Pulley End Leak Repair

[patkeefe]

I decided best to start a new thread, instead of continuing this one:
Pulley end oil seal

Pulley end leaks out the #8 bearing. Not the seal, but likely the internal o=ring. A substantial amount of work to repair. Since I have a spare engine, I figured I'd give this one try at repair. This is a substantial leak. not the common slow drip.

I stripped the rear of the engine in place in the car, and cleaned the case-nose bearing interface with sandpaper (a very small sanding block, sandpaper, brake cleaner and acetone on a toothbrush. It is pretty clean.



I put some schmutz (in this case, Black RTV, the oil proof stuff). I believe the only pressure driving this is crankcase pressure, so it is fairly low.





Then, I installed the little ring I fabricated from 6061 T3 plate. I made it a slip fit, so it has about 0.002" clearance around it. I drilled and tapped 3 10-24 set screws to retain the ring securely, and mounted it. Squished the RTV up between the case and the ring. I am letting it sit for 24 hours now. So, we will see if this works tomorrow morning.



I am not holding out much hope for this repair, but who knows.

[donbecker1234]

Hi Patrick, thanks for posting the step by step here of the fix, I hope it works out for you.

Would you mind posting the dimensions of the ring? I may try your fix if I end up having the issue myself.

[faapgar]

After assembly of the case just put a bead of red hi temp silicone around that point.The o-ring there is main oil galley pressure and todays seals are not as good.But you have done a novel repair.Good luck

[boyt911sc]

Pat,

That is the old aluminum collar-RTV trick made popular by Tom Amon from California many years ago. It looks ugly but does a good job keeping the oil from leaking. As you know, replacing the nose bearing requires engine tear down and could be avoided by doing this trick. Good job.

Tony

[patkeefe]

Don, the ring is 2.367" ID, which is the critical dimension . I think it is nominally 3" OD and 0.375 thick.

Tony, yes, I had heard about the Tom Amon trick, I believe Henry Schmidt has one also. I did find a thread by fishcop where he did the same thing, using a different sealant. I guess no one will ever see it with the pulley over it

[boyt911sc]

Originally Posted by patkeefe

Don, the ring is 2.367" ID, which is the critical dimension . I think it is nominally 3" OD and 0.375 thick.

Tony, yes, I had heard about the Tom Amon trick, I believe Henry Schmidt has one also. I did find a thread by fishcop where he did the same thing, using a different sealant. I guess no one will ever see it with the pulley over it

Pat,

I think using the right sealant is more important or critical than the machined collar. If the “sealant” could not prevent the oil seepage, the collar will not help at all. But installation of a collar makes it look very professional.

BTW, you were not having a good time with the oil seals. Last time it was the rear oil seal and later the front? Well, it is nice to be retired and have all the time in the world to get the project/s done later on. Stay safe.

Tony

[Walt Fricke]

Once I saw that Patrick had a lathe, and thus some machinist skills, I knew this would be duck soup for him. Don, why did you feel the need to ask? You must have at least roughly equivalent skills, and the one dimension which seems important - the ID - is easy to calculate on the nose bearing. I note from reading the old posts someone less lazy then me pointed to that one design uses a press fit tight clearance, others don't. To me it seems the ring's function is to take the outward pressure of the oil - using just adhesive strength to do that isn't apt to work if the pressure is more than just from capillary action or crank case pressure.

I've wondered about the pressure. So I got out a spare #8 bearing to look at. The regular crank seal faces the pressure of the oil which exits rear race of the bearing. That oil was at, say, 60 psi going in. But the space behind the seal has two large diameter drain holes back to the crankcase space. At the crank case, that oil would be at case pressure. But how does pressure decrease along the way, from a high were it exits the bearing?

The outer seal is different. Like the main bearings, the joint between the case oil gallery to the #8 is sealed only by the tightness of the metal to metal joint. And here it is not a crush seal. A thin piece of the #8 separates the incoming 60 psi oil from the O ring space. The part toward the crank case void is much wider. Oil which gets into the O ring space is supposed to exit via a slot across the bearing. If case sealant were to get into this slot - and harden - that would raise the pressure on the O ring, perhaps up to the 60 psi hypothetical. But otherwise, the slot, while not large, looks like it should be able to handle what works its way in from where the case clamps around the #8's OD. Which suggests to a mind which, alas, was not trained in fluid dynamics that the pressue on the O ring ought not to be high.

The guy who taught me engine assembly advised putting the orange Loctite case sealer on the outer raised lip of the #8 before bolting the case together. It makes final assembly a bit tricky, as you have to raise up the crank which has been sitting there peacefully awaiting der Tag so you can get the bearing off, apply sealant to the concealed lower part - but not too much - and then get the bearing back in place without smearing sealant where you don't want it, or removing it from where you do want it.

I haven't had this leak, knock on wood, but a dozen or so assemblies probably isn't all that statistically significant.

Still, it seems clear to me that the more pressure you can put on any external sealant, or suitable gasket, the better chance you have of it resisting whatever pressure is behind the oil trying to escape in that area.

An advantage of a suitable gasket is that its two sides are of uniform thickness to begin with, so the pressure is evenly distributed. Applying sealant which cures, then compressing it, probably means that the pressures on the sealant are variable from place to place, allowing the oil to find the weakest spot to try to escape. To be sure, the sealant should take care of surface imperfections. So maybe the best is both - coat the gasket with a thin sealant?

I sympathize with those who have had this problem. You work hard to do everything right, especially if this isn't new to you. And still you get a leak.

I built a tray which bolted below my RMS on one of my motors to catch the oil I had failed, despite two seal replacements, to catch the dripping. Sort of half assed worked.

So fingers crossed here.

[donbecker1234]

Hi Walt,

I just thought if Pat was currently there and had the number, taking notes just in case my future looks similar.

[patkeefe]

Gents, without reviewing the Porsche dry sump system, I would surmise that the o-ring around the outer diameter of the #8 bearing has only crankcase pressure on it. The pulley end oil seal, much the same as the flywheel end seal has a somewhat higher pressure, but this pressure is mitigated by the fact that it is no longer contained in an oil galley, and is free to expand volumetrically speaking. The pressure side of the pump forces oil into the galleys, where it lubricates the bearings and other parts, and when it has finished that job, drains back to the sump.

Crankcase pressure is order of magnitude lower than pressure galley pressures, unless there are extenuating circumstances. I once broke up a piston due to detonation at the track, crankcase pressure went through the roof, and blew out the flywheel end seal, causing massive amounts of oil smoke to enter the cockpit. I kind of learned the hard way about excessive crankcase pressure.

The fact that I built this long block 10 years ago, left it wrapped up in the basement, and never ran it likely contributed to the demise of the rubber seals. I don't remember how I assembled it, but it was probably whatever Wayne's book said. I took a chance sticking it in the car in 2020. Good news is that I wasn't playing Russian Roulette. If this doesn't work. I put the 3.0L back in, sans turbo. As Tony alluded to, I have time on my hands.

BTW, here is the oil system:

[patkeefe]

Small miracle...it stopped leaking. Another Band-Aid successfully applied.

EDIT: Well, upon further review, it is mostly stopped. It still weeps a bit. Regardless, I have a low confidence level in this repair. I suppose if I had no choice, I'd call it good, but I'm going to consider the repair. I have to go away for a couple of months starting in June, and I want to leave the car drivable for the wife. I'll instead spend June rehabbing the 64 VW Type 261 for my daughter.

EDIT #2:
I waited for the engine to cool down, and restarted it. It is again leaking; not as badly, but still leaking. Thus. I will consider this Band-Aid as a failure. Guess I'll put the 3.0L back in.

[patkeefe]

Originally Posted by boyt911sc

Pat,

I think using the right sealant is more important or critical than the machined collar. If the “sealant” could not prevent the oil seepage, the collar will not help at all. But installation of a collar makes it look very professional.

BTW, you were not having a good time with the oil seals. Last time it was the rear oil seal and later the front? Well, it is nice to be retired and have all the time in the world to get the project/s done later on. Stay safe.

Tony

Tony,
I would guess that a leak is basically a force driving a mass (oil in this case). The pressure on the inside the case (the force) is greater than the outside. Hence, a leak. The collar offers a greater resistive force than sealant alone can provide. I used sealant because it will conform to the irregularities of the surface better than anything else I have lying around. I would have used some of that high dollar sealer, if I had any. Any port in a storm, as they say. Being retired also means fixed income
Pat

[Walt Fricke]

Are we to conclude from all this that while the crank seal, part of which is exposed to the atmosphere and deals with a moving shaft, didn't fail due to age and non-use, the O ring, immobile and sequestered largely if not entirely from the vagaries of the storage atmosphere, did?

I'd guess you used the thicker O ring? That's been recommended for quite a long time, I think.

All the more mysterious - and irritating.

[patkeefe]

You and I have concluded pretty much the same thing, Walt. Go figure.
I already have the engine back out. Gonna do the 3.0 swap pretty soon.

[boyt911sc]

Originally Posted by Walt Fricke

Are we to conclude from all this that while the crank seal, part of which is exposed to the atmosphere and deals with a moving shaft, didn't fail due to age and non-use, the O ring, immobile and sequestered largely if not entirely from the vagaries of the storage atmosphere, did?

I'd guess you used the thicker O ring? That's been recommended for quite a long time, I think.

All the more mysterious - and irritating.

Walt,

Was there an issue in the nose bearing O-ring spec? So far, I have not encountered this oil leak problem for my engine rebuilds. I test run the motors for several hours before turning them over. Plus the crankcase sealants available today are really performing well in preventing those pesky oil leaks.

Tony

[Walt Fricke]

Tony
The #8 bearing is, I believe, the same from the first 911s through at least the end of the 3.2s? Using the '74-89 parts booklet, from 74-77 the O ring was 50x4.0mm. From 78-89 it is 51x4.5mm. Someone (Anderson, maybe?) at some point advised all of us to use the 51x4.5mm O ring. I sort of think the change came not at the start of the 930 cases, but somewhere along the line.

One supposes that, if a seal kit is purchased for a 3.0 or 3.2, it would come with the larger ring. If a cheap guy like me buys a reduced set of seals, and reuses on his SC something left over from a 2.7, he could end up with the smaller ring.

We can all guess why Porsche increased the O ring size.

I've never had a leak in this area, but lots of others have. I have no reason to believe my engine assembly skills are somehow superior. This is a baffling area.

[boyt911sc]

Walt,

I never experienced this problem simply because I refrained from magnesium crankcase rebuilds. Finding the correct type of sealant for Pat’s leaking nose bearing is a good band-aid fix. I’ve seen motors with a blob of RTV and not leaking.

While I have your attention, I like to learn how to use an oscilloscope. Someone gave me an old professional grade oscilloscope and I have a $200 brand new one. Unfortunately, I have no idea how to use it. I simply like to read a square wave signal. Could I contact you for tutoring? Thanks.

Tony

[patkeefe]

Tony. there are oscilloscope tutorials on YouTube. I got rid of both of mine.

Lastly, my case cold be leaking just because it has bad seams, Magnesium. I can understand why they went to aluminum castings.

[boyt911sc]

Pat,

I will try it. I am old and not computer savvy. My grandchildren are more advanced and proficient than me in this subject. But willing to listen and learn. Thanks.

Tony

[Walt Fricke]

Tony - U-Tube stuff sure to be better than me as a guide. I'm half arsed self taught, and do stuff in fits and spurts, forgetting in between and relearning as need arises.

I have a monster CRT scope which is about 60 years old (doesn't have circuit boards inside, much less integrated circuits), doesn't always work, and probably should be in a technology museum. I have an inexpensive hand held digital, which is just about good enough for anything I might use it for. And an inexpensive computer program based scope, which works nicely when I remember how to use its controls. And I haven't used any of these in two years, since I was confirming what you see for tachometer signals, and CIS FV control signals.

These tutorials doubtless will teach you about the controls. You set a voltage (height of trace) scale, a time scale (the X horizontal axis), and a trigger voltage level (tends to mess me up). You also select the trigger type. Accuracy used to depend on the scope's time base. At one time that would be a rather large crystal module with a thermostatically controlled heater. Digital technology seems to have gotten around that somehow. But for the kind of checking we are apt to do to see if things are as they are supposed to be, lab grade accuracy isn't necessary.

With your engineer trained mind you will promptly know more than I ever did by a little study rather than just futzing around.

The trick these days isn't so much the scope, but reading the coded messages the microprocessors on the new cars send back and forth to each other on a bus. But our CIS systems don't have any of that. Just frequencies and pulse width modulation.

[patkeefe]

https://www.google.com/search?q=youtube+oscilloscope+tutorial&oq=youtube+oscilloscope&aqs=chrome.2.69i57j0l7.9677j0j8&sourceid=chrome&ie=UTF-8#kpvalbx=_PtHFXujNEYGe_QbtrZnYCQ26