jonathan909

I'm still trying to catch up with all y'all wrt what a "quick strut" is. Does that just mean "strut-with-spring assembly"?

In other words, you homebrewed a pass-through socket. You could do it that way. And if your experience was like mine, you'd be doing it twice - once with a 17mm socket and once more with a 21mm.

GD, I heard you and believe you. I'm just keeping my mind open at the moment to the possibility that this faked-up VSS signal is otherwise confusing the ECM and causing some other unpredictable behaviour unrelated to rev-limiting.

Easy enough to confirm with a scope whether in fact the speedo circuit is just doing a simple waveshaping/level shifting without otherwise altering the signal. Unlikely, I think, that it's "random". Random is actually pretty hard to do, and I don't see how it would be useful in this case. More likely that it's variable or adjustable or something that's easy to implement with a 555 timer or similar. You wouldn't happen to have the schematic for it, would you? Might be helpful at some point. And yeah, I'm an electronics geek, mainly embedded systems (like this one). I'll be interested in hearing the outcome of the test with the VSS fed through the speedo.

I specified a self-serve yard because they're a completely different animal than a full-service yard. They're utterly indifferent to the removal of the valve body devaluing the remainder of the tranny. As far as they're concerned, any money they make on any portion of a car is gravy in advance of the rest going to the crusher for metal weight value. I didn't pull that $50 (CDN) figure out of my butt; that's what they actually charge around here: http://www.picknpull.com/part_pricing.aspx?LocationID=76&NavItem=12&SearchCriteria=&#partpricing I'm not going to argue the point further - it's your money, spend it as you wish. I just want you to understand that I'm not blowing smoke here. On the reprogramming part, I don't know, but my guess is that it wouldn't be necessary.

Disclaimer: I know nothing about ATs; the extent of my experience is replacing the 4EAT in my '99 Forester. It was suffering from a neck-snapping 4-3 downshift under power - if you stepped on it for passing, to get power on a hill, etc., it felt like you were being rear-ended at speed. No codes. Xmission repair shop estimated $2400 (a nonstarter). Found a matching transmission in a junkyard car; it had body damage (implying it was mechanically sound when taken off the road), so I took a flyer and spent $130 and an afternoon dropping it. When I bolted it up it worked perfectly. Having said that, if I were in your situation I'd get my butt to a (preferably self-serve) wrecking yard and pull a valve body. Around here they charge $50 for them.

Oh, I get it. You're feeding it a pulse train slower than normal to make the ECM think the car's groundspeed is lower than it is. I can't speculate on what other implications that might have. Any idea what normal is, in Hz/mph or something? What do you think your top end speed should be? [Edit] That's interesting - the xmission's speed sensor output doesn't go straight to the ECM. It gets fed into the speedometer circuit (in the gauge cluster) which in turn feeds the ECM VSS input. That could either just be a wiring "convenience", or the speedo circuit is conditioning that signal somehow. Seems to me the easiest approach to debugging this would be to unplug your VSS hack and put a gauge cluster in to see if it behaves differently.

I really don't know $h!t about $h!t, but I'm kind of curious about this. Assuming that by "We built a circuit to simulate the 5 volt signal on pin 83 to the ECM for the VSS." you mean you built an oscillator that feeds a pulse train to the VSS input, why would you go to all that trouble instead of just connecting the transmission's speed sensor output to it as usual? (I was also briefly confused by this, as "VSS" has a totally different meaning in my world - it's a power supply rail, usually ground.) And is p.81 grounded to tell the ECM that it's connected to a 5MT?

Here's the thing: That's what I found while disassembling the strut - that I was able to get a (cheap) offset onto the stock 17mm nut. However, no way no how was there enough room to get one onto the 21mm nut that the aftermarket shocks came with. That's why I'm getting a pass-through set.

Yup.

To be fair, I've only replaced one pump (out of the six cars I've owned - not thousands!), and it was in the '02 Forester that I rescued from a country junkyard after it'd been sitting there for nine years. Fools hadn't drained the tank when they parked it, so after I replaced the motor the fuel pump was the next step. It was seized solid, there was corrosion all over the bracket, and the rubber mount had dissolved into a sticky gob resembling half a dozen packages of well-chewed Black Cat gum.

Why, you're lucky. Around here, we walk uphill both ways. (For anyone wondering - and those who weren't - Monty Python were ripping off Stephen Leacock when they did that bit - Cleese admitted it.)

Well, or another carrier (USPS?), or leave a note saying "sorry for the blocked driveway, please leave package in car", or deliver to a neighbor, or...

Good. With the right tools, though, you won't even have to do that.

For good prices on parts, I haven't found any better than Rock Auto (other than a wrecker), though others here may differ. The hoses you just have to fight with. That it's turning is a good start; beyond that, sometimes slipping a really skinny little screwdriver between the hose and the pipe will help loosen it. And if you can push on the end of the hose rather than pulling on it, it'll tend to expand and slide off rather than squeeze tighter. I suppose squirting a little lube like WD-40 between the hose and pipe may help as well, without causing any problems.

My pleasure entirely. Sounds like you've got it well in hand. I've only messed with these things in 99-02, so I don't know which style of pump the 95 has. So are you now stuck at home, bingeing a TV series, until the UPS man shows up with a package from Rock? Or do you have some local support that might give you a lift to a local parts source?

That's precisely what I'd take it to mean. Sounds like it's time to pull that cover and get the old one out and confirm it's dead. What's your plan for getting a replacement?

Hubba-hubba! Zackly what you need.

I empathize - I'm way out in the country too, but with the advantages of a ridiculous stockpile of stuff to work with and (usually) an extra car to get about in for things like this. If you're CAREFUL, you can run a quick test (that's all you need right now) with that alligator clip on the lower pin (as shown in your photo) and just a piece of wire for the upper. If the wire is something like AWG 16 or heavier, cut the end flush. The insulation will give you reasonable protection from shorting against the clip, but the exposed end will still make contact with the pin. Remember, you only have to make contact for a few seconds so you can hear the pump run (or not), so you don't have to be elaborate about this. As for anything under the dash, yeah, that's how it is. I'm sure it appears in one of Dante's circles of hell.

I have the wagon too, and I still wouldn't chase wires. There's no way you're going to jam two alligator clips of that size in there without shorting them together, so you should definitely snag a couple of the small quick-disconnect crimp connectors. (Photo credit: My daughter Becky)

Sorry, I can't help on the rumoured-wire front - I'd just find a way to connect to those (male) pins. If you can lay your hands on some small (e.g. 1/8") female "quick disconnect" crimp lugs, they may push onto the pump pins. Otherwise, I'd disconnect those three hoses and pull the cover (it's only eight nuts, so you'll have it off in a minute) and lift out the whole fuel pump and sending unit assembly. Then it'll be easier to test "on the bench". Personally, I'd rather do that than start pulling seats and chasing wires.

Looks like the original nut was a predictable 17mm, while the replacements are an annoying 21mm. Also appears I'll use this as an excuse to buy a pass-through socket set. (For the canucks in the audience, Princess Auto has a decent-looking Channellock set on now for $50 that goes up to 22mm.)

The pass-through socket is definitely the way to go - I just did a set of four in August. IIrc, the nut sizes varied between those that were on the car (presumably factory) and the aftermarket replacements I got from Rock. I think I still have the old ones in the scrap bin and can check a little later.

Stands to reason that it'd be carbon, but maybe in some weird annealed state. Not being a carbon chemist, I really can't speculate further, but I remain curious.

No - the valves don't have to come out to pull the shims, but I removed/reinstalled the one I had to grind the stem on. I suppose one could grind it in situ, but I don't think I'd try that myself (for the obvious reasons). But lacking Subaru's (expensive) fancy-shmancy tool, the cam has to come out for each shim change - or at least be loosened off to get the clearance to slide the shim out. Shim grinding strikes me as counterintuitive. Much more material to be removed than just grinding the valve stem, and unless you've got precision machine tools, impossible to grind true.