Mysterious Loss of Compression
 

This is going to be a long post. You've been warned.

Back in July, I installed a GT3071R big turbo in my Speed. At the time, I was fully bolted and also decided to up the ante running direct port meth injection.

After everything was installed, the car was hauling serious ass with the meth injection allowing me to run 16* of timing up top (got too scared to run more). Shortly after, I did a time attack event (dialed back timing to 12* just in case). I ran datalogs during my runs and the ECU pulled a maximum of 1* of spark advance. After that, the car was no longer the same. Something was up.

No matter what I did, the engine would misfire at WOT with the ECU throwing a P0300 (random misfire) code. I tried adding fuel, removing fuel, gapping plugs, changing plugs, installing a BAS, lowering boost, checking for boost leaks etc... Funny enough, at idle and putting around town, the car ran fine.

Eventually, I ran a compression test on the motor. Results are as follows:
Cylinder 1 - 190PSI
http://farm7.static.flickr.com/6137/...9c636077_z.jpg

Cylinder 2 - 180PSI
http://farm7.static.flickr.com/6026/...37bab890_z.jpg

Cylinder 3 - 160PSI
http://farm7.static.flickr.com/6139/...11244c98_z.jpg

Cylinder 4 - 185PSI
http://farm7.static.flickr.com/6022/...e7f9cb21_z.jpg

Cylinder 3 was somehow hurt but at this point, I didn't know why. It could've been a number of reasons.

Boost values at idle also read -8PSI vs. -10 to -11 when things were normal.

After a while, I couldn't live with the misfires at WOT. It was driving me nuts and having to pussyfoot everywhere in this car was unacceptable. I decided to pull the motor, figure out what was wrong and forge it. (More relevant info can be found in that thread)

Once the motor was taken apart by the engine builder, the problem was revealed...
http://farm7.static.flickr.com/6078/...619ef2b1_z.jpg

http://farm7.static.flickr.com/6180/...5675e17d_z.jpg

A 2" chunk of the ringland cracked and was held between the rings the entire time (I don't know when it finally decided to let go - maybe it was a progressive deterioration). I'm glad it didn't decide to gouge the cylinder wall or somehow get past the rings as that would've been incredibly expensive. The block is perfect.

Other noteable information is as follows:
- The top of piston 3 and cylinder head were not pitted. Pitting is a sign of detonation.
- The bearings along the crank and all of the conrods were in mint condition. Damaged bearings are also a sign of detonation.

Cracked ringlands are a sign of detonation but @Lex pointed out that generally, the top ringland is the one that gets damaged under detonation events. The one damaged on my piston is just below.

So, I think I'm the first one to report this issue on MSF and will subsequently update my sig to reflect the discovery of this fail. Haha.

At any rate, those with mysteriously low compression in one cylinder may have something to take a look at. Our motors may have caught a bit of the Subaru disease. Of course, mine could be an isolated case.

More pictures of the motor teardown will be posted as the build progresses. There may be more to the story of this motor; who knows. Check back often if you're interested.

Cheers,
Dave

Holy shit :eek2:

Dave, that crack is on the exhaust side of the motor from what I'm seeing. That is where the fuel is sprayed against the bore by the injector and the piston ringlands are completely black from poorly combusted fuel.

In certain cases, when running rich and in the case of our motors spraying fuel at the walls of the cylinders, the fuel won't completely vaporize and get trapped against the wall and against the ringland. Eventually it does vaporize and become ignitable. With the engine running very hot, it can ignite uncontrollably in that region.

You probably experienced detonation at that point breaking the second ringland.

I don't even want to speculate this was at WOT. It could very well have been on the highway and is independent of ignition timing at WOT.

EDIT:

The other possible cause is the ring gaps butting together. Have your shop measure the ring gaps on those rings.

So you're saying the fuel dribbled down the injector side? I would expect such a problem to happen on the exhaust side since fuel is sprayed across the cylinder.

Thinking about this a couple of more minutes - I bet this is how Subies break their ring lands as well. Since the pistons lie flat, unburnt fuel drips onto the bores and eventually ignites between the ring lands.

Hmm, got turned around there ... you're correct in that fuel is sprayed against the exhaust end - perhaps there is a leak issue with the injectors ... how was your KR before the failure?

I wonder if this happens after shutting down a very hot engine from say... time attack? Coolant temps were in the 220F region.

I am running direct port injection so meth could drip into the cylinder theoretically. It would have to overcome gravity though as the meth still has to travel upwards through the runner.

Non-existant. Haven't knocked in months beforehand. Save for your occasional up-shift KR.


They're on it.

Could even be combustion chamber design and poor fuel atomization. These cars have a lot of carbon buildup on the crowns and top ringlands of the pistons (they are black). That's an indication that fuel gets in there.

So are we ruling out oil consumption causing black pistons via weak PCV system?

I haven't installed my catch can yet... :bigeyes:

We don't burn more oil than other turbo cars. The valve deposit issue is different but oil consumption is not anything crazy. All stock cars recirculate through a PCV system.

Hrm...

So idle vac was -8psi....mine is also -7/8psi UNLESS i increase idle to 950ish then its back to -9/10psi ish.

I also have 2 cylinders with less compression (the middle ones), 18x-16x-16x-18x (cant remember the last numbers but I konw they were in the less than 10% range)

Any these else that you saw before you pulled the motor? I hope my engine isnt going to same way but I gotta do more investigating

Thanks for the disection of the motor, definately something else to lookout for now...might as well start saving :(

I don't understand what you mean here.

Sowwie! I should clarify, I mean is there any other symptoms that you found before pulling the engine. Like hindsight sorta speak. Alot of the symptoms you have had, I am personally having as well and I am just trying to figureout the likelyhood that I am in a similar boat as you.

-Low compression
-Idle Vac low
-Misfires (albeit mine is on start up with a slight hardstart at times, havent gotten any at WOT)
-Oil in intake tract (then after I did a PCV change, OCC install, Safeseal while mani was off...gotta double check my pictures of the seal install to see what was dirty and what seal was not dirty)

Dave, I know these were already cleaned, but did the shop take note of any one of the cylinder crowns looking different from the others? Maybe there's something to be seen from the top.

Nope, that's about it.

The pistons were cleaned but the carbon buildup still remains. That stuff needs to be scraped off and I doubt they'll do that since the pistons aren't going back in. I'll get some photos of the crowns when I can.

+1

It could also be a random mfg defect in ring gap or piston construction

I doubt detonation between the rings could do that.

Some more info. Copied from an email.

Some relevant background on this note.
- Time Attack Event.
- Ambient Temps: 35*C minimum that day. It was effing hot.
- Coolant Temps: Over 220F
- AFRs: 11.8
- Timing: 12 degrees max. Tune was below KR threshold without meth.
- Fuel: Petro Canada Ultra 94
- Meth: Tons.

As I would pull in after a session on the track, my WWF reservoir light would come on indicating I was low on meth. There is a chance that I may have run out on the track causing the engine to get hotter towards the end of a run but like I said, the tune was knock free.

Cheers,
Dave

I dug out my busted engine parts and took a look at the pistons. There was a big difference between the exhaust and intake side of the pistons in terms of carbon buildup as we discussed many moons ago when everyone was breaking rods. The exhaust side where the break happened is covered in black carbon indicating poor combustion and remnant fuel. In the world of port injection, if people see the top ringland covered in carbon they immediately suspect detonation because the uncontrolled flame causes the carbon buildups on the ringlands. Below is a picture of this exhaust side of a DISI piston. Note the carbon buildup on the top and second ringland - especially on the second ringland.

http://www.mazdaspeedforums.org/foru...1&d=1316186538

The second image is the intake side:

http://www.mazdaspeedforums.org/foru...1&d=1316186538

The entire ringland should have this brown color to it as seen on the intake side.

In spite of direct injection our cars are knock happy and run very hot. I would not be surprised if rings butted due to heat, or if the extra fuel on the hot exhaust side of the piston detonated and broke the ringland. In the picture above, you can clearly see that fuel/oil are getting to and covering the second ringland on the exhaust side due to the fuel being sprayed at the cylinder walls by the positioning of the injector. This is an engineering flaw of the motor.

Those images were huge - scaled them down. Dave, did the failure happen at the track event you mentioned above?

On the drive home after the event, I tried hitting boost. I started noticing the misfires then.

EDIT: What still throws me off about this whole ordeal is the condition of my injector seals. Looking at those, I would suspect that this issue was coming long before the time attack event.

What makes you think the injector seals were a tell-tale? What about them was indicative of the failure?

Have a look at them in my other thread and note the discoloration vs. the cylinder it was in. Cylinder 3 was spotless.


Sent from my iPhone using Tapatalk

Alot of the ring groove carbon buildup is going to be coked oil, its a big problem on DI engines because they run hotter than PI engines.

It could be possible that oil coke had been building up in between the ring gap and when you did that event the rings couldnt expand completely.

I havent looked but the oil squirters may be angled towards the exhaust side of the piston which would explain more carbon being there.

As far as the carbon being fuel, if 1700psi of fuel was getting in there it should be cleaning the carbon not adding to it. There should also be alot of cyl wall scoring on that side as compaired to the other side, which we arent seeing.

Did a test this weekend..

Cyl#1 160
Cyl#2 155
Cyl#3 143 Wet: 155
Cyl#4 170

Next i need to do an leak-down test to see if its either valves or rings in Cyl#3. I think its rings. FML

I wouldn't rule out the ringlands either.

Ringlands? Whats that lol.. Anyhoo seems like im going to build my motor..

The ringland is what I'm holding in my hand in one of the photos posted above.
http://farm7.static.flickr.com/6180/...5675e17d_z.jpg

Notice the piece missing in this photo.
http://farm7.static.flickr.com/6078/...619ef2b1_z.jpg

I see. I did read this full thread. Just didnt understand the word lol. Thx bud! Well done on such an inspiring build!