OB99W

Perhaps these will help: http://endwrench.com/images/pdfs/IgnitionCoil.pdf http://endwrench.com/images/pdfs/DirectIgnition.pdf http://endwrench.com/images/pdfs/IgnitionCoilSum04.pdf

Todd, I just came across this: http://www.scoobymods.com/forums/showthread.php?t=2213 I thought it might be interesting, although the symptom seems to be an increase ("flare") of a couple hundred RPM, rather than the momentary decrease you mentioned (assuming I understood you correctly) during the shift.

The link I posted was to an article about breakage of the reduction drive shaft in the trans. The input drive would still be intact in this case, and the fluid pressures would remain normal. Speaking of which, the first step is still to check the fluid.

One possibility, out of several: http://endwrench.com/pdf/drivetrain/4EATMovementInfoS01.pdf Can we assume that the fluid level has been checked and is okay?

Reference info: http://endwrench.com/images/pdfs/IdentifyInfo.pdf http://endwrench.com/images/pdfs/ABS.pdf

I found an interesting piece from "Motor" magazine: http://www.findarticles.com/p/articles/mi_qa3828/is_199810/ai_n8811652 This relates to a '93, so the wiring info may not be accurate for a '90, but the concept might be applicable. Here's an excerpt from the article (what appear to be "typos" are probably due to scanning using optical character recognition): -------------------------------- Subaru Engine stumble with a/c on. Zexel air conditioner-equipped 1993 Legacy models that stumble when the a/c compressor cycles can be fixed by doing a little ECM rewiring. First, start the engine and get it up to operating temperature with the alc on. Then, pull and replace the a/c relay in the main underhood fusebox several times to try to duplicate the condition. You may vvind up doing this 20 times or more. If you can't make the stumble happen, try doing the same thing with the la/c main fan relay If you get the stumble this time, disconnect the battery ground cable, then the green fuel pump relay connector (B30). Connect about 10 inches of 14gauge wire to the yellow 14-gauge wire at tenll 2 with a Scotchlok connector. Don't cut the wire at the relay! Route this new wire carefully to the ECM by pulling connector B48, then cutting the yellow/red wire at terminal 15 about 3 inches from the connector. Use another Scotchlok to connect the other end of the new wire to the yellow/red wire, then reconnect it to terminal 15. --------------------------------

Reference info: http://www.motor.com/MAGAZINE/Pdf/092001_08.pdf

Lucky guy! Some of us with 2.5, 4EAT, and/or rust-belt state envy you.

If you haven't seen these already, they might be useful:http://endwrench.com/images/pdfs/IgnitionCoil.pdf http://endwrench.com/images/pdfs/DirectIgnition.pdf http://endwrench.com/images/pdfs/IgnitionCoilSum04.pdf An exact sequence of what you checked/replaced relative to when the no-start problem began might help with diagnosis. Do you have spark on cylinders 1 & 2 and 3 & 4? A hot spark requires things be right on the primary side; is the battery in good shape, are connections clean? How's the cranking speed? You mentioned compression and spark; assuming air is getting there, is fuel?

If the pulley is wobbling significantly, it could cause noticeable vibrations, and that should be investigated further. However, increased load on the engine can cause roughness as well, and could be related to tightening drive belts. For example, stepping on the brake naturally turns on the rear brake lights, which the alternator/battery should be supplying current for; if the alternator belt was previously slipping, it wouldn't load the engine as much as when the belt is properly tightened and the alternator working up to demand. The battery is supposed to "fill in" for the alternator at low engine RPM, so if it isn't up to the job, even that can cause the alternator to have more demand at idle than usual; checking the battery terminal connections for corrosion is a starting point. Obviously, lots of other things can cause vibration from rough idle, etc.

You may need to get under the dash, but many of the ground connections tend to be on the other side of the firewall (in the engine compartment).

Nipper may well be right about the vacuum leak, especially if the misfire is continuous and not just a one-time event when the switching occurs. On the other hand, if you want to concentrate on possible electrical problems, I wouldn't look to defective back-emf-suppression diodes as the first choice. Rather, I'd suggest checking that all chassis ground connections are good (tight, low resistance). Since you seem aware of electrical/electronic terminology, I assume that you could use a multimeter and check for volage drops across connections, etc. A bad ground or two can sometimes cause "interesting" (and varied, depending which grounds are bad) symptoms.

I don't know which wheels interchange, but I do know that besides the obvious width/diameter/bolt-pattern, offset and some other things need to be considered. See: http://www.tirerack.com/wheels/tech/techpage.jsp?techid=95

It's been my understanding that the design of Subaru 4-cylinder horizontally-opposed engines is such that no harmonics requiring external balancing are generated within the normal operating range of these engines. This is due to factors like the rather short, stiff crankshaft, with cranks 180 degrees apart leading to mainly second-order harmonics (unlike V-type engines that have first-order harmonics and almost always require a balancer). This is not to suggest that these engines don't have significant torsional vibrations, which is a separate issue, and something that the rubber-isolated crank pulley is supposed to help deal with. Harmonic balancers differ in that they are "tuned" to work at a specific frequency that otherwise the crankshaft could excessively resonate at. Since what I said above is based partially on my reading of Subaru literature, if anyone has tech info to the contrary, please point me to a link.

Agreed; when the crank pulley rubber deteriorates, it can cause several problems. Damage to the front main seal due to excessive vibration (eventually causing an oil leak), is one possibility. Misalignment of the crank and PS/AC/alternator pulleys can even damage bearings and seals in the accessories. Slipping of drive belts is, of course, a possible symptom of failing rubber; that's why it's important to inspect the crank pulley before just tightening belts. Nevertheless, belts do wear/stretch/glaze a bit over time, and as long as the pulley has been cleared of blame, tightening the belts a bit may be all that's needed. As in most things, some judgement is required.

Okay, this is now sounding like the band servo is momentarily keeping the band applied as it would for 2nd (and 4th), rather than immediately releasing the band for 3rd. Take a look at http://endwrench.com/images/pdfs/4EAT.pdf . Scroll down a few pages to the section "Band Servo Operation". Read that and look at the diagrams to the right. It would seem that if either the servo is a bit sticky or the 3R accumulator is sluggish, that might explain the behavior you're experiencing; gentle acceleration might give sluggish things enough time to actuate, and hard acceleration might raise pressures a bit and force something sticking to move. I know you said the trans has "good clean fluid", but if it's not really fresh, maybe a flush would help. If there's a bit of "varnish", new fluid sometimes will break it down. That's about all the insight I can offer for now; hope it leads to something good.

Not to make too fine a point of it, but I seem to have said that previously; to quote myself (from back a few posts in this thread): "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."

Perhaps the following links will help provide some insight concerning the relationship between spark for two cylinders and the coil pack, as well as some troubleshooting info. http://endwrench.com/images/pdfs/IgnitionCoil.pdf http://endwrench.com/images/pdfs/DirectIgnition.pdf http://endwrench.com/images/pdfs/IgnitionCoilSum04.pdf

Under "normal" conditions, I wouldn't expect any shifting into 4th when the selector is in 3rd; it should be inhibited unless revs are so high that the shift is forced. This is again sounding as if the TCU (at least momentarily) thinks there's a higher engine speed than there really is, or something like that. Does the tach do anything unusual when the symptom occurs?

As long as you accept that answer, it's not a problem for the dealer. Have you been getting this response in writing, should a "lemon law" settlement be necessary in the future? Have you tried another dealer? You might want to contact Subaru directly: http://www.subaru.com/tools/contactus/index.jsp . Possibly ask to have a meeting with an area representative.

Interesting :-\ ; I based my comment on what I thought was a reliable source. According to Subaru, "The pulley on the front of the crankshaft does not serve as a harmonic balancer." See: http://endwrench.com/images/pdfs/AllSubaruInfo.pdf

Two questions:1) Does it do this when the engine/trans are cold? 2) Does it do it when warm with the shift lever in "3", not "D"? Both of those conditions should inhibit operation in 4th unless engine speed is excessive. I'm wondering if somehow the TCU is getting a wrong signal (or misinterpreting something) so that it temporarily "thinks" there is an over-rev situation, forces an upshift, and then corrects. Bad ground connections have been known to cause strange trans problems, so they might be worth looking into.

I don't understand why, if the part number for the replacement hydraulic unit matches the original, there is a need to replace the electronic control. I could only see the necessity for that if the new (or a used) replacement was not a match for the old, and had different control requirements. That, of course, may indeed be the case. Seems if lmdew can provide a part-number-matching unit, you're probably set. Of course, if an o-ring and bleeding is a cure, then that makes sense. Do what seems prudent. Best of luck.

You're welcome. As to bleeding, in addition to the info in the link I previously provided (which describes among other things getting the ABS to sequence), here's more: http://endwrench.com/images/pdfs/BrakeSystemWin01.pdf Unless you have more than a wrench, fluid, a hose and a jar, this is a 2-person operation that requires coordination and communication between the people doing it. However, as I said before, it's not rocket science. Try not to introduce any more air into the system than you have to while working. Keep the ABS hydraulic unit as full of fluid as possible. Bleed all 4 wheels, in the proper order. Then either actuate the ABS sequence control (preferable), or has been suggested before, get the ABS to actuate by hard braking on a loose surface (away from other cars, please). Then rebleed.

There's a "trick" I use to see if squeals are due to loose/glazed/slipping drive belts. I fill a spray bottle with water, and give the suspected belt a good "spritz". The water temporarily changes the friction, and the noise will either go away for a short while, or at least change significantly in nature if it was the belt causing the problem. If not, I look elsewhere.