Blown engine theory

[rd415]

PCV works in both directions - is only open under the right conditions - high pressure in crankcase or near maximum manifold vaccuum will close PCV. eg. closed at idle or over-run with high manifold vacuum also closed with high crankcase pressure. The PCV actually floats and early design valves were known to rattle.

Don't know whether this is relevant or not.

[Doug_MPS6]

RD15, Hi. Thx. That is correct. They do float - and rattle - as my current MPS6 installation does. But by "floating, I don't mean that the PCV valve permits aggressive flow either way (what would be the point?), which is the impression I get from your explanation (I may misunderstand you), and hence I have trouble visualizing your "closed at idle or over-run with high manifold vacuum also closed with high crankcase pressure", for over-run induces high manifold vacuum and hence will open the PCV valve and permit maximum crankcase evacuation, while high CP will induce the same effect and also open the valve. This is what I raised in my last posting when I suggested that excess boost air being returned to the crankcase (via the BOV bypass, turbo inlet offtake line and cam cover) would be expected to pressurize the crankcase (the condition I was trying to eliminate), and caused me to ask, "what happens to that ex-boost air once it has passed through the stock PCV valve into the manifold downstream of the butterfly" - logic tells me that it must be ingested by the engine, for it has no where else to go.

Also, further to my previous post, one can argue that the pipe linking the cam cover to the turbo inlet pipe is subject to intermittent venturi effect of varying degrees (depending on the angle of entry of the small pipe to the larger one - mine is non-stock and angled), thus pulling air from and thereby inducing a (partial) vacuum in the engine. This would apply even to stock and not just to my added PCV valve type of setup. Therefore, one could assume that the stock PCV valve at the manifold and the cam-cover to turbo inlet lines both pull air from the engine crankcase at different times and possibly simultaneously on occasion. Thus, regardless of which port the air is being drawn from, there seems to be a greater likelihood of the crankcase being under a vacuum than it is of being under pressure.

[Oldhead]

Hey guys ,to everyone who has inputed this thread,great stuff, objective /constructive/informative. A bloody good read. It would seem as though Mazda engineers are fully aware of previous probs, and have acted on Gen 2 PCV restriction. Typical for ist runout models to have gremlins and the bloody buyers do the R&D and suffer from design faults. Look at the costs incurred by Toyota . Money,Money ,Money!!!!!!!!!

[Doc]

Old head can you post a pic of your rocker cover vent port size.. I'm interested if that's the olny change made.. As I'm running a restrictor on my one..

[rd415]

hi Doug_MPS6
I read your response a few times, still don't grasp your explanation. (Maybe getting too old) Tying to explain my point is hard but it is that pressure either way actually closes the PCV. Therefore unless the PCV is faulty the crankcase should never have a positive pressure.The design of a PCV is to withdraw crankcase fumes thus reducing condensation in the crancase. It floats at cruise rpm to withdraw the fumes. One of the tests for a PCV is to apply vacuum to see if it leaks.

[Doc]

so under 20 inches of mercury or so (idle) it will be closed? hhmm I didnt know that

And only when things are close no neutral pressure it will be open?

Too much pressure or too much vacuum will close the bloody thing?

Damn I totally had the bull by the tail on this one if that is indeed the case

[rd415]

Correct - the reason for closing at idle is to prevent an obvious air leak into manifold at idle. Air leak - lean mixture - high combustion temperatures - PROBLEMS.

[kmh001]

My understanding is that the PCV on a boosted engine opens at idle and partial throttle (manifold vacuum) to create negative crank case pressure and to evacuate crank case gasses. This causes air which has already been measured at the MAF sensor, to flow into the rocker cover vent, through the crank case and into the inlet manifold for combustion.

Under boost (manifold pressure) the PCV must close to prevent the crank case becoming pressurised. Short periods of boost should be of little consequence because as soon as the throttle is closed the PCV opens and draws crank case gasses. If boost is held for an extended period, crank case pressure increases due to piston blow-by and/or boost air leaking past the PCV. When this happens, crank case gasses will flow out of the rocker cover vent into the inlet pipe. If you look inside the inlet pipe you'll almost certainly see a small patch of oily residue near the junction of the rocker cover vent pipe, which is deposited by oily crank case gasses during outflow.

Either way, the air entering the cylinders has been measured upstream at the MAF sensor and there are no mixture issues.

Problems arise when guys open the rocker cover vent to atmosphere (eg with a vented catch can) which allows unmeasured air to make its way via the crank case to the cylinders causing mixture problems.

I have read about guys in the US installing a second PCV in line with the existing one to reduce the effect of crank case pressurisation due to the single PCV leaking under boost.

The vent pipe restrictor modification is not without controversy and in my view it's bad news, because while it helps to create negative crank case pressure when the throttle is closed, it's going to work against you when the engine is under boost because it will cause crank case pressure to increase quicker and to be unable to dissipate via the breather pipe.

There are guys with blown DISI engines who are absolutely convinced that positive crank case pressure was a causal factor. And there are manufacturers who seem to agree, eg VW and Audi are fitting electric crank case pumps to DI engines (which eliminate the PCV and its inherent problems) to ensure there is never any positive crank case pressure, even under boost.

[Doug_MPS6]

KMH001, thanks for that explanation. My setup is not a restrictor per se and when open has the same effective bore as the unmodified stock line. It therefore permits normal venting under the conditions you describe and only closes intermittently under aggressive pulses from the bypassing BOV when it is activated by throttle closure, thus preventing crankcase pressurization via that vent. Cheers.

[rd415]

No PCV opens at idle - otherwise you have an air leak at idle. Just can't happen.

"At idle, the intake manifold vacuum is near maximum. It is at this time the least amount of blow by is actually occurring, so the PCV valve provides the largest amount of restriction. As engine load increases, vacuum on the valve decreases proportionally and blow by increases proportionally. With a lower level of vacuum, the spring returns the cone to the "open" position to allow more air flow. At full throttle, there is nearly zero vacuum. At this point the PCV valve is nearly useless." Quote. There are plenty of other explanations on PCV just look them up.

Please note second paragraph clearly states operation of PCV - first paragraph a bit confusing, but in essence means that the PCV is closed.

[rd415]

Quote "Either way, the air entering the cylinders has been measured upstream at the MAF sensor and there are no mixture issues."

True - mixtures won't be effected - but doesn't help efficiency or economy, and as there is very little blow-by at idle there is no point in opening PCV.

I still can't see a pressure rise in crankcase under any circumstances sufficient enough to bend a rod or cause any damage. It is more likely to create oil leaks from blown timing case oil seal or rear main way before any form of major engine damage. In fact the seals would act as a pressure relief valve if crankcase pressure ever got as high as suggested.

[Wardski]

The only way you're going to know exactly whats causing the kaaboomski is to run a series of tests, with a number of engines, fully stress probed & under xray.

I know theres lots of theories going round, but really, its all pointless IMO - especially if Mazda (the maker of this wonderful creation) has no clue.

[rd415]

The only way you're going to know exactly whats causing the kaaboomski is to run a series of tests, with a number of engines, fully stress probed & under xray.

I know theres lots of theories going round, but really, its all pointless IMO - especially if Mazda (the maker of this wonderful creation) has no clue.

I don't think it is pointless as it has produced an active thread. A lot of the theories are just a bit over the top and make me

It would be a fluke if you were able to reproduce the circumstances of the engine bending the rod under controlled conditions, because there is no indication of where to start testing.

It's obvious to me that some of the responses are hearsay and have not come out of a text book or from some-one with very little mechanical/engineering experience.
I don't know whether the actual cause will ever be found, especially seeing Mazda has no idea.
Whatever --- all the theories make interesting reading and seeing the forum is so quiet lately a bit of controversy keeps the interest going.

Don't forget OZmps National Relay - become involved

[Doug_MPS6]

The reason this issue was raised in this thread was because some workers noted large oil build-ups in the tops of the cylinders on blown engines. This leads to folks wondering about effective crankcase breathing, hence the discussion on PCV valves.

That said, after many strip downs of both auto and aircraft engines, I have yet to see an engine that destroyed itself from oil in the cylinders as a primary cause (except for one aircraft engine that was inverted, naturally far less surprising, and that was due to a seal failure and occurred upon startup after standing, not in flight).

If one moots "top oil build-up" as a cause of failure, one needs to ask how oil would get into the top of an engine. It would require catastrophic flooding via the intake system, which would in turn suggest that there was another primary cause and that engine destruction would be a secondary result. Oil accumulation in rocker covers (such as from over-filling or some other issue) and leaking down valve stems past (possibly crook) valve stem seals is one example. Hard to imagine happening without notice when in cruise.

When a blown engine is pulled down, there is often oil everywhere, but finding oil in cylinders then doesn't mean that it was a cause of the blow up. When an engine blows, lots of components can get ventilated with holes where there aren't supposed to be any, such as piston crowns. These effectively act as pumps in the dying throes of the engine, lifting oil into the upper sections, but this oil may not be the smoking gun.

It is almost as difficult, though maybe slightly easier, to imagine destruction by hydraulicking from oil accumulation below a piston in four-stroke engines, depending on crankcase and main bearing mount design. Two strokes are more noted for it as the case is sealed, being part of the induction system. Oil seeping in from remote two-stroke fuel-oil mix tanks or CCI systems can fill all or part of the available case space and initiate rod failure on piston down-stroke. This tends to happen on start-up after prolonged standing, not during running.

At this stage, and emphasising that I have no personal experience of it, I'm far more inclined to believe that rogue ECU and timing advance issues are possibly more to blame than oil hydraulicking for sudden and catastrophic engine failure with little or no warning in the majority of cases. Some folks point out that they have "never seen a rod bend in the (bi-axial) shapes" they have witnessed in some of these failures, but I think that is more a function of the space they have to fold up in, rather than an indication of the vectors of initial stress which might point to a cause, and their response is thus not an answer to the question.

The only other possible cause of failure I can think of in an otherwise healthy motor is runaway dieseling or sudden over-fueling of one or more cylinders. Given that electronics are involved (and that I have yet to find electronics that are 100% failure proof), I still tend to lean towards an ECU and/or DISI design issue (including perhaps excessive injector opening), as opposed to an exclusively mechanical failure. In this connection, one might even look at MAF sensors and the accuracy with which the ECU interrogates the MAF signal and responds to it. If it reads too much air for some bizarre reason (such as dust on a sensor?), it might just dump in too much fuel. Even in cruise, this could be the cause of instantaneous failure, especially if timing was also running amok, and even moreso in a high compression motor.

[Wardski]

It would be a fluke if you were able to reproduce the circumstances of the engine bending the rod under controlled conditions, because there is no indication of where to start testing.

I disagree

Don't get me wrong, I think all of this is interesting stuff - but doesnt really resolve the issue

What I do find totally interesting, is that Subaru had a bad stock tune on the very early 08 Subaru STi's which caused ring land failures on all 4 piston heads - something that you wouldnt even think would cause such a thing (eg, you'd instantly think something mechanical such as poor bore tollerances, etc).

Subaru then set about to try everything in house to rule out all other modes of failure and identified that they had developed a bad tune - amazing really considering there are thousands of things that can cause ringland failures. I often wonder that Subaru had the advantage of knowing everything there is to know about their Boxer design - as opposed to Mazda that borrowed a bit from Ford, volvo and others brands in the mix, and the 2.3 ltr has only been around for 4 years - where as Subaru are going on 15 years of development of the same block & design...

[rd415]

"At this stage, and emphasising that I have no personal experience of it, I',m more inclined to believe that rogue ECU and timing advance issues are possibly more to blame than oil hydraulicking for sudden and catastrophic engine failure with little or no warning in the majority of cases."

Totally agree - it is mechanically inconceivable that oil would cause hydraulic lock in this case as there would be oil "from arsehole to breakfast time" (excuse the french) through the inlet tract, in the air cleaner, etc. I'll stick with rogue ECU and timing issues which I believe would be impossible to replicate under controlled testing.

I have rebuild several engines and also dumped a few with gaping holes in block from Hydraulic lock - all from water, never from oil.

Great Post - Thanks

Remember the OZmps National Relay - become involved

---------- Post added at 11:58 AM ---------- Previous post was at 11:46 AM ----------

I guess your right in the 2.3 mazda engine being relatively knew but do you realise that Mazda have had 2.0 litre Turbo's since 1988 and a really weird supercharged diesel back in 1985. Not a great deal of difference between the 2.3 litre and the 2.0 litre in design. Point is - Mazda have plenty of experience in forced induction engines.

Boxer Subaru is only a water cooled version of VW dating back to late 1940's - old tired design

[Wardski]

I guess your right in the 2.3 mazda engine being relatively knew but do you realise that Mazda have had 2.0 litre Turbo's since 1988 and a really weird supercharged diesel back in 1985. Not a great deal of difference between the 2.3 litre and the 2.0 litre in design. Point is - Mazda have plenty of experience in forced induction engines.

Boxer Subaru is only a water cooled version of VW dating back to late 1940's - old tired design

.... maybe but not DISI. with DISI comes a lot of new challenges, especially keeping the motor cool - which is why I still dont understand the TMIC

The boxer maybe tired, but damn theres nothing on the planet that sounds the same with a decent exhaust, its reliable, and goes extremely hard!

[Doug_MPS6]

BTW, on the weird bent rod shapes that have been noted it is easy enough to model. The moment a rod fails under compression it normally produces a simple U or an S shape in a single flat plane, and in the plane of rotation, perpendicular to the crank axis.

But the engine doesn't stop there and continues to turn as it dies, driven by the forward motion of the car and/or by surviving functioning cylinders. Either at the moment the U or S is being formed or on the subsequent stroke, there is a time when there is no room in the case to accommodate that shape, and the mal-formed rod impacts the confining walls and is forced to bend in a second plane, this time parallel to the crank axis producing either a U or an S bent into another U OR a U or an S bent into another S.

Another BTW - A correction required:
The Subaru engine is a boxer engine but apart from that has nothing in common with the VW engine. I once rebuilt and drove Jowett Javelins, a well known 1950's English car which was quite revolutionary at the time in a number of areas and which won a number of records and competitions. It was pretty fast, very streamlined, had a flat floor with no transmission tunnel and adjustable torsion bar suspension. It was lovely to drive.

When Subaru were looking for a new engine, they bought a number of Javelin 1500cc engines for evaluation. The Subi motor is a direct copy or evolution from the Jowett engine. The Jowett engine was water cooled, as is the Subaru. In the Javelin the radiator was at the rear of the engine bay and the fan pushed air through it.

The Tatra T12 from the late 1920s featured a 1056cc flat twin air-cooled engine and is more closely aligned with the Ferdinand Porsche VW designed engine than the Jowett or resultant Subi engine.

[Wardski]

Boxer Subaru is only a water cooled version of VW dating back to late 1940's - old tired design

Actually no. Tatra used them back in 1926 (Tatra 30), and they were used in aircraft motors before that. Subaru started using them in 1966 in their EA series after licencing the tech. VW actually copied Tatra's design - as you do!

[rd415]

Porsche worked for Tatra. - 1926 even more tired than I thought

We can all play that design game - Gottlieb Daimler, Carl Benz and Rudolh Diesel have a lot to answer for. Reality there is nothing new, just improvements on an old design. Look at the 1920's Bently - four valves per cylinder X flow and supercharged