2 Stage Pumps

[Kealan O Carroll]

I'm designing the dry sump system for our f4i powered car.

I'm not sure about the stages of the pumps available.
Am I right in thinking that a scavenge pump and a pressure pump are physically different, regardless of plumbing?
I'm assuming a scavenge pump is designed to suck while a pressure pump is designed to blow?

With the pumps that have 3 ports, 2 on one side and one on the other, are they two "suck" inlets connected to one common outlet?

Is a scavenge pump sufficient for the f4i engine or is a pressure pump required if we're running the standard internal 4fi pump too?

Im thinking of running a scavenge pump with two inlets and one outlet, two inlets from the sump, the outlet running up to the oil tank via an oil cooler, then letting oil flow from the tank under gravity back down into the stock internal oil pump inlet and let the standard pump handle all the oil feed within the engine. Is one scavenge pump enough for this setup, i.e. will it have enough pressure/flow rate to scavenge the oil, pump it through the cooler, and into the tank, or are more pumps needed?

Thanks in advance!!

Kealan O Carroll
UL Racing

[Kealan O Carroll]

I'm designing the dry sump system for our f4i powered car.

I'm not sure about the stages of the pumps available.
Am I right in thinking that a scavenge pump and a pressure pump are physically different, regardless of plumbing?
I'm assuming a scavenge pump is designed to suck while a pressure pump is designed to blow?

With the pumps that have 3 ports, 2 on one side and one on the other, are they two "suck" inlets connected to one common outlet?

Is a scavenge pump sufficient for the f4i engine or is a pressure pump required if we're running the standard internal 4fi pump too?

Im thinking of running a scavenge pump with two inlets and one outlet, two inlets from the sump, the outlet running up to the oil tank via an oil cooler, then letting oil flow from the tank under gravity back down into the stock internal oil pump inlet and let the standard pump handle all the oil feed within the engine. Is one scavenge pump enough for this setup, i.e. will it have enough pressure/flow rate to scavenge the oil, pump it through the cooler, and into the tank, or are more pumps needed?

Thanks in advance!!

Kealan O Carroll
UL Racing

[Rex Chan]

What you describe sounds pretty much like our system, and it works pretty well (at least according to our datalogging via an SSI piezoresistive pressure sensor logged at 100hz in a DL1 logger).

The only issues I have seen with our setup is under braking: the engine is slowing down, but still spinning really fast; the forces are pushing the oil towards the front of the car. Therefore, if you have your scavenge pump pickups at the rear (we have them left rear and right rear), then there is very little oil returned to the oil tank. Since the engine is still spinning very fast, the internal pressure pump is still wanting to flow a lot of oil, and will run out. As a result, I sometimes see a drop in oil pressure under deccel, esp after a long time under constant sideways g-forces.

In short, if I was designing a dry sump pan, I would use 3 pickups (and thus 3 scavenge stages). Two is all you really need, but I still like reliability (even at the cost of weight). If you're planning to run this engine for a while, I'd do it so you'll never have to worry about it again.

Another you didn't mention, but will have to deal with is the pressure releif valve. You can return it to the sump, but this is extra oil volume for the scav pumps to return to the oil tank. We return ours to the tank direct, but some unis just plumb back into the feed, bypassing even more stuff.

Just to be clear: there is no need for an extra pressure pump stage. The internal one is fine, although you might be able to do cool stuff if you don't use that one and do it somewhere else.

[Rex Chan]

Sorry - one more thing: we use the stock water-oil cooler and have found it to be perfectly fine at keeping oil temps under control. We use a t-piece off the main water cooling circuit, so the oil cooler water runs in parallel with the water jacket. This keeps oil at 110-115C, when water temps are 85-92C.

Somethiing else to think about: Dailey Engineering sell scavenge pumps with oil-air seperators, which de-aerate the foamy oil scavenged from the sump. This reduces the need for the tank to deaerate, meaning shapes other than cylinders can be used. It also guarantees pure oil goes to the pressure pump and your engine internals. We don't have one, so I'm not sure about the quality of oil we're using to lube the engine.

[Steve_Chung]

We've had nothing but nightmares with the Dailey 2-stage pumps. Two blown engines in 1 week. For some reason, the 3-stage is golden.

Maybe this is attributed to our lack-luster powertrain team, but hey it happens.

-Steve Chung
University of Maryland
Aerodynamics Lead and CFD-Slave

[Jan_Dressler]

<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">Originally posted by Steve_Chung:
We've had nothing but nightmares with the Dailey 2-stage pumps. Two blown engines in 1 week. For some reason, the 3-stage is golden.

Maybe this is attributed to our lack-luster powertrain team, but hey it happens.

-Steve Chung
University of Maryland
Aerodynamics Lead and CFD-Slave </div></BLOCKQUOTE>
Your powertrain team must do something horribly wrong with the design of the dry sump system. We used Dailey pumps (2-stage) in our last 4 cars and never had one problem.
<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">Originally posted by Rex Chan:
Somethiing else to think about: Dailey Engineering sell scavenge pumps with oil-air seperators, which de-aerate the foamy oil scavenged from the sump. This reduces the need for the tank to deaerate, meaning shapes other than cylinders can be used. It also guarantees pure oil goes to the pressure pump and your engine internals. We don't have one, so I'm not sure about the quality of oil we're using to lube the engine. </div></BLOCKQUOTE>
We used exactly this pump (in the last 3 cars). Besides the advantages you mention, this is also (in my opinion) a necessity for the concept the thread opener wants to implement: "the outlet running up to the oil tank via an oil cooler". It would not be a good idea to run oil foam through the oil cooler, so you need the air-oil-seperator for this concept. A side effect is that you can use a (lighter) water radiator core instead of an oil cooler core (designed to handle the oil under pressure).

For the 2-stage-vs-3-stage-debate: If you place the oil intakes in your sump diagonally (maybe one left side front, one right side rear http://fsae.com/groupee_common/emoticons/icon_wink.gif ) you won't have any problems with a 2-stage, assuming that the flow rate of the stages is sufficient (with the "standard" Dailey 2-stage, it is by far sufficient).

EDIT: Another point: Make sure that your hose sizes are sufficient, and take care with the design of the oil tank...

[Rex Chan]

I'm going to guess that if you're able to use a 3-stage, then you have 3 pickup locations in your dry sump pan.

This raises the question of how you're able to run a 2-stage with the 3 pickups. Do you block on of the pickups off? If not, do you join 2 pickups into one scavenge stage? Because that would just let that scavenge stage draw air if either pickup was open to air, and ONLY work if both pickups are covered, and stop working as soon as one was uncovered.

Maybe I'm missing something here, but all scavenge stages on the Dailey pumps should be the same, so the only diff between a 2-stage and a 3-stage is the number. The only way a 3-stage works is if the sump is designed to work with a 3-stage, which probably means you shouldn't be using a 2-stage anyway.

[Kealan O Carroll]

Thanks for all the replies guys!

I've decided to run 3 pickups:
back right, back left, and the front centre.
The sump is kind of V-shaped with the point at the front, to encourage oil under braking to collect towards the front centre pickup.
Im going to run the stock Oil/Water exchanger from the 4fi engine as suggested by Rex Chan.

The oil tank is designed to act as a de-aerator, so could I plumb the sump - tank line straight to the resivoir, then mount the oil cooler on the feed pipe back down to the sump? Or would this require a pressure pump to pump th eoil through the cooler?

[Jan_Dressler]

Sorry for the late answer.
<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">Originally posted by Rex Chan:
This raises the question of how you're able to run a 2-stage with the 3 pickups. </div></BLOCKQUOTE>
If this question was directed to me: No, we are running a 2-stage with 2 pickups, as described above:
<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">Originally posted by Jan_Dressler:
(maybe one left side front, one right side rear http://fsae.com/groupee_common/emoticons/icon_wink.gif ) </div></BLOCKQUOTE>
A 2-stage with 3 pickups would be nonsense for the reasons you mentioned above (only if you "switch" one stage between 2 pickups in some way, I've seen solutions...).
<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">Originally posted by Kealan O Carroll:
I've decided to run 3 pickups:
back right, back left, and the front centre.
The sump is kind of V-shaped with the point at the front, to encourage oil under braking to collect towards the front centre pickup. </div></BLOCKQUOTE>
Why this way and not the other way round, with one pickup at the back? The engine is wider at the front, there is more oil coming down there (crankshaft, piston cooling, oil from the cylinder head coming down to the sump via "channels" at the front...). Another question: What pump flow rates do you intend to use?
<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">Originally posted by Kealan O Carroll:
The oil tank is designed to act as a de-aerator, so could I plumb the sump - tank line straight to the resivoir, then mount the oil cooler on the feed pipe back down to the sump? Or would this require a pressure pump to pump th eoil through the cooler? </div></BLOCKQUOTE>
First point: If you want to design an oil tank like that yourself, be careful and take a close look at commercially available tanks, especially the inner parts (baffles etc.). There are quite some failed dry sump systems where actually the oil tank was the thing to blame.
Second: I don't quite understand your question. If you intend to use the stock oil-water heat exchanger (maybe it works for the F4i, when we tried to use that concept with the PC37 / RR engine, we were a bit on the hot side with our oil), why do you want to use an additional cooler? Just leave the pressure circuit as it is (except of somehow relocating the PRV).

[Kealan O Carroll]

With regard to relocating the PRV, can it be left where it is in the circuit (After the stock pump), but instead of releasing into the sump, releasing into a feed back to the oil catch tank?
Im assuming the only problem with leaving it where it is in the stock setup is that it would relief its excess oil into the sump, and give the scavenge stages more unnecessary work to do?

I've a fair bit of experience working with engines, but none whatsoever with oil pump systems or anything akin to this drysump setup so I'm in way over my head:/

[Rex Chan]

you are 100% correct about leaving the prv to dump back into the sump: it just means the scav pump has to do more work. a few years ago, that was how we did it, but it got changed to a line into the oil tank before i got involved. some teams even plumb it back into the feed line for the internal press pump.

while i agree that more oil falls into the sump from the front of the engine, once it gets into the sump, its going to flow where the gforces are pushing it, which should be the back of the pan, and the sides - your placement of pickups is exactly how i would do it (our current pickups are at the back sides, but none at the front, which i would like).

Even though your tank is designed to de-aerate, i'm not sure how well it will work (as in, how are you going to figure out how well it works?).

I don't know anything about oil-air coolers.

[Jan_Dressler]

<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">Originally posted by Rex Chan:
some teams even plumb it back into the feed line for the internal press pump.
</div></BLOCKQUOTE>
That's how we do it. Less work for the scavenge pumps (as mentioned above) and you don't need another hose to the tank, everything is in the oil pan (with the PRV in a separate "channel" of course, feeding it back in the pressure pump inlet).
<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">Originally posted by Rex Chan:
while i agree that more oil falls into the sump from the front of the engine, once it gets into the sump, its going to flow where the gforces are pushing it, which should be the back of the pan, and the sides - your placement of pickups is exactly how i would do it (our current pickups are at the back sides, but none at the front, which i would like).
</div></BLOCKQUOTE>
I agree, but why only the back of the pan and the sides? During braking it will flow to the front of the pan, as you mentioned yourself?
As I said above, we use one pickup front left, and one rear right with a 2-stage pump. Works like a charm.
<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">Originally posted by Rex Chan:
Even though your tank is designed to de-aerate, i'm not sure how well it will work (as in, how are you going to figure out how well it works?).
</div></BLOCKQUOTE>
I can only second that. Or think about using an "active" de-aerator on the pump, as on some Dailey pumps.

[Rex Chan]

My thinking is that its more critical to have oil flow/pressure under acceleration than braking, as the torque produced/stresses on the piston-crank will be less under deccel/overrun.

At least that's what I tell myself...

On some logs last year (2010), we had some oil pressure drops under braking after long lateral accelerations. But never under accel.

[Jan_Dressler]

Hm...
How much oil do you have in your tank when the engine is running?
IIRC, we have about 3 - 3,2 l in the entire system, with about 1,5 - 1,7 l in the tank. If your engine is running at, lets say 10000 rpm, with a pressure pump flow rate of about 4 l / 1000 (crankshaft) revs, about 0,67 l/s of oil (or less if you use the PRV-plumbed-back-to-pressure-pump-inlet-system)are pumped out of the tank and into the engine. So you would have more than 2 s of time during breaking before the tank runs dry. But then you might be in danger that the tank is actually dry when you are accelerating again...

With the "one pickup front left, one rear right"-system we only once had a little problem (oil pressure decreasing during braking), but I blame a "not ideal" oil tank design for that.

[Kealan O Carroll]

Having a massiva amount of trouble gettin a pump ordered :@
Does anyone know if the return to the engine can be gravity fed or does it need to be routed through a pressure stage? Wanna have the tank as low as possible (CoG and Oil Pipe Lengths)

[Rex Chan]

By return, I assume you mean the oil line that goes to the internal oil pressure pump? In that case, it's gravity fed, and that's how I assume most other systems are done. On the stock bike (wet sump), the pump had to suck up about 200mm, so being fed from a tank with level above the pump should be fine (although the path is longer).

What's the problem with ordering a pump?

[Kealan O Carroll]

Hmm yeah I suppose, if the original pickup can lift the oil I may have a look at mounting the tank such that the pickup is above the bottom of the tank...the oil should level off (like the manometry experiments) and provide a supply to the internal pump....must keep in mind Braking forces tho...I hate fluid mechanics http://fsae.com/groupee_common/emoticons/icon_frown.gif

[Kealan O Carroll]

BTW, check out our sump depth in the ULFS Lab photos thread in the photos forum ;-)

[Rex Chan]

it seems to be only 10mm thick, which is very low. Any pictures of the insides? Or CAD images of what's going on the inside?

You can check my fb photos for images of our 16mm dry sump being made. You should be able to tell how it works from the photos, but I'm happy to explain further.

[mk e]

<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">Originally posted by Kealan O Carroll:
Hmm yeah I suppose, if the original pickup can lift the oil I may have a look at mounting the tank such that the pickup is above the bottom of the tank...the oil should level off (like the manometry experiments) and provide a supply to the internal pump....must keep in mind Braking forces tho...I hate fluid mechanics http://fsae.com/groupee_common/emoticons/icon_frown.gif </div></BLOCKQUOTE>
You want a tank that is tall and skinny. The oil surface will not be much over a 60 degree angle.....but the is a 2 to 1 ratio right? so if the tank is 5" diameter is must be at least 10" 10 to know the bottom is covered with oil. Also at high rpm windage can keep a quart or so of oil suspended in the air around the crank if you don't have an effective windage tray in place and other stuff oil not all being returned right away means you probably want to have about a 2-3 quart reserve below the point you assume is the bottom. 7 qt is the smallest commercial drysump tank I recall seeing so that might be a general guide....a 7 qt tank would be filled with about 5 qts so the oil never hits the top. You could probably get by with about a 5 qt tank, and fill with 3-3.5 and be happy.

[mk e]

Here's one I'm just finishing on my car, not an FSAE but will see autocross and track events. It a V12 I'm installing traverse so similar to what you guys are dealing with. Hopefully the pics work and are helpful.....

http://www.ferrarilife.com/for...n-v12-2012-02-05-003 (http://www.ferrarilife.com/forums/attachments/projects-rebuilds-modifications/40882d1328488765-308-v12-conversion-v12-2012-02-05-003)

http://www.ferrarilife.com/for...n-v12-2012-02-05-004 (http://www.ferrarilife.com/forums/attachments/projects-rebuilds-modifications/40883d1328488765-308-v12-conversion-v12-2012-02-05-004)