FA20 oiling problems, study and solution
 

Hi all, this is my first post but I'm a long time user of the wealth of information shared in this forum.

I'm co-owner of a performance shop in Lima - Peru and have a 2013 Toyota GT86 as a personal car. I wanted big power so I installed a Sprintex 335 supercharger together with many other upgrades and a full engine build.
Since then I struggled with oil starvation issues very similiar to the well documented problems in this good threads:

http://www.ft86club.com/forums/showthread.php?t=63723

https://www.ft86club.com/forums/showthread.php?t=131419

To sum it up, I had three engine failures with damaged bearings indicating oil starvation. Oil pressure was always good (I have a gauge with sensor tapped to the main oil gallery), engine build was clean and checked (we've built many performance engines). I chose to use 0.04 mm main and 0.06 mm rod clearance, and although a little looser than spec, by no means reason for failure. I'm also using Motul 300V 10W40 oil. For some reason there was not enough oil flow to the bearings (rod bearings suffer first of course), particularly at high RPMs: I could do a couple hundred km of break in below 5,000 RPMs full power without problems, but as soon as I started hitting my 7,600 RPM redline, boom. A stock engine can take 7,600 RPM all day long so :mad0260:

This felt a lot like oil flow restriction of some sort, having ruled out most of other variables. I had a much more detailed look at the oil galleries in this engine.

The first thing I noticed is the very long and restrictive oil galleries from the pump to the crank (including the main gallery which tapers from 12mm to 8mm) so I did this:

1. main oil gallery enlarged and straightened to 12.7mm
2. block entrance enlarged (shoulder modified for bigger oring)
https://i.imgur.com/6GBR70O.jpg

3. pump to block gallery enlarged to 12.7mm
https://i.imgur.com/wsMYccg.jpg

4. all restrictive parts on the timing cover galleries ported / enlarged to 12.7 mm (including galleries to / from the oil filter)
https://i.imgur.com/WKbFQhb.jpg
https://i.imgur.com/o5ZjHhg.jpg
https://i.imgur.com/Ah11MqE.jpg
https://i.imgur.com/pqSF5CO.jpg

5. blocked small return hole in oil pump with epoxy and shimmed the PRV 2.0 mm

After this work I had another oil starvation engine failure, at first I was at a loss and very frustrated.
Then I looked at the only part I hadn't improved: the inlet side of the pump. After all I know now, I think this is THE most important part to improve, even in stock engines:

After reading a lot about oil pump cavitation I had a much more detailed look at my oil pump's inlet parts. After three engine failures, cavitation HAD to leave a mark if it was part of the problem:

https://i.imgur.com/E99uMZg.jpg
https://i.imgur.com/Mb6f6qV.jpg

a smoking gun right in front of me!

I wanted some confirmation with the engine running (maybe the oil pump came cast like that? how can I be sure?). Cavitation is very hard to evidence in real life, luckily after a lot of searching I found this amazing research paper by E. FROSINA, A. SENATORE , D. BUONO, F. P. BOVE of the University of Naples, Italy:

https://www.gtisoft.com/wp-content/u...r_IUC-2017.pdf

Although the pump is not exactly like ours, it's very similar and the information I'm looking for applies: how inlet pressure (something I can measure with a little work) relates to cavitation and pump output.

This is the summary for that (taken straight from their paper):

https://i.imgur.com/zhFLF4q.jpg

So if you want good mass flow you need to keep your oil pump inlet pressure above -0.2 bar at all times.

I rebuilt my engine, tapped the oil pump's inlet and connected it to my vac / boost gauge and started doing tests:

https://i.imgur.com/avdWxbD.jpg

To my surprise, inlet pressure would drop to -0.23 bar at high RPMs. This was with the stock oil pickup but with the other oil inlet galleries already enlarged! With a completely stock oil pump inlet, pressures would surely drop below that causing heavy cavitation. I didn't test like that because I'd have to tear the engine apart to enlarge the internal inlet galleries after the tests while the oil pickup is easily changed.

I ordered a Killer B Motorsport Ultimate FA20 WRX Oil Pickup and modified it (the WRX FA20 pickup has a different length and orientation, besides, their pickup tube has aprox 19.4 mm inner diameter and I wanted at least 22.0 mm):

https://i.imgur.com/B52AtgX.jpg

We used a TIG welded 1" pipe that had 22.4 mm inner diameter:

https://i.imgur.com/psJnaar.jpg

After the install I repeated the tests and these are the results:

https://i.imgur.com/BEK0Abk.jpg

Big difference in inlet pressure but small difference in outlet pressure (which ilustrates how oil pressure in the outlet doesn't tell you 100% of the story).

After about 1,000 km of very heavy (almost track-like) use with clean oil and oil filter inspections I'm confident we've solved the problem. I'd love to take a look at the rod bearings but it would be hell in this engine.

I think stock engines survive because they have very tight clearances that keep oil flow low and the inlet and outlet galleries are *just enough*. Low viscosity oil also helps in the cavitation department. Even so I think that stock engines that are regularly taken above 7,000 RPM are cavitating, just not enough to destroy the engine (and sometimes yes, look at all the "spun rod bearing" stories). All FA20 engines will benefit from a much less restrictive oil pump inlet. Taking into account that the stock oil pickup elbow is THE most restrictive part of the whole inlet, a good enough solution might be to just install a Killer B improved pickup. They even have one correctly fabricated for the GT86 FA20 now (I will still ask them to make a 1" tube version for my engines).

The car has a lot of power and there's room for more :thumbup:. I had a lot of upgrade paths paused while solving this issue.

My happy car:
https://i.imgur.com/smp6NDS.jpg
https://i.imgur.com/T7ZfCBE.jpg

Timing cover pressure test
 

After reading the threads that mentioned the possibility of having oil leaks through the oil gallery covers in the timing cover, we did an air pressure test (after assembling the covers with anaerobic sealant and an impact screwdriver used *carefully*):

https://i.imgur.com/aP7qoQo.jpg

There were absolutely no leaks in any of the oil gallery covers or in the pump itself. This for me settled the issue (this test was done before beginning with the modifications mentioned in the first post).

@ETM_Shaman Have you considered doing this evaluation of flow/pressure with a bone stock block with just the oil pick-up upgrade?

Your work is brilliant. Thank you VERY much for sharing. I wish you and your shop all the best. Upstream vacuum gauge is the big winner here.


...despite using a Sprintex. ;)

Great work. I have also been wondering if the oil pump plate is flexing around the oil pump and opening up the pump clearances causing lower pressure . Most oil pumps have a steel backing plate or they bolt to the block or both. Another thing I have been thinking about is the expansion of aluminum compared to the steel oil pump gear.
What is your oil tempature? I have 83lbs at 7500, tempature 200f. That is with vr1 20w 50 conventional oil. Have not tried redline yet. Was also thinking of using BMWs fix for spun bearings 10w 60 tws. Great work and thanks so much for sharing.

IIRC oiling problems at high rpms are one of main limiters to upping redline on FA20. I wonder, how fast otherwise stock FA20 might be spun if oiling is enhanced. Highly doubt that to S2000 Mk1 levels .. but i can dream, no? :)

Interesting you installed spacer shims on the oil temp sensor. I have suspected that running my oil pressure sending unit there is giving me false cavitation readings from poking too far into the galley.

This is awesome work, thanks for sharing. I'm wondering how much of an improvement one can get from doing the two least invasive things here even on a stock engine? Relocate or shim the temp sensor and replace the oil pickup. With how big of a difference the oil pickup made on your inlet pressures, I imagine that it should still make a significant difference without removing the other restrictive areas.

Seems like a no-brainer to me for how easy those mods are.

It would be very interesting but it's fairly involved because you have to remove the timing cover to tap it properly. Then do the test with stock oil pickup and then remove the pan, change the pickup and do the tests again. I'd love to see these results but wouldn't be willing to do all that work on my account to tell you the truth.

Thanks! Your thread was what made me really look into cavitation.
I have a JDL oil cooler and use Motul 300V 10W40. For the first 100km or so I used 15W50 mineral oil for break in and got higher pressures overall.

When driving slowly my oil temp is around 90 deg C (I get 0.6 bar at idle) When I drive it hard it can get to 110 deg C (aprox 0.4 bar idle pressure, 4.0 bar at redline). At first I was a little scared because of the low pressures but after having oil starvation engine failures with higher pressures than that I'm learning to live with it. Just look at the guys with stock engines and less than 3.0 bar oil pressures at redline while tracking with perfect UOAs after every oil change (there's a couple of good threads).

I still have a last test which is taking the car to the track (I used to track my car monthly before doing the rebuild). I'm curious of how hot my oil and how low my oil pressure will get :cry:

Yes and I think this also applies to the FA20DIT. A bigger 14mm pump with the same restrictive oil galleries, oil pickup and more power looks very bad for me.

The thing with just spinning it faster is that you're not producing good torque high in the RPMs. This COMPLETELY changes with head work (already done) and a good set of camshafts (one of my planned improvements). We'll have to wait for enough people to build their engines with this mods and explore to start seeing if our engine ends up being a nice high RPM screamer (n/a or boosted) like we all want to see :popcorn:

The problem is the stock oil temp sensor has a thick part that ends up blocking the gallery a lot without shims.

solidsnake11's post illustrates this very well:

https://www.ft86club.com/forums/showthread.php?t=131419

Absolutely! Probably almost all of the improvement coming from the oil pickup but still worth it to do both.

Someone that has a stock engine with an oil pressure gauge installed, that tracks the car and is already familiar with how his oil pressure behaves should try this and report his findings!

Great piece of applied research! - :thumbsup: :thumbsup:


humfrz

I don't see drawback in that. Even that lesser a bit torque is more due lower gearing multiplication then more engine torque at lower rpms when upshifted, when it comes to actual torque at wheels. Also wider usable rpm range may mean less time lost at some specific track configurations, where otherwise one may loose time for short upshift and then immediate downshift prior turn .. or loose time not shifting but also not accelerating due lower redline cutoff. And, and .. i simply love how high-rpm race engines sound :)

I have a question iirc everything I read said that the wear would be on the high side of the pump, it is where the bubbles implode and wear the metal. If that is correct, your pumps erosion is on the low side. It would seem that it is building pressure on the low side before it goes into the gears. I would guess this is from squeezing oil into the gears at high speed. The little scallop on the high is to slow the pressure rise of the oil as it transitions from low to high pressure this reduces the implosion of cavatation.
Maybe the ramp on the inlet of the gear is not a great idea at high velocities. I remember reading about engineers trying to break the sound barrier and right before the engine would stall. It was discovered that the funnel to the jet turbine needed to be inverted because so much air got crammed in that the air would stall.

Could you name the anaerobic sealant you used on the oil pump plates? Thanks

Having peak torque at redline doesn't really matter. If torque is falling off, it may still be better to have more revs and to shift at redline:

[ame]https://www.youtube.com/watch?v=zZBqb0ZJSwU[/ame]

So what you have found is that the inlet is inadequate because sustained rpms leads to a drop in inlet pressure, and that leads to cavitation, right?

A few points of confusion:

--Shouldn't cavitation lead to a more significant drop in outlet pressure if that is the cause of the oil starvation, and you showed only a small difference in outlet pressures?

--If cavitation is happening then couldn't that metal debris be a cause of bearing failure? Moreover, shouldn't it show up in an engine oil test as metal in the oil?

--If this is the cause of oil starvation then couldn't this be replicated on a dyno or something, meaning, if it is just sustained rpms, and not sustained lateral g's or something else commonly related to track oil starvation, then couldn't it be replicated in a static environment? Is it common for oil starvation failure to happen, for instance, during a top speed run (ie long sustained rpms)?

We use Permatex 51813 / Loctite 518 in all applications that have machined surfaces and less than 0.5 mm gap to seal. Great product, doesn´t leave any solid residue that can clog anything.

Yeah, I don´t see why anyone would want an engine that has peak torque at redline, makes no sense.
By "good torque" in my previous post I meant "enough torque to make more power than in the next gear", maybe I should´ve explained it better.
That of course is for straight line acceleration and doesn´t apply to the situation that churchx mentions (specially in low gears) where a higher redline can save you from having to make additional gear changes.
And yes, there´s also that urge to just rev the crap out of your engine, no matter the torque or power hahaha :burnrubber:

it is one of the multiple reasons our engines are not good at high rpm's. This is awesome work and maybe just modifying the oil pickup will help a lot of people.

No. cavitation ALWAYS happens in the low pressure parts, actually the part with the lowest pressure is where it starts. It´s the low pressure that makes the disolved gasses convert into real bubbles. As soon as the pressure increases a little, the bubbles implode and cause the metal erosion (this is still in the pump´s inlet).

What it does in the high side is kill mass flow and cause huge pressure oscillations:

(again taken directly from the research cited and linked in my first post)
https://i.imgur.com/YGOCnAv.jpg

THAT is what destroys your bearings by oil starvation.

1. Yes I also expected to see a bigger increase in outlet pressure, it´s still a little mistery to me.

2. The ammount of aluminum it erodes is insignificant (probably less than 1.0 g). I don´t think it would show in an UOA or cause any damage. The importance of finding it is that it´s hard evidence that cavitation is happening. I´d love to see a close up pic of that part of a brand new oil pump.

3. Yes absolutely. My bearing failures where never at the track with high lateral g´s like to think about that. They happened accelerating in a straight line, at high RPMs. It could´ve happened at the dyno, I just wasn´t there.

Well let me rephrase, the erosion from cavatation happens on the high side along with the heating of the fluid.

If someone with a mostly stock engine (boosted but stock internally) wanted to try a better pickup tube, would you recommend they look for something with a one inch OD as well, or would killer b's tube be the correct choice?


Absolutely fantastic research. Thank you so much for sharing!!

You mentioned it somewhat, but is it possible the 10w40 oil is also the problem?


I am running 5w30 for the simple reason that Subaru runs 5w30 in all their FI applications, and I am boosted, but I could imagine running 10w40 or 10w50 on the track. With that said, I think I would drive the car for 10-15 minutes at low speeds/rpms to get the oil to operating temperature, with the concern of starving the bearings running the high viscosity oil.


There seems to be a debate these days whether higher viscosity helps to protect the bearings with a thicker layer of oil, or if it is bad. The argument is that pressures will be higher, which is good, but if oil flow rate is slowed then oil temps can creep up; thinner oil will flow faster, so it will deliver cool oil faster. Total system oil temps may not raise because the oil cooling measures post cooling could be adequate, but locally to any one part of the engine there could be a spike. Thick oil could also cause back pressure, creating resistance against restricted flow, which could be why outlet pressure is static in your chart after increasing inlet pressure.


Also, you mentioned thinner oil having less problems with cavitation. Why is this the case? Intuitively, it seems like warmer oil and thinner oil would result in more bubbles.

This is the same conclusion solidsnake11 came to in the other thread.

Based on this thread I ordered a Killer B pickup tube. Worst that happens is it does nothing, best case that happens is it's a simple change that improves the oiling situation when I'm on track.

The erosion in cavitation happens in the intake (low) side, where negative pressure takes gas out of solution in the fluid, bubbles form and immediately implode eroding the material in the process. In ship propellers or fire truck pump impellers this is in the trailing edge of the blade, for example.

@CSG Mike


Maybe Mike can chime in. Pretty sure he is on a stock engine with decent turbo'd boost/power, and he is probably tracking the car more than most on here. Pretty sure he is running 0w20 oil too, if I remember correctly from other threads.


Poiseuille's equation shows that the resistance to flow will be inversely proportional of the radius to the power of four, meaning a small change in diameter should result in a huge change to flow, so increasing the diameter of the oil pickup and oil galleys should result in a large increase in flow, but there is an assumption that oil starvation was a feed problem. Let's say it was then what would happen if the feed diameter in a system is enlarged everywhere except to the bearings or in multiple other locations? In other words, will increasing only part of the system's diameter be enough, or would the system still be limited by the smallest diameter? I believe the weakest link in the system will restrict flow, or in this case, the smallest diameter pipe will still restrict flow. This is like trying to attach a larger straw to the end of a smaller straw and expecting to suck fluid out of your drink faster. He clearly enlarged many pipes in the system, even the clearance of the bearings, but were the feed holes to the bearings all enlarged, specifically at the delivery end? I don't think so.


The other thing that can effect resistance is viscosity, which will change with temperature of course, but it is proportional to resistance, so the greater the viscosity the greater the resistance. Other factors not mentioned in the equation is turbulence, or other phenomenon like cavitation. The big thing to mention with the viscosity is, if the diameter of the feed is constant because the oil channels at the bearings are fixed, or the bearing itself has a feed diameter that is fixed, then viscosity will be the greatest influence to resistance in the system, which is why I mentioned it above.


The flow rate is also fixed and is proportional to resistance, but this would be good resistance, or rather pressure. Like sucking harder on the straw, or like increasing stroke rate or heart rate on a heart, more flow rate would be better for delivery. There is the Reimax solutions to replacing or increasing the pump volume rate, so this could be a solution, but the problem with increasing pressure or volume/flow rate is that the new volume could go to the path of least resistance, which could be away from the bearings. I have read before that improvements to delivery just results in more flow to the head and not the block


https://en.wikipedia.org/wiki/Hagen%...uille_equation


https://www.counterspacegarage.com/reimax-oil-pump-fa20


https://www.counterspacegarage.com/r...-oil-pump-fa20


Sorry this is long winded, but the other thing that could be mentioned is that there is an assumption like I mentioned above that oil starvation is the issue. Perhaps there is just an inadequacy in bearing size for the forces of FI or the forces from high piston speeds. A larger surface for supporting the loads and dissipating the heat might be advantageous instead of increasing viscosity.


http://www.ft86speedfactory.com/king...l#.XOgeS1iWymw

I would absolutely recommend to get the Killer B oil pickup!

I was going to use the one I ordered straight up, but then I noticed I couldn't because length and orientation where different in the GT86 FA20 and the WRX FA20DIT. At the time I ordered, Killer B didn't have a GT86 option. From the pictures I thought there were the same but go figure :bonk:.

As I had to modify Killer B's pickup anyway, I chose to go with the 1" tube that gave me an inner diameter that better matched the rest of the intake galleries I had improved.

If there's a real improvement between a 3/4" or 1" tube is up for discussion.
Even if it existed, it'd be far less of an improvement than going from stock (18mm, bent, restrictive mesh) to Killer B (3/4", straight, much better mesh).

I've contacted Killer B and Chris has been kind enough to offer to send me one of their new GT86 FA20 oil pickups to make a test in my car: stock vs 3/4" (Killer B) vs 1" (Killer B modified by us). We'll see how it goes, but I'd recommend to anyone with the stock pickup to get Killer B's asap.

The 3/4" vs 1" discussion is better suited to rebuilt engines with looser tolerances and higher flow, so even the objective results from the test we'll do will have to be interpreted correctly and probably won't apply to a stock engine.

OK, who will be the first to go into a parts department and take and post a picture of a NEW oil pump - :iono:


humfrz

I think thinner oil will have less problems with cavitation because it has less resistance to flow which results in more pressure on the inlet (or less vacuum). Interestingly enough, higher viscosity does seem to reduce damage done by cavitation: http://www.aun.edu.eg/journal_files/80_J_7365.pdf

I think the most important factor here for cavitation is the inlet pressure which is caused by major and minor hydrodynamic losses. Major being mostly friction, minor being mostly changes in direction of the fluid. So a thick fluid running through bendy pipe and small diameters is going to cause more vacuum on the inlet. A thin fluid running through straight large diameters is going to cause less vacuum on the inlet.

This is just examining the inlet side of the pump though with the purpose of reducing or eliminating cavitation and increasing flow rate.

You make good points about whether or not this is actually significantly helping on the other side of the pump. Even if most of the increased flow is going to the head and not the block, there should be some small amount of increased flow even through a bottleneck. Again, whether it's significant or not, or even the correct solution, is another question.

Edit: Removed a possibly incorrect statement about oil viscosity and its relation to cavitation.

I'm running a 100% stock engine.

Increased flow isn't always a good thing, when your return system/mechanism is unable to keep up under cornering.

I'm also tuned for 7800 RPM, but I don't make it a habit to stay up there.

I never actually expected a stock engine to last this long, but I also have more supporting mods than most, and probably more tuning hours (2 days with Bill and 80+ with Zach at @DeliciousTuning on the turbo tune alone) than most too.

So cavitation will be reduced by higher viscosity, but the vacuum will be increased on the inlet side due to increased friction losses. I think the question is which of the two effects is stronger. I think it doesn't matter either way, increasing the diameter of the inlet to accommodate the viscosity that will work best for the engine is what's important.

Dude this is a discussion that has existed for ages and will probably never be settled, you can read for hours arguments for both on BITOG from very reputable persons.

I think in the particular case of an inlet restricted oil pump, higher viscosity is worse because it will cause cavitation sooner with the same flow. High viscositý will have lower intlet pressure than low viscosity because it creates a bigger pressure drop with the same flow. Once the pump is cavitating heavily, the thicker oil film the high viscosity gives you (it's only advantage, actually), doesn't matter because the oil just doesn't get to the bearings. This also depends on the PRV on that particular engine, so it becomes a complicated discussion. One of my engine failures was using Motul 10W60 oil, I used it thinking of the same reasons you describe.

I recommend you pursue a solution, rather than keep trying to fix the symptoms.