The heads obviously fill on the 16 valve engines...go back and look at threads that talk about blown cam carrier gaskets, after running these engines at high rpms, for extended periods of time.

There are significant differences, besides the obvious difference in the number of valves.

The volume of the heads (how much oil can be stored in the heads) is obviously much smaller. There are no vents in the cam carriers, so the engine doesn't "vent" through the heads. Certainly, the "smaller" stroke doesn't allow for as much oil "whipping" around inside the crankcase and the fact that these engines do not make as much power, in the upper rpms, certainly means they get shifted at a lower rpm.

While certainly the crankcase is bound to get pressure, like the 32 valve engines do, what does that hurt, as long as the oil sump doesn't run low on oil? The heads fill up with oil and any additional oil is simply "pushed" back to the crankcase through the head oil return paths.

Does that mean that these engines don't get lowered oil volumes in the oil pan and not suffer from oil starvation at high rpms?

Certainly not. It simply means it isn't as severe or frequent as on a 32 valve. The smart racer would keep a really close eye on the oil filter and on frequent "oil reports" and change the rod bearings at the first sign of "problems".

If I was running one of these engines on race courses that had frequent shifting (and thus lowering of the rpms) and I only ran the engine up to 5,500 or even 6,000 rpms, I think a pan spacer and perhaps the early style oil pick-up and screen might be sufficient....along with frequent oil and filter testing.

If I was going to go try and run one of these engines at 6,000+ rpms at WOT, for extended periods of time (like what happens on open road racing courses or at events like Bonneville), it would get an oil return system hooked up to the heads.

Blowing up an engine at high rpms and at high speeds is not going to ever turn out good.

 

Couldn't agree more...... I'd say an accusump is a good idea....

The 16V USA models only make noise past 5500rpm anyway....yet another reason not to spin big rpm....

I will press on with my "experiment" of seeing how long a $300 used 4.7L engine can last with only a pan spacer and OB pan..... 51 hours and counting....

 

Numerous cases of S/S2 engines failing in Europe on track. They seem to survive better than S4/GT/GTS until you put R-compound or stickier rubber.

The minute you add serious grip, stock engines fail. Simple as that. We had 3 928S (Euro 300Bhp) fail within weeks of each other in the historics before crankscrapers and accusumps were installed. And we are not talking 100k+ miles ****ters here, but freshly rebuilt engines... Similar problems in Germany too...

 

Rob and Greg, thank you for sharing. Excellent work!

 

That makes sense. If you read this entire thread, which is very long, you will see that I had oil pressure issues at 5,500 rpms with the engine on a steady state dyno test. Sure, this took over a minute of running to push enough oil into the heads to "starve" the oil pick-up, but regardless, it showed exactly where the oil was going and what was happening. Finally, after all these years of everyone blowing up 928 engines, we have some absolutely solid testing that shows where the basic problem lies....and one way to fix this problem.

If you add sticky tires to the car and throw the oil to the sides of the pan, up the sides of the block, and also "pack" the oil into the heads....you're going to see the problem even faster than if the engine was sitting still, on a dyno.

Certainly, anything that you can do to keep the oil in the pan, while you go around corners, is going to help (windage trays). Certainly, increasing the oil capacity with an Accusump is a good thing to do.

The basic underlying problem is going to remain. Eventually, if you improve the handling of the vehicle to the point where you are going faster and faster (which means you are turning more rpms for a longer period of time), you will push more oil into the heads and then this problem shows up again....I've seen this happen, for years.

As I've discussed, even dry sumped engines show this happening...although we didn't realize what we were seeing, when we saw it happen. Dry sumped engines will "suck" the level down in the tank, when run 5,000 rpms and above. The oil is being stored in the heads.

The obvious solution is to dry sump the pan and dry sump the heads, with a multiple stage "suction" pump....and I doubt I would ever attempt to run a track engine without doing this, again.

However, on a nice "street" driven vehicle, dry sumping is, at best, extrememly difficult. Finding room for a tank that is large enough to hold the required amount of oil and then plumbing that tank to the pump, the pan, and the heads is very difficult. Making that all fit neatly in a pure street vehicle is...cumbersome. (Very difficult/cumbersome means that it is possbile and I am working on it...but that takes time and money.)

The immediate relief, for street vehicles that are run at frequent high rpms and for some "race" customers, is to install the oil removal system. We know that you can them run the engine at above 5,500 rpms and not have oil "packing" into the heads.

 

Have a few minutes and recently purchased the workshop manual to see some better views of parts I do not have available to inspect directly.

The oil entering the oil feed annulus for the lifters is often highly aerated. When the oil bleeds down to the top and bottom edges of the lifters the dissolved and/or entrained air will create a foam with lower bulk density than neat oil; aerated oil will also exit the cam bearings. This will flood both the intake and exhaust valve seals in extreme conditions. Oil will be consumed through the intake valves by being drawn in.

 

When I ran my 2 valver on the dyno and at full revs there has never been any smoke, not even a whisper of it. The engine has a new bore and Kevin's scrapers. After about 2000 miles there was roughly 1 ml of oil in the u bend. It would have done at least a dozen dyno pulls. it hasn't been used on the track though.

 

It was well established in the early 1970s (and 1930s and 40s and surely earlier) that aeration levels of oil rise with rpm (and other parameters). See, for example, published SAE research from Mercedes Benz in 750051 on their V8 concurrent with the development of the 928 V8. Since Porsche (and Mercedes) performed -- minimally -- 600 hour static dyno tests at peak power and rpms the problem in a static dyno run is NOT that the pickup was being starved but rather that the oil being ingested had more air in it. This is not seriously controversial.

This is true but surely also known by Porsche during the development of the 928 -- if by some boggling stretch of the imagination they did not know previously then they certainly knew in 1975. Given that the first development mule was powered by the same Mercedes V8 in the SAE paper prior to its publication I suggest that Mercedes tried to help them, at least informally. The SAE paper from Mercedes reveals mappings of aeration levels from tilting dyno tests. Mercedes published a chart in the above paper with compound angles of aeration levels of a diesel truck engine. I am sure the (compound) mapping for the V8 at .75G was a bit too scary and inappropriate for a paper trying to (perhaps) embarass pushrod V8 manufacturers.

It would be nice if the problem were so simple as starving the pickup. Consider that Mercedes went so far as to consider (and build and test) a head oiling pattern similar to Porsche's. They then used a hydraulic lifter lubrication feed design that drew oil from the end of a gallery after it had been consistently lowered in pressure and deaerated by the bearings for the cam. It bears mentioning that Louie's video shows this sort of aeration being released from the cam bearings.*

Jumping to the present. What is the likely state of the art?

The Kibort Effect references the effect of the attitude of the engine on oil draining and collection in the sump in a system under lateral acceleration. Published videos of computer controlled multi-axis dynos from multiple OEMs and and engineering firms are suggestive. However, these dynos do not directly address or model the Kibort Effect. What they DO enable is the comparison of track obtained information (which drives the articulation of the dyno) from specific motors and platforms with the data re-obtained on the dyno. This THEN allows the construction of models, with correction from actual empirical data. At some point (already reached presumably) the models are good enough to reliably predict real life behavior.

Edit: *It has since been found through research that when ingested bubble diameters exceed a certain range in a system they will not be fully dissolved under the Henry gas laws. Louie's video provides graphic evidence of the large bubbles that exceed this by orders of magnitude. These bubbles then go on to destroy the rod bearings by well known methods in the crank. The number two main passages in the 928 and 944 receive a disproportionate local extreme percentage by virtue of the centripetal acceleration and stratifying of the air/oil mix when entering the main gallery from the pump and then immediately being available for the same acceleration into the number two main passage. It is an artifact of the passage pattern.

 

soooooo state of the art is......I got lost in the verbage...

 

bubbles

 

There are no comparisons between "dyno Pulls" and running on a brake dyno at load so this comment is nearly meaningless..."When I ran my 2 valver on the dyno and at full revs there has never been any smoke, not even a whisper of it. The engine has a new bore and Kevin's scrapers. After about 2000 miles there was roughly 1 ml of oil in the u bend. It would have done at least a dozen dyno pulls. it hasn't been used on the track though...."

Dyno pull is a FEW seconds.

 

Dry sump.

If not possible, whatever it takes to make an individual engine model work at the specific conditons that are required.

If you had not asked for an engine that ran at high rpms for extended periods of time, the 928 world would still not be aware that oil packs into the cylinder heads at high rpms....which explains why so many of these engines have blown up, over the years.

Real life testing trumps theory, every single time.

 

State of the art in testing new engines. The Kibort Effect refers to empirical data obtained over many years on the race track by a driver who has consistently made the the various wetsump 928 engine(s) live. That other drivers could make the same engines fail on the same track is missing the point. What you can learn from studying the former can benefit the latter.

The development cycle for new engines is so much shorter now that the use of computer modeling is unavoidable.

 

Dyno dynamics are a load dyno, the other 928s smoked liked they had a bad habit and given the dyno area was not properly ventilated everybody moved out of there as soon as the run was made however no smoke was evident from my engine. Given were were also running the car at different rpms, this was thousand rpm intervals and loads to check the fueling since the engine power had been increased by approx 1/4 to 1/3 and I wouldn't say that 2000 miles of highway driving and hard running is a few seconds.

 

Jim's point is easy to understand.

Go run it at WOT for 60 seconds at high rpm and report back.

We all know it will run/live for 10-15 seconds at WOT.

The people that have been blowing these engines up (with and without windage trays and scrapers) run their engines for extended periods of time at WOT and high rpms. Go run an hour or so at Willow Springs or Road America, continuously.