jonathan909

I'm happy with the progress on my (EJ25D) block, now it's time for the heads. I don't know what the best way is to clean carbon buildup from the valves, and googling the question just turns up endless spray-this-crap-into-your-intake links. Wire wheel? My old dip tank of Gunk carb cleaner? Ultrasonic cleaner? Combination of the above? Other?

Figured I'd close the loop on this. Issue resolved. There was a PEBTAC (Problem Exists Between Tool and Chair) component in that I (inexplicably) had the top and middle rings swapped. Yeah, stupid, but I don't think it caused the trouble. It was simply that there was a little tiny ridge on the chamfer that was catching the ring, and with each tap of the piston the ring just made it a little bigger. I still don't get what originally caused the ridge to form, but once I dressed it off with a little emery cloth (and corrected the ring order, of course), all four seated in just fine.

You're not missing anything. The parts store doesn't have any magic, just a voltmeter like yours. I mean, if someone had a reason to really characterize an alternator, he'd have a test jig that could spin it at selected RPMs and a variable load so a determination of the power (that is, both the voltage and current) the DUT (device under test) can deliver at specified speeds and loads could be made. But you don't need to do that, and presumably the parts store can't either.

Sure, but overcharging can cause battery failure, and it sounds like you're nearing the upper limit. To answer your earlier speculation, there really aren't any "components" within a regular lead-acid battery other than plates and acid. But charging at too high a voltage will mess up the cell chemistry, alter the plates, and boil away the acid. If it were me I'd be fetching an alternator from the boneyard rather than risking messing up a new battery.

It's worth mentioning too that an excellent long-term investment (in addition to the basics, like a DVM and battery charger) is a load tester. Not just useful at home for diagnosing these sorts of problems, but the only way to find a good battery at a self-serve junkyard. I haven't spent full pop for a battery in many years, and we need good ones, since -40 is not unusual around here in the winter.

My understanding is that there are very significant differences between Subaru transmission filters and anything aftermarket, esp. engine oil filters. There's a huge difference in price (and I'm about the cheapest SOB around), but this is a case where I'd stick with OEM.

Aftermarket manuals (e.g. Haynes), while broadly useful, are never accurate enough for this kind of work - there's just too much subtle deviation in the wiring from year to year. Look around here and you'll find links to free downloadable sources for the authentic Subaru TSMs.

Well, 13.5 (or not too much more) is okay - that's what a 12V lead-acid requires for charging. Thing is, rp2813 is reporting somewhat intermittent behaviour, so while the voltmeter may say it's fine while it's in the driveway, 20 miles down a bumpy road may be a different matter. A tanked alternator is easy; to manually induce the failure in an intermittent one not so much. That's why, if it's reading >13.5V in the driveway, the best way to determine whether it's at fault is to replace it, keep driving, and see if the problem reappears. But there's no disagreement here - all of the above answers are correct.

Perfectly valid point. My experience with alternator failures has been "sudden death" rather than intermittent behaviour, but there's no reason it couldn't be the latter (e.g. a flakey regulator). The only problem is that it's harder to "force" alternator misbehaviour than it is to wiggle a bad connection. Best (cheap and easy) way to debug it is to drop in a replacement alternator (esp. from the junkyard) and see if the symptoms recurr.

That all these functionally-unrelated systems are failing at the same time points to a global electrical fault - as the previous posters suggested, portions of the circuit that are common to pretty much everything. Get the FACTORY TSM schematic, follow the lines, then (and I know this is going to sound really unscientific) start wiggling stuff. That the radio is failing should be a big help - you can turn it on, lift the hood, and push wires around (and remove/reinsert relays and fuses) until you replicate the failure. As suggested, the problem is probably in the main power distribution area. (A few months ago I chased down a similarly transient failure in my 2002 Forester ABS - it was caused by a corroded fuse termination in the fuse/relay box.) As you remove each fuse and relay, look down into the terminals, and if they aren't clean and bright you're probably getting close. This is very much a DIY thing, because chasing down intermittent electrical failures is more a job for time and patience than it is of skill, and you could easily get into thousands worth of shop time if it's a nasty one.

Oh, you're good...

I think I have the picture. I'm not doing anything wrong - I just think I have to finesse this thing a little by dressing the chamfer and adding some lube. Heh - I think that dates us. My first assembly course was PDP-11 (useful when I started working with 68000s later), a night course I took a year or two after I got out of high school. Our instructor was a hard case who wouldn't actually let us use the assembler - he made us hand-assemble all the code to make sure we understood the relative branch calculations. Then the object got entered via front panel bit switches. Oh, happy days...

Perhaps I should word the question differently: If the taper is normal, and your compressor sits on top of the block, how come your rings don't hang up in the chamfer when you insert them? Is my chamfer too rough and in need of a little polishing, should I be greasing it up with assembly lube so the ring will slide down it into the cylinder, or both?

You're missing the point - the chamfer at the top of the cylinder is a larger ID than both the cylinder wall and the compressor, and thus forms a space for the ring to snap into as soon as it leaves the compressor. That's where I'm getting stuck. That's the style of compressor I'm using - with the little clutch. I'm wondering whether surface roughness of the chamfer is the problem, and maybe whether smoothing it a little with emery paper might be enough to let the ring slide down.

No, not ratcheting, just a couple of cheapies. I'm thinking that may be the problem; at the same time, even if the compressor is perfect (e.g. the Wiseco GD mentioned) I still can't see why the ring wouldn't snap open as soon as it enters the chamfer and get trapped just like it is now.

Of course - I'm just puzzled, not stupid. Gap is well within spec (.018), and with it sitting in the chamfer I can see the gap i.e. it's not pinched shut. Hell yeah! Those are a couple of old legacy PCs (DOS, NT, 2K) I keep around for various not-unreasonable reasons. It's too unfocused to see, but the silver+blue plate over the drive nearest the cylinder is the nameplate from the service console for an ETA Systems supercomputer (CDC spinoff company circa late 80s). You can see an ETA-10 get destroyed in the big computer room shoot-em-up in the first "Die Hard" - it's the Fluorinert-cooled box with the cool plastic dome on top.

Don't know whether this shot is good enough to be useful, but you can see the chamfer (it's about .100" top to bottom) fairly well. A quick+sloppy measurement with my dial caliper (where did that digital go....?) says that it tapers out from the wall about .010".

I'm sure that this is just me being an idiot, and that Someone Smart (tm) like GD will straighten me out. I'm trying to install these EJ25 pistons with new rings. Squeeze the rings in the compressor, position it, tap it down, and it stops dead before the piston and block tops are flush. The problem is that the cylinder has a slight chamfer at the top (about .010" larger than the cylinder bore) that extends down about .125", the compression ring is hanging up on it, and not ramping into the bore. Is there some trick I'm missing, like using a special thin ring compressor that will seat down into the chamfer rather than sitting up on the block's mating surface?

Because of the increased friction from the teeth constantly engaging/disengaging? It's a theory...

You're quite right - without religiously changing everything at the same time (or keeping precise track of when - in operating hours, or miles, or whatever - the various components are changed) we can't get statistically significant MTTF/MTBF numbers. But I suggest that "real world" numbers - derived from the results of people just changing the parts they think they "need" to - can be just as useful. What I find really interesting here (and everyone's experience seems to agree) is that the toothed idler fails at a much higher rate than the non-toothed ones. They probably aren't replaced any less often, their construction is the same (I'm assuming that any given manufacturer will use the same bearings for both idler styles), and they all rotate at (more or less?) the same rate. So what would account for the predominance of toothed idler failures?

Of course, we all know we're supposed to do periodic timing kits, but we also know that procrastination runs high. Last summer (about a km from home - could have been a lot worse) I suffered my first timing failure on a '99 EJ25D. We were just starting out on a trip down to Montana/Idaho for camping, sailing, and the eclipse. When I did the post-mortem, I found it was the idler sprocket that had seized, shattering the right-side sprockets and cover. So I'm curious what the dominant failure cause is, or what the proportions of failures are across the idler pulleys, idler sprocket, water pump, tensioner, and the belt itself. A straw poll would be welcome.

On the other hand, you can lay down a bead, break for lunch, and then assemble the halves, since it won't start curing until mated. I don't know that stuff - will look into it (always nice to find a local disty). Gotcha. I normally only use assembly lube where the stickiness is really need to keep things from dropping, but get your point. Would a suitably viscous oil (e.g. 20W50) not work as well? And now that we're on the subject of lubes, I've seen mentions of using ARP Fastener Assembly Lube on the rod bolts in order to get more accurate torquing. What's your take on that vs. using regular engine oil (e.g. 10W30)?

An OT but related question, if I may. What's your preference for sealing the block halves? I think the TSM calls for an anaerobic flange sealant, but I've seen some speak favourably of Ultra Grey in this application. I've long had nothing but fear of and loathing for silicone RTVs in general, but the specialized gasket variants tend to be a better breed.

GD: Thank you - I appreciate the detail of your reply and will follow your advice. A question, though: This approach (and your description of the ring manufacturing in particular) is predicated on the cylinder not wearing out of round over time. If true, that's very impressive. I'm curious what the tolerance spec for out-of-round wear is, and what your experience is in this regard, i.e. how many miles will a block typically have to see before going out of round far enough to need reboring?

As I indicated, I noted your frequent strongly-expressed opinions on the subject. My understanding is that the "glaze" isn't a deposit (as you suggest), rather the mirror finish resulting from hundreds of thousands of miles' worth of wear against the old rings, and that new rings (that still carry microscopic surface roughness) need a similarly slightly-roughened cylinder wall to wear against in order to bring the two into conformance with each other during break-in. The foundation of this notion is that a new, rough ring will not seat in properly against the cylinder wall previously "mated to" the old ring. If this understanding is incorrect, I'd welcome a more thorough explanation of the present case (without the gravel, drywall mud, etc., hyperbole). I'd also like to hear dissenting opinions, should any be present. Just looking to get at the truth, not for a dustup.