Wrong!

I've done this analysis and the oil will pool at the rear of the pan and go directly into the crank. The slope of the pan is too little. Consider your number, 0.3g. The surface will be sloped at 16.7 degrees. That's a lot more than the slope of the pan. The oil is not flowing down the pan but instead being driven to the rear and piling up into the crank.

Shorter straights will reduce this effect. Think about it.

 

I get that, and did you notice how far the front of the crank is from the edge of the oil area? its about 3". there is almost enoug room for all the oil to be up front, and none of it would be hitting the crank if the car was standing on its nose, or pulling 2g 's stopping.

I know this is crude, but it has something to do with the reality of why i dont see any issues. I have the same stuff everyone else has. Im beating it on the track, i have constant oil pressure, i dont burn a lot of oil, and Im just trying to understand why the engine seems to be living with no issues!

 

The GM LS 2 and 3 engines appear to have the same problem with oil packing in the heads. Some people have "solved" it with restrictive pushrods and dry sumps. I don't know if the restrictive pushrods is a good idea because the valve train may need more oil than those pushrods allow. The problem may have caused some dry sumped engines to failed.

http://forums.corvetteforum.com/auto...the-track.html

 

thats an interesting point. so under mild acceleration, all the oil coming down from the heads, ends up pliing up to the rear of the pan, butthe stuff above the sump area, feeds it. I certainly undersand that the oil will slosh with some high velocity to the front or the rear under decel or accel. that accel g loading would only be for a couple of seconds, and go down proportional with speed, best case, and in reality, much less.

mk

 

Mark ... I'm glad you're taking Rob up on his generous offer. I'm looking forward to seeing the data ... I'm sure it'll add a lot to the knowledge of this issue and your driving style.

There was a very successful driver here in Australia (Peter Brock), who was able to blow the doors off other drivers in all types of events, and yet look after a car mechanically. Part of his secret was his inate ability to drive with a smooth seamless flow, and pick up and interpret all the mechanical signs the car was giving him. On countless occasions his cars survived where others of identical build were in pieces.

Maybe your reliability record comes down to not needlessly throwing the car around, which is reflected in lower wear rates on many components, and no oiling damage to date. Perhaps you have the issue .. but you haven't pushed it over the edge of the cliff. It would be informative to have G force data logging along with the engine data. I don't know if that's a possibility for you (available/affordable), but it would give a great deal more insight into the other data. There's a world of difference between deceleration/acceleration smoothly from A to B, and jumping on the anchors and throttle, in terms of G force spikes and consequent component wear and lubricant behaviour/flow.

 

Ok, I think I solved one member's issue with his Accusump there. Back to thinking about the 928.

 

Similar comments have been made about (Sir) Stirling Moss. Jenks knew that Moss was at 10/10ths because he was clipping the road/street posts with his wheels in turns. My old friend, Jean-Claude, was in that race and remembered Moss roaring by him.

 

It would be helpful if the data logging were able to indicate the attitude of the car.

For example, in the multi-axis dyno recently performed for Porsche by Schrick*, the simulation cannot accurately show oil flow patterns in a system that is itself under acceleration with a centripetal component. Moreover, the attitude of the car, thereby the engine, is still a further complication. This is a limitation of static water displacement/angle-of-repose/surface-angle tests in pans, as well. The dynamic surface will generally be a complex concave curve/shape.

http://www.youtube.com/watch?v=mfuleS9rnzc


* Schrick, member of AVL Group http://www.youtube.com/watch?v=C_SoR...eature=related


BMW: Check out ~2:30 Toll, nicht? http://www.youtube.com/watch?v=8_T_r...eature=related

When philosophers of logic and mathematics and language are really bored they think about aircraft propeller design. Hint: Hero of Alexandria rocked too.

 

The point of the link to the Corvette forum is to show other engines have the same problems as the 928 engine under track/race conditions. It seems some might think Porsche engineers were unable to solve the oiling problem when they were more likely limited to the solutions they were allowed to try due to cost. I'll bet the Porsche engineers do know the effectiveness of dry sumps and scavaging the heads. Very few people track their car, so why build an engine to withstand that using and drive up the cost and reduce the profit on that product?

 

I agree with you. At the time when this was a relatively "young" problem, Porsche's cash woes are well known. TAG sponsorship of F1 efforts was a godsend to them, for example.

It is impressive that they continually sought to improve the situation. I did not like hearing, however, that they (some department, exactly which I do not know) threatened to withdraw general support from teams that wanted to race the 928 here in the USA. This is more in line with carefully avoiding mention of trailer hitches: DNFs on race day. The towing capacity of the 944, an aside, is in the FSM and the same oil consumption figure of (up to) 1.5 liters per 1000km is given.

Gregg has a lot of background with them trying to solve aeration/oil-supply issues in the four cylinder cranks. Later there was a large amount of industry attention given over to addressing aeration issues and the physics of rod bearing damage by this means. Any review of Porsche's extensive patent portfolio cannot help but notice their attempt to sledgehammer the aeration issue.

 

How do the new posrsche engines get around oil packing in theheads....if the ls2 or 7 do it, and drysumping doesnt address the oil packing issue like gregs research concludes maybe porsche should be quietly alerted to gregs findings

andy

 

Has Gregg investigated running a depressed atmosphere in his drysump experiments? Say, 15" Hg. That level is managable and well tried. Not too practical for an oem though, with consumers who are ingenious in their ways of neglect and destruction.

That is probably the most common way in racing of reducing air entrainment and resulting misadventures with Henry's gas law.

 

It certainly seems like there is plenty of oil in there...but I keep going back to the question of why the 944S models all "******" the chain tensioner off of the head and turn everything into scrap metal, with virtually the same exact design.

When I build an engine, I have to make sure that engine isn't an 'experiment". It needs to survive. Removing the oil doesn't seem as big of a risk.

 

Try and keep in mind that Greg has "Greg's" budget...not Porsche's budget. There is one of Greg...not an building filled with engineers, designers, fabricators, and assemblers.

I'm pretty stinking happy I figure out what I do figure out!

It takes a lot of time and a really big effort.

Don't forget...I could have told Andy: "Well, it looks like we have found the limiting factor in running a 928 engine at high rpms...just like Porsche must have found out. Good luck with this engine. You want me to start on the replacement engine, now?"

 

Here are some thoughts.

At idle speed, the spec for oil pressure is 2.5 bar. We already know from Project 928 that the valve train was designed with a spec of less than 9% dissolved air per bar. Given that used engine oil already will easily have 9% air dissolved in it (at one bar atmospheric pressure) and that Porsche will likely want a one bar safety margin then it can be reasonably surmized that, at idle, the 944S will be drawing from the sump, oil that has up to 13.5% free air bubbles in it by volume.

Using similar reasoning, and a spec of minimum 4.5 bar pressure at 4000rpm, it can be surmized that Porsche anticipated up to 31.5% by volume, free entrained air bubbles in the sump. This is for the passenger car.

Under race conditions, these percentage free air by volume levels could easily be exceeded. Moreover, by reducing the prevailing pressure in the head to a maximum 3 bar, Porsche locked in an aeration issue that would exceed the design spec of the hydraulic valve train if the actual aeration level exceeded the anticipated level fed to the pump, sans the safety margin provided by the 4.5 bar min pump spec (at 4000 rpms, mind you). At idle, disaster could easily ensue much more quickly.

What disaster? With hydraulic components supplied with aerated oil there would be eratic, spiking load rates on the camshaft that would be transmitted to the tensioner which was never designed for these loads. The serpentine belt on the 928 links together twice as many cylinders and shock loading would be further damped by this mass, presuming the loading was largely random.

The contemporary Dodge Shelby Turbo 2.2 was experiencing valve train failures during competition due to high aeration levels of sump oil. Ed Peters, a failure analysis engineer at Dodge, designed a windage tray that reduced the aeration level and stopped this. This was as a result of joint research by Dodge and Mobil.

The 1986 Mercedes OM603 has a very high failure rate of the belt tensioner. My suggestion is that this failure rate is caused by eratic combustion during hard starts in cold weather (perhaps caused by owners failing to keep the glow plugs in good order) and by eratic shut-down conditions when the shut-off valve deteriorates but is not promptly replaced.

Lots of common threads in failure analysis.