jonathan909

Have you tried http://jdmfsm.info/Auto/Japan/Subaru/Impreza/1998/ ?

GD, thanks for that - very interesting and informative. Puts the lie to what we've been told all these years about the importance of friction-modifying additives in ATF. Fwiw, Walmart's house brand Dexron VI / Mercon LV specifically mentions compatibilty with Subaru ATF and ATF HP, so if one insists on playing by the rules it isn't that hard to find after all.

Oh. Trademark lawyers. That sheds a whole different light on it. Ask me sometime about why Intel switched to naming (rather than numbering) their chips when the 486 made way for the Pentium... Trivia: Yup, sperm whales, in fact. And it started out as a pretty high percentage, as I recall. They had to get smarter about it when it became obvious that the whales would become extinct if the fluids weren't reformulated.

So the group's considered opinion is Dexron III or Dexron VI? Any further confirmation or dissent?

Have to look at the '98 schematic. Something like that draws so little power that there's little sense (from a design standpoint) in dedicating a fuse to it alone. So it's likely that it's on a fused supply rail that also powers other stuff. Better to track down and disconnect the module itself.

Ahem... My rather extensive diversion into the world of the Alpine system revealed that as the ignition switches get old, they get noisy and the counts become NFG - that's why you can't get an old car to program itself to a fob. So (assuming this is an Alpine) I wouldn't consider the three-cycles-to-turn-off-alarm to be reliable.

Does the '98 use the Alpine remote? The '99 does - there's one module for the door locks and another for the security. I had a '98 but never bothered to look at it. So if yours does as well, you just have to unplug the security module. They moved it around from year to year, so it's either up underneath the dash or under the parking brake lever.

I was thinking of something less complicated, like a broken separator plate. I've never had a reason to try, but can you poke a probe down into the block vent and touch the plate?

Isn't that what the oil separator plate is about? Keeping the oil from splashing up into the PCV valve, which is right above it?

Fwiw, faxvin.com says 4-speed.

Good question - not my car, largely unfamiliar. I'd like to say "I dunno", but I think it may have been a US market car - speedo has big-number miles and little-number kms.

So you're saying standard-issue, off-the-shelf Dexron is okay for this one? Which Dexron (there are so many...)?

Now I don't know what the hell is up with ATF anymore. Can someone tell me which fluid I should be using in this '06 OBW? I mean, something I can actually find around here?

Absolutely. Nobody here thinks any less of you for asking. For my part, I too first came here with a "stupid question" that I couldn't get a straight answer to elsewhere, and I stayed because this is where I found clarity. So forget the greasy kid stuff. Get over your nerves and do it the way they tell you to do it. If you start winging it, that's when you're going to be stepping outside of established parameters and getting into trouble.

Thanks, I was just being too lazy to pick up a book and get the name right. But I probably wouldn't call it a dogbone because sailboats use "dogbones" too (though I favour other styles of trapeze rings myself) and I hate namespace collisions.

Well, I'll be off to the wrecker - your best dietary source of just about every car part - to get a replacement, but since I don't have the original to compare to I just want to make sure I don't waste a trip if they come in more than one length. Might be worth a call to the dealer's parts dept. to check.

Perhaps, but if so it sounds like a stupid shortcut to me.

If you're using the right gaskets, dry. If you're not using the right gaskets... you're not using the right gaskets.

Hmmm... maybe I'm just being fooled by the material change. Used to be an Al casting, then moved to FRE somewhere along the way, no?

This 2006 OBW has a few signs of clueless hands having been on it. One is that the brace between the bell housing and firewall is missing (the ends are there but the middle isn't). I'm guessing that the length changed over the years. If I'm right, which years might offer a replacement for this one?

Oh - and do you still do the (factory) initial torque steps (before backing them out again), or do you just dispense with that business?

GD: I hope I'm remembering this right - that a handful of years ago you spoke on the subject and said that you used the ARP lube. When I asked them about it, they cautioned that it was only to be used with their bolts, otherwise the reduced friction risked exceeding the tensile rating on Subaru's bolts. So you're saying that the stock bolt+Amsoil combination is a good one. I'd like that a lot. Torquing to an actual torque value instead of the torque angles would be a whole lot less PITA. [edit] As long as we're here, may I ask a couple of other questions? First, do you apply those same torque specs to both the SO and DO engines? And would it be safe to assume that a good degreasing (e.g. Brakleen) of the block holes, bolts, washers, and head-washer mating surfaces, followed by a coat of the Amsoil lube on all of the above, is appropriate?

I have the usual assortment of "good" and "just okay" torque wrenches, but the question was specific to torque angle wrenches.

Now a different question without actually going OT: I'm doing this one in the car, and these 45 degree torque specs are rather harder than the 90s specified for the older engines. I've tried using my cheap little dial torque angle indicator, but it isn't a good fit. Has anyone tried this (relative) cheapie or anything like it? I've looked at a lot of indicators that attach to ratchets, breaker bars, etc., and nothing's impressing me much. https://www.harborfreight.com/12-in-professional-flex-head-digital-angle-torque-wrench-56683.html

We get the same ones you do - they have the "Made in Indiana" stickers in the windows - modulo the required metric speedo, etc. In this case, though, your automatically reaching for the docs filed under "USDM" would appear to be a mistake. That's the set that has the radically different procedure, and without understanding why (or which engine it's intended for), I don't think it should be trusted.