Machine shops & STI engine rebuild...

[Dean]

I think we have at least two members who work at machine shops, and other's who have dealt with other shops locally.

I know about Scott's experience with Aims, but not sure what the resolution/root cause was.

Not necessarily looking for a hook up, but not rejecting one either. Just looking for a shop that I can trust with reasonable prices.

What I have:

'04 STI 100K long block mostly stripped down. Valves still in heads, rods still on crank at the moment.

The motor ran fine but ate/spewed oil under load when removed. My untrained eyes show only worn rings and ring grooves.

At minimum, I am looking to have the parts inspected by someone who knows something to determine next steps. Parts to be reused will need a good cleaning.

A cylinder bore/hone and new pistons/rings may be all it needs. Would likely want it balanced and maybe blueprinted and short block assembled.

Forged or cast pistons? CP. JE, other? Top/skirt coatings? Gapless ring(s)?

Depending on budget, might be interested in larger valves and upgraded springs/retainers, porting and polishing. A valve job requires valve lash/bucket adjust/swaps, right?

Or, just replace heads with Cosworth. Anybody know what the difference between Euro/JDM and US Cosworth Cams is? I have yet to find technical description of the differences.

I can't find any compelling reason to swap rods/crank unless they are out of tolerance. Anybody think of a reason? New bushings/bearings/seals make sense though I assume.

Also, what is the preferred oil pump these days? There appear to be a bunch of choices. Any other water pump options?

All timing idlers and belt obviously get replaced since they have 100K on them.

What have I missed, and who to use?

[sperry]

Alright, I skipped the ignore and read this thread since it looked like a good topic for a tech thread that I could contribute to.

First, has anyone at C&C talked to you Dean? Dave was looking for an STi long block to prototype with... I dropped off my shortblock over there, but it's probably not that useful to them. I told Dave that you had a long block that you might be willing to sell them... so that may change your plans for your spare motor.

Either way, the issue with the motor AIMS built for me had nothing to do with AIMS. I reused my oil cooler (the small oil/water exchanger the oil filter attaches to) when I reassembled my motor. Apparently, no amount of cleaning will get all the bits of bearing out of that thing... so even though I cleaned the snot out of it, I ended up with dirty oil and that's what demolished my motor during break-in on the dyno. It was a nice $2000 mistake. So, having had an otherwise good experience with AIMS, I would put them up there on the list of machine shops to go to for an STi motor build. You may also want to talk to C&C and find out who does their Porsche motors, as I get the feeling that a good Porsche shop would probably be a great currently untapped resource for Subaru motor builds, though they're probably expensive.

As far as building your motor: you basically get one (.020 over IIRC) bore out of the stock sleeves, which means buying some 100mm pistons. (Doing it twice means re-sleeving the block... which is why I ended up getting a new shortblock after my last one died because I didn't want to spend the money on potentially needing to sleeve my block and repair my crank.) I went with forged JE pistons and rings, ceramic coated, with stock base compression. That's what Mike at GST recommended. He might recommend going to a higher base compression these days if you're going to be doing the rest of the supporting mods like injectors, turbo, oil cooler, etc. I think that's where they went with the L at least. Higher base compression and a little less boost nets the same power with a more responsive throttle and earlier power.

I don't believe the valvetrain needs a ton of work unless you're going to go after a lofty redline. Even the Cossie heads use the standard style bucket lifters IIRC. They just up-rate the springs and oversize the valves, plus the port/polish. You can even run the stock STi cams in the Cossie heads and make 40hp over the stock heads. I'm not sure what the differences are between Cosworth's cam offerings, but I do know they're ridiculously expensive, and probably won't gain you much unless you're planning a GT53R or larger turbo, and a 9000 rpm rev limit.

Stock rods and crank are probably fine. People tend to break pistons, not rods or cranks. You'll just want to get them blueprinted and balanced, and have all the bearings replaced during the rebuild. I'm no expert on sizing, but apparently you'll want the motor to be a little looser than stock if you're going to run an upsized oil pump. This is where you'll probably want to trust your engine builder.

For oil pumps, I went with a JDM 12mm pump. Get it from JapanParts.com... there's a thread on NASIOC w/ the part numbers, and one of the members there works for JapanParts and was at one point doing a NASIOC member discount. The only issue is running the 12mm pump with stock bearing clearances... I'm not sure if the higher flow 12mm pump will end up just bypassing the oil because the crank doesn't flow enough. Either way, I'm not running a big redline, so I'm not too concerned. But if you're planning on going north of 7500 rpms, you're going to need to do some research about the larger oil pump, shimming it, bearing clearances, etc. Again, this is where your engine builder should be on the ball.

I believe the stock water pump is probably just fine. Though you'll probably want a bigger radiator and oil cooler.

Finally, you'll need to come up with something for the oil breathers... the GST tank I've got is nice, but messy. Crawford makes a decent, and smaller air/oil separator that breathes back to the intake. Plus you'll have to deal w/ the PCV and emissions crap (or just remove it like I did). If you're not planning on going too far past your current power levels, you may be able to get by with just catch cans... as the oil breathers will require you to change oil more often as condensation and exhaust pollutants don't always get removed from the oil and can drain back into the crankcase.

[MikeK]

The last I remember reading about cosworth cams (and this was years ago) is that they make a lot more power at high rpms, at the cost of spoolup, so basically move everything to the right on the power graph.

I had planned that if I ever went with cosworth heads I would stick to the stock STi cams.

What happened to the 07 engine you installed? Is it dead already?

[Joeyy]

DB Tuned!..

[sperry]

Originally Posted by Joeyy

DB Tuned!..

I'm pretty sure Paul doesn't have a mill/lathe/etc for doing engine machine work. He has to send the motor out to someone else just like the rest of us.

[Dean]

Thanks Scott, I really appreciate your input.

Nobody from C&C has contacted me. They are probably afraid I'd ask how the Mercury is doing...

On an oil pump note, what are the two or three plugs on the bottom and side of the thing? They look like they were put there for an oil cooler or some other attachment... very strange...

Sounds like I will stick with the stock cams unless I find some more details on the Cosworth. Anything big $ is dependent on job status...

I may have found a super deal on pistons and machine work. My Dad's neighbor in Gardnerville owns United Engine and Machine Company makers of KB pistons and they are very interested in getting into the performance import market. I'll talk to them on Monday but from what Dad talked to him, it sounds like they want to put hands on and see failures and build some pistons for me. (crosses fingers)

On my '07 engine.... OK, I'll fess up now... Something happened to one cylinder last Friday at partial throttle between turns 6 and 8 at Thunderhill. And yes, I have pretty good video front and rear, but it will have to wait.

My guess right now is that I had catastrophic ring and/or ring land failure on 1 cylinder. It pressurized the block, blew the hose off the crossover breather on the left side of the intercooler and then sprayed oil all over the engine compartment onto the windshield, etc.. and of course started billowing smoke. It still ran and I limped it back to to the pits. It still runs, mostly and is smoking less, but is very not happy.

I just got it back to Reno yesterday right before the meet, so have not had a chance to pull plugs or anything yet.

I had some initial overboost but slowed and pitted immediately and bumped all the wastegate duty cycles. I'll have to look at the video, but I think it was fine for a number of laps the next session before it blew. I was watching gauges carefully. Good thing I know this track very well.

I might have to blame this on tuning at altitude and driving close to sea level, but for the most part it was the same tune I had been running on the old engine and beating the crap out of at multiple "seaer" level tracks. And Matt and I beat the crap out of the new engine for the last 2 solo days. TGV deletes is the only real change I made between engines and I tweaked the intake maps for those after quite a bit of logging both open and closed loop. I had also backed some timing out in a few areas where it would occasionally knock under extreme load at low RPM and pull down the advance multiplier. It is largely a glorified stage 2 map damn it. Grrrrr....

At least the piston guys will have a long term wear and a likely catastrophic failure parts to put hands on... And yes, this is why my Nephew and I tore down the old long block earlier this week.

Unless I get a job, this car may be down for quite some time as I can't justify any more expense on it.

What is kind of interesting is that Scott has had mostly oiling and bearing issues, and I appear to have ring issues...

[sperry]

Apparently, those ports on the oil pump will not work for an oil cooler. I thought they'd make for a nice simple and clean cooler install, but apparently that's not what they're for. What they're actually for, I've got no clue.

A broken ring-land is pretty much the "normal" failure for these motors. That's what killed the original EJ257 I got on the dyno over at S-Squared... but that was due to running lean 'cause of a lack of fuel pressure. And usually the piston goes because of detonation. My issues with spun bearings was mostly because I would let the oil temps go sky high... some of that was poor oil breather/cooler setup, but a lot of that was because I'd ignore the oil temp gauge and run "just one more lap" and put retard wear on my bearings eventually leading to their failure.

Was there any chance of a FP drop due to fuel slosh in the tank? Or perhaps a burst of oil getting into the intake via the breathers? Anything to limit the amount of fuel and/or lower the octane significantly and quickly can cause detonation faster than the knock sensor can correct for, and will pop a piston. And of course, overboost will do a number on the motor... if you were having WG issues, that would be a likely guess. Even if you fixed the WGDC, it may not have been enough for the partial throttle condition that the car would only see like once per lap.

Also, have you ever run at sea-level at colder winter temps? Any idea what your AFR's and EGT's were?

[Dean]

Ran the old motor on very similar tune at sea level all the time. Tank was full. Oil in the breathers is a possibility but never was a problem on the old motor. Don't know what AFRs were that second. AFRs at WOT had been good, low 11s, high 10s. EGT close to or pegging the meter at 1600 which is the way the old engine ran and typical as I understand when running that rich as fuel is still burning as the exhaust valve opens.

Mid map WG cycles were probably OK as that part of the map is the oversafe Cobb stage 2 and the logging I had done had that portion of the range was plus or minus a very small percent of targets, but that was at altitude. The Duty cycle issue was on the high end where they were not stage 2 data.

I need to plug in and see what the multiplier is/was when it died. And nothing says the damage happened mid throttle, that is just where it failed. At minimum, this means more sea level tuning before track days. And damn, I need a way to mount my laptop to log on-track.

Are you running ball/spring in addition to the solenoid? I wonder if that would be a good safety valve? I have one sitting in the other room.

I need a catch can setup for the next iteration and definitely hose clamps on breather hoses. The oil mess was real bad and just an insult to go on top of the injury. Cleaned much up there, but I am sure I will find more after I pull this motor and there were a few drips in the trailer. it was everywhere on top and down the passenger side of the compartment.

[ScottyS]

Wow, and Oh NOES!

[cody]

Sucks.

Alright, grab your popcorn. You guys who keep popping motors need to consider having Ed tune your cars. He's got tons of experience tuning cars at this elevation and at sea level.

*runs*

[Dean]

Originally Posted by cody

You guys who keep popping motors need to consider having Ed tune your cars. He's got tons of experience tuning cars at this elevation and at sea level.

I should have done more logging and testing at that altitude, just as Ed would have. And I have no reason to reopen the road vs. dyno discussion or discuss hearsay about him. I am glad you are happy with what he has done for you and will leave it at that.

Since you chimed in, I'll ask the same questions i asked Scott. Do you have both a MBC and the standard solenoid or just the MBC? How are they plumbed? Is the MBC just for full boost/safety valve and the solenoid handles the midrange? Which MBC?

[100_Percent_Juice]

The MBC is used in place of the factory BCS.

[knucklesplitter]

This prolly has nothing to do with the failure, but...

There is an altitude compensation table in the ECU file. If the table numbers are stock or stock-like the ECU will automatically add boost as you descend in altitude (and vise-versa). If you were running high boost up here then you may have gotten too much down there. Always best to set the overboost fuel-cut parameter below failure-inducing boost levels.

Originally Posted by Dean

I had some initial overboost but slowed and pitted immediately and bumped all the wastegate duty cycles.

Did you bump them up or down?

[Dean]

Originally Posted by knucklesplitter

There is an altitude compensation table in the ECU file. If the table numbers are stock or stock-like the ECU will automatically add boost as you descend in altitude (and vise-versa). If you were running high boost up here then you may have gotten too much down there. Always best to set the overboost fuel-cut parameter below failure-inducing boost levels.


Did you bump them up or down?

I am aware of the altitude table. Very stock as I recall.

And yes to the bump direction. Yes, I know less duty cycle equals lower boost, or technically, I guess, makes it more likely to catch the rise earlier or less likely to overshoot and hunt. I actually just dropped the stage 2 low and high tables back in which were both lower and a smidge higher respectively as I recall and went for a couple test pulls on the highway and things were much happier.

Again, I am not blaming anyone or anything other than myself at this point, except maybe Subaru for a crappy piston design. I should have spent a couple hours logging and tweaking Thursday night and did not. I should have dropped by DB Tuned and had Paul run a few pulls on the Dyno and make some tweaks or taken it down to Mike at GST after I dropped the new TGVless engine in months ago and did not. Needless to say, I will not repeat that mistake. No tuning just at Reno altitude.

[cody]

Originally Posted by Dean

I should have done more logging and testing at that altitude, just as Ed would have. And I have no reason to reopen the road vs. dyno discussion or discuss hearsay about him. I am glad you are happy with what he has done for you and will leave it at that.

Since you chimed in, I'll ask the same questions i asked Scott. Do you have both a MBC and the standard solenoid or just the MBC? How are they plumbed? Is the MBC just for full boost/safety valve and the solenoid handles the midrange? Which MBC?

Ed has a dyno at his new shop in Fairfield in case you weren't aware. Obviously he doesn't bring it when he comes up here though.

I have removed the stock solenoid and replaced it with the best MBC in existence, the Hallman Pro RX. I believe it was $80 shipped from ebay after I haggled with the seller, new.

[cody]

Originally Posted by 100_Percent_Juice

The MBC is used in place of the factory BCS.

Ed used to talk about an optimal device that would combine a stepper motor and a ball and spring so that the ECU could control the "MBC" for the best of both worlds.

[AtomicLabMonkey]

That sounds as useful as building a computer controlled robot arm to press buttons on a control panel. It should just all be done in software instead.

[cody]

More like a computer controlled robot arm to maintain perfect target boost in every gear while the ball and spring offer better boost response, and believe me, 2.0L Subarus at this elevation have awful boost response in stock form.

[100_Percent_Juice]

Tell me about it.

[sperry]

Originally Posted by cody

More like a computer controlled robot arm to maintain perfect target boost in every gear while the ball and spring offer better boost response, and believe me, 2.0L Subarus at this elevation have awful boost response in stock form.

A good dual solenoid based EBC with proper programming would be more responsive than an MBC and totally programmable for target boosts in every gear. You're just unhappy with the stock BCS since it's designed with safe failure in mind (i.e. if the solenoid fails, the car runs pure wastegate boost).

Go drop $700 on a Greddy Profec or something similar, then drop another few hundred getting the boost controller tuned and the ECU maps retuned to work around the external boost controller. Or deal with missing your target boost by ~2psi and ~200rpm with the stock solenoid that came free with the car. Or slap the MBC on there for under $100 and hit your target boost all day in 1 gear, and get close in the rest.

There are plenty of options without going retarded developing a stepper driven ball and spring. If you're going to go that route, just bolt a linear actuator to the wastegate door and open and close it directly. Just hope it never fails, 'cause you'll end up with a permanently closed WG! *blammo*

/hijack

[cody]

Fortunately the ECU can be configured for fuel cut at a certain PSI. My car's set for 18 or 19PSI.

[Dean]

First engine diagnosis...

While I initially did not see anything other than blow by looking symptoms on the pistons of the 2004 block, while the piston guys were having an initial look at them a large chunk of ring land 2 on one pistons fell out. Thus it was more than just ring wear on a 100K mile engine.

After talking to them for a while, my understanding is that ring lands really only fail under detonation.

That says my tune likely killed both engines. I guess it is time to dial back the boost, use 100 octane and/or change to forged pistons the last of which was the plan anyway.

I hope to get the '07 into the garage and compression tested today and maybe get it pulled. Any bets on which cylinder is down? I think people say #4 is the problem child, but I still don't understand that as it is closest to the fuel input... Guess I need to look at the rail layout again.

I also determined that some idiot reset the ECU after the failure before pulling codes and checking advance multiplier, so all that data is gone and no misfire hint as to which cylinder went.

And did I mention that the '04 cylinder sleeves appear to be made of pretty tough stuff. 85K stock and 15K boosted and you can still see some of the machine/honing marks in the piston area of the sleeves.

[Kevin M]

Originally Posted by Dean

I hope to get the '07 into the garage and compression tested today and maybe get it pulled. Any bets on which cylinder is down? I think people say #4 is the problem child, but I still don't understand that as it is closest to the fuel input... Guess I need to look at the rail layout again.

My understanding is that it gets too much air, not too little fuel.

[ScottyS]

Dean, if a lean tune was the issue, is there anyway to catch this instrumentally early in the process? Even though you have a wideband sensor, was that not enough? Or, is it just a matter of tuning and building on the conservative side? I'm applying the lesson to my own scenario, where I have no real tuning control at the moment and am limited on the fuel side by injector capacity. I.E. how do you detect detonation conditions reliably?

[Dean]

My guess is detonation due to overboost, not necessarily leaned out, though I will likely get the injectors tested or upgraded before they go back in.

More logging/tuning at lower altitude would have likely avoided/prevented it. You have to be able to collect data, so boost, and wide band are a minimum and a way to log them. So if your ECU doesn't have these sensors and a way to log them, you need smart external ones like the LC-1 and Dephi gauges.

Misfire (It did rethrow a P0302) = Cylinder #2 Oil all over the plug. Compression apparently insufficient to even compress the valve on the gauge.

What is mildly interesting is that this is the same cylinder the TIC guys had problems with the end of last year that I was reading about last night.

I thought it was supposed to be Cylinder #4, Driver side rear, that fails...