fgf

bad ground to injectors? Or loose connector?

Be careful about the heater wattage. Oils convect away the heat much more slowly than water or coolant, so a heater intended to heat coolant can overheat the oil and damage it. No more than 4.5 Watts per square inch for oils versus up to 40 watts per square in for water according to www.hydraulicspneumatics.com/hydraulics-at-work/article/21886513/how-to-heat-hydraulic-oil-without-cracking-it

Not always the charcoal canister or plugged vent hose; in the case of my 03 Outback the problem was a failed vent valve. Best to test rather than guess...

per GD: Spiders or mud wasps nest in the charcoal canistervent line/fitting into the frame rail. Or an exploded charcoal canister. Or a failedvent valve. GD

Do you hear the pump running when you turn key to run position? Listen close to the fuel pump cover. If not, check for voltage at the connector - blown fuse?

When replacing a Legacy BJ I slather the bolt with anti-seize and also paint some into the female threads with a small brush. I also pack enough chassis grease into the BJ pocket so that it squishes out of the slot when I insert and tighten the BJ, then wipe off the excess. My goal is to fill with grease all the cavities that might hold salty water. With that approach I've found that years later the BJ is still easy to remove and find no evidence of rust. Our Rust Belt is an exceptionally nasty environment. My son brought me a 96 legacy from Wyoming - many dents underneath from the back country roads but not a spec of rust and every bolt came out clean. Wyoming certainly has Winters, but they don't use road salt...

May not be relevant, but on an '03 2.5L I did have the rubber gasket at the bottom of the radiator cap swell to the point it sealed the outlet to the overflow tank. This resulted in excessive pressure in the radiator which caused radiator failure (radiator cap crimp failed). Swelling was the result of oil contamination of the coolant (from head gasket problem) - visual evidence was a small amount of black "goo" floating in the overflow tank. Rubber gasket in radiator cap should not extend beyond its metal backing plate...

The road salt used during winter is pretty much off the road (and into the surrounding land) after the last snowstorm has melted away. HOWEVER, calcium chloride or magnesium chloride are widely used on unpaved roads to reduce dust in Summer and both are corrosive to metal. Application is usually done in Spring once dust becomes a problem, and sometimes repeated later in Summer if the problem returns. Most often applied as a 30% to 40% brine, but sometimes as a powder (flakes). Both are hygroscopic (meaning they absorb moisture from the air) and deliquescent (meaning they can "melt" into liquid form once they have absorbed enough moisture). This helps keep the road surface damp and thus helps bind the dust ("fines") into the surface. Mostly used in areas with good (high) warm-weather humidity; ineffective in arid ares. The concern is that some of that treated dust can get thrown up into your wheel wells where it can form a sticky and corrosive coating. Driving over a just treated road is probably a bad idea... I could not find any good info about the severity of the problem after the brine has soaked into the road surface. Periodic thorough hosing downs of your undercarriage during Summer may be wise if you drive in an area with treated dirt roads. Our addiction to roads that are snow-free in Winter and dust-free in Summer is expensive. Consider that in the Northeast a new car that could last 20+ years is pretty much rotted away within 10 years due to corrosion from salt - we are paying a "salt tax" of several thousand $ per year...

I don't know if the rear wheel well sheetmetal design in the Imprezas resembles that in the Legacy or Outback models, but there is a well-known issue with an improperly sealed joint in the 95... Legacy/Outback rear strut tower which results in rust breaking through in the lower rear corner of the doors. Salt spray goes up into the strut tower, leaks through the unsealed joint into the inner fender area (which it rots) and drips down towards the rocker panels. See: http://www.ultimatesubaru.org/forum/topic/119859-the-fantastical-rear-wheel-well-rot-thread/ especially the second page. Also: http://www.ultimatesubaru.org/forum/topic/135273-00-04-outback-rust-in-rear-door-openings/ Hard to inspect - one of those Harbor Freight "Digital Inspection Cameras" with its flexible shaft can help. The problem appears to be an overlapped sheetmetal joint which was never sealed...

Check wires to relay coil. I had exact same symptoms about 10 days ago and they were caused by cracked wire leading to the connector. The copper strands can fatigue crack inside the wire insulation.

Some test info: '96 Legacy, EJ22 engine, key off & removed mA meter between -battery post and post clamp. 65 mA steady current after 1 minute (to let ECU settle down) = 1.5 Ah/day = 47 Ah/month Group 35 batteries used in the '96 typically list ~ 40 Ah capacity, so could discharge battery in a month... Leaving my UltraGauge (OBDII viewer) plugged in doubled the draw - apparently it keeps checking to see if the key has been turned on yet - so only 2 weeks to discharge battery. Got bitten when I didn't start the car for several weeks followed by a sub-zero cold snap (discharged batteries freeze easily)...

Check http://rockforddriveline.com/catalog specifically: http://rockforddriveline.com/catalogs/11_Universal%20Joints_2012July.pdf see page 28 Determining factors are the bearing ODs and the spacing between the insides of the yoke. The inside surfaces of the U-joint yoke needs to be flat, parallel and centered because this is where the retaining clips fit. Some (e.g., 1995) yokes were unmachined, but I've had success carefully machining the surfaces. Key is to get the right spacing and to have it precisely centered on the yoke's centerline...

Drive shaft U-joints? In neutral with rear wheels off ground try rotating drive shaft by hand to check for any binding in U-joints or diff or rear axles?

Check the stabilizer (aka swaybar) bushings too - they can cause klunks. Also your ball joints...

When you have the trany off, check the back of the engine for signs of leaks. The usual culprits are the rear RH camshaft support and, as mentioned above, the oil separator cover. The camshaft support comes off with 2 screws and contains a thin O-ring. The O-ring in my 96 was hard, brittle, and leaking oil - an easy fix (get the O-ring from Subaru). The separator plate in some models was plastic and tended to crack. Replacement from Subaru is metal. Check the snout of the trany - that is what the throw-out bearing slides on and it can get worn. If badly worn you use a clutch kit that includes a sleeve that slides over the snout and a throw-out bearing with over-sized ID to match. Other items to check - the clips that hold the throw-out bearing to the clutch release lever and the pivot that the lever pivots on - both cheap parts worth installing if originals are worn or corroded.

re keeping track of parts: I buy various sized zip-lock sandwich and freezer bags (at supermarket) - in which I store each set of parts and label as I go with a Sharpie felt-tip. I ended up with at least 100 bags during my 96's overhaul - but that involved a lot more than just doing the clutch. Also take many pictures as you go - you might never need them, but if you do they certainly help! The hardest part may be separating the trany from the engine. The 2 steel dowel pins tend to corrode in the aluminum castings and can be very hard to break free. I ended up grinding some 1/8" steel straps to make a half-dozen very gradual thin wedges that I could tap into the split. Patiently working my way around (evenly) eventually got my 96 2.2L to separate. Before reassembling, clean the dowel pins thoroughly and slather hole and pin with anti-sieze...

Had the radiator on a 2003 Outback separate (and also blow the upper radiator hose) due to a damaged radiator cap. Failing head gasket had contaminated the coolant with oil - which caused the rubber pad in the radiator cap to swell to a noticeably larger diameter than the metal piston it covered. Apparently the cap got put on with the rubber pad swollen enough to seal the outlet to the overflow tank and the result was enough pressure to pry apart the radiator and explode the upper hose. Bottom line - make sure the radiator cap's rubber pad hasn't gotten swollen, especially if you've had head gasket issues... It should be the same size as the piston it covers.

I sent the following posting back in Nov 2012, but cannot find it in the USMB archives - so I'm posting it once again. The key point is that in 2012 there were counterfeit or defective IACV being sold at attractive prices; they may still be around. IACV (Idle Air Control Valve) on 2003 Outback, H4 (2.5l) engine I recently troubleshot some IACV issues in my 2003 2.5l Outback. I ran into some apparently defective replacements. Here are some notes about what I learned. Symptoms included stalling at idle, hesitation when shifting, and a P0519 code, all suggestive of a "sticky" IACV. Cleaning the air passages only helped a bit. Replacement IACV are horrendously expensive - $300 from most after-market vendors - so it was worth trying a bit harder to fix the existing unit. This model IACV consists of a small stepper motor whose rotor drives a lead-screw in or out. A plastic cone at the end of the lead-screw controls the amount of idle air allowed to bypass the closed throttle plate. The stepper contains 2 coils, which I'll call 'A' and 'B', each of which is center tapped resulting in 6 connections: A1, A2, Atap, B1, B2 and Btap. The two center taps, 'Atap' and 'Btap', are connected to +12 and the ECU grounds one of the sides in each coil. Grounding A1 makes the current in coil A run one way while grounding A2 instead makes the current go the other way, thus reversing the direction of its magnetic field. So there are 4 grounding combinations: A1+B1 A1+B2 A2+B1 A2+B2 sequencing through these left to right makes the motor turn one way, reversing the sequence turns the opposite way. To move the valve the ECU steps through this sequence at 200 steps/second. When you turn the key on to the Run position the motor takes 8 steps outwards (extending the lead-screw to reduce the idle air). When you turn the key off the motor first takes 176 steps outward at reduced (150/sec) speed then takes 156 steps inward at full (200/sec) speed. A stepper motor driven valve doesn't "know" its own position - it just stays where it was last left - so there is a need to place it into a known position. I think the shutdown logic is as follows: 1) first keep stepping outwards long enough that the valve closes completely and the motor stalls as it tries to step beyond. This ensures that the valve is completely closed in a known "home" position. 2) then step inwards to the desired idle position (mostly open). (the ECU 'learns' the valve's behavior so it might not always step inwards exactly 156 steps) If you cycle the key on and off with the IACV removed from the throttle body, the lead-screw never hits the stop (fully closed) so it will extend 20 steps further out after each cycle. Eventually it will run out of threads and the lead-screw, plastic cone and spring will 'sproing" out - they can fly several feet, as I now know from experience... This provides a way of removing the lead-screw for cleaning and re-lubricating. I strongly recommend enclosing the IACV in a plastic bag before trying this so you don't waste time searching for the flying parts... It also provides a way of reinstalling the lead-screw - while a helper turns the key back to the off position gently hold the lead-screw in place until it gets drawn back in during the second half of the shutdown cycle. Cleaning and lubing the lead-screw seems to have solved my problem. However I don't see any way to relube the rotor bearings themselves if they get sticky. If (when) that happens I'll try flushing the motor with solvent to remove the old lube, then flood it with a very light weight spindle oil; might even work... Before reaching this solution I tried buying a surplus stock IACV being offered for $128. The part looked absolutely identical to the original - same blue sticker, same part number, etc. It seemed to solve the problem when installed, but within a week the ECU was throwing codes again. Examination showed a critical difference. In the original IACV the shaft is not free to turn - it is keyed. The replacement is missing the key so the shaft can be turned, and unscrewed all the way out. The vendor sent several replacements, each had the same problem. I have a feeling that he had received a batch of defective units, or (more worrisome and the reason for this long note) that someone is manufacturing counterfeit units which may now be widespread. They can appear to work. If the motor takes only a few steps then the inertia of the lead-screw / cone may keep it from turning. However, if a long sequence of steps is made, or if the lubricant is cold and stiff, the lead-screw / cone will start to turn with the motor and thus not move in or out as desired. The attached picture shows: far left: part number sticker, identical on old and new middle: defective part, red arrows indicating the keyway and key which does not match and allows rotation and a crude red sketch of the missing part on the right two views of the original piece with a proper key - which appears to be a stainless steel shell around injected plastic. Bottom line - if you buy an IACV. check it before you leave the store. If you can turn the cone and unscrew the lead-screw you've been given a defective or counterfeit part...

Similar issue on my 96 legacy. Some mechanic had mostly stripped the hole in the bracket that the pulley bolts into (but left it as is) and had also stripped the left side tensioner bolt hole - and solved that one by drilling out the tensioner and running in a short 3/8 self-tapping bolt. I had to mill out the oval slot in the new tensioner to fit the 3/8 bolt, tap the bracket for a full-depth 3/8-16 thread, and helicoil the pulley bolt hole (lucky to have 10 x 1.25 helicoils on hand). All went together OK - with a bright orange note nearby about the 3/8 bolt ;-).

Camshaft seals can leak substantial amounts of oil into the timing belt area - my 98 2.5L Outback blew one. If SOHC there are also O-rings sealing the "camshaft supports" to the head - maybe 1/16" thick X 2" OD - which can make slow leaks.

Don't know about '05, but my '96 2.2L SOHC needed a couple new camshaft O-rings - one at the back that I did when I had the trany split from the engine and the other this week at the front while doing the timing belt. Little thin O-rings, maybe 1/16" thick and around 2" OD. Probably were original and had the gotten baked to the consistency of uncooked spaghetti - hard and brittle. Slow leaks in both cases, not gushers, but enough to cause drips and, in the case of the front one, swell and soften some dust gaskets. In that engine there are "camshaft supports" at both ends of each camshaft which seal to the head with that O-ring. The front supports also have seals where the camshaft passes through on its way to the sprocket; while I had the support out I replaced the seal even though there wasn't evidence of it leaking. Also had a camshaft seal blow in my 98 2.5L - that was a major leak...

Starting system wiring for 95 Subaru Legacy - which should be the same as the 94. See

I got most of mine from: http://www.clipsandfasteners.com/ Other sources: http://autobodyclips.com http://www.indexfasteners.com http://automotive-hardware.com

I've taken to using cupro-nickel brake line for such replacements - easier to bend and flare and it doesn't rust. More expensive than plain steel tubing, but I've found it worthwhile. Ribet Auto Parts <http://www.RibetAutoParts.com> carries it - FMSCN3-25 $24.95 for 25 feet of 3/16 - as do others.

The weak spot in my 95-03 Legacy and Outbacks seems to be the brake lines which exit the body under the back seat and then run above the gas tank to the rear wheels. Salt and sand park up there and eat the lines (and the gas tank)... I've taken to using cupro nickel brake lines. They are said to be much more resistant to corrosion and do seem easier to bend without kinking. There is a good summary at <http://www.copper.org/applications/automotive/brake-tube/brake.html> I've ordered mine from <http://www.ribetautoparts.com/>, about $25 for a 25' roll of 3/16. About 2X the cost of plain steel line, but that hardly matters when compared to the high labor and time costs of installing new lines.