Originally Posted By: ZeeOSix
Originally Posted By: dnewton3
Being a "micro analysis" example, this was just one vehicle with one operator over several years of use. About as "controlled" as one can get without being in a lab. And just about as good a real-world example as I can hope to see.
Even though it might have been a "well controlled garage test", it still might not have been controlled well enough since we are looking for small differences. I'm sure if you could absolutely control all variables perfectly and only change the filter's efficiency in testing, you might see slightly stronger correlation, but I get that it's still in the noise level.
And again, the suspended particulate size range used in UOAs doesn't really even lead itself to being useful data to correlate filter efficiency - so it' kind of a bad source of test data to even use for filter beta correlation. You would need to have a UOA that measured everything below 60 microns instead of everything below 7 microns to see the effectiveness of an oil filter in the true range they are meant to filter the most particulate.
Given that, to me it's still better to use a higher efficiency filter because it certainly can't hurt anything unless someone is absolutely pinching every penny they have, or other aspects of a particular oil filter is more important to them beside filtering efficiency.
Well - this is why we don't see any SAE studies proving the correlation; they likely isn't much, if any.
What I can say, and it's a generally accepted concept in the world of high-dollar large lube maintenance programs, is that the presence of small wear partciles can come from two sources. It either comes from "normal" wear or "abnormal" wear.
It is VERY HIGHLY UNLIKELY that some small but catastrophic event would cause ONLY large particulate to shed. If a bearing were plowed up from a big offending particle or a cam follower gouged, and large particles were created, they would not be exclusive; it is reasonable to believe that small particles would also be formed at the same time. When you dig up dirt with a shovel, you may be able to get a big clump, but there are always small bits falling off; it's never a clean cut with no remnants. Or, to put it in more of a medical viewpoint, it's not an incision but an avulsion. When a bad event occurs inside an engine from a particle large enough to cause damage, that event is typically a violent one that tears away the weaker material. It's not a slice, but a tear. And therefore, small particulate comes off along with the big stuff. So while you can see particles large enough to be caught in the filter, you'll probably also see particles small enough to be discovered in the UOA.
In contrast, "normal" wear is most often from the abrasion of metal to metal, or simple sloughing due to fluid shear, etc. This typically will NOT result in large particles, but only small ones. Again, UOAs can see this, but they'd never show up in a PC.
The real problem is that there's simply too much variance in the response of UOA data to correlate any one event relative to what a catastrophic response would be in "normal" views. Each large event is somewhat unique and unpredictable, whereas the small events are very predictable. Think of the difference between a typical windy day versus a violent storm with a tornado. It's easy to predict normal weather changes, and we can predict the POTENTIAL for violent storms, but we can NEVER know ahead of time EXACTLY where the funnel cloud would touch down, or just how long it would last, or how wide it would be, etc, etc. The relationship between UOAs and PCs is very similar. I can tell you with certainty what "normal" wear should be like, but I cannot predict large particle events. Large particle events are fairly rare; if they were not our engines would suffer quite quickly. Many decades ago it was not uncommon to have engines wear out in less than 75k miles (or at least degrade to a very appreciable degree), but today that would be a rarity, as most well designed, made and cared for engines can EASILY go 3x or 4x that distance.
UOAs most often, but not always, can show elevated numbers from a big particle event. And PCs can show some response to "normal" wear. A PC or UOA will never be able to predict a true acute failure (broken con rod for example), but they can show tell-tale signs of impending doom, if you know the real wear data and variance for your application. UOAs and PCs can speak to chronic issues.
Much of the consternation comes from the fact that, while there may well be an incidental coincidence between some event causing both large and small particles, there's not any steady relationship between the percentages of those separate results. It's not like 770 particles at 20um will exist with exactly 2560 particles of 3um of Cu; it' just never is that clean of a line.
And that's why finding a true causational relationship between the two eludes most everyone. It's not like we've stumbled onto a new topic never discussed before, guys. This is old hat in true Predictive Maintenance circles. Most people here just play with UOAs as toys; that's OK. But there are folks out there that make their living running huge lube programs, and if they cannot find a direct link between UOA and filter efficiency data, what makes you think we can? Most anyone would agree that filtration will never hurt the lube system. But what many really entrenched folks can say is that really efficient filters often don't pay for themselves, once "clean enough" is established, in terms of equipment wear.
Where ultra-efficient filtration pays off is in the lifecycle of the LUBE, not the equipment. If you think back to many of my former posts, I've always said that premium filtration isn't about making an engine last longer, it's about making the lube last longer in service. I'll never say that "better" filtration is a waste, but I will say that "better" filtration will be wasted if you don't understand what it's to be used for.
