OB99W

A primary difference can be the amount of catalyst. Since a major component of the catalyst uses precious metal (platinum/rhodium/palladium), it simply costs more to have more. Some aftermarket manufacturers cut corners. The less catalyst there is, the faster the converter can become inactive (''poisoned'') by contaminants in the exhaust. Those can be from antifreeze getting into a combustion chamber, certain oil and gas additives, etc. Also, too much unburned fuel can cause destructive melting (however, your reported 24 MPG suggests that isn't likely happening with your car).

The EndWrench SRS link already given above is good information. The black 6-pin B82 diagnosis connector (for ABS/SRS/AT testing, depending on model year) and ''diagnosis terminals'' (ground leads) are often fully taped to other wiring, so check closely. For a better idea of what to look for, see: http://www.ultimatesubaru.org/forum/showpost.php?p=668884 Once you get the airbag light flashing, the codes are 2-digit, with ''longs'' of 1.2 seconds indicating tens and ''shorts'' of 0.3 seconds indicating ones, with 0.3 second spacing. If there's more than one code, they'll flash in sequence, with a 1.2 second space between. As the EndWrench article states, the ''normal'' code is an on/off sequence of 0.6 seconds for each state. Be sure to note if the ''normal'' code is outputed following the trouble code(s), as that has meaning. If you don't have the code list, post what you read and I can tell you what they refer to.

Is the front O2 sensor OEM? If not, it might be the cause of the problem. Pre-sensor exhaust air leaks can trigger the P0420 code. Any problem that causes the engine to output hydrocarbons sufficiently excessive to overwhelm the cats can cause the code. Aftermarket cats might not be as efficient as OEM ones, making it more likely that an out-of-spec mixture could tax them. If you have access to an OBD-II scan tool (not just a code reader), some real-time data could help pinpoint the problem. Fuel trim numbers, etc., might reveal something. If you can't do a scan, let us know how the engine seems to be running in general. Specifically, what gas mileage are you getting? See: http://www.motor.com/magazine/pdfs/102006_09.pdf Also, not Subaru-specific, but informative: http://www.wellsmfgcorp.com/pdf/counterp_v7_i3_2003.pdf http://www.wellsmfgcorp.com/pdf/counterp_v7_i4_2003.pdf

It seems that your question was already asked and answered here: http://www.ultimatesubaru.org/forum/showthread.php?p=826130

Although I can appreciate your concern, I'm not sure it merits getting so bummed. The question is whether the ''match within 1/4-inch of circumference'' requirement is totally Subaru being protective of your car, or partially something suggested by their lawyers. There's no doubt that a significant size mismatch can eventually damage the transfer clutch pack (auto trans) or viscous coupling (manual). However, probably very few of us always keep our tires evenly inflated, properly rotated, etc., so that it's likely the AWD is seeing some slight mismatches on a regular basis -- and yet the failure rate of the transfer system apparently isn't all that high. I'd also like to point out that for reasons other than the AWD, many of us try to keep a reasonably matched set of tires on our cars. I personally prefer tread design and depth to be close so that the car doesn't do anything ''unexpected''. Therefore, it wouldn't just be due to having AWD that I might replace more than a single tire if one were damaged.

Trouble codes of the form P1xxx are manufacturer-specific, which is why you're supposed to ''check service notebook'' (that is, look up the code definition for Subaru P1143). I already provided that definition, ''Pressure sensor circuit range/performance problem (low input)'' (which is a barometric problem), in post #9 of this thread. Since you said the other code you're getting is also baro-related, it would seem that's the area of the problem. The cure might be what Shawn suggested in post #10.

A bad valve in a single cylinder obviously wouldn't cause all the cylinders to not fire. Whether to repair or replace the engine will depend on what's really wrong. You might want to start with compression testing. If only #4 measures low, that would suggest the original claim was probably valid, and that something else (perhaps not too serious) has occured. If more than that one cylinder is low, maybe the timing belt has jumped, etc.

Unfortunately, a 1/4'' difference in circumference is equivalent to only a little over 1/32'' (actually, about 1.25 32nds) difference in radius (tread depth). Besides that, the tires would have to be otherwise identical in order for the tread depth measurement to be a valid means of comparison. Circumference may still be the better way to go, and is really the only way if the tires aren't an exact design match. The tires should, of course, be properly inflated when measured.

Beside the EndWrench link already given, here are more that might help: http://endwrench.com/images/pdfs/2.2Liter.pdf http://endwrench.com/images/pdfs/TBeltEWWin05.pdf Motor Magazine series; starts with 2.2, then moves on to 2.5: http://www.motor.com/magazine/pdfs/072001_08.pdf http://www.motor.com/magazine/pdfs/082001_08.pdf http://www.motor.com/magazine/pdfs/092001_08.pdf http://www.motor.com/magazine/pdfs/102001_08.pdf http://www.motor.com/magazine/pdfs/112001_08.pdf

I didn't mean to imply that insufficient oil delivery was the only possible cause of HLAs not pumping up. Since you secured the back plate screws previously, they may well still be tight. If the screws and plate are loose, there's often leakage at the front crank seal -- if you're not seeing that, it's more likely that the screws are still tight. Some general info on the 2.2, including HLAs: http://endwrench.com/images/pdfs/2.2Liter.pdf

Besides the pump being worn, there's the possibility of its back plate screws being loose. The following might be of interest: http://endwrench.com/images/pdfs/LoyalInfo.pdf http://endwrench.com/images/pdfs/Valve07MayEW.pdf

Based on the P1143 code, ''Pressure sensor circuit range/performance problem (low input)'', the problem might be related to the barometric pressure sensor or solenoid switching valve. If the codes were there with the 2.5, it could be that the baro sensor or solenoid valve is bad, etc. They're mounted on the right (passenger side) strut tower. If the codes weren't occuring with the 2.5, I'd concentrate on related vacuum hose connections. Look at the emissions label pasted to the hood underside. Trace the vacuum hose back from the strut-mounted parts, through the filter (make sure it's not blocked), to what's engine-mounted, and make sure a hose isn't off or misconnected. It might be easier to advise you if you would give all the Pxxxx codes (not the code definitions) that have come up.

You're welcome, seems like you'll have no problem passing.

Be sure to check that the back plate screws of the oil pump(s) are tight. Also, besides checking roughness/play, verify that the idlers can't be spun freely -- if they can, that would indicate a lack of grease in the bearings.

As well as the '92, is it correct that your '97 has a 2.2L engine? Besides replacing the timing belt and possibly idlers/tensioner, what else did you have in mind to do?

My comment is in the form of a question: What problem(s) are you experiencing that you'd hope a grounding kit would resolve?

Did you resolve the front parking light problem yet? That problem might be related. Are all the turn signal lights working, and properly bright? The flasher speed can vary with load, so not only will lamps being out cause a more rapid flash rate, but so will poor connections, wrong lamps, or varying battery/alternator supply voltage.

If the correct procedure is followed, less than 30 miles should be adequate for I/M readiness. See: http://endwrench.com/images/pdfs/OBDInfo.pdf Note that if you turn off the engine before testing, the monitors will reset to ''not ready''. However, that shouldn't be a problem at any testing facility that knows about this, since the '96 models are typically on an exclusion list for OBD-II monitor readiness. See: http://www.nj.gov/dep/bmvim/bmvim_obdvehicles.htm Also see Appendix D of this EPA document: http://www.epa.gov/OMS/regs/im/obd/r01015.pdf

Steering wheel shimmy (oscillation) can have a few different causes. Sure, at some point buy a dial indicator and you can check rotors and wheels for runout. But before you do, definitely rotate the tires (swap fronts with rears) and see if the problem is eliminated or at least changes in nature -- if it doesn't change at all, it isn't likely to be wheel balance, a shifted tire belt, or a rim being bent. If it changes significantly, then check for those things.

Since you're doing relatively frequent changes, I'd suggest that if you decide to have analysis done, do it twice. Choose a time when summer heat is the worst, and a second time when winter is coldest. Oil is stressed differently during those two seasons, and it's possible for a single test to be misleading for the untested seasonal conditions.

I think Gary has summed up the situation fairly well. Mostly, analysing oil is helpful in determining its useful life. That's much more important to fleet owners and those running very expensive rigs than to individuals -- it can positively impact their bottom line if they stretch the oil change interval without risking the hardware. There's no doubt that finding significant contaminants could be a tip to impending failure. Certain metals or fluids could point to the likelihood of specific parts having worn or are otherwise going bad. However, what would be done with that information also probably depends on ''cost'' of a breakdown. The engine in a vehicle where the consequences (financial or otherwise) of failure are high might be taken out of service and torn down a lot sooner than the one in a family's second car. I suspect you'd have to be lucky for an oil analysis be very helpful in deciding about purchase of a used car. If things were bad enough, they'd probably be obvious in other ways. If something was just beginning to fail, maybe you'd get a clue -- assuming the seller hadn't recently changed the oil. EDIT: Gary posted while I composed, so some of my points echo his.

In broad terms, analysis will tell you if the oil was still doing its job at the time the sample was taken, and if the engine is exhibiting symptoms of certain types of problems. Examples: Readings for metals can inform of wear of things like bearings, cylinder walls, etc. Viscosity of multigrade oil can divulge whether the index improvers are still intact -- along with fuel/antifreeze/water percentages, it can indicate whether the oil is detrimentally contaminated with those things. TBN (Total Base Number) can determine whether the oil's alkalinity is sufficient to neutralize acids that form when an engine is run. Etc. EDIT: Here's a comprehensive explanation: http://www.analaboratories.com/commonrootie/oiltests.asp

Looks good. The molybdenum is assumedly from the oil itself. Fuel dilution at under 0.5% and no water suggests that you don't do a lot of short trips. Unless you want to be conservative, it definitely looks like you could run the oil and filter a bit longer, as Blackstone suggested. If anyone else is interested in using Blackstone's services, I found a link for the test kit: http://www.blackstone-labs.com/free_test_kit.html . There are other labs that do similar testing.

When the ECU can't properly determine vehicle speed, it defaults to a ''limping'' mode that includes rev limiting to help ensure that nothing bad happens. The symptoms you're experiencing fit a VSS problem. The sensor itself might indeed be bad, but connector/wiring problems are sometimes the cause. If you can provide the ''Pxxxx'' code, it might be possible to give further advice.

Since I determined which fuses were applicable from a non-Subaru diagram, I rechecked using the FSM. The FSM wiring diagrams show the same fuses as I previously posted affect the parking (side marker and clearance) lights, but furthermore that the circuit powers the rear as well as the front ones. Therefore, if either fuse was blown, the rear lights would be out as well as the fronts, so the fuse issue is moot. If your car is wired differently than the FSM shows, any additional info I provide may not be valid for you. Assuming a 1995 Legacy that's wired as the FSM shows, pin #1 (red colored wire) of connector F40 is where the front parking lights (high side) connect to the rest of the circuit. Since that pin is at one end of the connector, it's possible that if the two halves aren't correctly mated, the front lights won't be powered. I've attached a diagram of the connector.