James Bailey

On the 944 does it run on the load side or tension side of the crank ?? The forces on the pulley bearings being far diffferent ......

Lizard928

Unloaded side of course.

James Bailey

That would concern me somewhat.....designed ,built and tested for a far different task.

GregBBRD

Plastic pulley with a whimpy bearing on the load side is a bit concerning.

VDW.MS

Good Day.We do not need the orginal pump.We have a different pump.We need in order to do what they see in the pictures.The Oil hole 2/6 is of course reworked.They can also see what we use for pistons an connecting rods .The connecting rods are not Carillos.We are sorry that we can not ansewer many questions, unfortunately,very little time.For Info.service@mirowmotors.de Uwe

Imo000

Before we start undermining the idler set up, why don't we first let VDW to explain what he has and why?

ptuomov

+1.

The car has a lot of advanced stuff. Therefore, the default assumption should be that things like idler pulley bearing load capacity are thought out. Of course, I'd want to see all that analysis or durability test results before running it on my car, but that's not the point.

I was gone for two years. Is it just me or has this become a place where the first response to new stuff is always a put down?

Mike Simard

Uwe first posted in a transaxle thread, he created some kind of special trans that was very impressive and way beyond a normal effort and quietly dropped a pic of it in an otherwise mundane thread.
I dare say, if he posted his entire 928 project it would probably be the best thread in all the internets.

Thanks for reviving this thread Tuomo!

Imo000

He bought your ITBs so you know who he is, ask him to post some more.

docmirror

Having suffered a 2/6 rod bearing failure under duress, I can commiserate with the team. I did not do any engineering, so maybe this can be ignored, but here's several ideas that I've done to minimize the problem, and a way to basically eliminate it that is used in the aviation field.

First, although the link showed some interesting findings in oil entrainment control, there was limited discussion about avoiding or minimizing oil entrainment to start with. This was one of my goals, to remove as much air as possible from the oil in the sump. Knowing that part of the problems stems from oil retained in the heads, and the rest of the circuit, I decided to add a dual breather to both cam covers. I was less concerned with blowby, or oil ingestion in the engine than I was with high crankcase pressure causing added windage within the lower crankcase. Allowing a suitable venting system for the case will hopefully reduce the amount of air retained in the case.

Next, I consulted with a long time oval track racer who investigated the block with me and his finding was a great deal of casting flash on the heads, block, and cradle. The areas of the casting flash were commonly found along the oil return path from the heads down to the crankcase. This casting flash, or sharp edges on the casting are a serious cause of aireation while the oil attempts to return to the case from the heads. This flash will retain oil in the heads, and also along the lower edge of the crankcase.

To resolve this issue I spent about 5 hours with a die grinder, and some small files to remove all the casting flash from the heads and case and cradle. After the edges were smoothed, and all the filings removed, I then did a bit of polishing where the flash was removed. Again, I'm hoping that this will allow the oil to return to the case without being retained along the way.

Lastly, I added a windage tray below the crankshaft which is predominantly to minimize oil splash from the pan up into the rotating mass. This has been well proven to reduce the amount of aireated oil/vapor inside the spinning engine and will be primarily useful on tracked cars.

Now, the solution that is used in some air cooled aircraft engines. The aviation operation is different than the auto operation, but it's still an internal combustion engine with pressure oiling. The cases often have a problem with windage as the boxer design tends to maintain a large quantity of oil suspended around the recip mass. A Weir oil tank was the solution to allow the entrained air to separate from the oil. Weir tanks have been around for hundreds of years to perform various separation processes. Anyone with an anaerobic septic system prolly has a Weir tank to separate solids from the liquids which are then pumped out into a field, while the solids remain behind in the separator tank.

In the case of the aireated oil, the scavenge pump would be the existing 928 pump driven by the belt. The entrained air/oil would be delivered through the filter, then the cooler, and then rather than returning to the cradle under pressure, it would be sent to the delivery side of the Weir tank. As the oil travels up the tank, it is reduced in pressure by the weight of the oil in the delivery tank until at the very top, the oil is unpressurized as it passes over the Weir wall, and falls into the collection side. The top of the tank would be vented via a simple hose that could return the air/oil vapor to the intake, or it could be dumped out into the air(this is what old airplanes do). This area of low pressure allows the air to separate out, and be removed from the oil, so that the oil in the collection side of the tank has significantly less air entrained. Google 'Weir tank' to get an idea of what I'm talking about. I can't upload any photos anymore.

There is one other factor that I've never seen discussed concerning the oil circuit of the 928 and that is the potential for cavitation and oil vaporization on the low pressure side. I realize that the vapor pressure of the oil is prolly very low, but we must consider the possibility that the oil reaching the pump gears under high RPM and high temperature operation could be cavitating and/or reaching it's vapor pressure. The two concerns about this are that 1) cavitation of a fluid is instantaneous, and 2) it is completely fatal to a fluid pumping system. Seeing that the low pressure side of the pump requires going through a strainer, and making several bends before reaching the gears is another factor in potential cavitation/vaporization.

hairywithit

the most interesting thing - to me at least - is why isn't the air falling out of entrainment at the oil filter?
The output side feeds the rather small dia hole in the girdle. On the input side the oil flows from either the oil cooler or the block and flows down and around the filter element. Largest volume by far on the pressure feed system & with the filter element to detach any bubbles, so I would expect the air to fall out of suspension at that point.
Could it be that it is (falling out of suspension) but air fills up the volume in the oil filter mounting plate to the point that the oil level falls under the pickup tube that extends into the filter?

It might be possible to put a simple adapter between the oil cooler return line and the mounting plate - seems to be the highest point in that air space - with a bleed nipple to confirm how much air is being trapped and held in that space.

ptuomov

I am not an engineer, but I have suspicion that a man who builds/integrates to his 928 a sequential transmission can figure out an idler pulley load capacity question. Maybe not, but somehow it just seems that the magnitude of the tasks is quite different and if you can do the former you are probably going to be A-Ok on the latter...

ptuomov

Here's my current thinking on the root cause of some of the problems, drawing heavily on other people's experiments and logic. Apologies to everyone if this has become "a solved problem" and I am wasting people's time or if someone has already proven my views false. Recall, I was out of this world for two years. I am not an engineer, and my advice to everyone is to think with your own brain like you were capable back in school and not blindly follow anyone's proclamations, including mine. With those caveats, here it is:

Let's abstract to two separate pressure zones, the head and the crankcase. The crankcase and head are poorly vented. At high rpms, there's blowby into the crankcase from the combustion chamber, pressurizing the crankcase. The heavy piston rings and low volumetric efficiency of 928 are a bad combination in terms of ring flutter especially in the normally aspirated engines. Usually, pressure equates between the head and crankcase by gas flowing from the crankcase to the head via the drain channel. This flow creates some resistance for oil draining down, but that's usually not a big deal. Just a little bit of headwind.

However, if a drain channel temporarily becomes filled with oil, I believe it's possible that it becomes completely blocked. Air bubbles don't easily flow thru viscous oil in the narrow channel. The pressure differential required to support the oil column in the drain channel is surprisingly small. The vertical distance of the drain channels in the 928 engine is about 8 inches. The density of oil is about 80% of water, therefore, 6.4 or so inches of water in pressure differential is enough to support a drain channel completely full of oil. A small oil "plug" in the channel can be supported by much less and oil return to pan can stop earlier, but at that pressure differential the heads can start filling. Six and change inches of water is a completely believable number for the crankcase pressure even if the rings don't flutter.

As long as any of the channels connecting the crankcase and head remains open, the gas pressure should equate to some extent (with the caveat of average pressure possibly not equating between crank throw bays, discussed below). However, there are two ways in which I can see the oil drains being plugged with oil. First, if the pressurized oiling system of the heads produces enough oil that the oil can fill the entire cross section of the drain channel and then becomes suspended by the pressure differential in the drain channel. Second, when the car goes thru a 1g or so turn, all the drain channels temporarily fill with oil as the cylinder bank is effectively level. Now, when the cornering force is removed and the engine effectively turned back up and drain channels to 45 degree angle, I think it's possible that all four drain channels are simultaneously plugged with oil if there is a significant crankcase-head pressure differential.

Regardless of why the oil plug was formed, once it's in, oil can start accumulating in the drain channels. Because of the oil plug, the crankcase pressure and the crankcase to head pressure differential starts growing. The increasing crankcase pressure leads to a worse ring seal and more blow by, which in turn increases crankcase pressure further. This is a positive feedback process that spirals until the oil plug actually starts blowing up into the heads.

One necessary condition for this to happen is six inches of water or so in pressure differential between the crankcase and the head. That's a testable hypothesis that one should be able to test with a pressure differential sensor connected between the dipstick or filler neck and the cam cover breather vent. We'll test it when we get around to that, probably this summer. If someone else has the time, funds, equipment, and interest to run that test, I'd be interested in seeing the results before this summer. Maybe someone in Europe who likes to both perform experiments and is willing to share the actual data and experimental data with the community? I predict that one is going to occasionally register more than six inches of water pressure differential.*

Does this theorizing have any practical implications? There's basically two ways that the stock and modified wet sump 928 engines evacuate the gasses: First, thru the head vents which are in the cam covers and second thru the oil filler neck vent which connects directly to the crankcase. If my hypothesis of the problem cause is true, directly venting the head to atmospheric pressure makes things worse, not better. This is because reducing the gas pressure in the head increases the pressure differential between the crankcase and the head. Venting or scavenging the crankcase should help, and increasing the pressure (slightly as there are cam seals and whatnot) in the head should also help.

The absolute worst thing one can do is not to vent the oil filler neck and vent the cam covers only. I would think that with the filler neck blocked off with a plate or effectively blocked off with a restrictive "oil scrubber", one virtually guarantees that shortly after blowby starting oil fill flow out of the heads.

Moving to the active scavenging world, with enough brute force and pump scavenge pump capacity, active scavenging of oil from the heads will still work. But since 99% of 928 people will not install a second oil pump, I think understanding how passive breathing works is very much worth it.

*One more caveat, referred to above: It's also possible that the average (over two crank revolutions) pressure doesn't equate between the four crankcase bays. In a static setting, they would of course equate. However, at high rpm, I am not sure that they do, given that only the front bay is vented externally thru the filler neck. I can't rule out the possibility that the average pressure is significantly higher in the rear bay than in the front bay. This would make it harder to draw conclusions from the experiment I proposed. If this is the case, opening up the venting channels between the bays or maybe even drilling venting holes thru the valley next to the knock sensors might help a wet-sump engine. I also don't know how to test the hypothesis that the average pressure is different between bays without drilling holes etc.

ptuomov

This sequence reminds me of an old joke about a mathematician speaking about a competing mathematician: "His new theorem is clearly false. And it's not new, I proved that theorem long time ago anyway."

AO

Hi Tuomo,

I think this is a well thought out and reasoned hypothesis. I think there are many people here that would also like to see the results of a test to understand if this is indeed the case. I believe it may well be.

I don't think you went quite far enough in your hypothesis to connect it to what is happening at the 2/6 bearing. It's implied. Because the heads may fill/plug with oil due to inadequate ability to equalize head/crankcase pressures, the sump becomes oil starved and the 2/6 bearing suffers as a result. At least that's my understanding.

I have a little anecdotal evidence and possibly another data point. First, I agree that blocking the oil filler plate may be the single worst thing you can do. I also believe it is insufficient to merely provide a "crankcase vent" through the oil filler plate to help alleviate excess crankcase pressure like the factory did. In fact, when I had my original oil-filler "vent" from my SCer install, I seemed to get a lot of crankcase pressure build up as I would find evidence of leakage all around this plate and the hoses that connected to it. The theory back then was that the crank would fling oil up against the oil filler plate - potentially blocking off any vent you may have put in to equalize pressure. The solution was to build a small baffle to knock down this oil and provide a passage for the air to escape. (see pics below.) Thanks to Tony Harkin for this.

Since installing this, I believe I had a 3/8" ID copper pipe connecting this to the passenger head and then to my Provent. The 3/8" seemed to get plugged with oil from time to time as my Provent would fill with oil (not sure if it came from the heads or the oil filler plate, but I suspect the heads.) Recently, I increased it to 1/2" ID and the problem seems to have gone away. At the time I had thought about connecting the driver's side head to this "circuit" but packaging was a constraint. I may revisit this at a later date, however.

The other data point. Louie Ott, doesn't seem to have issues with this either and his solution appears to be to have a 1" ID (or thereabouts) pipe connected to the oil pan a-la-early oil pans (see third picture below). This also has potential issues, but has a much longer vertical distance for which the oil to travel to be impacted by plugging.

All-in-all, I agree with you and believe the first-step to protect from oil starvation and 2/6 bearing damage is better ventilation of the crankcase (and heads) so as to not create an environment where the crankcase has a higher pressure differential vs. the heads.