OB99W

P0304 is "Cylinder 4 Misfire Detected". If that's the only code, it's hard to tell if gas quality is an issue; since there apparently aren't codes for the other cylinders, it might be less likely. Does the problem seem related to the tank level (is the problem any more or less evident at full or near-empty)? Does it show up any more frequently going up/down hills or on curves? How many miles are on the car? When was the most recent tuneup done (replacement of spark plugs, possibly ignition wires, etc.)?

The spin-on filter may not "require" routine replacement, but if you don't change it, at least inspect it periodically. On my own '99 OB, the darn thing rusted through at relatively low miles; fortunately, I caught it before the leak became significant. It annoys me that Subaru markets cars for use in snow, but doesn't design certain parts to withstand salt.

The restriction prevents rapid changes in pressure from impacting the transducer in the sender, thereby minimizing the gauge needle bouncing around; the pressure readings are smoothed/averaged/damped in the short term.

An engine that starts could still run poorly due to just a one-tooth misalignment; more than that, and with an interference engine, you can wind up with bent valves, etc. The two links below have some very good information. The first has general info about Subaru t-belts, while the second gives very detailed instructions on getting the marks lined up correctly on a DOHC 2.5L; I suggest both be read fully (ignore the SOHC info in the first link if working on a DOHC) and followed carefully. http://endwrench.com/images/pdfs/TBeltEWWin05.pdf http://endwrench.com/images/pdfs/2.5Timing.pdf

If the misfire is consistently happening "every few revolutions", it probably doesn't have an electrical cause. A rhythmic miss like that reminds me of "hot" cam behavior; valve overlap causes a pulsation in the intake manifold that makes sensor readings erratic and can cause the ECU to get really "confused". Steve455 previously reported that the belt had jumped, and was replaced. How about a simple explanation, like the timing is off on one cam? That could have an effect similar to hot-cam valve overlap and explain things if the miss is truly rhythmic. Okay, I've got my Nomex/asbestos suit on, so flame away.

The search function doesn't like things that are only three characters long. Try "iac*" (without the quotes).

Ohhh, Canada... sorry, eh!

You might find some useful info in one of my previous posts: http://www.ultimatesubaru.org/forum/showpost.php?p=586934

You probably don't need to buy an aftermarket part. First, see this thread to better understand the situation: http://www.ultimatesubaru.org/forum/showthread.php?t=61890&highlight=31099AA091 Then, buy from an online dealer, and pay about $20 including shipping. Here's one: http://www.subaruparts.com/cart/?pn=31099AA091 . That's just an example, and I'm not endorsing any particular dealer; there are several others whose pricing will be about the same.

Congratulations on finding the problem, and condolences on what has to be done to resolve it. A loose crank pulley can be a real problem and cause significant damage if not caught soon enough. It doesn't tend to occur unless the pulley bolt wasn't sufficiently tight, and that typically happens when the torque spec isn't met during timing belt work. Before things get too bad, there's sometimes a clue: the engine-driven accessory belts (alternator, PS, A/C) will loosen, often causing strange noises or charging/steering/cooling problems. Since the ECU needs timing information from the crankshaft angle sensor, if the sprocket is loose the pulses don't relate to actual crank rotational position and the timing of both ignition and fuel injection will be off. At best, the ECU uses the data it gets or defaults to a fail-safe/limp mode. That explains the symptoms you experienced. I suppose I should add "timing light" to my suggestions of "old time" test gear that's still useful when OBD codes don't do the trick.

The electrical connection is important, but so is the condition of the reluctors, sensor (especially the tip), and the gap between the tip and the reluctors. If the gap is excessive, for whatever reason, the generated signal will be weakened.

You might need to verify that the reluctors (those ridges/teeth at the back of the sprocket) are okay as well. A small deformation or piece missing could result in a weakened pulse that might not trigger the ECU consistently.

Let's try a little math, and see what it tells us... Assume the engine is idling at 600 rpm. Since there are 60 seconds per minute, 600 rpm translates to 10 revs/sec. Each cylinder has one power stroke per two crankshaft revolutions, so at 600 rpm that would require 5 (10/2) injector pulses per second (at each cylinder). If the misfires are about four seconds apart, at 600 rpm there would be around 20 injector pulses per cylinder between misfires. Since there are injector codes for two cylinders, there could be around 40 pulses on average for the two coding cylinders between misfires. That suggests the problem is marginal, and that most of the time the ECU is properly pulsing the injectors. So, what could cause the ECU to mispulse an injector or two on occasion? The pulses are triggered by info from the cam angle and crank angle sensors, and a faulty signal might be to blame. Those sensors are dependent on the reluctors that are on the back of the sprockets (that is, their condition and position relative to the sensor). Perhaps something unkind happened to one of them: http://www.ultimatesubaru.org/forum/showpost.php?p=594070 ...Just a thought.

Another thought -- besides exhaust restriction (which was already checked for), intake problems can of course severely limit the engine's ability to "breathe". Blockage or even holes can be a problem. See http://endwrench.com/images/pdfs/EWMAFAug05.pdf , for example. A vacuum gauge (yes, I'm beating that horse again ) might tell you a lot about what's going on. See the links in my previous post in another thread: http://www.ultimatesubaru.org/forum/showpost.php?p=596419 . If you do vacuum testing, let us know the results at idle and with the throttle opened further. One other possibility is an electrical problem. Corroded or loose connections (grounds and otherwise) can sometimes cause strange problems with the engine management system.

The ECU could be defaulting to a "fail safe" mode (although it would usually light the CEL and there'd be codes); some of those modes cut off one or more injectors. If you haven't already seen http://endwrench.com/images/pdfs/Fuel.pdf , I suggest you read it through. A few questions: 1) Did the parts you swapped in come from the same model/year? 1) Does the CEL light with ignition on, before starting the engine? 2) Does the engine stumble if you try to accelerate, or just refuse to rev? 3) Is there any difference between cold operation and when the engine is warm?

If anyone's interested in Subaru pics (including action, non-stock paint jobs, etc.), Google Images is a place to look. Searching there for "subaru rally" will turn up quite a bit.

Have the engine/trans mounts checked. If you were in a rust-belt state (or the car came from one), I'd suggest having the whole underbody looked at for corrosion-related problems.

Really? Can anyone else confirm that there were MY99s built wihout ABS?

Mickey Mouse? Maybe, or... http://www.sfondideldesktop.com/Images-Animals/Frog/Frog-Common-Gray-Tree-Frog/Frog-Common-Gray-Tree-Frog.Jpg

The proper procedure when pushing caliper pistons back is to open the bleed screw at the caliper. Otherwise, you risk forcing contaminated brake fluid back up the lines into the ABS unit and elsewhere. Once contaminated fluid is in the system, all kinds of problems can develop. If you're lucky, a complete flush with fresh fluid may resolve the problem. See: http://endwrench.com/images/pdfs/BrakeFluidSumm02.pdf http://endwrench.com/images/pdfs/BrakeSystemWin01.pdf

It's apparently been replaced, but code for it is reportedly still there (see post #48 of this thread). Obviously, there could be circuit problems beyond the sensor itself, which we've already mentioned. I suggested clearing codes and seeing if it returns, to verify whether the problem is current.

Your mechanic is mostly correct, for the reason you proposed; the tire lugs will bite into the snow better. However, I said "mostly" because there are other aspects of handling and load-bearing that limit just how narrow it's prudent to go.

Most codes will clear by themselves, but some take several completed drive cycles for that to happen. I like to clear codes and see which come back, but it's your choice. The reason I provide links is that it's easier and shorter (often, much so) than typing out the procedures. The vacuum readings you got would seem to indicate that the basic engine health is at least adequate. (The 18 inch reading at 2000-2500 rpm could have been a bit closer to the 21 inch reading at idle, so there might be a bit of exhaust restriction.) To the point... ...If the ECU thinks the engine is always cold, it will always call for too rich a mixture. Until you get the coolant temperature sensor code to not return, I'd suggest skipping other work. Once you get past that, if there are other issues/CEL/codes, you can attack them.

There's no need to apologize; I agree with you -- this is a great forum, and it would be even better if people would let us know how their problem was resolved.

Have you seen this: http://endwrench.com/images/pdfs/Emissions.pdf ?