Several thoughts: I've experienced a broken rocker, and the broken pieces just sit in the cam housing without doing much. In my case, it seemed to be a poor part failure, it broke near the adjuster and I started missing on one cylinder. It took some troubleshooting to find out that it was the rocker, rather than ignition, or fuel.

In another case, about 30 years ago on a 2.4 engine, the shaft backed out far enough that the rocker fell off and sat in the cam housing until I opened up the valve cover. No apparent damage. So I put it back together, and drove several thousand miles further before swapping it for a 3.0 engine. From that I concluded that when the rocker starts shifting, it may get loose on one end, but the other end is well-supported by the other side of the bore, until it gets far enough over that the rocker simply gets pushed off. For the purists, I did not mic the inside of the rocker bushing to see if it had been rounded out or worn excessively, I just visually looked it over. Maybe there was some damage to the bushing, but I put it back together, ran it for several thousand miles, and I don't recall having to make multiple valve adjustments.

In your case, the shaft moved to the side that has the thicker bore, and would have been well supported as a cantilever. You also said you were able to put the rocker and shaft back into the bore and they went smoothly. Therefore, I really doubt that the rocker shaft moving to one side caused the failure, or that the rocker had anything to do with the failure. So don't kick yourself for not catching the shaft earlier. I'll bet you had an oil leak on that side of the engine too.

Hint for everyone else. If you fairly quickly develop an oil leak from the area of a rocker shaft, check to make sure it didn't loosen up.

That level of damage to a piston usually does not come from a dropped valve, at least not initially. That kind of damage occurs when the piston fatigue fails, or the valve end breaks off and gets jammed sideways into the piston, breaking it, then the broken piston parts get thrashed around and smashed further. Even after that, the bottom of the piston and wrist pin boss remain on the wrist pin. To me, a key sign is the broken rocker at the adjusting nut. I think that was caused by the piston hitting the valve while it was still open against the cam lobe. In most cases, this would be caused by the cam losing timing, but then we would expect to see damage on multiple valves and rockers, instead of just one. The failure was confined to one cylinder on one bank, right?

Piston failure: This I think is the most probable hypothesis. Pistons that hit the broken-off valve end do not typically fracture into a bunch of pieces like yours did--even the meat of the piston around the wrist pin broke off. Instead, I suspect that the piston fatigue cracked and failed and separated into several parts that got crammed up against the head and valve, causing the rocker to break (as designed), and successive cycles beat the remaining valve stem sideways, breaking the boss in the port that holds the valve guide. Running the engine to high RPM just before the failure may have been the "last straw" that caused the piston to break, but it would have failed soon anyway.

NOTE: Carefully inspect the breaks in the piston parts to identify whether they were cracked from many cycles of fatigue, or from one big impact. Maybe get a mechanical engineer friend of yours to look at them. Pictures here don't have the resolution to diagnose from afar.

Valve failure: Another hypothesis is that the valve head broke off the stem, fell into the cylinder and got hit by the piston, breaking the piston into several pieces, and then causing the rest of the damage. Valves breaking off are infrequent, but not rare failures. It's not uncommon on racing engines that get thrashed at high RPM. But if that happened, why would the rocker have broken? Thus I rate this as a lower probability hypothesis.

More to think about.

Thanks for the insightful analysis. To answer a few questions directly.

"but then we would expect to see damage on multiple valves and rockers, instead of just one. The failure was confined to one cylinder on one bank, right?"

The cylinder head on #1 and #2 show touches from those valves. Here is #1 and #2, respectively:
http://forums.pelicanparts.com/uploa...1752092633.jpg
http://forums.pelicanparts.com/uploa...1752092633.jpg
However these ones did not fail - their rockers were also fine.

But #3 was the only one significantly damaged. Also the spark plug was smacked very hard, so hard that its wire flew off the back.

http://forums.pelicanparts.com/uploa...1752092772.jpg
http://forums.pelicanparts.com/uploa...1752092772.jpg

With regards to the progression of the failure event - it started with a loud banging for several seconds. Then it lost power. So maybe the piston disintegrated, but the engine continued to cycle until the cam jammed and locked everything down? However when I was disassembling, the crankshaft was still jammed even after I removed the camshaft sprocket. It loosened up without obvious reason later during removal of the crankshaft housing and cylinder headers. So maybe there was a secondary jam.

Once I have the crankshaft open (hopefully before the weekend's camping trip) and all damage is revealed, I will do a careful analysis of each failed component in the #3 cylinder neighborhood and bank 1 more broadly.

I switched the yoke to the other side, and while I'm waiting for my triple-square socket wrench bits, I laid out all my cylinder 3 goodies and took better pictures.

PelicanParts hosts at a lower resolution, so here is a gallery with all of the pictures in high res: https://imgur.com/a/vR29rU6

Hold down Ctrl on your keyboard and you can then use your mouse wheel to zoom in and out on these pictures.

Cylinder 3 Cylinder Head
https://i.imgur.com/aYgG6NA.jpeg
https://i.imgur.com/6f1l1Lr.jpeg
https://i.imgur.com/DwX0XnO.jpeg
https://i.imgur.com/AXgr7js.jpeg

Valve Remaining Parts
https://i.imgur.com/MRLmkg2.jpeg
https://i.imgur.com/WQcckwT.jpeg
https://i.imgur.com/TZ6kzkc.jpeg
https://i.imgur.com/ufswpSo.jpeg

Here is the cylinder and rocker arm assembly for 3.
Again here you'll need to hold down Ctrl on your keyboard and use your mouse scroll wheel to adjust the zoom level.

https://i.imgur.com/fCuEO6y.jpeg
https://i.imgur.com/AMzhMSE.jpeg

The rocker arm shaft did not have any scoring perceptible to touch.
https://i.imgur.com/MbL6yoE.jpeg
https://i.imgur.com/7qRWuqP.jpeg
https://i.imgur.com/0WkP9gH.jpeg
https://i.imgur.com/2FH3LWR.jpeg

Finally, here is the piston, kind of reassembled.
Again here you'll need to hold down Ctrl on your keyboard and use your mouse scroll wheel to adjust the zoom level.

As another poster noted earlier, it appears as if the piston was killed by a strike from the valve - note the center blown out and the scarring around the impact point.

https://i.imgur.com/bmMStuG.jpeg
https://i.imgur.com/KsmWZe4.jpeg
https://i.imgur.com/5cKwt4u.jpeg

Some of the unmatched bits. I'm not sure if I'm seeing stress breaks or sudden shock breaks.
https://i.imgur.com/jJ2gvR3.jpeg
https://i.imgur.com/9PTMUmJ.jpeg
https://i.imgur.com/U773OoD.jpeg
https://i.imgur.com/5XBYKLn.jpeg
https://i.imgur.com/pPTcsMz.jpeg
https://i.imgur.com/yqkLns9.jpeg

Ascertaining the exact cause of this failure will prove to be a fools errand.
Two things are true about this engine based on the pictures and story supplied.
First: the initial "rebuild" at 200K was a low quality effort.
Second: This engine suffered far more damage than necessary by driving it after the initial failure.

What are some things from the first rebuild that were done wrong that I can do better?

That is a longer conversation but new parts when appropriate is a good place to start.
Bloviating about "why" is a giant waste of time.

Pictures are deceiving but even the exhaust valve on the exploded head looks to be burned. That would suggest this was a 5 cylinder engine even before the catastrophic event occurred.

As I'm almost done teardown, I'm looking around for Porsche shops in the Milwaukee area to do the assorted machining work, cleaning, precision measurement, etc.

Any recommendations?

Hi
Did you find the Valve Guide piece(s) ?
Photos ?
Cheers

Is it one of these? This is most of the small stuff I've fished out. I'm at the "tap here and there" stage, trying to split the crankcase.

http://forums.pelicanparts.com/uploa...1752195791.jpg
http://forums.pelicanparts.com/uploa...1752195791.jpg

Someone experienced in metal fatigue and failure analysis will be able to tell you whether the piston failed first due to fatigue cracking that eventually caused it to scatter, vs. it being beaten to death by the valve and failing.

I don’t know of any in Milwaukee, but Craig, cgarr on here did my heads and rockers and I was very happy, not hard to send them, just bolt to a piece of plywood or individually box and box the set up.

Here’s a thread. http://forums.pelicanparts.com/911-engine-rebuilding-forum/908017-pelican-who-does-head-work.html

I messaged Craig, likely will use him for my top end work. Am hoping to work with a reputable Porsche team down here (Joe's Garage) to identify what is needed in the bottom end.

I'm going to need 1 new piston, which I suppose requires me to get a totally new set of 6 pistons, which I suppose requires me to get a totally new set of cylinders.
Is that right?

Or can I get my existing cylinders cleaned up and matched to new pistons? One cylinder is scored, but I'm not sure if there is some kind of lining or what.

See whether you have Alison or nikasil cylinders. Alusil have a mixed reputation for mixing and matching parts, and even re-ringing.

These cylinders are 100% Alusil and if the original pistons were used in the "rebuild" at 200K, they are cast not forged.
Extending the life of a cast piston to 280K is begging for disaster.
This is one of the challenges I offered about to 200K rebuild. I suspect that the 200K pistons and cylinders were reused in the "rebuild".
Looking at the whole pictures suggest the a valve retainer broke or a piston cracked.
One again, fixing blame is a waste of time. Evaluating what is usable for the next rebuild is paramount.

The pistons are done. If you want, you can have you existing cylinders Nikasil coated and then install a new set of forged pistons.
If you are so inclined you could have the cylinders bored to 97mm and have custom pistons made to suit your needs. At 97mm you can use Goetze 930 Turbo rings to finish a very nice conversion.
If it turns out you have damaged to rod, forgo the arduous effort to match the rod. Source a new set of rods. 911 SC rods are readily available and not very expensive.

Thanks for the detailed analysis Henry! I dug through the records and found the pistons are pistons and clyinders were replaced as a set during the rebuild, 9.3:1 "KS".
http://forums.pelicanparts.com/uploa...1752345628.png

Number 3 cylinder has some ugly gashes in it, not sure if a larger bore would solve that issue. Is it possible to use 5 existing cylinders plus a new one?

I'm thinking about getting the heads, cams, and all that top stuff up to Craig Garrett at G2 right away so I can get in line.

Using the 5 existing cylinders is completely dependent on how they measure out for diameter, taper and oval. If they are all within acceptable tolerance and don't have any defects that will catch a fingernail they can be re-used IMHO. Finding a single "in spec" Alusil cylinder of the same height and tolerance group should be pretty straightforward. When I did the top end on my '78SC I took the existing P&C and a second set of 6 used Alusil P&C and mixed/matched a set of P&C that were all within the dimensions given in the spec book as "in tolerance." The tech who did all the measuring (and rebuilt the heads) recommended using a silicon ball hone with only 4-6 passes in each cylinder to prepare them for new rings. This was based on feedback he received years ago from a tech who works for a long-established 911 race team/performance shop. (YMMV) The Goetze rings I used seated quickly in the ball honed cylinders and so far (touch wood) I'm very happy with how things came together. If you have deeper pockets then having Alusil cylinders bored and Nikasil plated for new forged pistons is the way to go IMHO.

The darkened area inside the valve scallop and corresponding area on the stem itself would seem to indicate a preexisting failure point/crack.


http://forums.pelicanparts.com/uploa...1752419846.jpg