Topic: Low compression, suggestions?

[BoMbS]

2001 EJ207 has been running good and pulling strong, but has had poor vacuum at idle. Usually between 11-15 inHg. I can sorta feel a roughness in the idle, but barely. 

I searched up and down for vacuum / boost leaks and found nothing.  Next test was compression test.  Both cylinders on the right side were 110 psi, rear left cylinder was 30 psi, and front left was 100 psi.

Obviously, I am shocked since it’s still running good other than at idle.  Suggestions on what the cause could be, or another test so I can confirm the issue? 

Thanks in advance.

[VTAKLOL]

Cylinder leakage test should be your next step.

[BoMbS]

Okay, forgive me, I’m new to this process, but I went and bought a cylinder leak tester, read the instructions and watched half a YouTube video, so I'm basically an expert.

Side note, also bought a new digital compression tester to verify if my million year old one was lying or not.  Same low compression unfortunately. 



Then I disconnected an intake manifold vacuum line, removed the coolant cap and the oil dip stick, tested each cylinder.  The first task was finding top dead center of each cylinder, which I found difficult.  Ultimately, I accomplished it by hooking up the leak tester and applying about 10 psi to the cylinder under test, and then slowly rotating the crankshaft.   I could hear the air escaping through the intake manifold vacuum line as I cranked it past the intake stroke, so I knew the next stroke is the compression stroke, and TDC. 

It was obvious on a good cylinder when I’d hit TDC, because the leak rate needle on the tester would move.  Then I’d have to hold the crankshaft still while I turned the pressure up to 100 psi.

Cylinders 1, 2, and 3 all showed pretty much 0% leak rate as seen in the first picture below, cylinder 4 had 90% leak rate.



When testing cylinder 4, the leak rate gauge needle never moved at only 10 psi, so I was worried I wasn’t testing it properly at TDC.  But after rotating the crank so many times I was sure I had it.  I cranked up the pressure to 100 psi and then listened.

Nothing from the intake manifold.  Nothing from the coolant system.  Nothing from the oil dipstick (well a little bit, not much).  Then went around to the back and the exhaust seemed to be leaking a lot of air.



Does this confirm a leaking exhaust valve?  What else could it be?  What else can I test?  Why does it still run so good?

Thanks for any insight.

[RockThePylon]

Could pull the cover and check valve clearances.

But it could very easily be a cracked/burnt/bitchy exhaust valve.

When CanehdianJ's exhaust valve took a dump, it kinda ran okay.
http://www.westernsubaruclub.com/smf/index.php?topic=9738.775

[Dr Beans]

Seems like you did everything correctly. BTW, 100 PSI across the board is on the low side.

Your Subaru is being a Subaru. Time to take the engine out.

[10secdream]

Are you sure you were at the right part part of the stroke for TDC? 90% leakage sounds like the valves are open.

[BoMbS]

Are you sure you were at the right part part of the stroke for TDC? 90% leakage sounds like the valves are open.

This. 

I'm still sort of worried that I missed TDC, but with spending 5+ hours on this yesterday, I must have found it properly on at least one of my tests...  Also, since it's only making 29 psi of compression, I figured that it must have an actual serious problem.

As I mentioned before, the car feels like it's running great...  Is there any possibility that the cylinder is still firing?  I figure I'd hear some backfiring if I had a hole in my exhaust valve?

[10secdream]

How do the spark plugs look?

[Mason]

It's safe to say that I would not even start the vehicle unless I was running dignostics. There was something wrong which lead you to doing the compression test in the first place and the leakage test verfied that the compression test wasn't a null.

As for where to continue, I would be tearing it down

Someone correct me here, but your head gasket could be leaking into the other cylinder? Which the leakage test results were never given. That's just one theory. 10sec had a part of his piston (?) And still drove autox on it

[10secdream]

It's safe to say that I would not even start the vehicle unless I was running dignostics. There was something wrong which lead you to doing the compression test in the first place and the leakage test verfied that the compression test wasn't a null.

As for where to continue, I would be tearing it down

Someone correct me here, but your head gasket could be leaking into the other cylinder? Which the leakage test results were never given. That's just one theory. 10sec had a part of his piston (?) And still drove autox on it

Correct one piston is missing a piece near the top but the rings are still there somehow. My results were a bit different though as it has low compression but the leakdown results were perfect.

Still runs to this day with with a tired engine and very low compression on 1 cylinder. Has the odd misfire on the highway but it still oddly runs. I wouldn't have any worries putting another track day it on.

[Dr Beans]

90% seems high, but I think it makes sense in relation to a CT of 29PSI.

It may be a timing issue (cam gear could have jumped some teeth on the timing belt) so you think the cam is on the EXH Close but its not, or a billion other reasons mentioned already.

If you don't want to pull the engine just yet, take off the exhaust manifold and double check that way. You will be able to see the valves close. Once they do, up the pressure and see if air still passes them.

[BoMbS]

Thanks for the help everyone, much appreciated.  Sorry, long post ahead.

First things first, some answers:

It's safe to say that I would not even start the vehicle unless I was running diagnostics.

You are right, my car has been parked for repairs.  Sadly.

How do the spark plugs look?

The plugs all looked fine, cylinder 3 looked a little whiter and oilier than the rest, but nothing telling on cylinder 4 where the low compression is. 

Someone correct me here, but your head gasket could be leaking into the other cylinder?

Yesterday while cylinder 4 it was leaking at 90%, I could feel some air coming out of the spark plug hole of cylinder 2, so I thought the same thing, leaking head gasket.  But the compression in cylinder 2 was 100 psi, and it's leak rate was near perfect, so that didn't make sense.  I later figured out (at least I think this is right), that when cylinder 4 is at TDC, cylinder 2 is in the exhaust stroke, so if my cylinder 4 exhaust valve was leaking, it would leak into the exhaust manifold and into cylinder 2 from the open exhaust valves... 

It may be a timing issue (cam gear could have jumped some teeth on the timing belt) so you think the cam is on the EXH Close but its not.

First I removed the timing cover and checked the timing, looks good.  Bottom of the picture, but it was correct. 

If you don't want to pull the engine just yet, take off the exhaust manifold and double check that way.

Thanks, this was a great idea.  When I first removed the exhaust manifold I saw that cylinder 4 was black and cylinder 2 was white.



With a fancy little camera I was able to look up the exhaust ports, and both valves in cylinder 2 were white-ish, with a tinge of orange, but mostly white (not photographed).
When I looked up cylinder 4's exhaust port, one exhaust valve was white and one was black. 



Upon closer inspection of the blackened valve, I saw the top of the valve is very reddish-brown.  I then took a similar photo of the other exhaust valve in cylinder 4 just for comparison as it is white throughout.






I didn't redo the leak test with the exhaust off, does anyone think I still need too?  With the black soot coming out of cylinder 4, I'm pretty sure this valve is my problem. 

But the thing is, I don't see any physical damage to the top of the valve, and I viewed the top of the piston through the spark plug hole, and there are no marks on it, everything looks smooth and tidy.

Does poor compression / leaking exhaust valve in one cylinder create a low vacuum scenario in the engine?  I just want to be sure this is the original cause of the low vacuum.   I wonder if this problem is old, I can't remember my car ever having strong vacuum since I installed the boost gauge.     


Next step?  Have the heads rebuilt?  I could remove and install them myself...      Would that be cheaper than buying a new (used) EJ207?

But it could very easily be a cracked/burnt/bitchy exhaust valve.

SIGH.

[Dr Beans]

I'd redo the leak down test now that the manifold is off just to verify that the exhaust valve is the issue (based off the color, itshay got really hot) and it's most likely not seating correctly on the valve seat. As for the vacuum issue, it can. There's less energy pushing the piston down on the compression stroke which means less air being sucked in on the intake stroke.... but I think the vacuum issue is something else. These cars have a million vacuum lines.

However, I do have a suspicion that there's more damage in that cylinder than just the valve.

BTW, based off the heat differences between the two cylinders I'd say make sure and flow test your injectors or check the little screens for debris.

That all said.... just pull the engine man.

[Mason]

That all said.... just pull the engine man.

Dr.beans just wants to see this engine pulled, make his dreams come true.

I wouldnt be surprised if you are sporting a vac leak as well. The pcv system is a rats nest of dried up 15 year old hoses. I wouldn't bother searching for a leak and just replace all the hoses while working on your other issue.

I believe you have collected all the info you can and it's time. I would get that engine split before I decided what to do next.

[BoMbS]

Thanks Masib Doctor.

I did another leak down test tonight on cylinder 4 with the exhaust manifold off, and the rear most exhaust valve is indeed the culprit.  I still can't see any explicit damage, but it's definitely leaking, while the other exhaust valve in cylinder 4 is not leaking.  What would have caused this?   

BTW, based off the heat differences between the two cylinders I'd say make sure and flow test your injectors or check the little screens for debris.

Just to be clear, the two exhaust valves pictured in my last post are from the same cylinder, cylinder 4.  What would lead you to think I have temperature differences between two cylinders?

However, I do have a suspicion that there's more damage in that cylinder than just the valve.

I stuck my camera into the cylinder via the spark plug hole while the piston was at the bottom of the stroke.  The piston head looks fine, and from what I could see of the cylinder walls, they were smooth.  Piston 4 top and bottom side pictured below.

   

Is it stupid just remove the heads and have them rebuilt, leaving the block alone?  I really don't want to split the case, but I'm thinking if I have to remove heads, I should service the short block too...  But $$$ are in my way of doing everything.

[Dr Beans]

Rotate the crank until C#4 exhaust valves are fully opened. I'm thinking one of the valves isn't fully extended (or is slower to open) which is letting more exhaust gas out the other valve causing it to see more heat. Could be anything from a bent valve, out of bore guide, seat that isn't flush, or shim/lash issues. Hard to say at this point without taking the head off.

Temp difference between cylinders is most likely an injector issue.

I can't tell from the pictures, but on the bottom picture, on the right side of the carbon build up there looks to be a bunch of small pieces of carbon that have been knocked off (kinda look like white dots) which looks like pre-det. Other than that it looks good.

[RockThePylon]

Pull the valve cover off and check clearances.

Too tight and it won't fully close. Super loose would suggest it's bent.

[BoMbS]

Ok, what the duck?

I’ve never checked clearances before (I’ve been learning a lot here, thanks ladies), but I did some more YouTube learning, bought a feeler gauge, and got the following results.

FYI, I check the clearance between the cam lobes and the “buckets” while the lobe under test was pointing exactly opposite the bucket.

Cylinder 4 good exhaust valve = 0.15mm fit, 0.18mm didn’t.
Cylinder 4 bad exhaust valve = couldn’t get the smallest feeler in.  Less than 0.04mm.

Why would this happen? From my Haynes manual the exhaust clearance should be 0.35mm.

Any ideas?  Think it’s possible this valve might seal again if I just adjust valve clearances? Save me from removing the head.

[RockThePylon]

Not sure what you wanna do from here... I guess it depends on method of adjustment and parts availability. Could just be a tight valve for some reason, or there's damage to valve/seat.

[WAGOWN]

Could it be a ringland?

[darthekai]

What's the tolerance on those exhaust valve clearances?
You sure the tight one the cam lobe wasn't pushing the valve?
If it's tight, worn seat, worn valve, failed retainer, etc etc etc. Look and make sure the spring retainer etc all look normal and are sitting in the right spots. Too tight might be the cause of your compression problems? Sorry I'm just jumping in with dumb advice and didn't give the full thread a read just skimmed. It would leak because it wouldn't be tight on seat on base circle.

If the valve leaks, I imagine the valve will get hot because expanding burning gas will blow past it every ignition event.

[BoMbS]

Could it be a ringland?

Can't be a ringland because the leak down test would have pressurized the crankcase, not the exhaust.

OK, update:

I measured all the cam lobe clearances and the right side head is perfectly in spec. The left head intake valves are in spec too.  The exhaust valves on both cylinders 2 & 4 were out of tolerance, with the leaking valve being the worst. 

I just removed the exhaust camshaft and the top of the leaky valve looks normal, so I'm hoping maybe it was just being held open.  My plan was to get thinner buckets and re-install and see if I get lucky, but I was just thinking, with no camshaft in the way, can I re-pressure test the cylinder now to see if it will leak?

That should confirm if just getting thinner buckets will work? 

[RockThePylon]

I was just thinking, with no camshaft in the way, can I re-pressure test the cylinder now to see if it will leak?

So, I was gonna suggest this, but thought better because it's sketchy. If any valves are being held open with no timing belt, and the crank moves suddenly with 100psi in one cylinder, you might crash into some valves. That would suck.

I'd try it myself, just be very damn sure the crank isn't going to move on you.

[BoMbS]

But these engines aren’t valve to piston interference engines, only valve to valve interference.  With the timing belt off all valves are fully closed, so there should be no concern I would think.

I was wondering if I should hold the piston at top of the stroke or if if matters if it sits at the bottom.

Thoughts?

[BoMbS]

Well shift, look at that.

Just did the leakdown test again on cylinder 4 and achieved a 10% leak rate. Not ideal, but pooploads better than 90%.  Now I really want to try putting it back together after properly adjusting valve clearance. 

Problem is, in order to know what size bucket to order I have to know the existing clearance, and it was actually a negative clearance so impossible to measure...  Hmmm.

Also, I will admit my screw up, to help others learn from my mistake.  I decided to hold the piston at the top which I knew would fight my power bar so I braced myself and started turning up the pressure.  Since the crank could rotate either direction, I used a non-ratching power bar on the crank bolt, but what I didn't expect was the bolt to unscrew....   The piston slammed down pretty quick, but with no drama.  The pressure just built back to what it was quickly, meaning the test didn't require the piston at the top in the first place.