idosubaru

I’ve swapped 95-99 legacy FWD and AWD TCUs and noticed no difference. Wouldn’t surprise me if 91 is the same TCU failure is not common. There’s 100 others things more likely to fail before a second TCU failure in three Subaru’s.

I would expect a crank or cam sensor or variable valve gear (if equipped) code if there was an issue with timing. you can probably freeze frame the data while starting and someone who knows could tell right away wether the cam and crank are out of synch. But I don’t know how to do that.

Load dependent is classic sign of clogged radiator. Make sure it flows great with no resistance. It should flow all the water supply of a standard well pressurized 45psi home hose. If you have a well or long runs or otherwise weak volume it may not be a good test. It is not the most common scenario, But I’ve seen replaced EJ25s do the same. Pretty common for them to only do it under load for a few months. Get slowly progressively worse - they’ll exhibit overheating at lesser loads over time and then water loss from internal consumption. issues get worse quicker over time. If you’re that confident in your HGs then just focus on the radiator since you already have classic clogged rad signs.

I use OEM, but have used them a couple of time and i would use them again if there was a reason too. there are shops that have installed gobs of them on EJ25s over the decades without issues.

I like data. Being around hundreds of EJ25s over long periods of makes it incredibly obvious as a fact that replacement headgaskets have higher rates of additional failure modes. Discussing why ends up in dreadfully anecdotal opinions which I can’t stand, or long winded nuances of what exactly we are talking about to make any real progress - history of decades old vehicles, prior ownership, repair choices, and then design elements. I find that efficacy of that limited and too filled with anecdotal commentary. But I would start the conversation by qualifying what we are comparing: 1. Newly cast materials that have never been heat cycled being assembled by robotic instruments calibrated on tight schedules in near clean room environments developed by QC driven departments focusing on 6 sigma, ISO 9001 (which in some ways don’t really mean anything sometimes but as an illustration it still stands) And: 2. Decades old heat cycled multi owner unknown history and maintenance materials being assembled with facilities and tools that are unlikely to be as good, accurate, clean, well maintained and calibrated as the former.

Good grief the website is freaking out.

Smoked up

Cracked out

I had a 89 TCU fail a year or two ago. Used ones are everywhere for pocket change. They’re so worthless I wouldn’t offer to sell the extra legacy TCU I have, it’s not worth my time for as little as they’re worth. They fail so rarely that right now they’re easily obtainable. Www.car-part.com

Radiator clogging will over heat under specific conditions like hills, temps, AC, towing. headgaskets will as well - On that hear, particularly if they’ve already been replaced. check or replace radiator check for hydrocarbons in the coolant.

Okay, I understand what you mean now. Test the alternator output, preferably immediately after the transmission issues happen again. There’s a chance the alt is overcharging (and maybe under charging) and I would want to rule out overcharging right away. Then test battery condition and connections. If it was only a poor connection it would most likely cut out while driving, not just at start up. So something else is the culprit or there’s an additional constraint on the power issue (like a TCU start up routine). I’d check the alt output first.

Awesome glad that helped. Now you know to try that if it happens again. Which it probably will. Did it happen more than once or did you drive one time like this? Im not sure what your question there is - if there’s an active TCU code, the transmission related light will flash.

Yes the power light that sounds right for 80s and early 90s. Subaru has given that transmission light a few names over the years - Power, AT, AT Temp or Temp. If there’s a code, itll be the only thing flashing 16 times at start up. If you start the car while looking at the instrument cluster, it’s the only light that will blink 16 times (or really at all - so 16 is quite obvious - if there’s a noted code). it is also the light that will flash, and you will count, if you try to retrieve any codes. good to check to make sure the light is working, in which case you do need to know the name.

Sounds like the TCU codes need to be pulled. I haven't looked in a long time whether these have a "memory" mode or diagnostic mode or not but I would wonder if you can learn something even if it's not tossing an active code, from a memory functionality or pending code like the ECU. No there's no fuse. Sorry if you already know this, I don't know if you do or not. Are you sure the TCU light isn't blinking - it only does it immediately at start up, really fast and is easy to miss. It's not like the check engine light which stays on all the time. If you don't look right away at first start up, you won't see it. It seems unlikely to go into fail safe mode without a code but stranger things have happened. Have you tried turning it off and restarting it multiple times - does it do it every single time? I assume the fluid level is fine? Check the trans pan and make sure it's not dented at all, even a tiny bit. There's almost no clearance between the pan and solenoids, wiring, and sump. Rodent damage is unlikely but I've seen it more than once. Chipmunks, even in squeaky clean residential neighborhoods, can crawl into areas and gnaw on wiring that's unseen. I've seen wires chewed under carpeting that seemed impossible to access. But this is, of course, unlikely.

Sure thing. Hope you find something simple. Custom wiring is so commonly problematic as a car ages that it’s hard not to triple check that.

That’s how the trans operates without any TCU input. Or, yes, it could be a hosed trans. Does the AT light blink 16 times at first start up? Any check engine lights? Id read the TCU codes. If there are none, I’d heck the trans connectors for power or rodent damage.

bolt the 03 intake and all the plumbing to the longblock and swap the ypipe.

There’s a relay for right side lights and one for left side lights. There’s a relay for left side motor and one for right side motor. Same with fuses - they are left and right side specific too. Fuse 7 and 8. no relay for highs. No relay for high beams and unlikely for both relays to fail, but lows work anyway. FSM is online somewhere with a little digging. Section 6 page 60 has the headlight diagram I’ve always owned an XT (or multiples) since high school in 1992. Never kept track, but I’ve had more than 20. They weren’t common back then but they weren’t hard to find with some effort.

Front can struts interchange. Rear is totally different and can not. Many Outback wheels and tires won’t fit under front legacy struts, they hit the lower spring perch. outback front struts on a legacy lifts the car 2” and works fine. Front legacy struts on an outback you need to make sure the wheels tires aren’t too big to clear the lower spring perch of the legacy. If you run legacy wheel and tire size then you’re fine to install legacy front struts on outbacks.

I’m a little uncertain on your details, symptoms and custom install but I’ll take a stab since I’ve owned 20+ XTs and worked on more. when you turn the lights on does the high beam indicator light turn on in the instrument cluster? low beams work fine and normal? in general custom wiring is always suspect. Check the wiring and any added relays. XT light wiring flows through a large connector just above the pedal assembly and in manual transmission vehicles this is prone to short out. It gets hit more due to little clearance between pedal and wiring and in areas that see moisture rain or snow where your shoes are wet when hitting it. Cut it out and splice in new or used connectors. Obviously they’re not available from Subaru so you’re on your own for that. the high beam switch itself can be problematic. You can look up in there and play with the mechanism to get it to work, or even attach a paper clip to it to turn them on and off. But be careful messing with the steering column covers. They will break very easily. check the custom wiring and relays again.

Vacuum hose isn't reconnected or one of the plug isn't plugged in/fully seated Verify all connections are fully seated and proper - O2 sensor from dropping exhaust, trans connector, any engine connectors disconnected. Check for vacuum leaks - intake manifold hoses and the brake booster vacuum hose is particularly easy to forget or not have fully seated.

i would guess radius rod bushings or tie rods - get a long digging bar and get some serious prying going on from multiple angles underneath and make sure you're subjecting this to more than just hand pulling and pushing.

idle air controller/valve. clean or on these models it usually needs replaced. they fail rarely enough that a used one is a reasonable option. but that's only a guess, which isn't ideal - answer lmdew's questions above

People have said they leave the rings on but I’ve never tried it, they’re so easy to remove it takes 10 seconds.

Fronts are interchangeable. Something else is wrong - maybe they’re not original knuckles, or the splines are just rusty dirty and corroded and interfering. That’s not uncommon. All front 1995-2004 axles are physically bolt in interchangeable. 90s Foresters ABS tone ring is on the hub so you can install *ANY* front 95-04 legacy outback or forester axle your want and the bolt right in. The issue is something else I guarantee it. If you’re installing 90s axles into 00-04 knuckles, they still physically bolt in but you need to verify ABS style if you want ABS to work. But you’re not doing this so it doesn’t matter.