EDIT: In reference to a 2000 Legacy L, 2.5 SOHC, 164k miles, 3k miles since head work. So, I just did head gaskets in November, got the heads milled at DG and used STi gaskets, both per recommendation from some of our friends at Morrie's Mtka. EDIT: During the job, I busted the front O2 sensor, which I just replaced this past weekend. So it was running in default all winter. So in diagnosing a misfire this weekend, I did a compression test on all four cylinders for comparison, and got some surprising results... According to Chilton's, you're supposed to let the engine do at least 7 compression strokes to get an accurate reading on a gauge. It also said the high end of spec compression was 185psi. In following those directions, I got 210psi on every cylinder. Dead nuts on with each other too. So my questions are: Did I just plain old do it wrong? A rough calculation suggests that equals 11.2:1, assuming 185psi = the OEM specified 10:1. This sound right? Would this require higher octane than 87?
Short of adding oil to the cylinders it's hard to make compression readings artificially high. Mistakes in method tend to lower them. Depending on how much you had milled from the head and how thick the gasket is you may have raised the compression some. I just searched and found the 2.2 is 9.7 and the 2.5 is 10.0. I'm pretty sure that's towards the upper end of what should get 87 without a knock sensor. Try throwing some premium in and see if that solves the problem. My guess is you didn't raise it enough to need more than 89, but in diagnosing it you probably want to try high test.
AFAIK, there is no good way to correlate dynamic compression pressures to static compression ratio - usually the only good way to measure the ratio is to CC the heads and block while they're apart. If a fresh rebuild showed me 210psi and was even across all cylinders, I would be smiling - nothing less. Do a few test tanks with 87 and 91 - if the 91 seems a lot more powerful, then you may have some knocking with the 87. Don't worry, if the knock sensor picks up excessive knocking, it will retard the ignition timing to prevent it. This just weakens the power output/efficiency slightly to prevent the knocking from wrecking the bearings in the block. Not sure if you can read actual knock values in real time with your ECU, but it may be worth a few dyno pulls to remove all doubt about whether 87 is kosher (without having the ECU pull too much timing that is). Maybe Ronnie or those more well versed in Subaru engine management can chime in.
Oh, sorry, 2.5 SOHC. It's a 2000 Legacy L, so I think it's an EJ251 but I'm not sure. Well, I don't know how much they milled from the heads, but I do know the STi gaskets are 3 layer metal but still thinner than the original paper-metal-paper ones. I'm sure there's no way to find out how much was milled off without pulling them off and measuring, which is completely out of the question for satisfying my curiousity. Well, a certain StatGSR from here is a buddy of mine from UMD, and he also told me that my math isn't quite right to extract an accurate compression ratio from that. Regardless, I'm quite happy with the compression I'm getting, especially the consistency. Not too shabby for a 164k engine on the original internals, eh? As far as an actual problem, there isn't one yet. I just got the thing running correctly for the first time since the head work, being that my HO2S was dead all winter. I'm currently just setting in to my first tank of 87 on the setup, and it's running great, so there's no real problem to speak of. What does concern me is whether the management system will take action against knocking without throwing a light, or whether the increased compression could prove to be damaging to the rings. Something tells me the rings will be fine though, as long as I'm diligent in timely maintenance. If it's not running at its peak, and the fix is as simple as "run 89 octane instead of 87", I'll gladly start buying 89.
Those compression test results are not abnormal for a new or freshly rebuilt engine - absolutely nothing to worry about, and it won't be hard on any part of the engine. The way most knock sensing ignition systems work: If the knock sensor detects engine knocking over a certain threshold, the computer will retard the ignition timing slightly or progressively to a certain point until the knocking goes away. After a set length of time it will advance the ignition timing back to normal, and the process starts over again. It does all of this without the driver ever knowing. The only thing a driver could pick up on is very slightly reduced power - most of the time imperceptible in an average car. If the knocking is very excessive or the knock sensor is reporting erroneous values or is open/short circuited, the Check Engine Light will come on alerting you that the problem is more than just a slight temporary mixture/fuel issue. If you run 87, the engine may suffer from knocking due to pre-ignition, but the computer will adjust the timing accordingly. Knocking every once in a while is not that harmful to an engine - only extreme knocking and/or knocking for long periods of time will hurt it. In most cars, the only side effects of running 87 is reduced power and reduced fuel economy (due to reduced efficiency).
Directly answers one of my questions. Thanks! This is the other thing I'd like to avoid. Just wondering if the car would throw a CEL if performance is not optimal, or only if the engine is in danger. I think I'm gonna pester a buddy of mine to see if we can read indepth data from the computer, look for knock counts...
