ptuomov

Here's detailed engine doc for a N-55 BMW engine. It's a turbo diesel and it evacuates the crankcase gasses in a very sensible way:

http://www.e90post.com/forums/attach...1&d=1303136356

Page 14:

"Five oil return ducts on the exhaust side (4) now permit oil to return from the cylinder head into the oil pan. These oil return channels extend into the bedplate up to below the oil deflector. They help reduce churning losses as the returning engine oil can no longer reach the crankshaft even at high transverse acceleration. "

"Five oil return channels on the intake side (5) also ensure that the blow-by gasses can flow unobstructed from the crankshaft area into the cylinder head and to the crankcase breather in the cylinder head cover."

Application to the 928: Try to connect the chimney to both valve covers with a large enough hose. This will allow the piston pumping pulses to communicate to the heads thru those hoses. That in turn will slow down the gas flows in the oil drain holes on the exhaust side and will thereby help the oil drain from the heads.

Page 34-37:

"Crankcase Ventilation"

"The blow-by gasses flow into the settling chamber of the cylinder head cover through an opening located in the rear of the cover. Here, the blow-by gasses are directed through holes on to an impact plate, against which the oil impacts at high speed, and drains off. The blow-by gasses, cleaned of oil, flow via the pressure control valve (depending on the operating mode) through the non-return valves into the inlet pipe upstream of the turbocharger, or via passages in the cylinder head ahead of the intake valves. The separated oil is drained via a return flow duct into the oil pan."

"Naturally Aspirated Mode"

"The standard function can only be used as long as a vacuum prevails in the intake air manifold, i.e. in naturally-aspirated engine mode. With the engine operating in naturally-aspirated mode, the vacuum in the intake air manifold opens the non-return valve (15) in the blow-by duct within the cylinder head cover."

"This draws off blow-by gasses via the pressure control valve. At the same time, the vacuum also closes the second non-return valve (12) in the duct to the charge air intake pipe. The blow-by gasses flow via a distribution rail integrated in the cylinder head cover, through the intake passages (16) in the cylinder head, which lead directly into the intake ports, ahead of the valves."





"Boost Mode"

"As the pressure in the intake air manifold increases in boost mode, blow-by gasses can no longer be introduced via the passages in the cylinder head, otherwise, the boost pressure could enter the crankcase. A non-return valve (15) in the blow-by channel within the cylinder head cover closes the connection (16) to the intake air manifold. This protects the crankcase from excess pressure. The increased demand for fresh air creates a vacuum in the clean air pipe between the turbocharger and intake silencer. This vacuum is sufficient to open the non-return valve (12) and draw the blow-by gasses via the pressure control valve."

fraggle

Thanks! Great info to think about.

ptuomov

I've also recently found some crankcase windage simulation results from a major car company. This is relevant for strokers, I believe.

There's a recent SAE paper on simulating the crankcase gas (not oil) flows at 6000 rpm. The simulated engine is pretty similar to the typical stroker 928 engines. It's a V8 with 103.9mm bore and 94.5mm stroke. It has approximately similar head oil drains, in that it doesn't have inside drains. The cam chain cover of the simulated engine creates an opening not too different from the 928 oil filler neck chimney. There are many differences, too. However, I think it's close enough that it can be informative about what is happening inside a stroked 928 engine. I've got some data not published in the paper, but basically at the similar level of detail.

Take a look at the following picture, for example:



This picture shows the flow velocity vectors in a field for the front bay at 315 crank degrees. The large with circle signifies the crank pin. The driver side piston is pulling up and wrecking havoc in the crankcase. This engine has a windage tray with drain holes in the bottom, and the windage thru the front bay's hole is rather extreme.

You can also see the small flow fields near the crankshaft caused by crankshaft rotation. It's immediately obvious that in a 6.4L V8 spinning at 6000 rpm, the windage caused by the crankshaft rotation is completely trivial compared to the windage caused by piston pumping. Consequently, I don't see much benefit from a crankshaft scraper in a big V8 in terms of gas flow. Maybe on a small four-banger the pumping effects are smaller and the crankshaft rotation effects are larger, and scrapers matter for gas flows. But on that 6.4L V8 that's being simulated here, the gas velocity vectors around the crankshaft are tiny compared to the gas velocity vectors caused by piston pumping effects, the latter at times completely overwhelming the former.

The main benefit I can see from a scraper in a big V8 is that a scraper can be designed in a way that it prevents the oil being thrown off the rod bearings from rebounding and hitting the crank again. Without a scraper, the oil will come off the crankshaft, hit the pan or crankcase wall, and then get reflected back to the crankshaft. With a scraper in the correct angle, oil comes off the crankshaft (on it's own, not because of the scraper), then hits the pan or crankcase wall, and then get reflected to the scraper and not back to the crankshaft. If I am right, then the angle of the scraper is much more important than trying to place the scraper very close to the crankshaft.

The data also covers the gas flow in the oil drainbacks. The simulated engine is vented from the valve covers. This is the flow rate over the full crankshaft rotation in the drainback 1, that is the front drainback on the driver side. Positive numbers signify flow up from crankcase to valve covers and negative numbers flow down. The three different lines are for different windage tray designs:



At 6000 rpm, this drainback flows gas up to the heads most of the time. The oil draining down is fighting against this gas flow. Based on the simulation data, the drainback 2, second from the front on the driver side, is even more extreme and flows gas up at all points of the cycle at 6000 rpm.

What one can conclude from these drainback gas flow numbers is that if one is having a problem with oil not draining back from the heads or a problem with oil aerating the in the drainback channels, then one has to provide an alternative path for the gas flow and piston pulses from the crankcase to the valve covers. The BMW N-55, for example, has a second set of channels on the inside to just flow gas.

There's also an extreme pressure differential between the front bay and the back bay at two points in the cycle:



If one wants to mitigate the windage problems, one needs to provide a clear path for gas to flow from the front bay to the rear bay, and back. This can be thru block breather holes on the bearing sides or in the main webs. It can be thru the bottom of the oil pan if one can figure out how to increase the volume of the pan. It can be thru the crankshaft mains, although those holes are usually so small that they only help on the margin. Finally, it can be thru an external hose or manifold that connects the front bay to the back bay, for example in the block valley.

Thought I'd share some of this.

GregBBRD

As always, it will take time to digest the information you've provided. Just had a quick thought about connecting the two valve covers together.

blau928

Has anyone tried connecting the valve covers, the oil filler, and tapping the rear of the block to have all the lines on one circuit?

This way, the rear of the crankcase, the front of the case, and the heads are connected. The pressures can then equalize across the entire motor. (Especially the front and rear bays.)

The N55 motor seems to address the issue this way by connecting the chambers with holes in the girdle across all bays, as well as inner and outer drains, asn of course the valves that prevent pressurizing the crankcase and heads..

Just a thought...

ptuomov

Just connecting the two valve covers together sets up a circular gas flow motion from the passenger side valve cover to the driver side valve cover, then down the driver side oil drains, oil pan, up the passenger side oil drains, and finally back to the passenger side valve cover. It'll be more or less a constant flow without significant pulses, because the pulses average in the valve covers per bank. Whether this is desirable or not depends on the situation, I've got the impression that German car manufacturers like to set up that sort of venting motion.

Just connecting the chimney to the valve covers with a large hose will mostly equalize piston pumping pulses between the front bay and the valve covers. Usually, the front bay piston pumping pulse (a two punch combination of driver side pumping first and the passenger side 90 degrees later) causes a strong flow up or down the front bay oil drains. I believe that connecting the chimney to valve covers with a hose that is about as large as the front bay oil channels would significantly reduce the extreme gas pulses up the front bay oil drains.

GTS has many issues, but I believe those aren't caused by the breather system and in any case can't be cured with the tiny stock hoses. I believe that the main problem in the GTS is the crankshaft. First, the increased stroke increases the pumping pulse energy by about 18.5% based on a crude back of the envelope formula. Second, the GTS crankshaft counterweights are very poorly designed. They not only cause high bearing loads but also block the crankcase gas flow with the addition of the center counterweights and the large fan angles of the counterweights.

ptuomov

People have thought about it. For example, I mentioned in my earlier post. People have also tried it, but I haven't seen anyone trying it with a large enough hose.

I think you should first understand what problem you are solving. If you are trying to just increase power by minimizing pumping losses, then connecting the front bay to the back bay and also reducing the crankcase gas density will work. It can't be just any connection, but the diameter has to be very large.

One way to do this would be to have an external hose from the chimney to the last bay in the valley. For example, drill a large hole to the block valley. Fabricate a plate with a hose end sticking thru it and insert the plate inside the crankcase on the top, such that the hose end sticks thru the hole in the block. Epoxy in place. The plate epoxied there should more than make up for the weakening of the block due to the hole. Then connect a hose from the hose end in the back bay to the chimney. I am talking about >=1" ID hose or manifold here, it needs to flow a lot of gas and change direction 15000 times per minute at 7500 rpm.

Another alternatives for reducing pumping losses between front bay and back bay are:
- Drilling GTS-style breathing holes to the main webs and bearing sides, or starting with a GTS block. The center main web is the most important place to have holes on the bearing sides.
- Porting the existing breather holes to flow the best in the directions that the CFD simulation suggests.
- Making the pan larger, for example, by using a pan spacer
- Pulling a vacuum to the crankcase
- Shaping the crankshaft counterweights correctly such that they don't inhibit flow thru the breathing channels are critical points of the cycle. Use center counterweights only when the bearing loads require them and make the counterweight fan angles small by, say, using heavy metal liberally.
- Gun drilling mains, although the benefit there is marginal
- Keep the stroke short and instead boost the engine

If one wants to solve the oil drainback problem from the heads, then one should connect the chimney and possibly the back bay valley to the valve covers. Now, the pulses communicate to the valve cover sthru these connections, so less has to happen thru the existing oil drainbacks. There's a side benefit here in that the front and back bay pressure differentials will also equalize better thru the valve covers. Again, the added connections have to have a cross-sectional area that is at least similar in size to the head oil drainbacks.

If one wants to solve problems with the oil being blown out of the sump or oil hitting the crankshaft draining back from the heads, then the solutions are yet different. However, I believe that the main thing to worry about again is gas flow inside the crankshaft and what it does to oil flow, not just oil flow itself.

Strosek Ultra

Much of what Tuomo is talking about has been solved by Ford in a very smart way regarding the Coyote Engine. Outer oil drain-back passages like for the 928 but going deep down the oil sump. A much deeper oil sump compared to the 928. In the valley large ventilation ducts connecting all bays together and chimneys going up to the cylinder heads.

Something Ford has identified as important in its new 6.2 truck engine and the Coyote 5.0 is crankcase bay-to-bay breathing. This is managing air pumped by the pistons sliding up and down in their bores. This constantly changes the shape of the crankcase volume, creating powerful pulses, especially in the area where opposing cylinders share a "bay" between main bearing bulkheads.

Research shows either sealing the bays to minimize breathing, thus forming "air springs," or opening the bays to allow liberal communication have advantages. The Coyote team chose liberal bay-to-bay breathing, with limber holes strategically placed in the main bearing bulkheads and credits this as an important power builder. It no doubt has a positive effect on ring seal.

http://www.mustang50magazine.com/tec...e/viewall.html

Akw

ptuomov

First, let me note that I've benefited greatly from my email exchanges with Ake on this topic. In particular, he brought into my attention the BMW engines that have solved some of the pumping problems with a second set of passages on the "inside" of the bank.

Continuing on the pumping losses, here's a page from a Ricardo presentation that shows how completely sealed bays work very well, but if they leak at all then the losses are large. A huge, open oil pan for car engines is almost always close to the optimum. For very high rpm engines, such as F1 or motorcycle engines, the required breather area would be so big that sealing the bays and scavenging each bay with a pump is the best solution.




About the Coyote solution, where there's a breather manifold about int he place where a pushrod V8 has the cam tunnel:

One could fabricate the Coyote solution into a 928 block on the outside in the valley. (One being someone other than me, I couldn't do it.) With stock manifold, packaging is a challenge so this may be most relevant option for max effort ITB engines.

Perhaps a more practical alternative would be to run two 1" hoses from the chimney to the front of the valve covers and run two 1" hoses from the back bay valley to the rear of the valve covers. This way, the valve cover cavity itself becomes the manifold. The flow between the two center bays is not the largest problem, so I think it's probably enough to just deal with the end bays with separate passages and allow the center bays breath thru the existing drainbacks.

Lizard928

Then you also have the option of a vacuumed crankcase too......

The way I run my system is as follows.

I run a VP102 GZ motorsports vacuum pump in place of where the stock air pump normally sat.
This pulls from the chimney via a 3/4" line (all that is needed with this vacuum pump).
From here it goes through the pump and into a custom air oil seperator which is of centrifugal design.
The free air vent on here could be returned to the stock intake suction, which would help to increase pump efficiency at higher RPMs. On a boosted car it would need to be sucked in before the turbo/SC inlet.
The oil which is seperated out by the pump/seperator would be then trapped in the bottom of the seperator. Instead though, I have a -3 hose which connects the bottom of the seperator to the 1-4 bank valve cover. The vacuum from the pump pulls the oil back into through the valve cover which then creates a natural flow of gasses (and oil) back down into the block. The benefit of this is that when the cars are being driven with high cornering (g) forces the oil will have a significantly reduced (if not prevented) ability from travelling up the 1-4 bank head drains. This then in turn leaves more oil in the sump for better oil control.
Since doing this the oil consumption of the engine disappeared entirely. I did notice a small increase in power from doing this as well due to the aided sealing of the rings which the vacuum provides.

PorKen

What about the S4-up rear cam plugs?

Upper vent, lower drain?


I've been wanting to find some purpose for those openings, but havent' gotten there yet.

ptuomov

Another thing worth considering is windage trays and scrapers.

The simulation data in the SAE paper shows, fairly conclusively, that one needs relatively large openings in terms of flow area in the windage tray, or otherwise the pumping losses increase. If you look at Figure 9 of the SAE paper 2013-01-0580 by Chrysler engineers, it's clear that shrinking the windage tray holes below the size that they have in current production "new hemi" engines increases pumping losses dramatically.

On that score, I find the below data quite unexpected. Here's what you can see on 928 Motorsports web site on a page that describes their scraper/windage tray kit:



Based on what I know about pumping losses, I had serious doubts that this 928 scraper system would increase power. It may improve reliability, and one of my engines has a version of the I-J kit for that reason, but it's unlikely to increase power.

When I asked the company selling the kit about how exactly they measured these data, they told me that in fact it's not data about their scraper kit or even from a 928 engine. According to the supplier of the kit: "That test data was provided us by Dodge motor company relative to one of their engines. I cannot speak to what engine it was on, or the parameters of the test." I didn't see this noted anywhere on the page, so I misinterpreted that the dyno results were from a 928 engine.

Now, looking at these results in the paper by the Chrysler engineers and the additional data that I've gotten, I don't understand how they could have gotten this kind of data from a V8 engine that is in any way comparable to the 928 engine. The scraper system that blocks the air paths between bores and bays will likely increase pumping losses a lot, and in a V8 engine the windage caused by crankshaft rotation is second order compared to piston pumping losses.

As a side note, I have another concern about this dyno data on the kit vendor's web site: http://www.928motorsports.com/parts/.../TestChart.jpg.
Perhaps I am missing something, but the torque and hp values do not behave in an expected way. Torque in ftlb = power in hp * (5250/rpm), d Torque in ftlb = d power in hp * (5250/rpm), holding rpm constant.
rpm, reported hp change, reported torque change, hp implied torque change
4000 0 3 -
5000 2 7 1.90
6000 3 8 3.43
7000 5 10 6.67
7500 7 13 10.00

Unless I am making some very fundamental error here, both the torque and power increase columns can't be true in that table.

ptuomov

There are a number of sensible ways to use those vents. The main problems is that they are basically too small. I have the elbows drilled open to the hilt, but just by looking at them it's clear that they are too small to negate the piston pulses.

I am considering just cutting a hole and adding breather port to the cam cover on top of the end oil drains. If one can get the piston pulse directed from the hose into the drain, then it should give on the most bang for the buck reducing the gas flow up in those end drains that are the most important for oil draining. The middle drains can in my opinion do what they want as long as the end drains flow oil down without piston blowing the oil back into the head.

PorKen

Sorry, I wasn't clear, I'm talking about the plugs in the rear of the head, where the cams were chopped off, S4-up?

Big holes.

ptuomov

Very clever. Except with the nasty cams I'll have, I was planning to use those for extra support. Still, worth considering. Where would you hook up the other end in the block?

The 928 head oil drains in the block are actually quite large compared to the oil drains in many other engines. I don't think you need to add any drains as long as you can equalize the piston pumping pulses some other way.