2.5L DOHC Head Gasket Failure *PICS*

[WAWalker]

The point I was trying to make here is..........................Head gasket replacement on 2.5L DOHC Subaru engines before complete failure (over heating) is good maintanence. And as a rule, should be paid attention to and considered a possible maintanence replacement item after another 100k or so if it is desirerable to keep the car on the road.

 

If the open deck block design is the major contributor, then failure of any gasket design is a possibility.

If repeat failure is even a remote possibility, with the updated gaskets, this would pertain to vehicles with rebuilt engines also.

[WAWalker]

Yeild is the point at which a metal becomes elastic (stretches) and does not come back to original shape. Using this defenition, yes the gasket is deformed after being torqued, and never goes back. The question is the bolts. Using the argument that SOA never told you to replace them means nothing. Just because they never told you to, doesn't mean that you shouldn't. The fact is that you strech the bolts in the installation process. The real question is if it's to yeild.

 

 

I think everyone discussing this topic should read this:

 

http://www.rv8.ch/article.php?story=20051227112926716

 

It has some really good information.

 

I have read the info in this link. I think you may be on to something.

This is why I look for imput from others. I don't have the patiance to search for all this stuff. I'm still learning and there are many people smarter than me who can help.

 

I still don't think that Subaru uses "Torque To Yeild" head bolts. There is nothing in the '99 FSM (lastest DOHC service manual) about replacing the head bolts. The torque proccedure specificly states, to be sure that finale torque angle does not exceed 180 deg. This could very well be to prevent torqueing these bolts to the point of "YEILD".

 

The info in your link explains the difference between "torque to yeild" and "torque to angle" and the advantages and dissadvantages of.

 

From the link............................................................................................................................................................

 

"With the classic style head bolt or stud, the tensioning below yield does

provide a very steady clamping force. The steadiness of the clamp is the

problem. With the use of aluminum cylinder heads, the needs for clamp

changed. Instead of a maximum clamp applied being optimal, a steady clamp that can allow for the thermal expansion of the aluminum head without exceeding the compressibility of the head gasket became more important. The classic head bolt or stud does expand and allow for some thermal expansion, but the flexibility was not quite optimal. When the aluminum head warms up at full operating temperature, it grows in size, putting more bolt tension on the head gasket, and on the aluminum itself. This results in brinelling of the head surface where the fire rings of the head gasket are located, overcompression of the HG itself, and distortion of the aluminum around the head bolts. If the HG cannot handle the excessive pressure that occurs, it will remain permanently thinner; so that when the engine cools down, the HG does not provide as good a seal between the head and block. Over time, this will lead to HG failure.

 

Enter TTY. By stretching the bolt beyond the yield point, and into the

plastic range, the maximum clamp applied is reduced; but by being in the

plastic range, the bolt can and does give more with the thermal expansion of the aluminum head. The clamp is not as great; but it is steadier throughout the temperature range...a very important thing when using aluminum for head casting material. HG longevity is increased; and brinelling and distortion of the head is reduced."

 

................................................................................................................................................................................

 

This could explaine the "wear" that I noted between the head gasket fire ring and cylinder/cylinder head matting surfaces. As I said the "gass cutting/erosion" crossed my mind and made sence, but on closer inspection there was an absences of discoloration from combustion gases leaking by. The discoloration was there, but not over the entire area of "wear".

 

So if the head bolts and torque proccedure used on these engines truely are not "torque to yeild", it is very possible that torque to yeild bolts and torque proceedures would help.

 

And a "torque to yeild head gasket" would only make the problem worse. As described above, that is the problem with to much clamping force on a head gasket. Once the gasket "yeilds" it can't maintain the seal.

 

Bad day when you don't learn something.

[Petersubaru]

just curious here.. would it be advisable to tighten up a little more the bolt with the torque spec of 77.6 and bring it line or closer to the other bolts(81-84) ..to a layman it seems like such a wide spread in numbers

Ok, it's a done deal.

 

Here are the finale ft lb. torque numbers on the head bolts.

 

Left bank Right bank

1 & 2 @ finale 180 deg of torque angle.

 

1- 84.4 1- 77.6

2- 82.6 2- 81.4

 

3, 4, 5, & 6 @ finale 135 deg of torque angle.

 

3- 72.5 3- 66.5

4- 71.5 4- 66.9

5- 76.5 5- 66.3

6- 66.7 6- 68.3

 

 

Now to find out at what point these bolts yeild. I have no way of testing this. Anybody else?

[WAWalker]

No.

 

Angle torqueing provides for a more even tightening of the bolts. Turning the bolt x deg. is going to thread the bolt in so far, the thread pitch on all the bolts is the same. If you turn all the bolts in the same amount of deg. they are going to all be tightened more evenly. The ft lb. torque numbers are not even do to factors beyond our control. Friction between bolt threads and block threads, bolt head and washer, washer and cylinder head, mainly. It may take more force to turn one bolt in a quarter of a turn than the next bolt. Make sense?

 

Go to the link that RallyKeith posted. This is explained very well there. Something like 90% of the torque applied to a bolt is used to over come this friction. It is a good read.

 

Thanks, RallyKeith:)

[robm]

Regarding WaWalker's feeling that head gaskets should be just another item to replace as part of good maintenance:

 

I can understand tires, brakes, clutches wearing. Same with alternators and waterpumps. Radiators can get worn out, as road salt eats them from the outside in. Axles and wheel bearings, now and again. Hoses, seals, cam belts and other rubber parts too. But in the great scheme of things, these are cheap, easy to replace, and often easy to inspect, so you can TELL when you are living on borrowed time. (Cam belts are the exception, they will fail without notice, but they're cheap and easy to replace on a mileage basis.) If they do fail suddenly, due to poor maintenance and inspection, they generally don't destroy the car (again, cam belts excepted).

 

But putting head gaskets on this list really sucks. They can't be inspected to see if they are ready to blow, and there is no mileage after which they are likely to blow (stories on this list range from <50,000 to >200,000 for the same kind of engine). They are not cheap and easy to replace (+/-$1000).

 

IMHO, a "300,000 mile engine" does not contain expensive, impossible-to-inspect components that have a proven history of failing without warning, with the potential of destroying the engine.

[forester2002s]

Yeild is the point at which a metal becomes elastic (stretches) and does not come back to original shape.

 

Just to set the record straight:

 

- below the yield point, metals are ELASTIC (i.e. they return to the same shape after the load is removed);

 

- above the yield point, metals are PLASTIC (i.e. they don't return to the same shape after the load is removed).

[WAWalker]

Edited due to lack of thechnical content.

[MDW25gt]

As I look at the photos again, this is my observation. On the head the 2 cooling passages at the leak area have LESS than half the bearing surface than there is between any of the other passages. If I was the engineer in charge of solving this, my approach would be like this.

1) Run FEA heat transfer to failure paying close attention to this area. (baseline)

2) Run FEA structual/sealing analysis to failure, paying close attention to this area. (baseline)

3) Make design changes to the head at the leak area. maybe shorten the 2 cooling passages by 10mm. gives 20mm additional bearing surface.

4) re-run step 1 to failure. compare results to baseline.

5) re-run step 2 to failure. compare results to baseline.

 

Continue iterative process until balanced improvement is achived. My guess is this has been done. I hope! And the result is the heat transfer needs in this area are governing the cooling passage design. In turn the failure due to structure and seal show up in our cars.

 

just my 2 cents Mike

[WAWalker]

As I look at the photos again, this is my observation. On the head the 2 cooling passages at the leak area have LESS than half the bearing surface than there is between any of the other passages. If I was the engineer in charge of solving this, my approach would be like this.

1) Run FEA heat transfer to failure paying close attention to this area. (baseline)

2) Run FEA structual/sealing analysis to failure, paying close attention to this area. (baseline)

3) Make design changes to the head at the leak area. maybe shorten the 2 cooling passages by 10mm. gives 20mm additional bearing surface.

4) re-run step 1 to failure. compare results to baseline.

5) re-run step 2 to failure. compare results to baseline.

 

Continue iterative process until balanced improvement is achived. My guess is this has been done. I hope! And the result is the heat transfer needs in this area are governing the cooling passage design. In turn the failure due to structure and seal show up in our cars.

 

just my 2 cents Mike

 

The distance from the inside edge of the coolant passages to the inside edge of the fire ring is ~4-5 mm at every passage.

[idosubaru]

wow he dropped the F bomb. FEA-that's Finite Element Analysis for those that are lucky enough to have never had to do it.

 

But putting head gaskets on this list really sucks.

the point is getting the information out there and letting everyone make their own decision. he's suggesting it, there's nothing wrong with suggesting it and letting people make their own decision. to emphatically declare you must do it one way or the other is short sighted and doesn't serve everyones interest. if you rarely take long trips and don't care to take the risk and don't have the money, then no matter. some people rely heavily on their vehicle for their work or traveling long distances and road trips and don't like to worry about it. and many americans don't want to watch the temperature gauge...or they just don't. in reality $1,500 is not that bad to pay for reliability. someone to spend a couple car payments on an item that will get them another 100,000 miles seems financially reasonable if you've got it. $1,500 really isn't that much money to many people, particularly people that know little about cars and don't want to have to worry about it. and also, people that don't know much about cars basically have the alternative of $20,000++ for a new car. sounds like a logical choice to me.

[cookie]

Frankly I think the more you know about cars the more likely you are to change the gaskets if funding permits. Some folks like nipper have purposely run them to see what happens and test thier personal therories (and he was right, the lower end let go before his gaskets).

In my cars I really like reliability and I reckon this saves me getting a new car.

Maybe it cost me $1500 or so but I have three reliable years behind me (Subaru kicked in bucks later too which I won't count for this purpose). This means reliabilty has now cost me $500 a year. Every year it does not leak reduces the cost furthur.

[Tech1967]

Bump...plus something I ran across while looking to see where the performance guys beef things up. The prices are insane and beware of the pictures. The semi-closed and closed block pics are identical and there are a few sohc heads with dohc labels on them. I'm trying to find the gasket makers site. There are stud kits and t-sleeves. I don't know if the botched info is due to them or their web guy so I would want to know a lot more about them before buying.

http://www.xcceleration.com/index.htm

I suppose I can take the time to peel off my water pump and make a detailed description on coolant flow ...the engines on a stand waiting for some love anyway. Interesting or waste of time? One response and that's how I'll spend my next couple coffee breaks.

Ooops...link above is their home, link below is the page the search engine gave me. The home page is more impressive.

http://www.xcceleration.com/engine-packages-raceengines.html

[Tech1967]

These guys seem to have some beefy stuff.

http://www.cgperformance.com/index.htm

 

Found one of these performance builders that included reworked coolant flow as part of their mods but it was limited to the transfer pipe/upper rad connection. Damned if I can find the link now. It also may have been just for custom engine/rad mounting.

 

Here's the link to Cometic gaskets

http://www.cometic.com/

 

I figured if the performance builders get good endurance under harsher conditions it MAY help us get more bang for the buck under normal conditions.

[RallyKeith]

I was talking about head gasket failures once with a guy I work with who drag races a very serious tube frame camaro. He said all the import guys he knows use the Cometic gaskets. Interesting thing, The EJ257 (STI 2.5 Turbo) uses the same gasket from them as a regular EJ25.....

[Distance Commuter]

As I look at the photos again, this is my observation. On the head the 2 cooling passages at the leak area have LESS than half the bearing surface than there is between any of the other passages....

 

I'm also thinking cooling passages. The pics of block near the start of this thread show a much shallower cooling passage around the failure area. The machined/milled groove is much deeper on the non-failure side.

 

Not having been inside one of these motors myself I don't want to try to guess where one could cut away more metal to create a larger passage. I'd be very wary of weakening the structure. My feeling is that a larger block casting would give everyone more to work with.

[Distance Commuter]

As I look at the photos again, this is my observation. On the head the 2 cooling passages at the leak area have LESS than half the bearing surface than there is between any of the other passages....

 

How about radiator fan switching? Most fan switching is based on readings of coolant temperature. In other words, the fan won't come on till some sensor detects a temp outside its allowed limit. If designed too close to the limit, it could allow excessive temps in some local areas - like the 2.5's failure area which sees less coolant anyway.

 

My own car has fan switching by temp sensors, PLUS fan operation triggered by the brake light circuit. That way when the car is stopped the fan is running on low speed. This gives a constant breeze over the radiator. The temp gauge is far steadier than it ever was before I added the brake switch-fan circuit. The circuit includes a delay to reduce on-off cycling due to brake pumping, stop and go driving, etc.

 

Remember, the rad fan enables the thermostat to do it's job. If the coolant coming out of the radiator is too hot the t'stat can't keep up with the engine's heat output.

 

Yes, the temp sensors are designed to meet the engine's needs - but improved cooling might just be enough to overcome the potential for failure. I've always felt that a warme3d-up engine idling without any airspeed over the radiator is building up potentially harmful heat.

 

And, that's a mod that can be added at fairly low cost. Ideally, Subaru would add switching to run the fan in park and neutral if the engine is running. I think that's a bit tougher to do as a DIY modification.

[johnceggleston]

i like the idea of the overheating or lack of proper cooling contributing to the head gasket failure. that would give me hope for the future of my engine.

 

but it begs the question, if this helps eliminate the problem, why hasn't subaru changed the cooling system. even if they didn't admit the fault in the earlier versions, they could have made changes at the phase II point. a larger cooling system / radiator wouln't have been too expensive.

 

as a non-engineer, i'm more inclined to believe the movement theory. this seems like it would be harder to correct. i don't know what it cost or what you would give up with a closed deck design, but it might be worth a shot.

 

question: some of the later 2.5 have turbos, are they closed deck? (like some(?) of the ej22 turbos) do they have the same head gasket failure rate as the open deck design? (assuming they are in the years with the problems).

 

 

How about radiator fan switching? Most fan switching is based on readings of coolant temperature. In other words, the fan won't come on till some sensor detects a temp outside its allowed limit. If designed too close to the limit, it could allow excessive temps in some local areas - like the 2.5's failure area which sees less coolant anyway.

 

My own car has fan switching by temp sensors, PLUS fan operation triggered by the brake light circuit. That way when the car is stopped the fan is running on low speed. This gives a constant breeze over the radiator. The temp gauge is far steadier than it ever was before I added the brake switch-fan circuit. The circuit includes a delay to reduce on-off cycling due to brake pumping, stop and go driving, etc.

 

Remember, the rad fan enables the thermostat to do it's job. If the coolant coming out of the radiator is too hot the t'stat can't keep up with the engine's heat output.

 

Yes, the temp sensors are designed to meet the engine's needs - but improved cooling might just be enough to overcome the potential for failure. I've always felt that a warme3d-up engine idling without any airspeed over the radiator is building up potentially harmful heat.

 

And, that's a mod that can be added at fairly low cost. Ideally, Subaru would add switching to run the fan in park and neutral if the engine is running. I think that's a bit tougher to do as a DIY modification.

[Distance Commuter]

...but it begs the question, if this helps eliminate the problem, why hasn't subaru changed the cooling system. ...as a non-engineer, i'm more inclined to believe the movement theory. this seems like it would be harder to correct...

 

Why Subaru didn't upgrade the cooling? Dunno. I think it doesn't matter.

Whatever is needed to remedy the hg leaks, they didn't do it - I think there's no value in us assuming that any solution we think of has already been thought of and properly tested by Subaru. Obviously they missed something somewhere. I kinda suspect that poor switching of elec. rad fans is responsible for troubles throughout the industry. Back in the "old days" the fans were mechanical, and spun always, even with a fan clutch. And I think there were fewer hg failures - but that's just me guessing.

 

Yes, my thoughts also. The block casting looks pretty much like a shell surrounding a series of explosions. Also as a non-engineer, it looks like it needs more metal. But as you say, that's a harder fix to add, and also harder for Subaru to add, having already designed, tested and built the engine.

 

But shucks, the fan modification should be easy to do. Testing would need a bunch of Sub owners to do the mod, and even then we'd need to hang in there for a year o so before seeing results.

[WAWalker]

Improved cooling may change the effects that heat expanssion have on the the design. But on the other hand it may also have negative effects on the efficency of the engine. Engines need to opperate at a certain temp. in order to run clean and efficent. The oil needs to get hot enough for moisture to evaporate....................Things of this nature. Also not being an engineer, I don't know how black and white or gray this area is.

 

Toyota had lots of years of recalled 3.0 head gaskets, and were running machanical fans. FWIW.

 

I have another theroy on the block design. It appears that the block castings of all the EJ engines are basicly the same. They started with 1.8L then 2.2L, then 2.5L, then there is the 2.0L (which I have only had the heads off of two). It seems that every time they went up in displacement the material around the cylinder walls got thinner. Thus less support for the cylinders in an open deck design. Thus more cylinder vibaration. More cylinder movement with the expansion and contraction.

 

I will be taking measuerments of this material, including cylinder sleeve, and posting what I find.

 

I do know that there is a LOT more "meat" around the cylinders on an EJ20 than an EJ25, not just my eyes playing tricks on me.

[Distance Commuter]

Improved cooling may change the effects that heat expanssion have on the the design. But on the other hand it may also have negative effects on the efficency of the engine. Engines need to opperate at a certain temp. in order to run clean and efficent. The oil needs to get hot enough for moisture to evaporate...

 

I, too, fear and suspect that vibration and lack of metal are hurting the engine. But anyway,...

 

Don't misunderstand me re. the cooling thing - I'm not suggesting to try run the engine colder than the intended design. I'm suggesting we try run them closer to intended temps, and with more even temps around the block, by slightly improved cooling.

 

If during idling the water coming out of the radiator isn't quite cool enough to cool the block, the temps will go up.

 

The thermostat still mixes hot and colder coolant as needed as it adjusts itself, trying to maintain the temp balance. But it can't achieve the balance if it only has hot water to work with. When the block temp goes up, the fan kicks in, and then the t'stat gets what it needs - but there was a delay. During the delay, things got hotter. "Pre-emptive" or pro-active cooling at the radiator while idling makes sure the t'stat always gets water that's been cooled somewhat.

 

Dig?

[cookie]

Yep WA, that's what happens when you stretch an engine. Their entire production line is set up to make these blocks and the cheapest thing to do is bore larger. After you have the crank with as large a stroke you can get this is the only way to get bigger. Due to smog laws the engine have to run much hotter than they used to and much leaner.

Closing the deck or O ringing would fix it but be more expensive. They seem to have been going the cheapest way which would be a gasket capable of taking movement.

This also explains why engines often fail on a long trip. You get into another whole level of heat soak then and my thoughts are that the liner moves then.

[WAWalker]

I, too, fear and suspect that vibration and lack of metal are hurting the engine. But anyway,...

 

Don't misunderstand me re. the cooling thing - I'm not suggesting to try run the engine colder than the intended design. I'm suggesting we try run them closer to intended temps, and with more even temps around the block, by slightly improved cooling.

 

If during idling the water coming out of the radiator isn't quite cool enough to cool the block, the temps will go up.

 

The thermostat still mixes hot and colder coolant as needed as it adjusts itself, trying to maintain the temp balance. But it can't achieve the balance if it only has hot water to work with. When the block temp goes up, the fan kicks in, and then the t'stat gets what it needs - but there was a delay. During the delay, things got hotter. "Pre-emptive" or pro-active cooling at the radiator while idling makes sure the t'stat always gets water that's been cooled somewhat.

 

Dig?

 

I Dig.

 

What the coolant temp senor on these engines does is it turns the fans on when the engine temp rises or the A/C is running (putting more heat in front of the radiator via the condensor). It does this. The only time the temp rises enough for this to happen is when the car is not moveing to get the nessasary air flow through the radiator. If you drive one you should have noticed this. Though setting inside the car the fans are hard to hear and if the radio or heater blower motor is running you won't hear the cooling fans. But they do come on when the engine is up to temp., car is sitting still(no air flowing through radiator), temperature does rise, cooling fans do come on. It works, I've seen it happen.

[nipper]

I Dig.

 

What the coolant temp senor on these engines does is it turns the fans on when the engine temp rises or the A/C is running (putting more heat in front of the radiator via the condensor). It does this. The only time the temp rises enough for this to happen is when the car is not moveing to get the nessasary air flow through the radiator. If you drive one you should have noticed this. Though setting inside the car the fans are hard to hear and if the radio or heater blower motor is running you won't hear the cooling fans. But they do come on when the engine is up to temp., car is sitting still(no air flowing through radiator), temperature does rise, cooling fans do come on. It works, I've seen it happen.

 

I can actually see it on the temp gauge by the slight movement of the needle in heavy traffic.

 

nipper

[Petersubaru]

Not sure? .. but I think an easier way to go would be to install a " programmable dual fan controller" from "dakotadigital.com" ...search for "pac-2000"

How about radiator fan switching? Most fan switching is based on readings of coolant temperature. In other words, the fan won't come on till some sensor detects a temp outside its allowed limit. If designed too close to the limit, it could allow excessive temps in some local areas - like the 2.5's failure area which sees less coolant anyway.

 

My own car has fan switching by temp sensors, PLUS fan operation triggered by the brake light circuit. That way when the car is stopped the fan is running on low speed. This gives a constant breeze over the radiator. The temp gauge is far steadier than it ever was before I added the brake switch-fan circuit. The circuit includes a delay to reduce on-off cycling due to brake pumping, stop and go driving, etc.

 

Remember, the rad fan enables the thermostat to do it's job. If the coolant coming out of the radiator is too hot the t'stat can't keep up with the engine's heat output.

 

Yes, the temp sensors are designed to meet the engine's needs - but improved cooling might just be enough to overcome the potential for failure. I've always felt that a warme3d-up engine idling without any airspeed over the radiator is building up potentially harmful heat.

 

And, that's a mod that can be added at fairly low cost. Ideally, Subaru would add switching to run the fan in park and neutral if the engine is running. I think that's a bit tougher to do as a DIY modification.

[Tech1967]

We have to remember not to think of this in a simple one cause kinda way. I am sure there is more than one cause all by itself that results in this failure. I'm sure some movement plays a part. Some chemical cause too. I did find out something interesting when I looked over my cooling system and will post pics and explanations tonight. Walker has a point about the walls getting thinner with each increase in displacement and that will have effects on both movement and cooling.