OB99W

These should help in identifying which ABS system the Legacy has: http://endwrench.com/images/pdfs/IdentifyInfo.pdf http://endwrench.com/images/pdfs/ABS.pdf

I believe that the '95 Legacy uses the ABS-2E system. From your description, the main part you need seems to be the "hydraulic control unit", or otherwise called the "modulator valve" assembly. From what I've seen, with the ABS-2E the hydraulic and electronic units aren't integrated, and can be purchased separately. Online sources (I'll let others make recommendations) appear to sell each for a bit over $700, so the $1400 price would appear to include the electronic control module (in addition to the hydraulic unit), which by your explanation wouldn't seem to be needed. A used ABS hydraulic unit may be worth considering, but could present some pitfalls; be careful. You also may be able to obtain the other needed hydraulic parts for less than the $450 quote, although I'm basing that on a "gut" reaction, since I don't know precisely what's required. The service rep's description sounds like detail being told in order to justify the large expense. Yes, the brakes will have to be bled carefully after the parts are replaced, and the procedure should involve putting the ABS system through cycling in order to help purge any trapped air, etc., but it isn't rocket science.

Toyota 4Runners and others also have that driveline problem. See: http://www.4x4review.com/discussions2/robots/3003.asp I didn't mention it because the original poster (jpsnowbumb) said the thump sometimes occurs twice, which didn't seem to fit, but maybe that can happen.

Although Subaru apparently doesn't use the crank pulley as a harmonic balancer (some other manufacturers do), it still needs to function as a torsional damper and to help isolate belt-driven accessory noise. Making it a solid unit would eventually cause grief . I didn't realize the noise was so loud. Based on that, dealing with it soon would probably be prudent.

I've been giving the problem a bit more thought, and it seems the following might be of interest: http://endwrench.com/images/pdfs/IgnitionCoil.pdf http://endwrench.com/images/pdfs/DirectIgnition.pdf http://endwrench.com/images/pdfs/IgnitionCoilSum04.pdf Assuming a '91 has the type of coil pack described in the links above, it appears unlikely that a bad coil in the pack would cause only one cylinder to be affected. If an internal resistor or connection in the pack is bad, a resistance check should reveal it. It would also seem that could be checked by swapping plug wires for cylinders 1 and 2 at the pack end (if they reach); the problem should move to cylinder 1 if the pack is the cause. If the coil pack passes, then I'm feeling even more strongly that an injector working "overtime" is a likely culprit. Under any circumstance, don't run the engine more than absolutely necessary until the problem is fixed. Otherwise, you may be buying a new catalytic converter soon, too. EDIT: Considering the mileage, there's possibly a valve problem. A compression/cylinder-leakdown test should reveal that.

I don't know what you feel comfortable doing. Here are some previous threads worth reading that discuss the issue and related things: http://www.ultimatesubaru.org/forum/showthread.php?t=50129 http://www.ultimatesubaru.org/forum/showthread.php?t=49761 http://www.ultimatesubaru.org/forum/showthread.php?t=51084 http://www.ultimatesubaru.org/forum/showthread.php?t=47978 http://www.ultimatesubaru.org/forum/showthread.php?t=38872 My '99 uses Dexron III, but you should verify what yours needs. Some people think a synthetic is better, others don't find an advantage to it. I'd also suggest that you consider that the problem may be sensor-related. Even a bad ground connection can cause strange symptoms in an electronically-controlled trans. Some of the threads listed above go into that (although torque-bind is the main concern touched on). Are there any dash lamps flashing, especially for a while after the engine is started? Maybe a trouble code has been set; you could get a scan done and see if anything comes up. Some of the auto parts suppliers will do that for no charge.

Hmmm, this happening more than once per stop makes it "interesting". Make sure the trans not only is sufficiently filled, but also not overfilled; be sure to check it with engine running, warm, on level ground, after going through all gears, etc. Since you don't know how old the fluid is, even if it looks fairly clear and red (and especially if the fluid is dark or smells wrong), and the level is correct, perhaps it's time for a full flush with new fluid. That will sometimes get rid of varnish, etc that may be causing a valve or solenoid to stick. Make sure the correct type gets used; the wrong one can cause strange problems due to different frictional characteristics.

A couple of years ago, a woman who was driving a Ford product pulled off the road in front of me, then followed me home, accusing me of having rear-ended her while we were stopped in a line of traffic. Of course, I hadn't done so! It took a cop and the obviously undisturbed road-salt residue (it was winter) on both my front bumper and her rear one to sort of convince her that our cars hadn't "kissed". I suspect that the car, which she said was new to her (although not itself new), was doing something similar to what you're describing. I'm not sure, but you could be having a problem with the torque converter not locking/unlocking correctly. The symptom isn't unique. Does it do it when cold, or only after the trans is warm? When stopped in drive, the idle doesn't change at all when the thud occurs? EDIT: Here are some things that might help in understanding trans operation: http://endwrench.com/pdf/feb2004pdf/4EAT.pdf http://endwrench.com/images/pdfs/4EAatPh2Win04.pdf http://endwrench.com/images/pdfs/4EAT.pdf

Don't get paranoid about this. Cookie's right, inspect things. If the crank pulley is okay, just tighten the belts a bit if they're slipping when cold. Don't forget that the steering power assist is variable with respect to engine RPM, so what you're experiencing with that may be normal. EDIT: You could put a mark on the mounting bolt and another near the edge of the pulley, aligned with each other. After a few engine starts and trips, see if the two marks are still aligned. If so, it's unlikely that the pieces are parting company, and certainly not enough to explain the steering concern.

Dry soot typically indicates an overly-rich mixture; if it's only happening in one cylinder, probably that injector is staying open more/longer than it should. Backfiring is when there's ignition of the mixture at the intake. You probably mean "afterburn", which occurs in the exhaust system due to unburned fuel being ignited by the heat there. With only one cylinder misbehaving, afterburning occuring and a catalytic convertor, you probably won't see much unusual at the tail pipe. In my opinion, the problem is somewhat less likely to be ignition-related, since a misfire due only to bad spark tends to not result in as much soot, but that unfortunately doesn't completely rule out the coil pack. When an ignition scope isn't available, I've sometimes used a timing light and see if it will trigger the same on the suspected plug wire as it will on the others. If not, ignition becomes a stronger suspect.

Offhand, I don't know what the "click" signifies, but it may be a clue to what's happening. (With my '99 Legacy OB, a click I hear from the left area of the dash seems to occur when the cruise control is engaged.) Darkening of the fluid often indicates it's getting "burned", or some worn material is contaminating it; it can look somewhat "purple" before it progresses to brown or black. It's also probable that something has been going on for a while, since it's less likely that what just happened would immediately darken the fluid. How does the fluid smell? Also, try putting some fluid on a clean white paper towel or napkin, and wait a while; if there are particles, they will stay on top of the paper while the fluid seeps in and disperses. If you have some fresh fluid, compare the odor, etc.; that makes it easier to detect burned/contaminated fluid. When was the fluid last changed? Full flush, or just drain/refill? Pan dropped, filter/screen checked? You probably know this, but make sure that the fluid level is checked "by the book"; engine running, trans warmed, on level ground, run through gears, etc. Under- or over-filling can lead to slippage, and eventually worse. Best of luck; let us know if there are any other clues. EDIT: Although the problem seems like it's mechanical in nature, isn't this the vehicle that recently got the ground connections improved? Were they previously bad enough to cause the trans to be erratic? Any possibility that something else got disturbed during the grounding work?

It seems that you've got master cylinder trouble, probably related to or compounded by contaminated brake fluid. Modern brake systems use dual-diagonal configuration, so problems in one diagonal circuit are usually isolated from the other one, which could explain why a different rear wheel at certain times seems to be affected. Check valves and compensating ports in the master cylinder determine how hydraulic pressure is applied and released, but contaminants in the fluid can cause valves to get stuck or ports to get blocked. That can result in reduction of pressure to wheel cylinders, or even line pressure not being released when you remove your foot from the brake pedal (causing what might seem like intermittently stuck calipers and "really hot" brakes). Contaminants can also cause excessive wear. The master cylinder should likely be replaced, and the rest of the brake system thoroughly flushed with fresh fluid. There may still be things sticking at the wheels even after that, but it's a starting point.

My question is: Are you still seriously considering this OB, or is the discussion just to decide whose mechanic needs to go back to school ? For now, I'll assume the former, but touch on the latter later, Peter . Fumes coming from the trans dip tube isn't a good sign under any circumstance, even after a hard run "up steep hills". That could certainly indicate coolant in the trans fluid. Tunered is right in suggesting that the trans fluid would likely be "milky" (sort of translucent pink, not clear red) after a hard enough drive; pull the trans dipstick and check. We all know about the HG issue, but I suspect that the owner is misquoting his mechanic. It's more likely that the oil/coolant pressure differences were stated as the reason the oil winds up in the coolant, not as the reason for HG failure. Your mechanic has a good logical argument, while the owner's mechanic has statistics to fall back on. However, the OB may be suffering from either or possibly both a trans cooler breach and bad HGs. The "cat related" CEL, depending on the exact cause, might indicate poisoning of an O2 sensor or the cat by coolant components, perhaps supporting the bad HG theory. So who needs to go back to school? Since apparently neither mechanic did enough for a solid diagnosis, I'd say both of them. A third opinion from someone who'll take the time to better verify things is probably a good idea.

As some of you may remember, my first post here was a request for the meaning of an ABS trouble code. The problem was evidenced by the ABS light often (but not always) remaining on after starting, or sometimes coming on after driving a few feet. In my case, the defective wheel-speed sensor apparently was marginally bad, with its resistance at about the threshold needed to fail the diagnostics at startup. It seems the times that the bad sensor wasn't caught by the "static" check at startup, the low tonewheel-generated "dynamic" output was detected as soon as I got moving. I hope relating my own experience helps. Thanks for the PDFs, Ferret, and I agree that reading the code(s) is the first step.

I suspect it's less important who does it then how they do it. The heads can be cleaned, checked for flatness, and if out of spec, machined. Do make sure that whomever does the job uses the latest gaskets, and precisely follows the current head bolt tightening recommendations.

Air has a variable moisture content ("humidity"). Oxygen can cause (amazingly ) oxidation. So, using unfiltered/un-dried air can lead to some eventual damage to the inside structure of a tire or rim. Most tires on passenger vehicles will wear out, or fail due to road hazards, before those things become an issue. If there's concern about moisture, an air compressor can be fitted with a "dryer". In addition, a filter might be useful if compressor oil getting into the air (and tire) could be a problem. Nitrogen is relatively inert. It also doesn't migrate as readily through tire rubber. If a car owner keeps reasonable tabs on inflation pressure, that probably isn't a major advantage. I've seen claims that using only nitrogen limits increase of tire pressure with rising operating temperature. That's not likely to have much of an effect, since volume expansion with temperature increase is independent of which gas, and air is 78% nitrogen anyway. The advantages of using nitrogen might make more sense to the military and airlines. Maintenance will be a bit lower, tire failures have greater consequence, etc. On the other hand, using only nitrogen to inflate passenger tires won't hurt anything, and might be beneficial in some cases, so use your own judgement.

Air is about 78% nitrogen, 21% oxygen, and the remaining 1% mainly argon with carbon dioxide and some trace amounts of other gases. So if the nitrogen tends to remain behind while the oxygen leaks out, just topping off a low tire a few times would leave it filled mainly with nitrogen, and the leakage should slow dramatically. Unfortunately, I've never seen evidence of that when dealing with micro-cracks or corrosion-caused rim leaks. Nitrogen is typically used for other reasons, and most of them are of debatable value for automotive application. As to the relative size of nitrogen versus oxygen (the other primary component of air, see percentages above) molecules, a periodic table or other data might shed some light.

Is it really all that? The Subaru database seems to use that description as a catch-all for the Legacy Outbacks of that model year. If it's not Glacier White in color, it's not an SSV. If it doesn't have a 30th Anniversary badge, it's not one of those. Does it actually have the Limited package? See http://www.cars101.com/outback_archive98-99.html to better identify what you're looking at. From that site: "Special Service Vehicle available only to fire, police and other public safety and emergency agencies, for $20,490." If it really is an SSV, that would concern me, since the mileage you quoted could be hard miles if it was in emergency service. Also, make sure the one you're considering hasn't lived in a rust-belt state in its previous life. I'm not going to get into whether it's worth buying beyond the above concerns, other than to say that my own '99 OB has had problems that I wouldn't have expected for its low mileage/usage.

Sorry, didn't realize that was you! I'll stop now. As are you, sir, even if we don't always agree.:cool:

Just to be certain, I reread my previous posts on this topic; I can't find anywhere I said "that any distance smaller than the actual perimeter of the tire is covered in a single revolution of an 'under inflated' tire". It's the wheel speed that changes. I found yet another article touching on wheel speed versus inflation pressure:http://www.findarticles.com/p/articles/mi_qa3828/is_200209/ai_n9119387 Excerpted from the above: >>The wheel-speed-based (WSB) or indirect method of monitoring tire pressure uses the wheel signal from the ABS sensor. As the tire's inflation decreases, so, too, does its radius. The tire then has to speed up to keep up with the other tires. Since most late-model vehicles already have ABS and a body computer, the main cost of adding this system is the need for a display device on the dash to give the warnings. NHTSA estimates the manufacturing cost to be in the area of $13. In contrast, the direct-reading system is expected to add as much as $200 to the cost of a vehicle.<< Since I enjoy this banter (to a point limited by my free time), I'll try once more to convince the remaining disbelievers. I'm not willing to get my CAD program running right now, so this is an exercise left to those who can picture the following in their heads, or are willing to draw it themselves. Here goes: Draw a circle (tire tread). Mark its center point (axle). Draw an angle like this /\ with the apex at the circle's center point and the opposite ends at the circle's lower circumference (tread). Let's call that angle a partial rotation of the axle. Note the distance around the circumference (tread "segment") that the angle intersects (distance traveled). Now draw a second circle of the same size, with its center point vertically aligned with the first one, but placed just a bit higher than the first one (axle offset from center, as when a tire goes down). Note that the angle now intersects a smaller segment of the tread. In other words, even though the axle is going through the same angular change, the bottom of the tire doesn't move as far. Therefore, to keep up with the other 3 wheels, the low one has to speed up. I hope that was detailed enough to evoke a "picture". (Yes, I know that my drawing "exercise" doesn't take into account the flattening of the tread at the road contact patch. That changes the ratio of the distance/speed change a bit, but not nearly enough to overshadow my point.) To further complicate things (believe it or not), even when a vehicle is going straight down a flat road at a constant speed, the tire tread isn't; the tread's tangential speed is always changing (accelerating and decelerating). Why? Because the effective radius is smaller at bottom of the tire than the top due to weight bearing. Interestingly, due to the ply orientation, even when radial tires are properly inflated the "squat" they take increases the tread-speed variation over stiffer-walled bias-ply tires. Rereading Carl's (1 Lucky Texan) link in light of what I said might clarify things. If not, I don't fault US schools, it's not an easy topic to get your head around. But this time I really mean it , no more .

Interesting link, but I think you've made my case, not yours. First, we've got the fact that low-inflation detection can be done by monitoring wheel speed, as I've previously mentioned; the low tire has to rotate faster to make up for its lessened effective radius relative to the others. Excerpted from the link you gave: "The effective radius (shorter than the ideal radius) has nothing to do with distance travelled per revolution. It will affect the effective gear ratio, however. ... Also, the greater the difference between the effective and ideal radii, the greater the tangential speed of the tread at the top of the tire." As I said, the low wheel has to speed up, with the attendant drive train concerns. Secondly, have you ever noticed that if you've got a solid axle with wheels of different diameters at each end, that it doesn't want to roll straight (in the direction perpendicular to the axle)? Well, anyone who's had a tire go down (and isn't totally oblivious to it) has experienced the car pulling to one side. Want to guess what the primary factor is that causes the pull? If anybody is still unconvinced, I give up . By the way, Jacek, physics doesn't baffle all of us schooled in the US.

It's because a certain percentage of change in tire inflation pressure isn't reflected as nearly that percentage change in effective radius (and hence wheel speed). In other words, it's not as sensitive a method of detecting the problem. However, that doesn't mean that the change is trivial. Keep in mind that for purposes of the original discussion, a fraction of an inch out of about 83 inches (very roughly what my OB's tires measure in circumference) is enough to be outside Subaru's recommended maximum difference.

Perhaps my usage of the word "effective" preceding "radius/circumference" was missed. In no way did I suggest that the "perimeter" varies much with inflation, just that the vehicle weight causes the tire to deform/deflect and have an effective change in radius. A "non-inflated" tire would have a significantly lower effective radius once it was bearing several hundred pounds. Olnick, does that do it for you?

Interesting theory, but it doesn't work quite that way. The belts in a tire aren't like rigid metal bands (segmented or not), and the tread and sidewalls are even less so. Neither is a tire running on a road surface similar to a rack-and-pinion. Rather than get into a long-winded discussion of how a tire deforms and squirm/slip/scrub vary with inflation pressure, let me give you some food for thought:http://auto.howstuffworks.com/question97.htm http://www-nrd.nhtsa.dot.gov/vrtc/ca/tpms.htm If the effective radius/circumference doesn't vary with inflation pressure, then how does one suppose the wheel-speed discrepancy needed for these flat-detection systems to function occurs? By the way, I chose the post to respond to somewhat at random; my reply is meant to be a general one on the topic, not directed to anyone personally.

Being of a practical as well as scientific nature , I decided a little research was in order. Choosing the rear driver's-side wheel of my '99 OB (with Michelin XW4 tires), I put a height reference mark very near the center. I then took measurements from my garage floor to that reference with 24, 28, and 32 psi inflation (recommended pressure and +/- 4 psi. The variation (change in radius) was about 1/16 inch for each 4 pound change, or 1/8 inch over the range of 24-32 psi. That would yield an effective circumference change of 2*pi*r, over 3/4 inch difference between 24 and 32 psi inflation (naturally, 3/8 inch for a 4 psi change). These are static determinations; a rolling tire will behave somewhat differently, but probably not by much. I'm in the rust/salt belt, and my alloy rims leak at the tire bead area (yes, they've been "resealed", to no avail). The above is one reason why I check tire pressure frequently. I don't know to what degree a Subaru AWD drive train suffers due to tire size/inflation mismatches, but I'd prefer not finding out through personal experience.