FredR

To be clear, the conundrum I refer is that oil pressure [on my engine at least] hits 5 barg at about 3k rpms. If the engine then goes on to spin at 6k rpms the oil delivery will double and thus the pressure drop from the pump discharge through the system should quadruple- i.e. 5 barg should increase to 20barg but oil pressure seems to flat line at 5 barg from 3k rpm to 6k rpm and yet there are no elements in the system to control this- or are there?

ptuomov

Suppose clean oil (no horrible aeration of the pickup sucking air). If with your current oil you hit 5 bar oil pressure at 3000 rpm, then you'll hit 8 bar pressure at 3800 rpm. If you run thicker oil and hit 5 bar at 2000 rpm, then you'll hit 8 bar at 2530 rpm. This is because the flow resistance is approximately constant, and the flow rate is proportional to square root of the pressure for incompressible fluid.

Along the similar logic, if you hit 8 bar at 3800 rpm, then the pump relief valve bypasses 37% of the pumped oil at 6000 rpm and 46% at 7000 rpm. If you hit 8 bar at 2530, then the relief circuit bypasses 58% of oil at 6000 rpm and 64% at 7000 rpm.

I don't think there's anything controversial here. This is how the period oil pumps work. The pressure relief circuit bypasses about half the oil at redline rpm. Only recently, Mahle, BMW, etc. have introduced variable geometry oil pumps that more closely track the pressure and flow demands of the engine instead of just moving volume proportional to the engine rpm.

I think the pressure gauge only sends up to 5 bar in the stock sensor/gauge setup? So it doesn't tell you when you hit 8 bar, but knowing the rpm where you hit 8 bar is a matter of simple computation (again with clean oil).

All bets are off when the pickup starts sucking air, in which case the pump inner circuit (that also lubes the pump shaft) can remove some air from the oil but not much. The pump can furthermore compress some air into the oil, which reduces the oil's ability to lubricate and conduct heat but is nevertheless better than big bubbles. The pump may be compressing a fair bit of air at high rpms before system pressure drops under 5 bar, by the way.

ptuomov

If you get 5 bars at X rpm, then you'll get 8 bars at sqrt(8/5)*X rpm = 1.26*X rpm and the relief valve starts bypassing. At 2*X rpm, you're bypassing back into the inlet 37% of the oil that the pump pushes out. That's a very good rule of thumb and shouldn't ever be too far off, if the supply oil is clean.

I don't believe that the basic cavitation of clean oil is happening. Simard engine feeds the stock oil pump with a hose from a dry-sump tank and there is no evidence of cavitation at 7500 rpm, for example. I don't believe you've seen any evidence of cavitation at those rpms with the stock pump in your engines either, but correct me if I'm wrong. Cavitation of clean oil (or water or gasoline) is a violent phenomenon and you'd see badly pitted crescent gears in the pumps if this would really be happening. If broken (junk, aftermarket, non-Bosch) fuel pumps with cavitation. Let's see if I can find a second with a spreadsheet to do a ball-park computation that rules out cavitation at the pump inlet with clean oil, it's a flow/area/gravity head computation to find if the inlet pressure declines below the vapor pressure.

However, what may be happening (and what Bill Daley may also be talking about, but I don't read minds) is what happens when the oil is badly aerated. Now, the relatively low pressure (that is however above the vapor pressure of oil that would be enough to cavitate clean oil) separates air from oil. Some of the air can be removed by the pump thru the bleed holes that oil the shaft, but not a huge amount. The rest will be compressed in big bubbles into the oil supply, which is not great. It's especially not great if the girdle oil passages and the crankshaft oil passages have not been revised, because the stock crank and stock girdle will send all that air to the 2/6 rod journal. Importantly, though, tf the oil pickup is submerged in clean oil, I believe none of this happens at rpms such as 7500 rpm. This is educated speculation, not facts, of course.

I also believe that the stock 928 S4 could happily live with much lower idle oil pressures than what the factory specifies. They left a big factor of safety there, I believe, to not trigger any warning lights etc. Furthermore, one could up the idle rpm to make it even safer. For example, one could up the idle from say 675 rpm to 900 rpm, which would amount to approximately 25% pump idle capacity requirement reduction even at the factory margin of safety. Higher idle and smaller pump would in net be better for the cams, as the cams by my logic wear the most at lowest rpms.

FredR

The oil pressure sender does not max out at 5 barg- how do we know this?- cold starts tell us this is not the case as the gauge pegs out at full scale on cold starts. I advised earlier that on cold starts the relief valve lifts but this does not happen once the motor is warm and it seems GB agrees with this. When the motor is warmed up my oil pump hits 5 barg at 2800 rpms. As the engine spools up the flow increases proportionately [positive displacement] and if the flow doubles the pressure drop quadruples but this is not what we see on the gauge and therein lies the conundrum. The relief valve is not lifting and the oil flow is doubling as we double rpms- hydraulic facts- nothing to debate.

GB also opined the same about what happens to oil pressure on cold starts- it is all very logical. When I first gave some thought to this some years ago like you I assumed that the sender was maxing out at 5 barg but this is clearly not the case. That the gauge is registering pressure higher than 5 barg at a cold idle tells us that when revving the engine what we see is what we get and the pressure is flat lining over 3k rpms. My assumption is that they all do this- maybe you or Greg can confirm that happens consistently on all 928 motors as I assume it does.

There is nothing magical about the oil pump and as far as I am concerned it works- if it did not work as it should then all journals would be having trouble never mind the 2/6 journal [under more extreme conditions]. I have no reason to believe the oil pick up does not work or will pick up air but start doing things the car was never designed to do and I dare say all bets are off. Throw oil into the cam galleys and not too surprisingly the level in the sump will drop.

So, unless you can find a flaw in the logic presented why does the oil pressure flat line at 5 barg instead of taking off proportional to the square of the flow until the relief valve lifts at 8 barg?

ptuomov

"The oil pressure sender does not max out at 5 barg- how do we know this?- cold starts tell us this is not the case as the gauge pegs out at full scale on cold starts."

I wouldn't read too much into this. Who knows how the sensor and gauge behave near a cold start.

Once the car is warmed up: Suppose that you "observe" a sensor reading of 184 Ohm or 5 bar or whatever at 2800 rpm and continue to observe that at 5600 rpm. Three options:
(1) the engine leak at given pressure doubled (not possible)
(2) sensor doesn't read anything sensible past 184 Ohm or 5 bar or whatever
(3) your relief valve opens at 5 bar and has a large enough bypass circuit to regulate the oil pressure to 5 bar.

I'd bet on (2) if you're using the stock sensor and instrument cluster, as my vague memory says it's a 0-5 bar system and the manual says the relief bypass looks to control to 8 bar. But I've been wrong before.

FredR

Therein lies the difference between our relative positions.

The pressure sender does not care whether the system is hot or cold- a given pressure gives a specific output and the gauge on the instrument cluster transponds the signal it receives into an analogue display. Common sense should tell us that when the display pegs at full scale on a cold start the pressure is significantly more than the 5 barg the display is intended to convey when the needle is vertical- interpolation tells us that the pressure is likely to be somewhere in the 7 to 8 barg region.. That Greg is advising the same should also tell you something and it is a fair bet that he has had a bourdon gauge mounted on the system as per the one shown in the WSM and if he has not, I can pretty much guarantee that if such were fitted it will go straight round to 8 barg and quite possibly a tad more.

FredR

Greg,

Section 17 of the WSM shows a pressure tester with a Bourdon gauge that can be used to observe oil pressure - have you [or anyone else for that matter] carried out such a test? If so, interested to know what pressure you noted at 4k rpms on a warm engine. Not too surprisingly the WSM states the pressure should be higher than 5 barg and I would be particularly interested to know what happens to the oil pressure if the revs are upped to 6k rpm if by any chance you have such info in your records and are willing to share such.

Rgds

Fred

Bulvot

Here's a little fuel for the fire...

When I bought my 1987 928 S4, the shop that had been maintaining it (a well known 928 shop at the time and one of their previous mechanics is a regular poster on here) recommended that at oil changes I should put in one less quart of oil than called for, and then put in a quart of Lucas Oil Heavy Duty Oil Stabilizer to top it off: https://lucasoil.com/products/engine...oil-stabilizer

My car has an aftermarket oil cooler, so it takes a total of 9 quarts. That meant 8 quarts of regular oil, and 1 quart of Lucas Oil.

GregBBRD

We've run mechanical gauges on many engines....including all those that go onto the engine dyno.

Pressure varies a bit, but I opened a book of dyno charts and Andy's engine came up first. On his engine hot, pressure is 96-98 psi at 3,000, is right at 100 psi at 4800, and 106-109 psi at 6700. If the bypass opens at 8 bar, some of the above thoughts don't make complete sense.

FredR

Greg,

Thank you for the data points- it confirms exactly what I have summised must be going on. To all intents and purposes the oil pressure is more or less "flat lining" between 3k and 6k rpms.

As the speed of the oil pump varies so does the oil flow [not too surprisingly] so, as the speed doubles so does the flowrate. If the flow was through a fixed resistance conduit the back pressure at 3k rpms would increase four fold and that would mean the relief valve would start lifting at 8 barg and would be at full rated flow by the time the pressure reached 9 barg. That the relief valve should lift at all during normal operation is illogical and your data confirms without doubt that this is just not happening.

There is only one explanation as to how this happens and that is the bearings themselves have to be acting as a pumping ring. The main oil pump being a positive displacement machine will move the amount of oil displaced at whatever pressure is required to move it. If the flow resistance across the bearings were fixed then the oil pressure would rise dramatically with engine speed as noted and clearly this is just not the case. Thus the plain bearings act as a variable resistance to flow with the resistance effectively dropping as the engine speed increases. This should be no surprise as the bearing spins in an elliptical orbit effectively sweeping the oil out of the bearing and the faster the engine spins the more it does this keeping oil pressure more or less constant. All in all a very clever design.

The oil cooling circuit tries to keep the oil at an appropriate temperature to protect the oil from thermal degradation and to ensure the bearings are cooled correctly. The bearing clearances are very critical in this regard- too tight and oil flow will be reduced and too sloppy in some bearings will change the oil distribution. Not too difficult to understand why oil viscosity is critical to long term stable operation. Change the viscosity and invariably the oil distribution within the engine will change. Thus why I advocate sticking with the 20W50 oil it was designed for unless of course one is in an exceptionally cold climate as defined by Porsche.

Lubrication systems are very smart pieces of design and not to be dicked around with lightly!

ptuomov

The idea that the engine leak is approximately constant at a given pressure up to 3000 rpm and then doubles with rpm from 3000 to 6000 rpm is completely asinine.

How about someone replaces the bypass valve spring with a steel tube section or otherwise forcing the pump not to bypass? Then let’s see what happens on the way to the unattainable 300-400 psi oil pressure at 6000 rpm as the steel pump gear and timing belt duke it out — I’ll take the steel pump gear winning long before that.

FredR

That is an absurd statement- what post number was it made in? Must have missed something.

ptuomov

That’s an (absurd) implication of the the hypothesis that oil pressure stays approximately constant in the 3000-6000 rpm range without the bypass diverting a large amount of oil at 6000 rpm.

Therefore, it must be the case that the bypass diverts a large amount of oil at 6000 rpm, right?

ptuomov

Some oil pump shots:

ptuomov

The above photos of the oil pump show that it's indeed a positive displacement crescent gear pump. If the pump spins so fast that the inlet wants to cavitate, then one might be able to help the matters by porting the inlet channel a little large in terms of cross-sectional area. If the oil has air mixed in, the centrifugal force pushes the denser oil to the outside and pressurized air remains on the inside. Because of this, pressurized air goes first thru the small channel that oils the shaft. The groove around the shaft will be filled with air and the air will then be sucked back into the inlet side. It's not clear whether some of the air will exit thru the seals. The bypass mechanism is entirely built into the block.