nvu

With the spark plugs out, crank the engine until you see oil pressure. It'll crank pretty fast and build pressure after a couple tries. Also don't prefill the oil filter, if the oil pump wasn't preprimed it'll take very long for the air to overcome the oil stuck in the filer. If you're not sure, just crank it with no oil filter until you see it spit out oil, then the pump is primed. Put back the filter and crank for pressure. I'd crank until the oil light goes out, let it rest while I button up the some things, come back and test crank again to confirm the oil light instantly goes out. Then you're sure everything ready to fire up.

Engine size doesn't matter as much as SOHC vs DOHC. SOHC belt is same on 2.0, 2.2, 2.5 SOHC engines. DOHC belt matches all DOHC EJ. You'll still need to make sure the tensioner matches the one you have: old style vs new style. And water pump matches: side exit or bottom exit.

If anything, inspect the donut flange area. if it's rusted out I'd just replace the entire thing. A sometimes leaking and sometimes not spring loaded donut flange makes it hard to chase the inevitable p0420 code. The muffler end can fall off and it wouldn't set codes. No resonator wouldn't bother the neighbors, it'd bother you with highway droning if you're sensitive to that.

Great job finding the issue. It just occurred to me that if it happened to one window, it might be the same on the other ones.

Before you order, go to your car and pinch all the hoses connecting to the pcv valve. Chances are they're already hard and brittle, probably crack or won't seal up if you take it apart. You might be better off ordering the pcv assembly with new hoses attached already.

it does sound like bearings. try spinning the wheels while holding on the the spring coil. it amplifies the vibrations. if there is a bad one, you'll feel nominal vibrations on the other wheels and really different vibrations on the one with a bad bearing.

Probably overthinking it, I've taken them apart and put them back randomly with no issues.

I don't know about newer axles, but for 2007 and earlier were all made by NTN. Look for the NTN stamped on the cup. If the axle doesn't have a name stamping or serial number, it's an el cheapo.

Actually I realized the pics I posted above have the 770 on top, and 670 below. The top center area is different, but for your use case either is fine. They are both MLS designs.

Ah, I copied wrong, it should be 11044AA670 gonna edit the original. Thanks

Did you have the heads resurfaced at a machine shop for this rebuild? If so, go with the MLS ones, 11044AA770 or 11044AA670 Look up how to press back the tensioner with a c-clamp for reuse. Going too fast can pop the internal seals. This video looks like the correct way.

There's only one high pressure oil passage on the top of the gasket. If it's flipped backwards that would explain the leak. It's the topmost oval one. It should match to the top port

You mentioned you're doing it in car right? At this point, might as well remove the radiator to get the best possible outcome for tapping a hole... especially if it might break through into block.

I don't know if there's a definitive procedure for creaking bolts. I could tell the if a creaking bolt has gone beyond it's yield, it just feels springy when you tighten. A good bolt can creak too, it's sometimes due to the threads in the block not matching the threads on the bolt. Good bolts that creak still feel as stiff as ones the didnt creak. I run them in and out a couple times lightly with a drill and clean/relube the threads. then try again. Be aware MLS gaskets deform as part of sealing, and then bite a bit into the aluminum at final torque. You shouldn't loosen the bolts all the way once you started, if you need to work on a problem bolt, back off the other bolts as little as you need to and always leave them in tension. At the torque wrench stage, try and keep everything creak free. Keep cleaning and relubing if needed. After that is the 1st 90 degree stage, some likely creak, you need to remember the feel of the good bolts and decide. It's the last chance to back off and redo. The second and final 90 deg stage, if it's a bad bolt you will definitely feel it. Never go more than 180deg total for those two stages, if it felt loose after the two stages... it's overstretched already.

It's not recommended to reuse headbolts, but I've reused headbolts before. You'll know you have a bad one when you do the final torque and it squeaks too much or just doesn't feel tight. I have a bunch of used ones laying around so it's not a big deal. It might be a show stopper if you don't have spares on hand.

Any silicate free coolant is fine. Subaru coolant conditioner was a rebranded stop leak to keep the composite gasket from flaking off. Subaru MLS shouldn't need any conditioners.

i find a table, bench, any flat corner with a L and taller than the entire axle. hold axle vertically into the corner with outer cv facing up; slam it down hard so the outer cv catches in the corner L. it's like a reverse slide hammer, mind your toes.

Angle looks bad. Get some more washers on the upper bolt to space it back in line. you might need a longer upper bolt. I did something like this when I still had the narrow wagon swaybar with the wider sedan control arms.

Does it start going christmas lights only after you start the engine? If so it's likely a broken alternator regulator sending ac voltage in the system. Try unhooking the red charge wire on top of the alternator and starting the car.

I'm installing TGV sensors again and forgot the twisting procedure.... Was it always install it with the connectors facing up and then twisting it into place for both sides?

You just need to jumper the clutch switch that goes off on full depress. The clutch switch on slight press is the one that disengages cruise control, doesn't have anything to do with the starter. FYI

did you go from nonturbo to turbo? the p0420 is pretty tricky to catch even on the original car. even throwing on a different exhaust system could trigger it. don't try and chase the problem itself, but get other things dialed in and worry about p0420 last. the car will run fine with p0420 code and no other codes, at least anything before 2007. it just won't pass smog. with that, some things to check if you haven't done so already: your mass air flow sensor is probably okay. make sure the car runs strong in closed loop, no misfires or rough idling. part throttle should be smooth, put it in 1st and see if it can creep along smoothly in idle. creep along smoothly 1000-1500rpm. try 2nd gear if it's bucking or missing, then check your maf sensor, intake, vacuum leaks... on the exhaust side of things, make sure the entire path from engine to the rear cat has no leaks. the flex joint and gaskets just before and after the cat are usual suspects. test with whatever way you think will help, i don't know of any surefire way for me, i had one car that would pop a p0420 every couple months, even did pressure smoke testing and never found the anything. only by chance one year i decided to sand a rusty looking flange only to find that it was slightly bent, the other mating flange was bent in the same direction so everything mated up and sealed fine, but was C shaped ever so slightly, you can't even tell by eye. after getting everything flat again, the problem went away for good. still on original cat from 2002

I've never seen a solid sprocket like that before. All the subaru ones of that era look like the one in the first post you made. Clearance should be about 1-2 business card between sprocket and sensor. You timing looks good enough though.

Is the EJ251 like the old EJ22 sohc with only a single cam sensor? If so, I've had an EJ22 skip many teeth on the non-sensor side and not throw any code. Engine ran really rough on idle. The flashing cruise is an immediate misfire signal and doesn't set any code. I'd check timing again with just two covers removed.

next time you catch it happening, open the reservoir and shine a light in, have someone steer the wheel with engine on. if you see bubbles it's a hose leak.