Please school me on dry sump systems

[MadBill]

Discharging the bypass oil as deeply as possible below the oil surface would minimize splashing and aeration.

[Kevin Johnson]

Discharging the bypass oil as deeply as possible below the oil surface would minimize splashing and aeration.

But also give maximum opportunity for any discharged bubbles to become entrained versus surface foam (which dissipates more readily).

[Belgian1979]

Ok, that's doable.

I've taken a good look at that pump graph again. You'll remember I said that I reached 75 psi (relief/bypass valve setting) at around 3000 engine rpm. If you look at the graph of the pump (1500 pump rpm) you can see that the pump puts out around 8 gpm at that point. This has me doubting that the engine consumes 22 gpm (full pump output at 6000 rpm). I think it will be somewhere around 10-12 gpm maybe.

[MadBill]

Discharging the bypass oil as deeply as possible below the oil surface would minimize splashing and aeration.

But also give maximum opportunity for any discharged bubbles to become entrained versus surface foam (which dissipates more readily).

Great, Kevin!
You're the one that raised the specter of "internal bypass recirculation helps give any entrained air bubbles additional time to dissolve into solution because of the higher pressure plus it helps decrease the average bubble diameter which also helps the air to dissolve." and then, re an above-surface external bypass: "...prediction that the external bypass will worsen the situation (via splashing of the ejected oil into the reservoir and decreased ability to drive entrained air into solution)"
Sounds like we're screwed regardless! Maybe if we use a dry sump with the bypass routed externally back into the return line to the tank, through an air separator?

[Kevin Johnson]

Discharging the bypass oil as deeply as possible below the oil surface would minimize splashing and aeration.

But also give maximum opportunity for any discharged bubbles to become entrained versus surface foam (which dissipates more readily).

Great, Kevin!
You're the one that raised the specter of "internal bypass recirculation helps give any entrained air bubbles additional time to dissolve into solution because of the higher pressure plus it helps decrease the average bubble diameter which also helps the air to dissolve." and then, re an above-surface external bypass: "...prediction that the external bypass will worsen the situation (via splashing of the ejected oil into the reservoir and decreased ability to drive entrained air into solution)"
Sounds like we're screwed regardless! Maybe if we use a dry sump with the bypass routed externally back into the return line to the tank, through an air separator?

Welcome to my world.

The best solutions revolve around trying to minimize air entrainment in the first place.

[MadBill]

Slightly O/T Kevin, but what do you think of the theory that loose dry sump scavenge clearances exacerbate aeration problems by 'jetting' slugs of air back through the leakage paths and so forming smaller bubbles that are less likely to be dissipated in the tank? One mfg. claims great things from his coated 'zero clearance' pumps re 'happier' bearings with far less pressure. (70 Vs. 160 psi. on a blown Hemi)

[Kevin Johnson]

Slightly O/T Kevin, but what do you think of the theory that loose dry sump scavenge clearances exacerbate aeration problems by 'jetting' slugs of air back through the leakage paths and so forming smaller bubbles that are less likely to be dissipated in the tank? One mfg. claims great things from his coated 'zero clearance' pumps re 'happier' bearings with far less pressure. (70 Vs. 160 psi. on a blown Hemi)

That should be straight forward to test.

Tighter clearances are less forgiving of ingested trash, however. I might also be worried if a scavange pickup was not located in an area that received a steady stream of oil or oil mist. With minimal run times that is probably not as big a concern, though.

[Belgian1979]

A question : if given enough volume being pushed through a small enough orifce into a small pipe filled with fluid what the effect would be from that regarding air entrainment. And beyond the previous, if that the fluid gets repeatedly recirculated and thus heated up, reducing viscosity how would that effect air entrainment ?

Then another question : in a closed system, such as the oil system, where oil is being sucked from below the oil level in the pan and exiting through the external relief valve/bypass how can extra air get in there unless the pump would already be sucking in air ? The external bypass is not that different appart from exiting to the pan where the internal bypass exits into the inlet.

Then something else : I agree that oil should preferably be free of air, but with a crank wipping oil around, and lifterholes spraying oil over a cam turning at high rpm, that is just not going to happen.

Then another thing : a dry sump sucks in a generous amount of air and probably a whole lot less oil. that subsequently gets sprayed into the tank where the effects of air entrainment would be at an even higher level than what would be introduced by virtue of a bypass where no air is being sucked in. Not even 4 qrts of oil are going to prevent air in the oil in that system. Proven by F1 engines where due to the forces acting on the rods the big end gets deformed creating negative pressure in some parts of the oil layer causing the formation of cavition from air being pulled out of the oil and where after the fact bearing were found with pits in them. (This BTW was solved by using a higher visosity oil untill design changes could be made)

[Kevin Johnson]

You need to begin with the typical amount of air that is dissolved in motor oil at working temperature at one atmosphere pressure is 9% by volume. So that means every additional bar pressure will allow 9% more.

A question : if given enough volume being pushed through a small enough orifce into a small pipe filled with fluid what the effect would be from that regarding air entrainment. And beyond the previous, if that the fluid gets repeatedly recirculated and thus heated up, reducing viscosity how would that effect air entrainment ?

It depends on how much dissolved and free air was in the oil.

Then another question : in a closed system, such as the oil system, where oil is being sucked from below the oil level in the pan and exiting through the external relief valve/bypass how can extra air get in there unless the pump would already be sucking in air ? The external bypass is not that different appart from exiting to the pan where the internal bypass exits into the inlet.

Oil splashing, Droplets of oil impacting the surface of the oil at high speed, Vortexing (think whirlpool) can allow air to travel below the surface of the oil to the pickup opening. Etc.

Then something else : I agree that oil should preferably be free of air, but with a crank wipping oil around, and lifterholes spraying oil over a cam turning at high rpm, that is just not going to happen.

Yup.

Then another thing : a dry sump sucks in a generous amount of air and probably a whole lot less oil. that subsequently gets sprayed into the tank where the effects of air entrainment would be at an even higher level than what would be introduced by virtue of a bypass where no air is being sucked in. Not even 4 qrts of oil are going to prevent air in the oil in that system. Proven by F1 engines where due to the forces acting on the rods the big end gets deformed creating negative pressure in some parts of the oil layer causing the formation of cavition from air being pulled out of the oil and where after the fact bearing were found with pits in them. (This BTW was solved by using a higher visosity oil untill design changes could be made)

Typically a device known as a hydrocyclone is used to extract as much air from the oil as possible. And you now know that oil at atmospheric pressure typically has 9% air dissolved in it. Negative pressure will cause some of that to evolve out and then violently collapse as the pressure shifts.

[Billzilla]

Sounds like we're screwed regardless! Maybe if we use a dry sump with the bypass routed externally back into the return line to the tank, through an air separator?

What can often help is to pass the oil through an oil cooler on the way back to the tank as that helps the oil lump back into a more homogeneous mass.

Perhaps a bit on-topic the dry-sump system I use on my old racing car might be of a little interest as it's as simple as I can make it.
It's a Suzuki G13B twin-cam 1.3 litre in a small sports car in Australia. (somewhat bent & sad right now but will be repaired)
The entire system has just three oil hoses, one from the tank to the engine. I use the factory oil pump on the front of the crank as it works fine.




Then from the front of the sump there's another oil hose that goes to the scavenge pump that's tucked away under everything so you can't see it. Then from the scavenge pump the last hose that dumps the oil back into the tank. (The tank return isn't right, it was built that way for the old system and I haven't changed it yet. I don't like the 90° turn into the tank)



The oil cooler is from a Subaru WRX and it's a water-to-oil heat exchanger that sandwiches between the block and oil filter. I welded on a small fitting onto the end of the water pump to get a cool water source for the heat exchanger and it returns the water into the water pipe that goes into the water pump inlet, it can be seen in the first photo. The benefit of the heat exchanger is that it warms the oil much faster than a conventional air cooler and also helps keep the temperature at a more stable level.

That's as light and as simple as I can make it.
The tank is quite large for the engine but it gives plenty of room for the oil to centrifuge around to de-aerate itself. I also carry a bit more oil than I really should (~4 litres) but I inadvertently found out with that much oil in it, if the scavenge pump fails (which the drive to it did once) the engine naturally fills with oil but with the height of everything it's just enough to keep pumping a bit back into the tank and drip-feed the engine.
I now run a big-arse low-pressure oil light .......

[Belgian1979]

What you see here is the pressure drop over an orife.

http://www.womackmachine.com/engineerin ... fices.aspx

Let's assume that the pump puts out 22 gpm and the engine uses 12 gpm. The total volume that can not go into the engine is 10 gpm and only serves to push the pressure further up. That's bled off because the relief valve opens. Those 10 gpm have to go through the orifice in the oil pump to the inlet. From what i've seen on my pump that hole is maybe 1/8-1/4" big. For ease of using the table in that link, let's use 1/4 of an inch. With 10 gpm passing through that hole the pressure drop is a wopping 75 psi. So the pump pressure is reduced to 0 after the orifice.

Pump manufacturers have argued that a pressure drop of greater than 3 psi pulls the air out of the solution and creates cavitation. So all of that 9% dissolved air is being pulled out, effectively making the pump pull in air and cavitating.

If I make the excess pressure bypass over the external valve, it has a smallest orifice in that line of .450", almost half an inch. Although it would end up at around a 7 psi pressure drop, it is a far cry from the 75 psi with the internal bypass orifice. Add to this that the external bypass allows for the air to escape to the surface more and I think the problem would be less.

Now for the interesting part : a lot of people I've talked to said that a std pump with a low pressure relief valve setting doesn't drop in pressure above 5500 rpm. I can believe that, because the typical relief valve setting in a sbc pump is 45 psi + it pumps less volume so less has to go through the bypass. All that makes for a reduced pressure drop and no cavitation....

[Belgian1979]

Sounds like we're screwed regardless! Maybe if we use a dry sump with the bypass routed externally back into the return line to the tank, through an air separator?

What can often help is to pass the oil through an oil cooler on the way back to the tank as that helps the oil lump back into a more homogeneous mass.

Perhaps a bit on-topic the dry-sump system I use on my old racing car might be of a little interest as it's as simple as I can make it.
It's a Suzuki G13B twin-cam 1.3 litre in a small sports car in Australia. (somewhat bent & sad right now but will be repaired)
The entire system has just three oil hoses, one from the tank to the engine. I use the factory oil pump on the front of the crank as it works fine.




Then from the front of the sump there's another oil hose that goes to the scavenge pump that's tucked away under everything so you can't see it. Then from the scavenge pump the last hose that dumps the oil back into the tank. (The tank return isn't right, it was built that way for the old system and I haven't changed it yet. I don't like the 90° turn into the tank)



The oil cooler is from a Subaru WRX and it's a water-to-oil heat exchanger that sandwiches between the block and oil filter. I welded on a small fitting onto the end of the water pump to get a cool water source for the heat exchanger and it returns the water into the water pipe that goes into the water pump inlet, it can be seen in the first photo. The benefit of the heat exchanger is that it warms the oil much faster than a conventional air cooler and also helps keep the temperature at a more stable level.

That's as light and as simple as I can make it.
The tank is quite large for the engine but it gives plenty of room for the oil to centrifuge around to de-aerate itself. I also carry a bit more oil than I really should (~4 litres) but I inadvertently found out with that much oil in it, if the scavenge pump fails (which the drive to it did once) the engine naturally fills with oil but with the height of everything it's just enough to keep pumping a bit back into the tank and drip-feed the engine.
I now run a big-arse low-pressure oil light .......

Neat. What pump is that?

I see that the pump belt is fairly close to the block surface. One of the problems on a SBC is that the belt is pretty far away and in my car that will make the belt/pulley hit the front crossmember that has to be notched for that reason.

[Billzilla]

Neat. What pump is that?

I see that the pump belt is fairly close to the block surface. One of the problems on a SBC is that the belt is pretty far away and in my car that will make the belt/pulley hit the front crossmember that has to be notched for that reason.

We made it ourselves out of a big chunk of alloy. It uses two gears from a Holden 308 oil pump - they're similar in many ways to a SBC though I'm not sure if the SBC has an external oil pump.

[Kevin Johnson]

Ok, now the next level of complexity.

Aside from the velocity increase (delta P proportional to v or v^2 depending on the Re) of the exiting oil (where will it go and not cause a problem?), you need to consider the design of your circuit.

If you never plan to push the car hard in the turns and not accelerate or brake at extreme levels then you should be ok.

Imagine that you do engage in some of these behaviors and some additional air gets whipped into the oil. The varying density of the components in the flow means they will segregate as they pass around bends and so on.

The danger is that the plumbing to the external relief will impart a handedness to the flow and present the relief valve mechanism with a segregated input flow (say, bubbles on the interior of a curve or corner). If the relief valve expels primarily air bubbles, that is wonderful. Porsche used such a deaerator orifice within its oil pump in the 928. However, if the relief valve expels primarily fluid and the balance of the flow that continues on the galleries is mostly air bubbles then that is a disaster.

You will need to be very careful -- and very CONSISTENT -- in how you route your hose, the orientation of the fittings and so on. And then you would need to test or trust in the Lord that at the very time you need it most, your routing is appropriate.

This, in essence, is why the rod bearing failures in the Porsche 928 (and 944/951/968) were so problematic. The configuration of the circuit gave one branch a disproportionate amount of air bubbles. The advantage of using the internal bypass is that this engineering dilemma is more defined/controlled.

It would be nice if hydraulic circuits only used neat fluids. Tying in the dry sump to the discussion: this is a primary aim of that type of system.

What you see here is the pressure drop over an orife.

http://www.womackmachine.com/engineerin ... fices.aspx

Let's assume that the pump puts out 22 gpm and the engine uses 12 gpm. The total volume that can not go into the engine is 10 gpm and only serves to push the pressure further up. That's bled off because the relief valve opens. Those 10 gpm have to go through the orifice in the oil pump to the inlet. From what i've seen on my pump that hole is maybe 1/8-1/4" big. For ease of using the table in that link, let's use 1/4 of an inch. With 10 gpm passing through that hole the pressure drop is a wopping 75 psi. So the pump pressure is reduced to 0 after the orifice.

Pump manufacturers have argued that a pressure drop of greater than 3 psi pulls the air out of the solution and creates cavitation. So all of that 9% dissolved air is being pulled out, effectively making the pump pull in air and cavitating.

If I make the excess pressure bypass over the external valve, it has a smallest orifice in that line of .450", almost half an inch. Although it would end up at around a 7 psi pressure drop, it is a far cry from the 75 psi with the internal bypass orifice. Add to this that the external bypass allows for the air to escape to the surface more and I think the problem would be less.

Now for the interesting part : a lot of people I've talked to said that a std pump with a low pressure relief valve setting doesn't drop in pressure above 5500 rpm. I can believe that, because the typical relief valve setting in a sbc pump is 45 psi + it pumps less volume so less has to go through the bypass. All that makes for a reduced pressure drop and no cavitation....

[Belgian1979]

Isn't this problem present as wel in the stock system with internal bypass? It would seem to me that the pump would expell air as well, which has a tendency to seggregate in the corners, with the heavier oil on the outside and the air on the inside ?
If the pump cavitates in its stock form I bet some air is going to make it into the oil channels/lines as well.
Plus I think that this problem is present in a dry sump also. It uses oil lines as well.

It looks as if this is never going to get solved, unless maybe using a 0W20 oil.

An alternate solution would be to drill an extra hole in the existing internal bypass right across where the current 1/4 hole is an which exits to the pan. I think this is the mod guys like Schumann use on their pumps. However the reality is that the bypass channel is not big enough to put a large enough hole in it. Even the total area of a 1/2" hole is barely enough to reduce the pressure drop. 2 X 1/4" is not even close. It would help but not enough.

A std pump would be an option too. It would reduce the total amount of oil pumped and would therefor put the point at which the pressure drops of higher. I'm not sure where, but my guess is around 6800 which is still below the redline on this motor.