nelstomlinson

I would suggest that you consider 3M 5200 rather than silicone. I have Yakima Q-towers on a '96 Legacy, and they have worked well since '97. No problems in use, and no damage to the roof.

Not fixed at all! The problem is intermittent, and it's back. As before, the live data shows the TPS value falls to zero (the gas peddle is depressed and the throttle is definitely open) and then it jumps back up to a correct value, causing a surge. I'm back to suspecting a bad wire or bad ground. For now, the engine runs just fine with the TPS disconnected, which means the computer is getting a 100% reading. We can live with that for a while.

Sort of on topic: would a '99 5MT from a 2.5L DOHC fit behind a '96 2.2L? How about a '01 4EAT from behind a 2.2L to go into a '96 2.2L? More generally, is there any kind of chart that shows which lego pieces interchange? It looks

I tested the resistance between E13 pin 1 and ground, got 136 Ohms. That's obviously bad, so I found the other end of that cable at the E2-B21 connectors. There are three engine-bulkhead connectors mounted to the passenger side of the engine, down low. E2-B21 is the 12 pin connector in the middle of that stack. I unplugged that one, and put on a smear of dielectric grease before I re-plugged it. Then I read -5 Ohms. I figure the negative resistance indicates there is some stray voltage, but I'm not going to worry about that yet. I plugged the connector back onto the TPS and took a test drive. Same problem! Now that I have the scan tool , I could check that, and it was indeed the case. I probably was at 1/8th throttle before the TPS value got off 0%. With Key On Engine Off I twisted the TPS until the scan tool reported the smallest non-zero value I could manage, about 0.4%. I drove about 50 miles and I think it's fixed.

Thanks. I should have said in the original post that I did replace the fuel filter and test fuel pressure early on. That is an excellent thought! I'm accustomed to AC power, and I don't have the habit of thinking about power flowing through anything but wires.

I got a scan tool, an Autel AL629, and looked at some data. What I found is that the TPS reading jumps about wildly while the engine is bucking and surging and hesitating, and drops to a steady zero once the thing stops misbehaving. Unsurprisingly I now have a P0121 TPS range error as a stored code, along with the P0120. My best guess right now is that I have a bad connector or broken wire somewhere in that circuit, and when it makes contact and the signal gets through it's all noise, and when the broken wire isn't making contact at all, the circuit reads zero. I plan to test that hypothesis by disconnecting the TPS entirely, and driving for a while to see if that intermittent problem returns. I haven't put an analog vacuum gauge on it yet, but the scan tool shows that manifold pressure responds as expected to throttle operation. The scan tool probably wouldn't show the flutter that could show up on an analog gauge, and I should get my old analog gauge on there.

Thanks, heartless. I was able to retrieve that from http://jdmfsm.info/Auto/Japan/Subaru/Legacy_Outback/1996/Service Manual/ENGINE SECTION/ENGINE/MSA5TCD96L5971.pdf It does look as if a vacuum leak is a likely candidate for surging and hesitation. The throttle position sensor that I keep getting a code for doesn't even show up on that chart. The TPS is new, and its resistance measures OK, so I don't think it's causing the code. I have a vacuum gauge. As I recall from carburetor days, if the idle vacuum is low, or fluctuating, there is a significant vacuum leak. Does it work that way for fuel injection?

FerGloyale, mine is the three pin TPS. Is that the kind you're talking about that has the fully closed switch? Can you point me to a description of the procedure to set it? I still haven't found that. Also, do you know if that 4.6V open circuit reading on the center pin of the connector is OK?

OK, that makes sense. Where do I find that procedure? I have downloaded a factory service manual PDF, do you know where that procedure is in there?

I cleaned the throttle body as best I could, no change, still have the same intermittent problem that only shows up once the engine warms up. I can't clear the P0120 code until the engine cools down. I removed the connector from the TPS, and the center pin shows 4.6 Volts. Should that be higher? Obviously, it's the three pin TPS. The new TPS shows about 5kOhms resistance between the outside pins, and the center pin to outside pin resistance varies smoothly from about 60Ohms to 5kOhms. I found a couple threads on how to clean the throttle body. The instructions varied from ``Don't even try, you'll ruin it!!!'' to ``Spray it with Gumout.'' I cleaned it inside with a rag moistened with a little WD40, until I couldn't get anything more out. There was definitely a lot of carbon in there, but only a thin coating, not thick piles of crud.

No. How do I do that? I have a couple cans of carb cleaner at home; can I just spray some in there while I rev the engine?

I have a throttle position sensor error, P0120. After the engine warms up the engine will eventually start to hesitate and then surge, as if I was taking my foot off the accelerator and then pressing it again. After a bit of surging, the check engine light comes on and I can read the P0120 code. Warm or cold the engine starts and idles fine. If I clear the codes and drive some more, the surging and TPS error will return after a bit. It never happens until the engine has been warm for several minutes. I replaced the TPS and the problem didn't go away, so I don't think it's the sensor causing the TPS code. Does anyone have any recommendations for what else I should try?

Two more questions on this business. I have a line on a running 2014 2.5L engine, pulled from a wreck. 1) Can anyone tell me whether this would play nicely with the ECM in the '99 Legacy 2.5L DOHC? 2) Would it be reasonable to use this as a new bottom end with the '99 Legacy 2.5L DOHC heads and ECM? Or maybe just use the crank and con rods out of this engine to rebuild the '99 engine? Unfortunately, I still haven't pulled the bad engine out of the '99. Maybe this spring, after it warms up. It was -24F this morning.

I won't know until I open it up, but rod bearings is what the fellow who gave me the car suggested was wrong. Thanks for the suggestion; that might be a good option.

If I luck onto a pre-'99 EJ22 cheaply, I might go that route. On the other hand, I would enjoy doing a rebuild. Either way, the first step is going to be removing the current engine. The problem up here is freight costs. It would probably cost more to ship an engine up here than it would cost to buy it down there. I've never seen a Subaru engine on Craig's List here, and finding that '01 with a trashed body for $1,000 was a happy surprise that isn't likely to happen again.

I found the Subaru Technicians Reference Booklets and have been reading up. It looks as if the easiest, cheapest route to resurrect the '99 might be to rebuild the engine. Next step will be to pull it out and see how bad the bottom end really is. The 2001 has a trashed body, so it will have to become a parts car eventually.

If you can prove that they lied about the mileage, I'd suggest small claims court. Also, maybe file a fraud complaint with local law enforcement?

Wtdash, I want to drop the entire, running 2001 Impreza engine into the '99 Legacy, and bolt it to the Legacy's manual transmission. I'll toss the old Legacy engine. There won't be any swapping of heads or anything like that. Are you saying that the older Legacy flywheel won't fit the newer Impreza engine? Or the engine mounts are different? Or the bell housing has a different bolt pattern? Or maybe all of those and more? Is there a primer on here that covers issues like this? Things like Phase 1 and 2, EJ22 and EJ25D are greek to me so far, and I'd like to fix that.

I have a 2001 Impreza with an automatic and a running 2.2L engine, and a '99 Legacy outback with a 5 speed and a blown engine (2.5L, I'm pretty sure). I would like to swap engines to make a running Legacy with a 2.2L and a 5 speed. What sort of troubles am I likely to run into?

So, replacing the MAF sensor and clearing the error codes solved all the problems. The P0106 has returned, intermittently, but the car runs fine, and the mileage has improved a bit. Thanks for all the help!

Thanks, all. I wish we had a junk yard nearby, but if we do I haven't found it. The local parts store will get a new one down here tomorrow from the big city, and I'll let you all know how it works out. I'll hose it down with the MAF cleaner first; that's a good tip.

I have unplugged and re-plugged the MAF connector. Everything looks bright and shiny inside the connectors. I hadn't known that the P0106 might indicate a vacuum problem, so I'll check over those hoses. The MAF is expensive, but at least it's easy to get at. It's below zero here, so easy to get at is a big deal. Thanks!

Hello, my '96 Legacy has started running rough, dies at idle, loses power. It smooths out a little at higher RPMs, above 2000 or so. The error codes I am reading are: P0100, P0303, P0304, P0106, P0170, P0106 again. I'm thinking that these all point to a MAF sensor problem. Smoothing out at higher RPM suggests that I'm getting decent fuel pressure, and the MAF and fuel trim codes (100 and 170) together might explain the cylinder 3 and 4 miss codes (303 and 304). Am I going wrong here? Is it reasonable to replace the MAF as a starting point, or what would you recommend? I have a code reader but no scan tool.