Gator's 818R

[Sgt.Gator]

ORP 818R Test
We took the 818R to Oregon Raceway Park for it's season first testing after the extensive upgrades to the dry sump, ECU, datalogging, and display.
First the good news:
The new AIM display is fully readable in the sunshine in an open car. In the past this has been an issue but the screen brightness was fine.
The new 4 stage dry sump with HTD drive worked better than expected. At cold startup idle the vacuum was -3.1 psi (about -6-7" vacuum) Before with the serp drive this was the when we had max vacuum, usually around -8 psi (-16" vacuum). That's a lot of vacuum at idle, so the new HTD drive has fixed that issue. Here's snapshot of the typical AIM data:



At full tilt, 6K+RPM, 17psi boost, we still had -4.2 psi (about 9" vacuum) and 58 psi oil pressure. Out standing! The least our datalogger captured at high rpm was -3.5psi. Here's snapshot of the typical AIM data:



This means we will certainly be swapping the STI over to the 4 stage and HTD drive shortly.

<<>>
The second session the engine called it quits. In the video you can see the moment it happened in the small video box from the camera facing backwards. It smoked so much for awhile I thought I might be on fire and was ready to pull the fire system. So did the corner workers who called out the fire brigade. But it was all internal, no oil spill, no chunks of aluminum engine block and pistons sprayed across the engine (yes, it's happened that way before).

I have a new engine sitting on a stand, we'll pull this one out and swap over the accessories, and install.
This engine was used from a 2015 STI and bore the brunt of the learning process about dry sumping and oil issues. We intentionally did it that way because we knew it might be expendable, and saved the new closed deck/forged pistons block to be installed once we had the formula correct.

I'm re-evaluating my desire to hit all the ICSCC races this year. The tow costs have skyrocketed. I figure at least $500 in diesel for round trip to Spokane, plus $200 in gas at the track. I may only go to ORP Club and races at PIR.

[Rob T]

Gator: Good news on the dry sump system and the modifications you have made. Sorry to hear about your motor self-destructing. I'm sure you'll let us know what happened after the autopsy. Fuel prices are a bummer....

[Bicyclops]

Gator, I'm working on plumbing up my S. I'm following your lead with the Aviad dry sump and I'm wondering what type hose you'd recommend for the scavenge lines to the pump.

Also, is the vacuum regulator necessary? Does it pull excess vac without it?

Thanks,

Ed Holyoke

[Sgt.Gator]

Thanks Rob, I'll take pics when we tear it open.

Ed, We use BMRS lines now on everything. (Brown & Miller Racing Solutions). They are very expensive custom lines. You send them the specs and they make them and ship them to you. As you can imagine that can be a major hassle. We are getting a BMRS crimper so we won't have to do the custom thing so we can build a BMRS line ourselves.

Aeroquip and XRP make good hose and fittings you can build yourself. So do several other vendors. The main thing from whoever supplies the hose is that it is vacuum rated.

Yes and No on the vac regulator. If you want to take advantage of the benefits crankcase vacuum you need a Peterson Vacuum Regulator for sure. We have ours set at -7 psi. You don't want o pull more than -8 psi.

Otherwise you can run a vented setup. You T the valve cover vents together and run the single line to the top of the DS Tank. Another fitting on the top of the tank runs a line to vented Catch can.
Basically like this except you don't need the external oil filter. It doesn't actually show the line running back to the valve covers, but it's noted in the upper right. Image courtesy of ARE:



And if you incorporate a Spintric;



No matter whether you order the pump from John at Aviaid or Gary at ARE, you want Side In on the scavenge and Top Out on the pressure to the DS tank. They will not come that way unless you specify it.
Edit: I meant Top IN on the Scavenge and Side OUT on the Pressure!
The normal is bottom out on the pressure which is a tight fit.

[Bicyclops]

All the hoses that say they'll handle vacuum want crimped fittings. Sigh. Might have to pay the big bucks and get them made up.

Ed Holyoke

[Bicyclops]

No matter whether you order the pump from John at Aviaid or Gary at ARE, you want Side In on the scavenge and Top Out on the pressure to the DS tank. They will not come that way unless you specify it.

Hmmm....I could swear that your pictures show top inlets and side out. It looks as if side inlets would foul the intake manifold, at least the rear ones.

I found XRP proplus XS hose in -12 for ~ $22 and they say that it will work with reusable fittings.


Ed Holyoke

[Bob_n_Cincy]

Gator,
Let me know if I understand this correctly.
1. With a serpentine belt the two original pumps spin at about 120% of crank speed. So the pump is approximately 8000 rpm.
2. with the toothed belt the four new pumps spin at about 50% of crank speed. So the pump is at approximately 3200 rpm.

What is the root cause?
Is it that running a gear pump at 8000 rpm is too fast?
Bob

[Rob T]

One thing I learned from my system is that you want the spintric to be the last thing in the line before the tank. The spintric is basically a centrifugal separator and if the pressure drop on the separated liquid line is greater than the separated air line, some of the liquid will take the "easy" way to the tank, bypassing the cooler. (This comment is based on "physics" of the separator, and I believe directionally correct - how much bypasses has not been quantified) I send the pressure discharge from the scavenge pump through a cooler in the side scoop and then to a cooler in the back that is thermostatically controlled (the fan is) and then through the spintric to the tank - the liquid in the tangential inlet and the vapor in the top. My tank is vented to a catch can with a vent. Along with the issue of needing three scavenge pumps, I believe some of the earlier heating issues on my car were caused by having the spintric right after the scavenge pump before the coolers. Foamy oil does not seem to be an issue for the coolers, although not as efficient as non-foamy oil, and there is undoubtedly some surging that occurs back to the tank. However, my car has at least 8 quarts of oil and nowhere near that amount of hold up in the lines and coolers, so oil supply is not a problem. At this point I am only seeing 175F oil on 80F+ track days during 20 minutes sessions.

[Sgt.Gator]

No matter whether you order the pump from John at Aviaid or Gary at ARE, you want Side In on the scavenge and Top Out on the pressure to the DS tank. They will not come that way unless you specify it.

Hmmm....I could swear that your pictures show top inlets and side out. It looks as if side inlets would foul the intake manifold, at least the rear ones.

I found XRP proplus XS hose in -12 for ~ $22 and they say that it will work with reusable fittings.


Ed Holyoke

Yeah I realized that yesterday and edited my post to reflect Top IN Scavenge, Side Out Pressure. Sorry about that!

[Sgt.Gator]

Gator,
Let me know if I understand this correctly.
1. With a serpentine belt the two original pumps spin at about 120% of crank speed. So the pump is approximately 8000 rpm.
2. with the toothed belt the four new pumps spin at about 50% of crank speed. So the pump is at approximately 3200 rpm.
What is the root cause?
Is it that running a gear pump at 8000 rpm is too fast?
Bob

IMHO, yes. It spins too fast. I'm no fluids engineer, or engineer at all, so I have no expertise other than reading a lot about pumps and cavitation, and the 7 years now of real life experience. Over the years I kept using larger and larger pulleys on the pump, which moved the overheating and loss of suction issue higher up the rpm range. But I ran out of room for a bigger pulley so decided to do what every other dry sump system in the world does, drive it at 50% crank speed with a toothed drive. Running it off the A/C ring on the crank pulley was the most compact way to do it, and the cheapest, but ultimately I gave up on it.

However if you limit your crank RPMs to under 6K, and only very occasionally spin it above 6K, you can go for hours with pressure and temperature in great shape, and the crankcase pressure staying just barely in vacuum.... But in just a few laps of spinning to 6,500-6,800 RPM the temperature would spike up dramatically, the oil pressure drop off, crankcase pressure would go positive, which quickly filled my under hood catch can to spilling oil all over the place.
In the Cascade Enduro in the STI we had that problem in the practice session when one of my new to the car drivers went out to practice. He made it about 4 laps before it went wrong. For the race we limited our RPMs to under 6K and went for hours with great oil temps and pressure.
But trying to do that when I was swapping leads in a sprint race proved to be too much for me at this last Portland SPM class 30 minute sprint. I went from 2nd place to 7th when I had to back off and let the oil cool off. It also cools off and goes back to "normal" in a dramatic fashion too. For a couple of laps it stays hot as heck, then suddenly the oil temp gauge starts going down and everything is good again, in about 1/2 a lap!
That tells me the oil was foaming, once the foam is gone it's back to normal again. The big problem is how much damage has been done to the crank and bearings for those laps when the oil was foamy? I'm guessing quite a bit of wear/damage. We'll see what the next Blackstone report says about the STI. I have a feeling it will be needing a refresh next winter, if not sooner. I think that's why the 818R engine went south last weekend. I should probably pull the engine out of the STI now and rebuild it. I'm limiting how far I'm going to tow to races this season anyway so I won't be in contention for any championships.
That also makes for Happy Wife - Happy Life.

One thing I learned from my system is that you want the Spintric to be the last thing in the line before the tank. The Spintric is basically a centrifugal separator and if the pressure drop on the separated liquid line is greater than the separated air line, some of the liquid will take the "easy" way to the tank, bypassing the cooler. (This comment is based on "physics" of the separator, and I believe directionally correct - how much bypasses has not been quantified) I send the pressure discharge from the scavenge pump through a cooler in the side scoop and then to a cooler in the back that is thermostatically controlled (the fan is) and then through the Spintric to the tank - the liquid in the tangential inlet and the vapor in the top. My tank is vented to a catch can with a vent. Along with the issue of needing three scavenge pumps, I believe some of the earlier heating issues on my car were caused by having the Spintric right after the scavenge pump before the coolers. Foamy oil does not seem to be an issue for the coolers, although not as efficient as non-foamy oil, and there is undoubtedly some surging that occurs back to the tank. However, my car has at least 8 quarts of oil and nowhere near that amount of hold up in the lines and coolers, so oil supply is not a problem. At this point I am only seeing 175F oil on 80F+ track days during 20 minutes sessions.

I've spoken to Gary Armstrong about the routing. He says the Spintric should be before the cooler, he wants it mounted in front of the cooler because the cooler is more efficient with the air separated.
But if it works for you, that's great. And Gary has a new dry sump tank with the Spintric built into the tank, so obviously that has to be after the cooler. If I were specing out a new DS system I'd probably use his new tank with it built-in.

[Rob T]

Gator: You make a great point about the rpm. I looked back at some of my laps at CMP. My max RPM is about 6500 and I probably average 4500 - 5000 during my laps. Nowhere near where you are running during your sprints. That may make all the difference. I am not seeing the same, extreme conditions you do, but interestingly, suffered the same issues with temperature, solved with the extra stage on the scavenge pump.

[Sgt.Gator]

Gator: You make a great point about the rpm. I looked back at some of my laps at CMP. My max RPM is about 6500 and I probably average 4500 - 5000 during my laps. Nowhere near where you are running during your sprints. That may make all the difference. I am not seeing the same, extreme conditions you do, but interestingly, suffered the same issues with temperature, solved with the extra stage on the scavenge pump.

My current thinking on this is the 3 stage scavenge with an HTD drive will work for 99% of the 818s on track. I went with 4 scavenge stages because I could, and now I'm standardizing both the 818 and the STI as much as possible. The new engine going in the 818 will spin to 7K routinely so I'm happy to err on the side of caution and go 4 stage scavenge.
A 2 stage scavenge definitely doesn't work, even with an HTD drive. I had the Cosworth system which is 2 scavenge stages driven by their unique HTD drive and it couldn't keep up with the oil flow/bypass gases.

[Rob T]

Gator: Makes sense. Thanks for all of your trailblazing on the dry sump issues. It has help many of us get to a good place with our cars, and I really appreciate it. Your tenacity and willingness to share are outstanding.

[jforand]

Do you have all the data relevant to the oiling system up to and during the engine failure? It would be very power to look at the graphs for oil temp, pressure, etc over time. The snaps shots look good, but the graphs would expose any anomalies if they did exist. After all, this is the method used to expose and diagnose the issues with the original dry sump kit. It would be proof positive if the new system passes the exact same scrutiny.

[Sgt.Gator]

Do you have all the data relevant to the oiling system up to and during the engine failure? It would be very power to look at the graphs for oil temp, pressure, etc over time. The snaps shots look good, but the graphs would expose any anomalies if they did exist. After all, this is the method used to expose and diagnose the issues with the original dry sump kit. It would be proof positive if the new system passes the exact same scrutiny.

Not just oiling but knock and a bunch of other data that can help diagnose broken engines. The LINK ECU is cool because it measures knock at each cylinder firing event, so it can tell you which cylinder is the one with knock.
If you have AIM RS3 and LINK PCLink G4x I can send you the files to look at yourself. Much more powerful than snips of graphs.

I looked bit harder at the data, there's a weird oil pressure glitch that wasn't there before...occasionally the oil pressure drops from say 60 psi to 29 psi and shoots back up, but it takes place in 2 tenths of a second.. The kind of change that wouldn't even start to register on a gauge. I'm not convinced it's real. It may be a voltage and sensor issue. We're looking into that. We are running the same CRR/OEM pressure pump which doesn't show this before we changed the electrical wiring.

[jforand]

Interesting. I do not have the ability to read those files, but appreciate the offer. I agree that the pressure fluctuation is far too fast to show on a standard gage and even if it did you would likely not be looking a the exact right moment. Personally, I would be concerned about that anomaly. You have a lot of things in play and it does happen to coincide with a catastrophic event regardless of possible previous damage or not. Half the pressure could be a few things, but my first though would be aerated oil. You won't get the pressures if you are pumping aerated oil. The next time you are on a dyno I would be sure to repeat your experiment where you were running at say 7k rpm and look to see if that reservoir drops substantially. If it does, it could be pulling aerate oil in before it can separate. This would be your classic issue of scavenge flow being less than supply flow. The only way to know that one for sure is to watch the level at speed.

I know you have pretty strong conviction on the speed of the pump, so I will just throw this out there and leave it alone. If the scavenge plumbing to the pan is the same size as the original setup (~-12AN times 2 lines) and you spin the pump at half or less than the original kit design you will get less oil up those lines. You are going to get more oil from the heads which is great, but this becomes a very mathematically dynamic fluid flow problem. There is a speed at which you will get NO oil up those larger center lines. As a slug of oil is picked up and travels up the tube it will lose oil volume to the tube walls. As it progresses the slug will get smaller and smaller until it will essentially 'pops' and the air will flow right on by. This is inevitable with the pump up top, there just in not going to be full pipe flow (and you don't want full flow in this setup). The magic is what happens next. If there was NO air flow I think we would all agree that the oil would simply migrate/flow back down the tube walls to the sump right? It is the air flow velocity that drives the oil up those tube walls. The air velocity needs to be increased to the point that the oil coating the tube walls is forced to migrate up against gravity and still come out the top. This would be like an air compressor line spitting water. Short burst here and there are dry, you can't get water to run the tubing walls for considerable distances. It migrates a little and when flow stops it simply falls victim to gravity again. If you sustain the high flow like in say painting a car, you will flow water up the tube walls and spit it all into your paint job.

I like the idea of dedicated scavenge to the heads. I think you have increased that without question, but I believe you have decreased your scavenge from the pan, how much is a very hard thing to answer as this is not just a full pipe volumetric flow calculation, it is an induced tube wall migration flow problem (far more complex). I can't say for sure what the reality is, but I can say for sure it is possible you are scavenging less than before. Based on the iterations you have tried this might be an easy (possibly free) thing you to try. If you were to re-pulley so that the pump was again running at the faster speeds it would restore the pan suction back to the original state and you would also have the additional scavenge at the heads. In this instance I think it would be virtually guaranteed that the scavenge was as effective as possible and would math that of the supply. The proof would be a steady state level in the reservoir level. I would expect that there will be some drop on rpm increase as the supply flow is directly linked to engine rpm and the scavenge is linked to engine rpm, but through induced oil flow from the passage of air. In other words, the air flow has to pick up speed and then the oil flow will follow. This would be the same induced flow lag that you get in carburetors. The fuel accelerates slower than the air and thus you need an accelerator pump (the fuel squirt) to get you through that transition. I would expect a drop and then a recovery if the rpm is sustained.

[Rob T]

Along the same lines....my three suction lines are on the bottom of the pan. The two originals are -12 and are to the rear of the engine. The third and newest one is toward the front of the engine. It is a -8. Interestingly, the -8 always gets warmer faster than the -12 when I touch the inlet of the scavenge pump. Is this because it has less thermal mass than the -12 and/or because it is moving more oil, either because of it position in the engine or because a -8 makes a better "straw". Or it could be some other reason completely. Eventually, everything warms up. I looked again at my data from the last track day. At the end of the third session, my coolant was at 190 and the oil was at 155.

[jforand]

Along the same lines....my three suction lines are on the bottom of the pan. The two originals are -12 and are to the rear of the engine. The third and newest one is toward the front of the engine. It is a -8. Interestingly, the -8 always gets warmer faster than the -12 when I touch the inlet of the scavenge pump. Is this because it has less thermal mass than the -12 and/or because it is moving more oil, either because of it position in the engine or because a -8 makes a better "straw". Or it could be some other reason completely. Eventually, everything warms up. I looked again at my data from the last track day. At the end of the third session, my coolant was at 190 and the oil was at 155.

This is great info Rob. I believe it is the latter "because a -8 makes a better "straw"". I assume that you are sitting there at roughly idle warming up the car. A smaller tube will make a better straw as the flow velocity will be higher and there is less wall for the slug of oil to lose volume to. Smaller volumes of oil will full flow up the tube much further. Also, at lower rpms the pumps (gear pump stages) are running slower than at operating rpm so the total airflow is down from track operations. The slower flow will entrain less oil and will take longer to pull it up and flow along the tube walls. I would expect this warming progression to be exactly as you described across the different sized tubes.

Further, I do like the idea of the different sizes tubes because the engine operation does indeed change. At idle the smaller tubing will function better and extremely large tubes might not function at all. As you wind it up you start to need more flow than the -8 can provide and the larger tubes start to come online and/or increase in function. It is tough to make one system have very wide operating parameters. This would be similar to needing different 'jet' engine technologies to get to crazy speeds...jet engines (mach 1-2), ram jets(mach 2-4), scram jets (mach 4-6+), rocket engines (6 and up) if you are into that type thing.

[Sgt.Gator]

Interesting thoughts.
I'm starting to learn how to use RS3 Analysis, it's got some good tricks. I've come to the conclusion that the oil pressure drops only happened at a few places on the track, but the same ones every time. They share one thing in common: Turns with high Gs laterally and negative vertical Gs (decompression) at the same time. Under Compression it never happens, and under neutral vertical it doesn't either. Right hand turns were far worse, but Left hand turns with decompression also caused a bit of pressure drop too. Nothing compared to the Right turns though. We were running clockwise so I suspect CCW would have caused more Left hand turn issues.

Oregon Raceway Park is most famous for it's elevation changes combined with corners at elevation transitions. In this data I hit -.67 G of decompression, and 1.5G lateral. I probably could have run around Portland International Raceway or Spokane (now called Qlispé Raceway Park) and had no problem. And the STI with it's taller and narrower tank might not have seen it.

I've run the 818R around ORP for years now and not seen this particular oil issue. However we usually run more oil in the tank than this time.

In short, there was not enough oil in the 818 tank to keep the drain covered 100% with oil under those combined G conditions. Hopefully I'll never make that mistake again......

The LINK ECU has some very cool logic for setting alarms. Instead of the simple Alarm whenever oil pressure drops under 20 PSI type gauge alarm, we can set up to 50 conditional events that set off alarms such as:

If Oil psi <40psi @ >4,000rpm then alarm. If oil psi <45psi @ >4500 rpm then alarm....etc., up and down the rpm band. We just have to be careful not to set off a bevy of false alarms, which we will learn from experience and data reviews.

The STI should be ready to go with it's new 4 stage system for testing by June 13-14. It doesn't have a LINK or AIM data yet so it will be the old style gauge watching.

[Rob T]

Gator: Interesting thoughts about the negative g's in the vertical direction. I never thought about that as CMP is pretty flat. On the supply side, I always assumed that a tall column of oil in a tall, skinny tank was the "guarantee" that enough oil would be present at the suction of the OEM pump. Another additive thought....what side of you oil pan are your suction lines on? All three of mine are on the left. While the volume is small in the pan, a hard left turn would move oil away from the suction and a hard right would bring it to the suction. Negative g or "lower g" in the vertical would move it "less" toward the suction in any case. Do you think the two may be related? Low overall oil volume in the tank, but not so low that it didn't work most of the time, and a big air slug in the suction side of the scavenge system temporarily (very short order) keeping oil in the pan, thereby causing the tank to go empty for a short time until the liquid slug hits it and "refills" the tank?

I did a lot of work with positive displacement pumps in the food industry during my career. A working assumption for a gear pump is that it is a fixed volume device, especially when it is "wetted". Have you ever researched volume per revolution of the scavenge pumps, at least from a theoretical perspective? Pumping air or oil foam with a positive pump, while still constant volume, makes the calculations much more difficult because of varying pressure through the system. Think ideal gas law: Pressure*Volume=constant at the same temperature (pressure is absolute pressure). There is no doubt we are pumping foam at least some of the time. When I was talking to Phil, we always talked about running the pump fast enough to deal with the foam or air by essentially overspeeding the pump more than we needed on a theoretical basis. In either case, with my drive belt and pulleys or your new system, we are overdriven, I believe.

Also, I read somewhere that at the engine speeds we are running, the OEM oil pump is circulating around 10 gallons per minute of oil. If that is true, my entire tank is circulated every 12 seconds. Which also means, absent of return from the scavenge system (or a temporary hold up of oil somewhere), the tank is "draining" over an inch per second.

[Sgt.Gator]

Hi Rob,
Yes my scavenge outlets on the DS pan are on the left. My main problem by far was right hand turns. As you mentioned, right hand turns throw oil to the inlets. It was the right hand turns that would cause the 60psi to 30psi back to 60psi drops in the course of less than a second. Comparatively, left turns would cause the oil pressure to drop 3 psi at a time when the psi should have been going up because RPM was going up (exiting a turn). In the data it would go from 50psi to 47psi back to 53psi as the rpm was climbing. So I don't think it was the suction outlets on the pan.
It was probably related to terrain features when going clockwise. Although I'm not planning on running CCW with the same low oil tank condition to find out!

"Have you ever researched volume per revolution of the scavenge pumps, at least from a theoretical perspective?"
Yes, as GPM @3,000rpm. Remember you can order different widths of scavenge pump sections, so it's related to the width of each section. These are approximate:
10 gpm for 1.00" @ 3000 pump rpm
12 gpm for 1.25"
15 gpm for 1.50"
18 gpm for 1.75"
20 gpm for 2.00"
You can see that a couple of sections will move a lot of fluid.

"Also, I read somewhere that at the engine speeds we are running, the OEM oil pump is circulating around 10 gallons per minute of oil. If that is true, my entire tank is circulated every 12 seconds. Which also means, absent of return from the scavenge system (or a temporary hold up of oil somewhere), the tank is "draining" over an inch per second."

At 6,000rpm the 11mm pump is moving 16.6 gallons per minute!

Because we modify OEM pumps with Verne Schumann's patented designs I have the oem stats on them, straight from various Subaru factory service manuals:
These stats are standardized at an oil temperature of 176F.

10mm 15010AA300, 02-07 WRX/STi ; LGT 2005-2009.
Marked as "10". 78mm Rotor Ring,
600 RPM -4.6L/min - 14 PSI
5,000 RPM - 47L/min - 43 PSI
85 psi relief

10 mm 15010AA320, 08-14 WRX Single AVCS And many more 08-12.
Marked "76", has 76 mm rotor.
600 RPM - 4.6L/min - 14 PSI
5,000 RPM - 47L/min - 43 PSI
85 psi Relief Valve

11mm 15010AA360 08+ STi Dual AVCS 10-12 LGT; WRX 17-20
Marked "78", has a 78mm rotor
600 RPM - 6.4L/min - 14 psi
6,000 RPM - 63L/min - 56.8 psi
78 psi Relief Valve

JDM 12mm 15010AA310
Marked "12", has a 78mm rotor.
600 RPM 6.5L/min - 14 psi
6,000 RPM - 59L/min - 85? psi (This 12mm data is not from a verified Subaru JDM FSM and is likely not accurate)

I make no claim that our modifications increase flow or pressure, they are for reliability and longevity. I haven't started selling them yet, but I do have a few out in cars on tracks and the street. More on that later.

IAG has a porting procedure for the 11mm pump that they claim flows higher than the JDM 12mm. The description is here:
https://www.iagperformance.com/iag-p...lgt-04-13-fxt/

Their graphs:

...

The 818R has the short wide tank. the STI has a taller narrow tank. I have to remember to keep the 818 tank filled higher up. The 818R tank: 3 gallon capacity aluminum dry sump tank 7.5" diameter, 19" tall net. 22" overall including drain bung and filler cap.
I have the 3 gallon, but this 2 gallon image gives the idea:

[Rob T]

Thanks for the information Gator. I really appreciate it. As I said before, you have been awesome supporting us with facts, development, and real world testing.

[Bicyclops]

I knew that I held off buying my tank for more reason than not wanting to spend the money yet. Thanks for the info Gator. I've been thinking big, I see that I need to be thinking tall also.

Having only three sections of scavenge, I'm Y-ing the lines from my heads together. I've been wondering if the extra length of the right side line would be a problem. I'm not going to be tracking as much as some of y'all so it'll probably not cause problems is my guess.

As far as scavenge to pressure pump size goes, all the old radial aircraft engine manufacturers went 2:1. At A&P school I teach lubrication systems using, as one of my aids, a rear sump from a Wright 3350. That's a 55 liter air cooled engine! They list cruise oil consumption at 5.6 gph. Check the gas and fill it up with oil please. The pumps are huge. We teach that scavenge is harder to do because of expansion of hot oil and foaming. Sound familiar? For certification the FAA mandates that the scavenge capacity be sufficient to ensure that the sump doesn't overfill but they don't specify a ratio of pump sizes.

Ed Holyoke

[Bicyclops]

I knew that I held off buying my tank for more reason than not wanting to spend the money yet. Thanks for the info Gator. I've been thinking big, I see that I need to be thinking tall also.

Having only three sections of scavenge, I'm Y-ing the lines from my heads together. I've been wondering if the extra length of the right side line would be a problem. I'm not going to be tracking as much as some of y'all so it'll probably not cause problems is my guess.

As far as scavenge to pressure pump size goes, all the old radial aircraft engine manufacturers went 2:1. At A&P school I teach lubrication systems using, as one of my aids, a rear sump from a Wright 3350. That's a 55 liter air cooled engine! They list cruise oil consumption at 5.6 gph. Check the gas and fill it up with oil please. The pumps are huge. We teach that scavenge is harder to do because of expansion of hot oil and foaming. Sound familiar? For certification the FAA mandates that the scavenge capacity be sufficient to ensure that the sump doesn't overfill but they don't specify a ratio of pump sizes.

Ed Holyoke

[Sgt.Gator]

I've been digging in really deep to oil pump issues. I found a couple of interesting nuggets:

1) The weird less than a second oil pressure glitches I now see in my data could be a normal characteristic. It's called Pressure Ripple. I never noticed on my gauges because it happens so fast, but if you have a capable datalogger you can capture it. Here's a chart from Melling Oil Pumps:


Does this partly explain my data? Probably not because it should happen constantly, not in 3 places on the track. And Melling is capturing very tiny increments of time. But it's something to keep in mind as I get more granular in my data logging. In short, I don't know if it has anything to do with my data.

More importantly:

2) Pump gears whether spur or georotor start cavitating around 5,000 rpm and seriously cavitate above 6,500rpm. For a dry sump system that is using a scavenge pump off the a/c ring of the crank pulley that means I was in serious scavenge cavitation every time I took it 6500rpm. And why when I short shifted under 6,000 it was ok.

AND it was also cavitating at my oem oil pump. A double whammy. Here's from a 14 year discussion over on NASIOC about 12mm oil pump issues:

KillerBee posted this dyno plot of 10-11-12 mm pumps. He is an advocate of using the 10mm pump in most situations of standard oem bearing clearances. Sometimes he monitors the FF 818 forum and is welcome to jump in here. :



I kept scratching my head about why the 10mm flows better in his dyno chart above 5,800 rpm than the 11mm and 12mm.....it's counterintuitive. But now I think I know why.
A couple SAE papers explain that georotor pumps fall off of flow above 5,500 rpm, and especially above 6,000 rpm, and why the 12mm would be worse than the 10mm. The pumps start cavitating at those rpms. And the wider the pump gears, the worse it gets. Here's two screen shots of two pump charts, the only difference is the better flowing pump has 1.2mm narrower gear depth. Sorry for the quality and the second one is not in color:

It's from Gerotor Oil Pump Performance and Flow/Pressure Ripple Study, SAE Vol. 115, Section 3: JOURNAL OF ENGINES (2006), pp. 204-209 (6 pages). You can read it online for free at https://www.jstor.org/stable/4468729...nts?read-now=1



In short, since the 11mm and 12mm are trying to suck more flow thru the same size inlet port pathway as the 10mm, they start cavitating sooner.

LS Engines have the same issues we have. Here's a chart from one of the engine builder forums or magazines. I have no way to verify the accuracy of the chart though, I don't think it's direct from Melling or any other manufacturer:



It's interesting to note how Subaru pumps move a LOT more gallons per minute than the LS V8 pumps do! Our 11mm pump moves 16.6 gallons/min @6,000rpm.

How do you push the onset of cavitation higher in the rpm range? Reducing the restriction of the inlet tube, the restriction of the inlet pickup in a wet sump, porting and port matching the inlet route all help. But another way is to crack open the bypass valve to feed higher psi oil into the inlet side of the pump. Yes it heats the oil, but it actually pushes the onset of low inlet pressure cavitation higher up the rpm....which might explain why Subaru reduced the relief valve pressure of the 11mm to 78psi from the 10mm 85psi relief.
I found this mentioned in a couple of places, including the LS chart. So people who shim their pumps to a higher pressure are actually making their high rpm flow worse.

I'm running our Colonel Red Racing modified 10mm pumps on both the STI and 818R. I have an AIM PDM32 on the 818R now which has massive datalogging capabilities. We intend to install the AIM PDM32 on the STI next fall too.
Given the capabilities I might run 3 oil pressure senders, one in a sandwich plate between filter and block, one in the stock location next to the crank sensor (where we get it now), and one at the rear passenger side top port. Over time that might give us clues about wear and impending failures. Maybe not the sandwich plate on the 818R because it's pretty low to the ground, but certainly on the STI it will work.

I never stop learning.....

[jforand]

That is great stuff Gator and I am 100% on board with almost all of it. There are just a couple things I would mention:

First the pressure ripple issue. As you noted they are capturing serious small internals of time. I believe what you are seeing is the pulsing nature of a gear pump (style does not really matter). It is not a constant pressure device. With each passing cavity you get a pressure bump. The chart is tiny, but I tried this anyway. I counted roughly 50 spikes/vertical gridline or 0.1 of a sec. There are 8 - 0.1sec blocks represented for roughly 400 pressure spikes. In the lower right corner you see in red that the chart is a total of 43.33 total revolutions. If you divide 400/43 you get roughly 9, which would be the number of teeth (or lobes) on the pump if it were driven at crank speed as the Subi is. I think this lines up pretty dang close. I quickly looked at picks and found one of a gutted housing, but not the gears. I would have to go down and count them, but it is totally in that ballpark.

The final point would be cavitation. Everything said and noted is pretty spot on, especially about how to reduce cavitation. This is the point I would like to expand a bit and how I believe the High Speed Scavenge Kit is intended to function. In a full pipe, fully primed pump, flow scenario (like the pressurized delivery to the engine as everything in the post discusses) you are in danger of cavitation. It is hard to pull the viscous fluid through a relatively small port for the high flows. All very true. How do you reduce cavitation? lessen the restriction to flow on the suction side as both they and you indicated. What this really means is increase the net suction pressure. It must be maintained over the vapor pressure or the fluid will start to boil (cavitation). Remember you have to separate "boiling" from temperature alone. It is all about a respective energy state of the internal energy of the fluid (vapor pressure) relative to the pressure it is being exposed to. You can absolutely boil water at room temperature by placing it in a vacuum chamber. It has a given internal energy that is trying to boil it and the atmospheric pressure is preventing that. Drop the pressure enough and it boils without getting any hotter. Conversely, you can heat the water and raise the internal energy past that of atmospheric pressure and boil it that way (212 deg F). This is exactly why there are pressure caps on the cooling systems.

So I went down a rabbit hole there a bit, but the point I wanted to bring to the surface is you can effectively increase net positive suction pressure by reducing the strength of the vacuum being pulled....it is again all relative. So in the high speed kit the scavenge sections are never intended to run with full pipe, full primed (zero air) flow. The need to pull a hard moving fluid up through the 3 feet of tubing is removed as massive amounts of air are making the trip all the time. If you allow air to freely rush up the tubing and through the gear pump at whatever speed it desires it will effectively wash out the majority of the suction vacuum (relative to what you are seeing on full flow supply pumps in your post) resulting in the relative raising of the suction pressure and prevent cavitation. In order to cavitate the scavenge sections you would need to run the pump magnitudes faster than the kit drives it (1-1.2 times the crank speed) and/or greatly reduce the flow capability of the plumping by drastically reducing tubing diameter or the like. In the subi kit I firmly believe that its design premise is substantially different than all traditional dry sump kits. It needs to pull very far up hill and it accomplishes that by rapid airflow that carries a heavy oil spray within it, but definitely not full flow. This results in significantly less power to run the pump and that is why they can get away with a micro V belt. If you were to fill the engine and create full pipe/full primed flow at RPM the pump would require more power than the belt could provide/transfer and it would slip and shred the micro V belt. This is precisely why the HTD and other cogged drives exist. If you want to experience the difference if drive power from wetted gears to full flow you need only to manually prime an old V8 rebuild with a drill motor down the distributor passage. It spins nice and easy until it clears all the air out and then it will twist your wrist off when it primes if the drill is a strong one. Alarmingly, this is at drill RPM (say maybe 400-500rpm), take that to 6000 rpm and it goes up dramatically! The subi kit is designed to scavenge in the 'oil spray' or 'heavy aerosol' realm. If the pump was on the pan at the bottom then you could slow it down and the oil would have no choice but to go up (as a heavy aerosol or full flow) as it would immediately be on the pressure side of the gears (essentially acting as a one way check valve). The physical state of the flow would be dependent on the amount and speed of air coupled with the plumbing length and diameter, as well as, the fluid itself. With the pump up top the 'check valve' or more precisely the gears are very far up stream. you can draw oil up in slugs and it can run back down the tube walls to the crank case if the motive force (air flow) is not high enough. It just gurgles and never gets to the pressure side of the gears were it is physically prevented from back flowing.

[Sgt.Gator]

Not specific to the 818 but still Subaru Engine/trans:
The Colonel Red Racing Team took another podium at the 2022 Cascade Festival of Endurance 8 Hour race at Portland International Raceway. I teamed up with our own 818 forum team, Retro Racing, to share drivers and crew. Retro brought their K swapped Miatas and a K swapped RX7. They had problems with their Honda K motors, but my Subaru STI ran great; which has to be a first in racing history!
Oil Temps, Oil Pressure, Coolant Temps were all at or below "normal". Only the rear diff got a bit hot and was a concern at 240 to a max of 250 degrees. Someday I will put a diff cooler on it.
We made it the full 8 hours to finish 3rd in class. There's a lot more write up to come but here's a short video at the race end when I pulled into the pits:

[Sgt.Gator]

Some drivers love driving the night sessions, others hate it. You have to have confidence you know where the turn entry is even though you can't see it in the unlit portions of the track. If you've never been endurance racing, last lap to the checkered flag, the crews come across pit lane to the wall to wave to the finishers.

[Rob T]

Thanks for the video Gator. Glad to see you back on the track. I know you are very familiar with that track, but night racing adds a whole level of risk to the sport. I can't imagine driving in the dark at 100+ mph. It's challenging enough at 55 on roads I know. It's a real testament to your skill.

[Sgt.Gator]

The 818R is repaired, dynoed, and ready to go again. It will be tested throughout the summer with the intent of racing the 2-4-8 hour Cascade Enduro next October at Portland International.
I'm looking forward to hitting the track with it again.
This year I also will be racing a Subaru BRZ in T4 trim and I still have the STI as a backup.
Now if it would just quit snowing here in Central Oregon!

[Rob T]

Good News, Gator. There is a lot of room on the tracks in the SE and we have been going since February. Just saying....

[Sgt.Gator]

I'm heading to PIR this weekend in the 818. The plan is to do 2 x Sprint Races and 1 x Mini Enduro (1 hour).

Meanwhile last weekend I raced the STI at Pacific Raceways. This video is qualifying at Pacific Raceways for the Enduro in the Rain. My Subaru of Bend STI excels at rain racing when I have my Continental/Hoosier Wets mounted. I out qualified cars/drivers that I can never beat in the Group 4 race and the Group 8 Enduro in dry conditions, including cars in classes much faster than mine.
Earlier I qualified 2nd in the Group 4 qualifier. Only Igor Levine in his yellow Viper beat me there. This video is the Group 8 Enduro qualifier where I ended up 3rd behind Igor again and Todd Clarke in his BMW M3.
Lap 5 was my fast lap despite missing a shift at turn 6. And I frequently miss shift at turn 9, going from 3rd to 6th. It's something about the g force and my subconscious telling me not to go from 3rd to 2nd. I'm working on nailing turn 9 since it's the most important corner on the race track.
Unfortunately the rain stopped before the afternoon races and the SST- GT1 -SPM cars were back to their usual sports in front of me in my ST class trim.

[Sgt.Gator]

I took the 818R to PIR for racing in the ICSSC in sprint and the 1 hiour mini enduro. There is very good news and a problem that is pretty easily solved.

The good news: All my temps were excellent. Oil pressure was fine.
The problem was fuel pressure. Once the fuel cell got down 1/2 full I had fuel pressure cut outs on the two left turns. This was the first time I had the car on track and pushed hard since installing the fuel cell. The second half of the 1 hour enduro I basically had to coast around Turns 2 and 6 to keep the engine from cutting out.
I called Arcflash and discussed it with my tuner Joshua Murray. He realized where I screwed up on the fuel cell installation. When I ordered the cell from Fuel Safe I specified the collector box be placed in the center back. However I forgot that what they consider center back is not the way we installed the tank. We rotated it 90 degrees to fit in in the passenger seat space, so the collector box and pickup are now on one side of the tank!
The fix is to remove the plate and reposition the collector box. We are also going to run the fuel return line into the collector box, we think it now just empties into the main part of the tank.

Here's a pic from the AIM datalogger. These are the Average (Green), Low (Blue) and High (Red):

PIR Enduro June 2023 Full Hour.jpg

The low oil pressure of 10.9 psi was during the mandatory pit stop 1/2 way thru the enduro.

I'm so happy that all the cooling and oil pressure problems are behind me!

[Sgt.Gator]

We modified the fuel cell to fix the lack of fuel in left turns. The collector box was not in the middle back, but in the middle left! We moved it to the rear of the cell and modified the fuel pressure regulator return so that it dumps back into the collector box. It had been setup to just dump into the general middle of the fuel cell. A short piece of submersible hose fixed that. Fuel issue fixed.

We had the 818R at Pacific Raceways July 15 to race in the 1 hour mini enduro. During qualifying I had a bad shake in the steering wheel with slicks on the car so had to switch to my Toyo R1R street tires for the enduro. The shake was gone and still great fun even though the 200TW tires are much slower than slicks. I did manage to get the hard charger designation for most places gained. The good news is the temps and pressures are under full control, see the pic from the AIM data.

818R Pacific Enduro July 2023 Car Health.jpg

This weekend is three days of racing at Qlispe, the former Spokane County Raceway, with my Canadian friends from Retro Racing. Qlispe has the longest straight in the NW, so I should find the 818's terminal velocity. I'll turn up the boost a bit and back on slicks. 3 days of racing!🏁

[Sgt.Gator]

Two days into the 3 day weekend and it 's going well for the 818. Two wins in the Group 1 SPM-Pro3 group, and a win in the MEO class of the 1 hour Enduro:

Friday Group 1 Race
Qlispe Spokane 818R SPM July 28 2023.png

Saturday Group 1 Race
Qlispe Spokane 818R SPM July 29 2023.png

More importantly for the 818 builders here, my temps were excellent. Many cars were dropping out with heat issues from the 90 degree air and 140 degree track surface temps. Even our own Retro Racing had to drop their K24 Turbo Miata out of the afternoon's races because of the heat. He vows revenge on Sunday.
In contrast my Water temp (ECU ECT); Oil Temp (ECU OIL T); and Trans Temp were great. The Oil and Water temps never broke 200 degrees, and the Trans Temp was just over 206. This snapshot of my Aim data is 2/3 of the thru a 1 hour race in that heat:

AIM Data Qlispe Enduro 818 july 2023.png

One more race in Group 1 tomorrow.

[Santiago]

Woop-Woop!! That is just awesome to hear! Well done on the race front - and just fantastic news on the temp front. Hard work paying off.

[Dave 53]

Congratulations!!!

Did you log intake air temp?

Just got my engine tuned today and the tuner said air intake temp is what is holding me back from any more power.

[Sgt.Gator]

Congratulations!!!

Did you log intake air temp?

Just got my engine tuned today and the tuner said air intake temp is what is holding me back from any more power.

Got it. This was my fastest lap in the Enduro race. I have a speed density tune so this reading is post FMIC about a foot from the throttle body. Considering the track temp was 142 I think the FMIC is doing a good job.

AIM IAT RPM Enduro fast lap.png

[Dave 53]

Yes, the FMIC seems to be doing a great job. Can you share some pictures?

I have the standard 818 AWIC. Great on street, but not enough on track. I'm in the 170 degree range on a 90 degree day (don't know track temp). My tuner (Travis, Snail Performance, Auburn CA) won't give me a black and white go into limp mode temperature, and he has tuned in safety, but I think I'm touching it and looking for a solution other than slowing down.

And pictures of your radiator ducting that you alluded to in another post.

[lance corsi]

Dave and Gator, what size ballast tank are you running for your awic?