I was hoping to improve the collective insight on this subject. While it has been a good conversation, I am afraid I only succeeded in deepening the questions that cannot be answered. This is a very complex system, far beyond the questions I raised, and our narrow, simplified view of the situation is simply ignorant. Even the principles and 'laws' that we learn in school are idealized and not applicable to real-world situations as I tried to think through. Without getting into details, I discussed this at length with someone who works to advance the accuracy of molecular modeling for a living. We haven't a clue how complicated this is. What is needed is the empirical data that none of us have access to. Comprehensive well designed and executed experiments are expensive, and likely even the filter manufacturers do not have the scope of data that we would like to see. They primarily need to be able to market some new 'innovation' or hype up some irrelavent characteristic such as the hole count and continue to whittle the cost to cater to those who buy solely on price, not come up with the best filter.
I don't think that we are in any better shape than when we started. Fundamentally the filter and engine manufacturers do not provide concrete empirical information regarding flow vs. pressure vs. viscosity and dirt load. Publishing what they have would only open them up to critique and would not be beneficial for them.
I have always used the Amsoil EAO filters, but they are expensive, but worse they were unobtainable for a long time. So I looked around and found WIX for 2-3x less money! Amsoil sold them before they made their own, and still sells them, so I took this as a recommendation. So I considered if they were suitable, and before long this thread was born.
I believe that I am going to blindly stick with the Amsoil filter because of personal experiences with the oil and several other products. I have decided to trust them. They seem to base their business on making money by making superior products, bought by people who are willing to pay a little more. I may very well be paying $8 more per oil change ($0.001 per mile, $300 more over the life of my Subaru) more than I have to. If this extends the life of the car by 4 weeks over 10-15 years, it will have paid for itself. But am I paying more for the right filter? Worst case I bought a good, competent product and will have wasted $300 in a well-intentioned attempt to extend a $35,000 investment. I am betting that I will do at least a little better than that, and being a long lever arm, I hope significantly better.
Out of appreciation for the effort and for completeness of the thread, I have responded below to the posts since my previous posting.
"Millions of people use them" and "I ran ABC filter for decades and my cars never blew up" miss the point. Damage due to poor lubrication is never instantly catastrophic. If it goes another oil change or two, any failure is not going to be associated with the oil used a couple of changes ago. We all expect our engine to wear out eventually. If a wrist pin bearing goes at 187,000 miles, do we blame the oil we used? No, we just figure it wore out and it is time to buy a new car. Yet the oil, filter, and our oil change interval may have contributed to more wear per mile than would have occurred if the engine had been lubricated better. I would be willing to bet that if you replaced the filter with an empty can and changed your oil per the manufacturer's schedule, or even half as often, that the majority of the time the engine would get to the end of the warranty without a failure. Depending on the warranty length, oil change interval, and the price of the filters you don't buy, you would save a small few hundred dollars. But engines and cars are expensive, so I am trying to skew the odds in my favor. For me, if I extend the life of the car by a month I will have likely paid for all of the better filters, and if I extend it a few months, for the better oil too. Everyone has to make their own decisions based on their own set of incomplete information and personal intuition. I have decided that quality is a big lever on the rate of engine wear.
'Supply chain' issues for many filters: If we could not get ANY filter, we would be in big trouble. I will keep a stock in my garage, and a list of second-choice filters.
ZeeOSix in post #3, I am quite sure that we agree in our understanding, however it may have appeared from my words.
ZeeOSix: Absolutely, the curve is only one data point, but it is one data point of the missing information. And one is infinitely more than zero! If we had a family of curves vs. measured viscosity that would be awesome, then we could take our favorite oil at different weights and temperatures and see how the filter behaved at that point. Regarding "If you're worried about the filter bypassing too much on a high pump flowing Subaru, then use filters with the high bypass setting specified by Subaru." I will address this later, but I have not seen a filter with the Subaru bypass valve pressure. For now because there are two branches to the conversation I will overlook this and address your statement as though there were such a filter... Continuing: Yes but if it is set too high, that will leave less pressure to push enough oil through the engine. If we never go above 6000 PSI (Which would probably be wise) evidently the positive displacement pump moves 58.1 qt/m presumably out through the engine, and even more than that actually pumped but bypassed inside the pump because the output pressure of 46.8 PSI exceeds the 21.7 PSI of the first stage internal shunt/relief valve which MUST be open in this situation. But all we need to know is that it is moving 58.1 qt/min externally. Nominally 23.2 PSI of the pump output pressure is dropped across the filter (Like your resistor to use the electrical analogy), leaving 23.6 PSI to move that oil through the engine. Though not stated, what external "load" could be more appropriate for this experiment other than the intended engine, and FB25?
So if the WIX opens at 27 PSI that is about 4 PSI higher, leaving only 19.6 to push oil through the engine. But the irresistable flow determines the pressure with a positive displacement pump, not the other way around. More oil will be shunted inside the pump because of the higher output pressure forced by the 27PSI filter bypass valve, so the engine will get less oil. Is this OK? Geez, I don't know! If the engine with the Subaru filter has an overwhelming greater amount of oil than is needed, it is probably fine. If it is marginal under some conditions, it will not be fine under that condition. I do not know how the relief channel for the 21.7 PSI first stage flow rate will vary vs. a 16% increase in pressure, but it will be more, but probably not proportionately more.
As many have stated, there has been discussion along these lines before. In one of those threads someone stated that engines are never starved for oil at high RPM, starvation only occurs at very low RPM. This would be very interesting to explore. If the engine is developing more torque at the high RPM (What RPM is the peak of an FB25 torque curve?) then there may be more radial (lateral?) load on the bearing, requiring MORE pressure to maintain the film. We need a bearing expert. Some of the bearings are throwing pistons back and forth and the forces must get enormous (Probably why wrist pin bearings go!) It gets very complicated very quickly.
I see the superficial argument that a higher filter relief valve pressure is good in that it avoids having the filter dump dirt into the engine as frequently (as my original note projected to happen in the 3xxx RPM range), but it is not that simple. And metal on metal would be catastrophically worse than a little dirt. Again, lack of knowledge about how much of that 58.1 qt/min oil flow the engine really needs at 6000 RPM.
Getting back to ZeeOSix' graph of PSI vs. GPM for a Purolator PureOne which has synthetic blend media, this is the closest thing we have to a fact. Is this is representative of all synthetic blend media? How is completely synthetic different? How is all cellulous different? How is it different for different viscosities, like 0W20 or 5W30 at 0 or 10 or 20F? The answer easy directionally, but how much, and with what shape curve? We don't know, but at LEAST we see that for this case, and likely a clean filter and clean oil, the pressure is really quite low. The highest rate we talk about is about 60qt/min which incurs a 7 PSI pressure drop. It would not be hard to imagine that a little dirt in the filter and cold oil would have dramatically more resistance than a brand new clean filter and oil, opening the relief valve. Similarly it seems to be common knowledge that synthetic blend filters require much less pressure. Fram Tough Guard are marketed as being "synthetic blend" too. WIX 57055 is marketed as having "Enhanced Cellulous" media (whatever that means). Does that mean it requires more pressure? And paper is what synthetic is less than, so it must be more.
Brian533, my mistake! Good catch, thank you. The pressure is 5.8 PSI at 600 RPM, NOT 40 PSI, I read the wrong number off of the table. 600RPM pumps 6.1 US qt/min at 5.8 PSI so the valve will be closed. The Purolater PureOne at 200 F and 5W30 per the graph ZeeOSix provided will drop 2.3 PSI, leaving 2.8 PSI to force oil through the engine. Evidently that is enough or it would be higher. We really can't go anywhere with this because we (I) don't know how pressure varies vs. flow through either the filter or the engine. The positive displacement pump insists on driving that 6.1 qts: That is the independent variable. The 5.8 PSI is NOT a characteristic of the pump at 600 RPM, rather it is the outcome of *whatever it takes* to move that 6.1 qt/min through the engine. It happens to take (evidently) 2.8 PSI! And the engine DOES get 6.1 sq/min. So no, the pump bypass valve does NOT open, nor does the filter relief valve.
The 6000 RPM data I quoted was correct: 6000 RPM, 48.4 qt/min, at 46.8 PSI. Recognize that these are the pump specifications as a black box: Whether or not an internal bypass in the pump opens is immaterial: At the given RPM it moves this much oil OUT of the pump if this is the back-pressure. If the back-pressure is lower or higher, less or more will go through the internal pump bypass. That back-pressure and the resistance of the filter+engine determines how much oil takes that route, the rest will go through the pumps internal bypass (The qt/min number). The two routes (pump bypass vs. filter+engine) will find a balance based on their respective resistance to the oil flow. Whether or not the filter relief valve opened is immaterial to *these* numbers.
On the other hand, as a system we care what the pressure is because if it is above the filter relief valve pressure it opens and alot of dirt is flushed out of the filter and through the engine. There has to be some context for these specifications, and the stated "load" on the pump is a specific oil filter which we can safely assume is the Subaru filter. The only specifications that matter are the by-pass valve opening pressure which they give, and the pressure vs. viscosity and flow 3-D graph, which they do not provide. But we can say that for the 600 RPM case, the valve does not open, otherwise the discharge pressure of the pump would be something north of 23.2 PSI. For the 6000 RPM case we cannot know whether or not the valve opens because we don't know the pressure drop through the engine or the pressure drop through the filter (One of these would do since the total is given in the FSM as 46.8PSI).
WPod: Yes, the oil filter bypass valve opens at 23.2 PSI. The alternative route (That I was not talking about) are the bypass valves internal to the pump at 21.7 PSI and 82.6 PSI. I agree with your explanation of what is happening. Nobody makes a filter that is "proper rated" except for Subaru (Which is "proper rated" by definition vs. by specification because the filter is incompletely specified). When I use the Subaru filter with 0W20 oil the engine rattles alarmingly for a very small number of seconds then quieter until I drive it a several hundred feet. The Amsoil filter does not do this, it rattles for some fraction of a second and quiets down much more quickly after that. Using 5W30 seems to be markedly better with either filter, but follows the same pattern (Noisy with the Sube filter, much quieter with the Amsoil. I wish I had better records to quote, but that is the gist of it. The devil of it is that my son has a 2017 forester with the same engine that does exactly the opposite (!). There are also people on both sides of the fence on this topic on various forums with respect to the Subaru filter. What a zoo. When my engine sounds like there is a handfull of bolts rattling around under the valve cover for a couple of seconds then trails off, it is hard to be happy. Whatever else is true, that cannot be a positive thing. For me, the "easy button" of just using the Subaru filter is cringe-worthy.
ZeeOSix: I emailed WIX twice. They did not respond to either email. Regarding viscosity, I have been using Amsoil oil for twenty years. Whenever it gets cold oil had always been visibly thicker - until I started using Amsoil. I am sure that it is thicker when I change my oil on a frozen driveway when the oil has been in the unheated garage, but since I have been using Amsoil, not noticeably. And the vehicles don't crank much if any slower either. Is it that cars are different than they used to be? I supposed that may be part of it, but the oil has to be part of that too.
SubieRubyRoo, What do you base your statement on that the 57055 with a 27 PSI bypass meets the Subaru FB25 requirements? The Subaru FSM specifies a filter with 23.2 PSI, not 27. The WIX filter does NOT meet that Subaru spec. If that is an indication that the WIX filter requires more pressure than the Subaru filter to force a given volume of oil through, that means that at any given RPM there will be about 16% *less* pressure available to force oil through the engine until the bypass valve opens at 27 PSI. Once the total pressure is above 21.7 (Above about 21.7/5.8)*600 = 2244 RPM) the pump internal shunt is siphoning off some of the volume so the engine sees less volume to boot. Only some unknown fraction, presumably not a ton less. But less.
It is not my intent to continue this thread unless there is new and substantial information. If something is added that leads to a more concrete, fact-based analysis, that would be great.
Thank you again.