STI Mav

David - It's an 06 plate STI Hawkeye with PPP. I know the head gasket's are being replaced with the Cosworth version. I'm just keen to make abasolutly sure nothing else goes wrong once I have the car back.

handsome rob

David, since joining sn and finding this problem, every oil change i've done on my 2003 sti has filled me with dread, in your opinion does this procedure cure this problem , or is it something that we do just in case it might help to stop it from happening ? sorry if thats putting you on the spot a bit , Rob.

Splitpin

He's not right on this. You'll get a little bit of drainage back from around the crankcase and heads, but (most of) the pump, feeds, galleries and pickup remain pretty much full of oil. This is something you can test yourself by removing the oil pressure switch or turbo oil supply banjo after the car's been parked up overnight (or for a few weeks). Even turning the engine over by hand will see oil emerging almost immediately (and consistently).

This might not be the case on every engine he deals with, but the idea that the oil system drains down immediately/overnight/etc is wrong on these. The non-return valve in the oil filter is there specifically to prevent drainback. I suspect the location of the filter more or less at sump oil level helps to maintain fluid lock here too, as does the fact that it's a flat engine, and so there's less opportunity for gravity to force drainage than there would be on an upright inline.

Nope. As soon as you take the filter off you drain all of the system "above" it, as well as the line between the pump and filter inlet. This is where the difference occurs.

One of the difficulties connected with this phenomenon, and the explanation you're asking for, is that all the evidence to this point is anecdotal. It's not as though anyone out in the wild has gone to the trouble of slicing up (or instrumenting) an engine and performing extensive detailed analysis of the oil flow (altho Subaru obviously should have!). We are all, to a large degree, guessing on the basis of an incomplete and anecdotal set of data.

As such, the prime filter and unplug crank sensor recommendations are made, usually, on the basis that they might help prevent transient oil starvation, don't do any harm, and are, as you say, a no-brainer, so better to do than not bother.

Anyway, all that said, here's a theory that may or may not be correct, but it happens to fit the facts so far as they are known.

During the oil and filter change you will drain out the oil pickup and most of the content of the galleries. Even if you fit a pre-filled filter, you will be left with an oil system that will be filled in some places, while others, in particular the line between pump and filter, the mains to the galleries, and pickup, will contain a significant quantity of air. Somehow, that air needs to be got rid of in order to properly prime the engine.

Disconnecting the crank sensor will prevent the engine starting, removing combustion forces from it and allowing the starter motor to drive the oil pump round, pushing the air through and out of the system and repriming it in a slow and controlled manner at a couple of hundred revs per minute, while the only forces the bearings are subject to are those arising from compression, valve spring resistance, torquing exerted by the timing belt and so-forth.

However, if you start "as normal" after an oil change, within two seconds, you will have the ECU in startup mode firing the engine and running it (and the oil pump) up to 1500rpm or so.

This is going to result in any air in the oil lines being compressed to some degree by the onrush of new oil "behind" it, and forced through at speed, seeking the path of least resistance.

There now seems to be a very real risk of this air literally (albeit only for a very short time) blowing standing oil out of bearings which are now, don't forget, subject to combustion forces. Under those circumstances it doesn't seem at all impossible for enough metal to metal contact to occur for problems to be set up either immediately or further down the line.

Whether you actually get contact, and how serious it is, depends on a whole range of factors, not least how much air is in the system, where it goes (and so which bearings are most affected), and which cylinder happens to be firing at the time. This also explains why it doesn't happen to all engines, all of the time.

Oh - and before someone pipes up saying "you don't get this on Ford (etc) engines", that's a red herring. Other designs of engine have different oil systems, different oil pumps, different size bearings etc. etc. etc. This line of logic simply doesn't apply, not least because sat here, in a Subaru forum, we're probably not best placed to know whether other types of engine suffer from this problem or not anyway.

So, it's a theory that may be wrong, but it happens to fit what we know. Make of it what you will.

Splitpin

Don't want to speak for David but IMO, as above, prefilling the filter, removing the crank sensor and running the engine on the starter to prime it does everything you reasonably can to head this possibility off.

About the only further steps you can take are pulling the plugs (to take compression forces off the bearings during prime), but that shouldn't really be necessary.

What I'd be interested to hear about is any anecdotal evidence of engines that have suffered bearing failures shortly after an oil change in which the "extended" procedure was used.

handsome rob

Thanks for that detailed reply ,Rob.

Splitpin

Stop calling me Rob, darling.

handsome rob

STI Mav

I was looking forward to getting my car back this weekend. Now I'm thinking - push it home and just leave it on the drive way to look pretty

APIDavid

iTrader: (4)

Hopper, your car's here. Safe and sound, helpful guys at the dealers.

A few scuffs around the rear bumper and all wheels slightly kerbed. Otherwise wet and rainy/snowy, unable to see any other obvious new marks. So l guess it's as you took it there?

David APi

eggy790

iTrader: (20)

interesting thread, aplitpins post have got me thinking, ive never changed my oil by disconnecting the crank sensor or removing plugs (done around 10 changes at least).

but i suppose he is right better safe than sorry only takes a few mins etc.

p.s. david could i get a copy of these useful notes you have please?

cheers

splitpin: just wondering why it takes a while for the engines to fail and why it doesnt happen immediately at oil change? it may have been said, or i may have misunderstood.

Hopper

David,
Sorry it has taken so long to respond to your msg of December 17, I have (only) just discovered the little extra page button at the bottom of the tread!
Yes, the car is as you described. Most of the marks on the wheels being when my wife uses the car . . . .
BTW I have had an acknowledgment of my letter to IM, but not any answers yet.
Brilliant piece of analysis by Splitpin, too. Many thanks for that.
Thanks for keeping me/us informed.
Hopper

Splitpin

Sometimes it does happen immediately after the oil change - I recall a thread here not that long ago where somebody was a bit gutted to have done his first bit of maint on the car in the form of a change, only to find the engine knocking immediately he started it up. Needless to say he didn't know about filling the oil filter/pulling the crank sensor etc.

The reason why it doesn't always happen as quick as that is, if my theory above is anywhere near accurate, it's a lottery whether you get any metal to metal contact, and if you do, how serious it is and how long it lasts.

If the faces of a bearing become scarred due to direct contact, oil flow through it will be compromised from that point forward. This, in turn, makes it more likely for contact to occur in future, and every time it does, the viability becomes further reduced.

You get to a point where it becomes a vicious circle where damage makes it much more likely that further damage will occur - and at that point, it might take a little while, but it's only a matter of time before the bearing won't support normal usage loads, and you actually hear it.


No sweat. It might be anything but brilliant - it's only a theory, but at least by putting it up others have the opportunity to disagree as and where they know better.

Let us know where you get with the correspondence.

davefaedeen

can someone post a picture of the crank sensor so i know for next time where it is??? I have a 95 2.0 WRX sti and a 2006 2.0 WRX STI

thanks guys

joz8968

iTrader: (13)

Looking at the engine straight on from in front of the bumper... right in line with the central crank pulley... on top of the forward edge of the engine block... little black electrical connector/sensor combo screwed into block.