presslab

So have you measured any temperatures when the gauge is close to the red mark, with the 270 ohm divider resistor? How hot is hot? Based on my calculations an EJ sensor will be 1/(1/19 + 1/270) = 17.75 ohms at 120C. The EA sensor is 16.1 ohms at 120C. So that means the EA gauge will read lower than it should when overheating, with the 270 divider resistor. No idea, sorry. As far as I know all the EAs had the same thermistor, and the same goes for the EJs. The high speed rad fan comes on at 100C. If the gauge goes to 105C that means that the coolant system is not sufficient, and unless something is done soon the engine will be damaged. Yeah it might be a bit conservative, but I was just using an IR gun on the upper rad hose so I'm not 100% on the 105C. Maybe I should hook up my SSM cable. Like I said we'll see how the 330 ohm cluster resistor does... There are two pips that are red on the digidash, so I can still see if the temperature climbs further. 105C would just barely turn on the lower pip. It would be very similar I imagine. The resistor value will likely be a bit different. It's not hard to find though, just trace the wiring on the PCB until you find a resistor of similar value. On mine there was a 10k resistor, the 160 ohm resistor in question, and a small capacitor on the circuit that connected to the sender.

Quick update. I replaced the downpipe with an aftermarket WRX one from eBay. 3" stainless, and I did a few cut and welds to get some more clearance. I also put on a high-flow 3" cat; although my car would never pass CA smog due to the JDM motor (and doesn't need to) I am considerate of the environment. This solved my high RPM power loss that only happened when the exhaust was cold. An odd problem; after noticing that with a warm engine (cruising slow) it would occur, yet when I was in the gas for a minute it went away, I figured it was not the coolant temperature but something else warming up. The stock WRX cats (my old downpipe) when hot would apparently flow much better than when cold. Anyway now with 3" from turbo-back I don't think the exhaust will be a problem. :-p I might even bump up the shift points as it's very strong at high RPM. Over the summer it didn't overheat at all while driving, but I did notice if it was close to 38C (100F) outside the coolant reservoir would overflow. This would only happen a few minutes after I'd parked due to the radiator heat soaking. The WRX has a bigger reservoir, but rather than go that way I put in a higher temp thermoswitch (105C) that is enabled even when the key is off. So I still have the ECU controlling the fans, but if the thermoswitch kicks in it will turn the fans on high as well. This way when it would heat soak it would cool the radiator down before it would push coolant out the reservoir. Finally I got around to correcting the coolant temperature gauge. It read a bit low with the EJ conversion, and most importantly it wouldn't indicate overheating when I tested for this. There is always the common fix with the resistor in parallel with the thermistor. But this creates a voltage divider and although the gauge might read correctly when at cruising, it won't be a good indicator of an overheating condition. Modifying the current source is the correct fix. I traced out the circuit and found the resistor in the digidash instrument cluster that provided the current source to the thermistor. It was 160 ohms, and I settled on a 330 ohm replacement. This puts the gauge in middle at 80C (cruising) and when it lights the red mark it's 105C. I read the newer factory Legacys tripped the gauge into the red at 120C. This seems much too hot! However maybe a 270 ohm resistor would be ideal here. We'll see how it goes after I get some time on it.

http://en.wikipedia.org/wiki/Compression_ratio#Fault_finding_and_diagnosis

You could certainly dyno tune with the stock ECU, it's just more work than one you can change on the fly. Although a romulator would solve that problem. That's probably how Subaru did it.

There's "open source" hacking with the stock turbo ECU, so that's a plus. http://www.romraider.com/forum/viewforum.php?f=15 The EZ30D weights about 75 lbs more than the EJ257. Not a lot more, but the 914 is pretty flexy. Yep intercooler could be tricky; it'd recommend an air-to-water setup, that's pretty straightforward. Here's has some stuff for the 5MT, like the spool, rear cap, and adapter flanges. http://www.subarugears.com

You could hack the stock ECU. http://www.ultimatesubaru.org/forum/showthread.php?t=134517

Could be a bad ground. Put a voltmeter from the intake manifold to the car body. What are the voltages with the blower running and without?

Rotella T6 is the only thing I buy at Walmart. I use this oil in my air cooled motorcycle, turbo EJ20G and EJ25.

Take the lids off, and see if they have an EPROM socket.

If you're not going to fix the root of the problem and just want to raise the idle speed, you can just turn the large hex on the bottom of the idle valve to let more air through. At that point you'll be fighting the ECU, because it will be trying to lower the idle speed by closing the solenoid valve. So it doesn't really matter what kind of IAC solenoid you use. Without tuning, my motor ran super lean at idle with the cams. I doubt you'll pass smog like this.

I have pretty much your same setup on my Vanagon. Here's my thread: http://www.ultimatesubaru.org/forum/showthread.php?t=134517 The IAC is not the problem, the mixture and timing are.

The EJ22 is a decently powerful motor in a Legacy. If you're going "semi speed" it sounds like you have something wrong. Have you timed your 0-60? Are you getting decent MPG?

Sounds like you need more caster. The EA81 (and to some extent the EA82) seem to have alignment specs targeted for manual steering. This means little caster and positive camber. But when fitted with power steering it's too much. The spool valve in the pump has a lot of hysteresis - it allows full pump pressure until you exceed the valve pressure, and then the pressure is reduced significantly. It keeps from squealing the belt and blowing hoses. If you were to reduce the valve's operating pressure, it would not feel very good. Basically your power "assist" would briefly cut out and be almost like manual steering. This is how my steering felt on quick turns before my valve mod. The ideal way to reduce the actual assist is to change the stiffness of the torsion bar inside the steering rack control valve. That's obviously difficult. An easier idea is to introduce a small orifice on the outlet of the pump. There should be an orifice in there somewhere, maybe you can weld it closed and drill it smaller.

So now that I have an EJ20G conversion, I decided to increase the pressure on that engine too. The spool valve is a bit different than the EJ22; instead of being threaded together it's pressed together. There is a little screen on one side held with a wire clip, and then a plug pressed in to hold the spring and ball valve. The stock depth of this plug is 5.75mm. A couple taps with a punch and now it sits at 7.25mm deep. Easy as that, no shims needed. I noticed the added response this morning just pulling into my parking space.

Maybe apology is the wrong word. Who is he apologizing to, the owner of the telephone pole he hit? Why should he need to apologize to us?

You can see my thread here. I did all the work myself. http://www.ultimatesubaru.org/forum/showthread.php?t=132060

I ran my EA82T with a TD-04L at 14+ PSI for about 20k miles. I also added more fuel with larger injectors, modified MAF, and TMIC. It had a lot more power, and ran pretty good. Alas though, it was tired at 160k miles. Leaked some oil, a bit of blue smoke on startup too. I wanted more power and better mileage. So I did the EJ20G swap for around $2500 total and I'm very happy. But I think a mildly modded EA82T with under 200 HP will be okay, just give it plenty of fuel. I'd say an EJ engine would be more cost effective in the long run.

I haven't noticed so much of a problem in the front; there's a fair amount of steel between the struts where the spare tire mount is. However the factory rally prep manual for the RX shows strut bar B0110GA054, so I'm sure it will only improve things. In the rear of my wagon however, the body twists so much that my rear gate rubs the body by the tail lights and it keeps rubbing the paint off. I think some kind of triangulated brace between the rear struts would be beneficial.

There is a small spool valve in the pump that regulates the pressure. Unlikely but it could be getting stuck. It's located just behind the high pressure outlet fitting. I modified mine to bump up the pressure as when I would get sideways the power steering couldn't keep up with my steering correction.

Yes, I agree. If your system is designed for R134a, use that. If designed for R12, use the ES-12a. The reason EnviroSafe calls it R134a "replacement" is to skirt the laws regarding R12 retrofitting.

Exactly. Not sure how many got my reference to tacnodes... Could mean making out as well. :-p http://www.thefreedictionary.com/osculating

There is a remote possibility the regular fuel caused it. If it were pulling timing the transmission would think the engine were making more power and shift firmer than it should. My 4EAT does this when I have the A/C on. In any case you should always run premium fuel. These engines do fail occasionally at the ring lands, and it's said that using regular fuel and the subsequent detonation is to blame.

Yes, I use the "Industrial ES-12a" in both my vehicles that originally had R12. It works great. Nothing to complain about, been running it a few years now. Indeed it is a mix of propane and butane.

My girlfriend has an '06 Tribeca. The 1-2 shift was hard, especially when cold, and flashing the TCU update did help a bit. The 2-1 shift is also tight, but not uncomfortable. What kind of mileage are you getting? Hers is not much better than 15 MPG around town.

Yes, looks great. It won't look as cluttered with the air cleaner on. Maybe you can wrap some of the hoses (like bundle the fuel hoses) with some sleeving like this: http://www.mcmaster.com/9344K6