# Where to best increase car engine oil pressure if intervention is needed?

• Auto/Motor

## Summary:

If I experience low oil pressure in a high G situation, how can I best intervene to stabilize the oil pressure? Where in the path of oil flow would I want to provide supplemental oil pressure and flow?
Subaru flat fours (a boxer engine design) are notorious for losing oil pressure under high lateral G's. The usual solution is using a dry sump. Dry sumps are expensive and complicated. They are, based on currently available remedies, unnecessary until you exceed 1.5 G's. I am at that threshold. So far I have not had an oil pressure related failure. One such remedy to losing oil pressure is questionable as to efficacy. The oil accumulator, which is an Accusump in my case, provides oil to the engine when the oil pressure drops below a certain threshold. Two or three quarts of oil can be added to the capacity.

There is debate about whether it works. One engineer I know says it is a placebo and that the engine's oil pump is working to recover a reasonable oil pressure and having to refill the Accusump at the same time. This issue can and should be an integral part of the discussion as far as I am concerned. This issue seems equally relevant and a part of the equation (or possible solutions).
Here is a link to a post and illustration of the oiling system.

During racing I have seen a 10 to 15 PSI pressure drop from the pump's output to the point that the remaining oil is feeding only the cams and turbo. The consumption of the main and rod bearings have lowered the pressure by that much. The oil accumulator is designed such that oil can flow out of it when the pressure is below the desired setting. 10 to 15 pounds is recommended.
Here is where I get into the weeds. A lot of torque, low RPM's and/or low oil pressure is a recipe for disaster in any engine. Oil protection can vary by viscosity, oil properties, heat, bearing clearances and resistance to friction of the mechanical components. Perhaps, for the sake of no arguing about brands, we leave the lubricants out of this? Let's just assume a "normal" operating situation where we are not overheated and we are using a good racing oil.

So, there is a magic number, at below which the pressure will not support a protective film of lubricant. There is also the possibility that the pump has started sucking air and you have aerated oil.
Regardless, the pressure has dropped to the 10-15 PSI threshold of the oil accumulator and it starts to release the oil. Where would you like for it to go? If it's before the pump wouldn't it need a check valve? It could be after the pump. Might it then fight the pump by providing a pressure resistance? Do you inject the oil before or after the main bearings? If after, (I'm assuming) some will go to the main and rod bearings and some will go to the heads.

In this engine, like most pressure systems, the farther from the pump you go the smaller the passages become to maintain the pressure. Whether it's my choice of bearing clearances (standard for a racing Subaru engine) and/or exceeding the design parameters of the pump and oil galleries, I am seeing a significant drop in oil pressure as it transitions from the crankshaft bearings to the heads, with the current design. The design is such that it's the dynamics of the fluid and pressures (and possible additional variables) that I'm have trouble with. Where and how can I best supplement the oil supply? Other remedies?

jim mcnamara

ChemAir
Gold Member
I couldn't look at the diagrams. Apparently you have to be a forum member there.

Do you have a way to log your oil pressure? If so, why not run with/without the accumulator to see what it does. Accumulators can take a little adjusting of the preset pressure to get to work properly.

Regardless, the pressure has dropped to the 10-15 PSI threshold of the oil accumulator and it starts to release the oil. Where would you like for it to go? If it's before the pump wouldn't it need a check valve? It could be after the pump. Might it then fight the pump by providing a pressure resistance? Do you inject the oil before or after the main bearings? If after, (I'm assuming) some will go to the main and rod bearings and some will go to the heads.
An Accusump has a solenoid switch that only opens when the pressure is low, it should close when the pressure comes back. Some people have used these accumulators for purposes they weren't really made for. I've seen (and done) some odd stuff to deal with very specific conditions. I have seen them set to discharge oil to turbos, oil pan, pump, and other places. I believe the most common is to the oil gallery on the discharge of the pump so it will replace oil not reaching the pump, and it has the secondary ability to prime the engine with oil. How you can use it is really only limited to your imagination, electronic control capability, and pain  threshold if you make an error.

If the oil distribution is optimized, a dry sump is the best solution for this problem. It is expensive, but effective. Most racers I've seen trying to use accumulators tend to complicate the system and get something that "still doesn't work as well as a dry sump", but costs about as much, and more if you break something.

I was in this situation once, for really low oil pressure in the shutdown area for a drag car. Car would switch from about 8500RPM with 90 or so pounds of oil pressure, accelerating at 1.4-1.8g, to negative 1.4G when the parachutes came open and no open throttle. Oil pressure would drop off scale low (<5 psi) for about 4-5 seconds while oil was draining back and pooling in the front of the oil sump. I never had engine damage, presumably due to the engine RPM going low, and low motor load, but I changed over to a dry sump to fix this, and it also helped engine vacuum.

Other remedies?
I would go to a machine shop familiar with the internal oiling on this engine block, and address this prior to going very far with accumulators. I would do a little homework trying to find knowledgeable people for this (if they exist). Oil passage clearancing, restriction installation, and modification is a significant part of race prep machine work. Someone may have already solved this problem.

Last edited:
jim hardy
Gold Member
2019 Award
Dearly Missed
Is this a traditional hockey puck sized oil pickup with a screen on the bottom, connected by a rigid tube to the oil pump ?

I would find a way to add a swivel joint to the oil pump pickup tube
so it can swing left or right of sump center.
That way the same centrifugal force that hurls oil to one side of the sump also hurls the pickup out to where the oil is.

Swivel joint in the pickup tube is not unprecedented
In high school i had a '49 Buick with the 320CID "Fireball" straight-8.
It had a floating pickup that followed oil level up and down. That sump held 8 quarts and the floating pickup provided a few inches of vertical travel.

There has to be swivel joints in a tube fitting catalog someplace. Buy several - your friends will want one too.

Good Luck !

old jim

cjl
I'm guessing he's running this oil pan and pickup, given that he's racing a subaru and hasn't gone all the way to a dry sump:

https://www.killerbmotorsport.net/ultimate-oil-pickup-ej25-152.html.htmlhttps://www.killerbmotorsport.net/high-performance-oil-pan-ej-series.html.html
As for where to put the accusump? My immediate intuition is to put it immediately after the oil pump. Before the pump, I'd worry that you'll never see enough pressure to fill it properly, and anywhere other than immediately after and I'd worry you wouldn't adequately feed the full system. Immediately after the pump, it's in a position to keep the system fed from (effectively) the same place as where you were already feeding from, keeping the oil flow as similar as possible between accumulator-fed and pump-fed operation. You could potentially also look at adding additional baffles to the pan, though if you're running the Killer B pan already, it's already a pretty nice design.

That having been said, you clearly have quite a lot of money into this if you're getting that kind of G-force and performance, so really, I'd be tempted to just go all the way to a dry sump system if at all possible.

(What kind of tires and suspension setup are you running to see >1.5 lateral g continuously on a Subaru, out of curiosity? That's a really high number, though not unachievable with slicks, especially if you also have a decent amount of aero)

jim hardy
Gold Member
2019 Award
Dearly Missed
I'm guessing he's running this oil pan and pickup,

well - THAT's sure not gonna follow the oil from side to side.

1.5 g's ? arctan 1.5 is 56 degrees......
I dont know how wide is a Subaru sump... but by how much oil level in the center falls should be calculable with some measurements and grade school geometry.

Have Fun !

old jim

Someone said they could not see this when they went to the NASIOC link.

I enhanced an oiling schematic so I could better visualize what's happening. With it and the dimensions it is straightforward to calculate areas in the galleries. But what does it all mean?
I think I mentioned enlarging galleries and smoothing and radiusing bends and passage transitions. I port my oil pumps.

I've talked to expert builders and few will disclose their mods, but from seeing one engine case and reading I know I have done pretty much everything the big boys are doing. I am an artist, ex-machinist and owned a prototyping business. I didn't finish engineering school... I've built and modded a few engines over the course of 56 years.

I don't know anyone else using an external oil pressure regulator as I do. My system is 8.5 quarts if I include the two quart Accusump.

I cannot do sustained 1.5 G's, like on a skid pad. I do have some aero but it is not full-blown even though I run 285 Hoosier slicks. I have 470 AWDHP. It's a 6.5:1 P/W ratio car. I have 42 hours on this race engine and I bump up against 8K quite a bit.

I liked cjl's response as it was closest to the kind of an answer I was looking for. Yes, I run KillerB stuff and I should be data logging. I'd rather not do too much guessing and experimenting as I have a system that works! I'd just like to understand the physics of it and why fluids flow the way they do and react pressure wise to changes in the galleries sizes. I'm not a whiz at higher math.

Last edited by a moderator:
sysprog
... An Accusump has a solenoid switch that only opens when the pressure is low, it should close when the pressure comes back...
I would go to a machine shop familiar with the internal oiling on this engine block, and address this prior to going very far with accumulators. I would do a little homework trying to find knowledgeable people for this (if they exist). Oil passage clearancing, restriction installation, and modification is a significant part of race prep machine work. Someone may have already solved this problem.
The solenoid valve is an option and is usually connected to the ignition so that it opens the Accusump when you turn the ignition on and closes it the second you turn the engine off. The valve is open during running of the engine and otherwise has no effect on whether oil flows in or out of the Accusump. Only air pressure in the Accusump and the oil's pressure controls that. I control mine manually.

If you looked at my link you will read that I have done extensive mods to the galleries and am using an external, adjustable oil pressure relief valve and am using a ported and modified oil pump to allow using the external control valve. I feel I have exhausted my resources and am in rarified air where those that know more don't want their mods known or they have no advice and do just common bolt-on parts and procedures.

Last edited by a moderator:
ChemAir
Gold Member
I feel I have exhausted my resources and am in rarified air where those that know more don't want their mods known or they have no advice and do just common bolt-on parts and procedures.
You probably are, and I've been in that boat with racing electronics/controls, and very high power small block V8s.

There seems to be an inverse correlation between horsepower use in racing and number of people with the expertise to make and use it. If you want to make 1000 hp with a V8, you have lots of people to talk to. If you want to make 3000, that list gets short. In the US, for specialty import engines/chassis, I expect the pool of quality knowledge to be quite small, and horizontally opposed 4-cylinders to be a small subset of those, so you may have to do a lot of R&D yourself. I'd like to be able to point at which oil passages need changing and by how much, but without knowing a pressure profile across the engine (need data acquisition here), and looking at internals before and after teardown, I'd have no way to make a suggestion. This may be a good application of a one factor at a time R&D project, which could become expensive. This would involve data acquisition from multiple points to get a profile to compare before and after each change. There's not an easy way to calculate and recalculate oil flow/distribution as you speculate changes. One (Subaru specific) battle you will have is drainback from the heads when you are undergoing high lateral acceleration. I don't know how the oil return to the center of the motor will work in this situation. At some point a dry sump with head pickups will probably become mandatory.

Before going completely dry sump, can the block you have be modified to put on an external belt driven oil pump? This will give the ability to move more oil at higher pressure, change to a high pressure external race filter, let you run additional oil lines outside the block to critical areas, and would open up some options for potential baffling in the oil sump/pan and creative pick-up plumbing. This will require upgraded ability to drain back oil. Of course, if you go to all this trouble, a designed dry sump system may be a better alternative.

When I came up against a performance barrier I could not cross on my own, I'd usually find someone who knew how, and pay them to show me or make it happen. Be careful with this, many people that say they know what they are doing in high performance, do not. As you said, many shops are "bolt on" performance shops.

Unfortunately, most barriers in high performance are taken down by money or time/effort, and money trumps effort, which is the primary reason I quit doing it. The highest volume race/chassis shops learn the most, and lead the way. You can pay them to help your program, but they will probably not help you get faster than their car, unless you're willing to fund their program, too--and even then, probably not.

It can be very frustrating.

Last edited:
I'd like to be able to point at which oil passages need changing and by how much, but without knowing a pressure profile across the engine (need data acquisition here), and looking at internals before and after teardown, I'd have no way to make a suggestion. This may be a good application of a one factor at a time R&D project, which could become expensive. This would involve data acquisition from multiple points to get a profile to compare before and after each change. There's not an easy way to calculate and recalculate oil flow/distribution as you speculate changes. One (Subaru specific) battle you will have is drainback from the heads when you are undergoing high lateral acceleration. I don't know how the oil return to the center of the motor will work in this situation. At some point a dry sump with head pickups will probably become mandatory.

Before going completely dry sump, can the block you have be modified to put on an external belt driven oil pump? This will give the ability to move more oil at higher pressure, change to a high pressure external race filter, let you run additional oil lines outside the block to critical areas, and would open up some options for potential baffling in the oil sump/pan and creative pick-up plumbing. This will require upgraded ability to drain back oil. Of course, if you go to all this trouble, a designed dry sump system may be a better alternative.
Thanks for the input. To fill in a few blanks about what and why: I am 72 and though I had some sponsorship in SCCA racing when young, what I am doing now is track days in a race prepared (NASA ST2 level) car. I may chose to do a Time Attack one day before I hang up my spurs. I could be doing extensive data logging if I wanted to spend my money there, but for what? Dry sump would be cheaper and straightforward. I do share all I know so some probably won't talk to me for that reason.

On NASIOC I asked about modifying the oiling of the mains for running a Manley billet racing crank. They are completely different from a stock STi crank or even a stock one that has been cross-drilled. I didn't get any expert advice. BTW, I do all my own work, except I don't tune and I don't do the precision machine work prior to assembly. What I'm doing on a shoestring budget puts me right at the threshold for needing a dry sump system. I am building an 818R kit car and it will need dry sump, so I know quite a bit about both. KillerB, a big supplier of oiling components and exhausts has shared a lot with me and feels strongly that you need scavenging of both heads, as you have suggested.
Sometimes when I ask pointed questions I get excuses because they probably don't want to share what they know or they don't wish to confirm/deny my ideas as valid.

It seems I am doing a lot of guessing when I only have pressure readings from after the pump and after the crankshaft. I wonder though if I were to open the flow to the crankshaft would the crank bearings (and rod bearings) still be the pressure dam that leaves enough pressure and flow for the heads? I know now that without a reference point of knowing what's left at the end of the line I am not going to know what works till it fails or I tear it down and measure it.

BTW, the Subaru line of oil pumps are quite powerful. There are two companies about to introduce a "better" pump, but most just have a painted, warmed-over Subaru pump. When I went to an external pressure bypass my first running of their 12MM pump produced so much pressure that some seals leaked. After getting it properly regulated I can easily dial in 85-100 PSI with 10-40 oil.

The discussion has helped to clarify some things and even theoretical would have to be proven out with testing. I may just have to leave well enough alone.

Last edited by a moderator:
jim hardy
Gold Member
2019 Award
Dearly Missed
Summary: If I experience low oil pressure in a high G situation, how can I best intervene to stabilize the oil pressure? Where in the path of oil flow would I want to provide supplemental oil pressure and flow?

Subaru flat fours (a boxer engine design) are notorious for losing oil pressure under high lateral G's.

Here's all i was suggesting
Lateral acceleration changes pressure in a pipe only very slightly
it's a safe bet you're losing suction

What's going on in the sump ?

Actual orange line might be a curve like a catenary - i'm not sure - but you get the idea.
You gotta take suction where the oil is.
I had an old Chrysler that would sometimes lose suction in a hard turn (one direction only) . Adding a quart fixed that.

When running, oil level drops because a lot of it is high in the engine, trickling back to the sump.

I'd try overfilling crankcase until level is right at bottom of crank throws at idle. Zero cost experiment.

old jim

Here's all i was suggesting
Lateral acceleration changes pressure in a pipe only very slightly
it's a safe bet you're losing suction

What's going on in the sump ?
View attachment 242864
Actual orange line might be a curve like a catenary - i'm not sure - but you get the idea.
You gotta take suction where the oil is.
I had an old Chrysler that would sometimes lose suction in a hard turn (one direction only) . Adding a quart fixed that.

When running, oil level drops because a lot of it is high in the engine, trickling back to the sump.

I'd try overfilling crankcase until level is right at bottom of crank throws at idle. Zero cost experiment.

old jim
Thanks, "old" Jim. With age comes wisdom (for some).

A common recommendation by Subaru road racers and autocrossers is to overfill the pan. My feeling is that the Accusump accomplishes a similar goal of making sure that the oil doesn't run out, but the more I think on it the Accusump is a knee-jerk kind of response to a bad situation. Overfilling may have very little downside and you might never run out of oil/normal pressure. Downside might be churning the oil with the crank during hard braking or acceleration. I do try and run it very full, but not overfilled. Some windage trays, like Roger Clark Motorsport's (also look up RCM's "Gobstopper")

Here is part of what I'm running, the Killer B parts. If you scroll through the images you can see all the components. There is a windage tray and the pan is baffled to slow surge. Many images. RCM's may work better but I think the rubber trap valves have a finite life and Cosworth's have been known to fail.

I am very curious as to where the 56° came from! Is there a calculator? With this, it seems I can calculate the vertical distance and see when the pickup would be uncovered. The baffling slows the oil escaping, and there is the issue of how much oil is in the upper areas of the engine.

I once calculated that it would take 11 seconds to pump out what's in the pan, but I did not account for some of it being missing. The highest G turns only last a few seconds. There is the "lightbulb" at New Jersey Motorsports Park where it is well banked for at least 90° of the more than 180° turn. I can generate at least 1.3 G's there @ 90 MPH and it takes about 11 seconds in all to traverse. I've never had an issue there.

Last edited by a moderator:
jim hardy
Gold Member
2019 Award
Dearly Missed
I am very curious as to where the 56° came from!

In absence of lateral force the surface would be level as shown by yellow line
lateral force pushes oil to one side,
making the top surface of yellow rectangle high on one side and low on the other

exactly as if the car were stationary and being tilted to one side

at 45 degrees tilt
the forces across and up the pan would be equal

tangent of 45 degres is 1.0, which is equal forces across and up(EDIT well down i suppose, gravity pulls down - pardon my crossthreaded words)

to get 1.5X as much force across the pan as force up(oops - down),
which is the situation you describe as 1.5g lateral
the car would require more than 45 degrees tilt

SOOO,

go to windows calculator
atan function gives you the angle whose tangent is what you entered

enter in sequence
circled keys [ 1 ] [ . ] [ 5 ]
then
rectangled keys [inv ] [ tan-1 ]

56.3 deg ?

old jim

Last edited: