fuel (narrowing the injector pulse width). 2) MAP readings at idle are high. You can roughly calculate intake vacuum by subtracting the MAP reading from atmospheric (barometric) pressure. Average atm pressure at your elevation is about 29.5 inches Hg. MAP of 12.9" corresponds to roughly 16.6" intake vacuum, while the MAP of 17.4" suggests about only 11.1" vacuum. (The MAP of 8.2 at higher rpm would equate to intake vacuum of 21.3, a much "nicer" number.) I just looked at the MAP readings from your post #103, where the freeze frame shows an even higher MAP reading, 18.8" at just above idle (772 rpm). 3) The TPS (Throttle Position Sensor) readings are all "0" (zero), which is acceptable at idle with closed throttle, but not at 2000 rpm, obviously. The above two observations may be directly related to it. It's surprising that no code was set. 4) Varying engine timing. For a steady throttle position, this is typically stable. The knock sensor might be picking up something that's causing this. (EDIT: Perhaps more likely, the MAP is indicating to the ECU that the engine load is heavier than it really is, and besides overly enriching the mixture, the ECU is also retarding ignition timing.) So, where does all this lead? MAP and TPS are very closely related as far as the ECU is concerned. If you look at the diagnostic procedure for P0106 and P0121, that becomes obvious. In fact, for P0121 (which is for a TPS range problem), it indicates that if the MAP reading is over 15.75, the MAP sensor should be replaced. I would suggest, rather than immediately going there, you start with a quick voltage check of the TPS. Here's how: 1) Remove the air box from the throttle body so that you can back-probe the TPS connector. 2) Connect the voltmeter negative lead to a good ground. 3) Turn the ignition key on, but don't start the engine. 4) Measure the black lead voltage (should be very near zero, since it's a ground). 5) Measure the red lead (should be close to 5 volts). 6) Measure the white lead while operating the throttle (should vary as it's moved, make note of the reading at closed, one-half, and wide open throttle). Please get back to us with the TPS voltages, and that will determine the next step. Getting a couple more TPS and MAP live data readings with the engine at idle could also help. By the way, a MAP live data reading with the engine not running should indicate atmospheric pressure -- if not, tell us what you get. I have specific other things in mind, but I don't want to make this post any longer than it already is.

Things look good. As a point of interest, could you tell us what the mileage on the engine is, and whether the cat or front O2 sensor have been replaced (and if so, when)?

Since the data you quoted was freeze frame at the time the P0304 was set, I don't think we know when or under exactly what conditions that occured. What's interesting about the ECT readings is that in the freeze frame data, at 132 deg F the system was running closed loop, while in the live data, at 170 deg F it was open loop. That makes one wonder what the ECU was "thinking" (unless the "170" is an error, and was either "107" or "110"). Hopefully the closed loop live data and the additional data points I asked for will help clarify what's going on.

I don't have the time right now to do a detailed analysis of the data, but I do see some things that make me go "hmmmm". Rather than comment, I'd like for you to get the live data once more, but after the engine is warm enough for the system to be in "closed loop". "Fuel Sys 1" will indicate "CL" (not "OL", which is "open loop"). Once it's "CL", note all the live data at idle, and also the following: Observe the "O2SB1 S (V)", and note the readings over time. Let us know the lowest and highest, and if there's a tendency to stay near a certain range of voltage. Get a reading of the MAP at idle speed and also around 1500 and 2500 rpm (in neutral). Same for the fuel trims (STFT and LTFT). Hold at each speed for several seconds until the LTFT reading is somewhat stable. Find a helpful passenger, take a short road trip, and get them to compare "Veh Speed (MPH)" with what the speedo says at various road speeds. Let's see where that leads us. EDIT: In the morning, when you first start the engine, compare the ECT (Engine Coolant Temperature) and IAT (Intake Air Temperature) live data readings. They should be almost the same with the engine cold.

Yes, that sensor (the first two words of which are either in the order "Air Intake" or "Intake Air", depending on who's describing it, by the way ) is mounted to the air cleaner case/housing. Having it disconnected with the engine running would set the P0113 -- however, if previously you had it off but didn't run the engine, that might be why you didn't set the code.

Yes, "PD" indicates a pending code. Since the ones that were "PD" are for cylinder misfire, of course that's no surprise. P0113 definition is "Intake air temperature sensor circuit high input". That would indicate a poor or missing connection to the air temp sensor. If you had it disconnected while cleaning the IACV, that code would be expected. Otherwise, it could be a clue -- the air temp sensor was something I had previously asked about, since it can certainly affect idle adversely if it's not working correctly. It's one of the things the ECU looks at in setting idle mixture, and the IACV duty cycle is a primary control for that. Okay on the PCV system -- it seems that you've checked it sufficiently. And ... you're welcome.

The throttle body cleaner might have loosened things up a bit, but in my experience it has the most effect almost immediately after use. Since the PCV is designed to tolerate petroleum and the like, if you have mineral spirits or something similar around, that should be fine as a solvent. A light coating of oil isn't a problem, just make sure there aren't sludge deposits. A scan tool will hopefully make the diagnosis less hit-or-miss.

Code 38 indicates a problem with the "Torque Control Signal". This can be due to an open or short circuit, and may be affecting lines between the TCU and ECU, or a bad transmission ground connection to the body. It's possible for the problem to be internal to the trans, but I'd start by looking for corrosion or disconnection at external grounds. Did the problem show up without any obvious cause, or has the car seen recent service or been driven off-road?

Highway mileage should be higher, probably about 26 MPG or perhaps a bit more. (However, it's hard to get an accurate reading with just a few gallons usage.) If the mileage is actually down, it might indicate that the rich condition is not just a low or idle speed issue. Poor gas mileage can be the result of a failing front A/F (O2) sensor. If the stalling problem was worse after the engine was warm (and therefore running closed loop, using the sensor data) rather than when it's cold, then I'd definitely say the sensor should be changed. At least it seems that the FPR diaphragm isn't ruptured. Of course, that doesn't ensure that it's regulating pressure correctly. With the vacuum hose off, the pressure will go to maximum. Did you notice any difference in the way the engine was running with that hose on or off? I was trying to get an idea of whether the PCV was working. At idle there's less blow-by than at higher speeds, so the PCV should produce a minimal crankcase vacuum. The results are inconclusive. A previous post indicates that the PCV valve has been cleaned, but was the tubing also checked/cleaned? Perhaps a little theory will help with understanding things. When the throttle plate is closed, although it doesn't seal completely, there's certainly insufficient air for the engine to run. It's the function of the IACV to allow a regulated (by the ECU) amount of air to bypass the closed throttle. Assuming that the valve is functioning correctly, it's still possible for passages in the throttle body to be somewhat obstructed. Especially if there really was an improvement after using the cleaner (and not just wishful thinking ), it's possible that's what's going on. Even though you haven't reported the code, you might want to look at the diagnostic procedure for P0506. If you decide to remove the IACV and/or work on the throttle body, be sure to have new gaskets on hand. If you aren't inclined to do more cleaning, etc., at this point, then it might be time to invest in a scan tool. The live data from that might point more directly to the cause of the problem.

I don't see a reason to replace the fuel filter, and wasn't suggesting it. If anything, the engine might be seeing an overabundance of fuel. 18-21 MPG is reasonable for local driving, but if there's any significant amount of highway travel, it's somewhat low. The cause could be anything that can affect A/F ratio, including a lazy front O2 sensor. But let's keep our focus narrow for now. Cleaning the IACV and looking into (and smelling ) the FPR vacuum line are relatively easy/inexpensive things to do. As much as I like good tools, since money is a consideration I'd suggest holding off on the scan tool purchase a bit longer, until we exhaust things that can be done without one. In looking at the comparison chart at http://astore.amazon.com/equus.innova.gauges.reviews.best.price-20/detail/B000KID31K , I don't see an advantage in the models above the 3130 for your current vehicle. If you have something else in mind, that might change things.

Download the PDFs here: http://hp.auto.ru/Subaru%20Legacy/1998/Service%20Manual/DIAGNOSTICS%20SECTION/AUTOMATIC%20TRANSMISSION%20AND%20DIFFERENTIAL/

Yes, that's the flip side of the coin -- either there's not enough air, or there's too much fuel. We touched on fuel pressure briefly when someone suggested a failing fuel pump as a possible problem, but I don't think it was ever checked, especially at a rail. Also a good question -- in particular since a failed FPR could lead to excessive consumption. So, that leads me to a really wild theory. 2000 Legacy has said "I do notice a burnt smell that I can't pinpoint.", which might be from an overheated cat due to too-rich mixture. He also said "I had replaced the fuel filter that I got from NAPA at about 176,000 miles (I think, it's documented early on). I'm not sure but it almost seems thats when the problems started and have often wondered about changing the fuel filter again." What if the FPR was failing a while ago, but a partially clogged fuel filter had been limiting the fuel delivery anyway? Then, replacement of the filter with one that flowed correctly could cause the FPR's failure to become obvious. Or not!

In order to hopefully give you confidence we're on the right track, I "played" a bit with my own car this morning. Not that I expected anything much different, but my '99 has about 20.5" of vacuum at idle (with some needle "flutter"), and disconnecting the brake booster hose caused an almost instantaneous stall. Yes, we apparently can't be sure what the dealer did concerning the IACV, so cleaning the valve is a first step. It might be there are problems with it that are beyond the help of cleaning, but as you mentioned, it's "easy enough and cheap enough". Let's see where that goes before considering anything else. As to the O2 sensor, we usually suggest replacement of front ones with only OEM Subaru. But let's not go there yet. The Innova I have is the model 3130. It's hardly my only electronic tool -- there are digital and analog multimeters, scopes, exhaust gas analyzers, etc. -- but being able to see OBD-II live data means that the others get used less frequently. By the way, would you humor me and see if the idle is affected by removing the oil filler cap? (Considering how air-starved the engine seems to be, I'm wondering how it might affect the PCV.) With the cap off, can you feel a slight vacuum, or pressure, at the oil filler tube?

True -- I didn't mean to suggest otherwise, just wanted folks to know that 90/10 split wasn't all there is, and that using the FWD fuse can have detrimental effects on some models.