Mystery oil leak

[merc81]

Here is a real mystery for those readers that like a good challenge.

The van is a 1999 Olds silhouette, with all options but AWD.

41 months ago, I changed out the engine for a rebuild that came with 36 months, 100k warranty. Engine always made some rattling noise at idle, nothing serious, just louder than original engine ever was. Rebuild company (Hessco) told me it was common, not a problem. The engine worked flawlessly for over 3 years.
For about 18 months, maybe more, the low coolant light came on randomly, but never was it low on water. The overflow tank always showed about the same level, and engine never ran hot. I took it to Florida a couple times, it has pulled a boat, and always works well.

Recently its begun to lose a little oil, but not dripping anything--floor is clean where it parks, not a drop of anything. I was unable to see or smell anything in the exhaust that might be oil burning. Then, after a long trip, a transmission coolant pipe burst at the junction where I had added in an external tranny radiator. This was an easy fix, and I didn't get too concerned about it; that pipe connection had more than 100k miles on it, so I thought it was just worn from vibration. I replaced it, topped of the tranny and it was good as new.
While I was in there, I checked the overflow bottle as I always do and found it completely contaminated with oil; brown chocolate milk like stuff that was found all throughout the radiator too.
I flushed it all out, added new coolant and ran the engine for a few hours in the garage (yes I have an exhaust hose). The system seemed OK, with no oil leaks I could find. There was no water in the oil at all, and the oil was always very clean, it would have been easy to spot any gunk in there.
I took it for a ride of about 3 miles and the problem showed up right away by lighting up the low coolant light again.

I've done many tests, tried lots of things, and at this point I've found nobody that can explain to me a mechanism by which oil can pass into a fully pressurized coolant system. Before you say that its the LIM gaskets, think again—oil is not under pressure on the heads, its only squirted over the rockers via the push rods. Even with a completely blown LIM, oil could not enter into the coolant.

I thought it would be an interesting puzzle for this group. Let me hear your ideas and I'll tell you what I did and what the results were.
. . .

[lesterl]

Oil cooler in the radiator.....

[merc81]

Hmmm.
Sorry, there is no oil cooler in the radiator. The transmission lines enter the radiator on the side below the cap and level sensor, but no engine lines enter the system. I pressure tested the transmission lines anyway and they held 35psi for days with no leak down.

[j cAT]

Originally Posted by merc81

Hmmm.
Sorry, there is no oil cooler in the radiator. The transmission lines enter the radiator on the side below the cap and level sensor, but no engine lines enter the system. I pressure tested the transmission lines anyway and they held 35psi for days with no leak down.

I would put a plastic bag over the coolant cap and rev engine to see if you have exhaust getting into the coolant system.

cracked head / gasket issue.

if the coolant got into the transmission you would know by now.

[merc81]

That's the weird thing. There is nothing wrong with engine performance at all. The coolant system gets lots of oil in it, but only after it gets warmed up. About a half a quart per week gets lost in the coolant. Nothing that would indicate a head gasket problem such as back pressure into the coolant or water in the exhaust. Engine temps are good always.

I pressure tested the cylinders both for peak and leak down. The avg. pressure was 210 psi, leak down test was good, held for as long as I cared to wait for it. All 6 cylinders were good, all plugs showed even, good burn.

I'm posting this for interest of the forum members. Frankly I'm ready to replace the engine at this point. I'd love to hear your thoughts on what I should've, could've done. Maybe I missed something or came to a bad conclusion.

[j cAT]

Originally Posted by merc81

That's the weird thing. There is nothing wrong with engine performance at all. The coolant system gets lots of oil in it, but only after it gets warmed up. About a half a quart per week gets lost in the coolant. Nothing that would indicate a head gasket problem such as back pressure into the coolant or water in the exhaust. Engine temps are good always.

I pressure tested the cylinders both for peak and leak down. The avg. pressure was 210 psi, leak down test was good, held for as long as I cared to wait for it. All 6 cylinders were good, all plugs showed even, good burn.

I'm posting this for interest of the forum members. Frankly I'm ready to replace the engine at this point. I'd love to hear your thoughts on what I should've, could've done. Maybe I missed something or came to a bad conclusion.

so how did the plastic bag trick work out ?

[sverker]

I would say it's the oil cooler. It isn't located in the radiator but at the oil filter. A friend of mine had similar problem on his and that was the problem.

Sent from my HTC Desire HD A9191 using Tapatalk

[sverker]

Regarding the low coolant light, if you've verified that you indeed don't have low level then it's most likely the sensor that is full of gunk. It's located on the same side as the filler cap.

Sent from my HTC Desire HD A9191 using Tapatalk

[Tech II]

If you have oil in the coolant, then the coolant level sensor is contaminated if the level is ok.....changing the sensor at this point won't make a difference, it will jst become contaminated again...

I agree with Jcat.....head/head gasket problem....

[merc81]

Regarding the coolant sensor possibly needing replaced, I considered that also. Gunk gets all over the probe if the head gaskets leak or oil otherwise gets into the coolant. This means a flashing low coolant light is a pretty good indicator that something is wrong with the engine if the coolant isn't actually low. There is a TSB on this issue (http://cemrweb.cemr.wvu.edu/~mathews...tsbCoolant.txt ).
So, the coolant light tells me when oil enters the coolant-works pretty good for that it seems

I used a pressure gauge on the radiator with the engine running and found no leaks or fluctuations in pressure like a head gasket would cause. I never tried a plastic bag, but the gauge would have showed any fluttering or pressure changes in the system. No indications of any sort of gasket leak I was experienced with.
No matter how I thought about this though, it just seemed like it had to be a LIM gasket or a head gasket leak. I just couldn't see how that would cause oil to flow into the coolant like it was. A head gasket leak would let exhaust gas into the system (and pressure) but that wasn't apparent and there is no way oil could flow into the coolant system due to a head gasket failure since there is no oil in the cylinder in quantities like I was seeing (half quart/week or about 300 miles). Also the gas mileage was still good, and the engine performs very well as always.

[merc81]

I like the idea that its the oil cooler doughnut thing under the oil fillter. I haven't figured out a way to test that yet though.

I couldn't come up with any other explanations either so I tore down the engine and replaced all the gaskets. I used the expensive felpro gaskets for the LIM when installing things back and all new bolts for heads and manifolds all around.
I could find no evidence of leaks anywhere in any of the gaskets, at the heads or the manifold.
By the way, I did observe that the O ring seal under the intake throttle body that seals the steel coolant pipe out to the heater was leaking and should have been replaced; you don't get that O ring in a lot of gasket kits and you can't change it unless you have the upper intake manifold off everything else out of the way--just a heads up there, I think that leaks on a lot of these engines.

Anyway, none of this changed the oil leak. It starts after about half an hour of highway driving as indicated by the low coolant light. No other effects can be noticed, just the oil in the radiator and after a few days, the low oil light flashes when going around sharp turns.

[sverker]

So you already know the answer? Well, it will be intresting to figure out your challenge...


Couldn't you just pull the coolant lines and either plug them or bypass the oil cooler? Pretty hard to reach them though.



One of the lines goes up to the thermostat bypass pipe but I don't see where the other goes. Directly into the block maybe?

That's the only place, except for cracks in the block, where I see you would have enough oil pressure to push it into the coolant. The leak must also act as a pressure valve to only be open in one direction, otherwise you would have coolant in the oil as soon as you turn the engine off when hot.

As you already noted that could not happen at the LIM gasket and I doubt the head gasket could cause it neither.

[merc81]

No, I don't know the answer yet. I'm pretty sure what the problem is, but I don't really want to hear it. I'm hoping someone else will come to the same conclusion I did.

I thought the idea about the oil cooler was pretty good, I'd considered it before, but decided that wasn't very likely. Last night I pulled the cooler off the block and rigged up two hoses on the water outlets of the oil cooler. I put the whole thing down in a bucket of water and pumped it up to 50psi. No bubbles, no leaks, nothing. As far as I know, its still at 50 psi since I left the thing in the bucket when I left.

Getting current status of the engine, I installed the new gaskets as I said but the leak came back on the first highway trip. I've pulled the engine for a 2nd time and removed the heads (again) and taken them to a good shop. The shop tells me that they find nothing wrong with the heads in any way and there is nothing that could cause the heads to put oil in water jacket. I've still got to go back and pick them up, but at this point I've ordered a new (rebuilt) engine for the van.

I'm still trying to determine just how the oil gets into the water, and I can't figure out a way to find that pathway. Any ideas?

[sverker]

I see.. So you are suspecting cracks in the block then?

The behaviour you describe is that you need to reach a certain temperatura and/or pressure to "activate" the leak. Also you've described that you only get contamination oil -> coolant.

That means that it's not just a simple crack, since then as soon as you turn off the engine while hot you would get coolant in the oil since the oil pressure goes away but you still have pressure in the cooling system.

That means that you are not re-creating the same circumstances in your test. You put pressure only on the coolant side and is that water at operating temperature? Also with pure water you have the surface tensions that the additives takes away. It could very well be that the cold water pressure in your test helps keeping the leak closed, if it is there it is.

You would need oil flow at operating pressure and temperature and maybe coolant on the other side at the right temperature and pressure to recreate the same environment.

Would it be possible to bypass the coolant lines past the oil cooler and do a test drive? That or replacing it with another unit.

I just can't see anywhere else, unless it's the block, where you could get this behavior.

Sent from my HTC Desire HD A9191 using Tapatalk

[merc81]

sverker, you are thinking the thing through on the same path I did. I can't test on the road anymore since the engine cradle is on the floor and the heads are at the shop, but I am pretty sure the leak must be in the block somehow.

I don't know exactly what path the oil passage takes to reach the cam shaft, but I'm thinking there is a crack near the oil passage and a cylinder. I did pull the front cover off and look over the timing chain and water pump passages, but even though that area looked pretty bad (possible leaks), I cleaned it off very carefully and sealed with new gaskets but the problem came back on the first highway trip.

Maybe its the oil pump shaft seal somehow?
Changed that too, but no luck.