OB99W

It depends what you mean by ''nonexistent'' leakage -- I've never seen an engine that didn't show at least a few percent loss. However, if the leak-down percentage is very low and the same on both sides, that indicates good valve and ring sealing. Leak-down is a ''static'' test, while compression is ''dynamic'' (being dependent on when and to what degree the valves open, besides how well they close). As I and others have said, timing error is what usually causes compression differences when the engine is otherwise healthy. Other than that, significantly misadjusted valves (not opening sufficiently) could cause lowered compression readings, but that would typically be obvious in other respects.

Even if the two cams seem to be in time with each other, that doesn't ensure that the crank is as well. As to the cylinder #1 misfire, see the last section on page 30 of http://endwrench.com/current/spring04pdfs/InsiderInfo.pdf for a possible cause.

Yes, the main fuse box is the one in the engine compartment.

Check fuse #9 in the main fuse box, and try interchanging the left and right headlight relays (also in the main box).

Does the high beam indicator on the dash light? (Its 12 volt connection is from the same circuit that powers the right-side headlight.)

It's great that you accomplished the work, but for future reference, see the attached.

One thing that can affect voltage readings in some circuits is voltmeter ''sensitivity'', given in units of ohms-per-volt (ohms/volt) relative to the full scale reading for a particular range. Inexpensive analog multimeters often have low sensitivity, about 1,000 ohms/volt. With such a meter, low ranges in particular can significantly affect readings if the circuit has resistance of the same order of magnitude. Usually the meter's sensitivity (which can be different for DC and AC) is printed in a corner of the meter scale. Better analog DC meters have sensitivity of about 20,000 ohms/volt (20kohms/v), adequate for most automotive circuit testing. Electronic types (VTVMs, FET-VOMs, DMMs) often have input impedance of 10,000,000 ohms (10Mohms), independent of the range they're set to. From your description, I suspect the meter you're using has relatively low sensitivity. If that's not the case, then the circuit itself may have a high-resistance connection.

The serpentine belt on the H6 shouldn't have to be ''levered and stretched''. It seems that you might not have released the tensioner.

My first thought is that patience is a virtue. Based on what you posted, it seems that there is a heater circuit problem with an O2 (A/F) sensor. That could mean either the sensor will need to be replaced, or there might be a wiring/connection problem. The actual trouble code (Pxxxx, not the code definition) might be helpful.

Diffs with sealed VLSD units (and Torsen diffs, if anyone is wondering) don't require friction modifying additives.

There's a relatively easy way to verify or rule out such a trans problem. The next time a rapid stop causes the engine RPMs to drop, immediately put the trans in neutral. That should remove any loading that the trans might be causing, and the engine speed should recover almost instantly. If it does, then you have likely confirmation of a trans-related problem. However, if the engine remains with a low idle or stalls, the trans isn't the cause. By the way, can we assume that there are no dash warning lights illuminating (Check Engine, AT Temp)?

With the engine cold, does the idle change (speed or smoothness) if you depress the brake pedal hard and hold it down?

Yes, the TCM monitors and can output trouble codes for solenoids, certain data from the ECM, engine and vehicle speed sensors, the ATF temp sensor, and the TPS (plus a few other things, depending on model). Reading any trans code that was triggered could help with the diagnosis.

HP (''High Performance'') is required starting with '06 models. Initially there was confusion over whether only the 5EAT (but not the 4EAT) needed it. However, there were some internal changes in later 4EATs, and they also should get the ATF-HP. See these threads: http://www.ultimatesubaru.org/forum/showthread.php?t=98645 http://www.ultimatesubaru.org/forum/showthread.php?p=804132 In a word -- maybe. VLSD/Clutch-type/Torsen? Clutch type requires the additive in the oil.

I assume you're referring to ATF-HP (High Price ). You may find that trustworthy alternatives aren't easy to come by.

The switch is above the pedal, to the left of the steering column, where the pedal links to the master cylinder rod.

Is this the information you're working with: http://ww2.justanswer.com/uploads/ebrock63/2009-03-21_162915_testing.pdf ? That info seems to be flawed. I believe that step 8 should say ''If voltage is not 2.9-4.0 volts, go to next step. If voltage is 2.9-4.0 volts, go to step 12.'' EDIT: I checked further. Actually, the error in step 8 apparently is that it says ''Warm-up ATF to operating temperature''. The 2.9-4.0 volt range (what I have says ''3.45 +/-0.55 volts'') is for the ATF at 20 deg C, 68 deg F (not operating temp). The lower voltage range in step 9 is for the ATF at operating temp. You'll notice that step 9 says that if the voltage is 1.0-1.4 volts that the ''circuit has returned to normal condition''. I suspect your measurement of 1.6 volts is probably close enough, given that the sensor isn't mounted where it can reach the actual temp of the ATF.

No, '97 would be a phase-1 (DOHC), '00 is phase-2 (SOHC).

MY = model year Improvements after '96 to obtain more power, which result in the valves being able to hit if the engine gets out of proper timing.

Get the code read and see if it's P0440 -- that EVAP code is common when there's a hole in the filler neck (among other things).

Let me clarify something. Some of the EVAP system codes point to specific electrical faults in solenoid circuits (opens/shorts, etc.). However, P1443 just indicates that the ECU couldn't detect proper venting. A stuck solenoid can cause that, and so can a vent/drain tube obstruction. Since that tube is open to air, small ''critters'' can build webs, nests, etc. in it and block it off. I'd suggest looking into that first. The link I gave previously explains how to narrow the possibilities. There's a diagram in http://www.endwrench.com/pdf/fuel/FtEvapEmissionTestW00.pdf that shows interconnection of EVAP components, although there are model-to-model variations. Subaru has some sophisticated equipment to test EVAP, as in http://endwrench.com/images/pdfs/EvapSpring04.pdf, but the systems can certainly be fixed with a lot less.

Welcome to the forum. You might find http://endwrench.com/images/pdfs/CoolingInfo.pdf informative.

The P1443 code definition is ''Evaporative emission control system vent control function problem''. As nipper already suggested, ''The tank vent solenoid to the charcaol canister is dead or stuck.'' (or it's a problem directly related to that).

i-AVLS description from the horse's mouth: http://www.drive.subaru.com/Spring07_whatmakes.htmAt the bottom of the article is a link to AVCS info, too.

Yes, the pistons should be halfway between TDC and BDC. The crank is at 90 degrees off TDC, and the cams (which of course revolve at 1/2 the speed of the crank) are therefore 45 degrees off TDC. Valves not colliding with pistons is certainly important if dealing with one of the interference engines. The procedure that Subaru uses for timing belt installation on the SOHCs ensures that cam lobes are positioned so that the cams aren't prone to rotate due to valve spring tension, helpful even on non-interference models.