Originally Posted By: ZeeOSix
Originally Posted By: dnewton3
And, let's not forget that my 15,000 UOAs also show no correlation in selection of various filters in normal OCIs to wear control either. The filter efficiency delta (ranging from 80% to 99% at 20um) doesn't make enough difference that you can distinguish them in analysis.
It's been discussed before that normal UOAs don't have the ability to accurately determine how much effect the oil filter has. I thought there was some kind of special particle counting test up and beyond a normal UOA that would need to be done in order to see that effect. And just really how accurate is the typical UOA on "wear particle" counts?
Wasn't there someone who actually ran an OCI without an oil filter at all to see what effect that had on the UOA? Thought there was, but if not it would be interesting to see.
It's true that UOAs don't have the ability to see all wear; hopefully folks realize that. But it's also true that a PC cannot distinguish between all the different manner of particles in terms of composition; it can see size, but not origin. It would not know if at 10um particle was soot, AL, or Fe or ??? So while PCs help understand the quantity and size, they say nothing about "wear". A PC that saw 25 particles of 7um, would not know if they were 20 particles of soot and 5 of Fe, or 20 particles of Fe and 5 of soot. PCs can tell us how much stuff is in the lube stream, but they cannot tell us WHAT is in the stream.
Wear metals in a UOA are not all telling, but thet represent some reasonable ratio of wear that can be inferred. If you have a LOT of Fe and Al in a UOA (all 5um or lower by the nature of the test) then it's very likely you've got a LOT of wear particles that are also larger than 5um of Fe and AL. But the converse is also very likely. If you have very few ppm of wear metals, it's likely that wear particles of ALl sizes are very few. Would we believe that somehow and engine is ONLY going to shed large quantities of big, fat stuff, and never any small stuff? That's absurd! So yes, I agree a UOA won't show us all wear; not above 5um. But it will show us a snapshot of wear in a range, and we can make a logical conclusion as to the overall effect of total wear. Few ppm = low wear in all ranges. Lots of ppm = lots of wear everywhere. I admit I don't know of any study to show this, but it seems quite logical to me. Does it not to others?
It would break down like this; there can only be four possible conditions of UOA data and overall wear:
consider the delineation of a 5um break-point; what a UOA can or cannot see. Overall wear would be ALl wear particles of any size.
1) low UOA ppm and low overall wear; very easy to believe and likely to exist
2) high UOA ppm and high overall wear; very easy to believe and likely to exist
3) low UOA ppm and high overall wear; not very plausible and would be difficult to comprehend how this could exist
4) high UOA ppm and low overall wear; not very plausible and would be difficult to comprehend how this could exist
I know of no study or SAE paper that proves (or even discusses) the presence of wear above 5um, with an accompanying total absence of any wear data below that threshold.
UOAs will NOT catch the impending onset of a major catastrophe; they won't tell you if a con-rod is about to shatter, or a piston wrist pin will crack apart at 8k rpm in a drag car.
UOAs will tell you of chronic onset of problems. Even big metal particles that come from being plowed out by a soot particle, might likely become smaller over time as it not only abrades stuff, but also gets beat up and abraded itself. That's the dichotomy of wear; soot will get bigger after the add-pack is overwhelmed, but wear metals don't co-join (agglomerate). Wear metals will essentially not only do some damage if large enough, but also get damaged themselves. Shed off a 10um particle of Fe, and it's not likely to be caught by a typical FF filter; certainly not the first several passes. But while it's causing damage, it's also getting beaten into a smaller particle, and quite likely, some portion of it will end up small eneough to be seen in a UOA.
So the relationships I suggest are valid are examples 1 and 2 above. Example conditions 3 and 4 are VERY UNLIKELY to exist.
So while some folks would eschew UOAs as not useful for determining wear, I say that they are certainly helpful in understanding a likely ratio of wear. I cannot tell you how many ppm of Fe (below 5um) would equal X ppm of Fe at 10, 15, 20um. But I do believe that some ratio exists. And the ABSENCE of wear ppm in a UOA is VERY LIKELY to be an indication that there's little wear of any size. Otherwise we'd have to believe that wear is discriminate, and can somehow hide itself above 5um all the time. Just pure bunk.
The GM filter study used component mass analsysis to determine wear. They did not use UOAs. But that does not stop them from stating that normal field OCIs will not show filter performance differentials. GM knew at the time that because they omitted OCIs, they cannot claim any correlation would exist in real life.
My data simply comes to the same conclusion using a totally separate methodology. Filters have shown no ability to create a discernable controlling effect on wear in normal OCI applications. When I say this, I don't mean to say that filters are worthless; please do not misunderstand me. What I am saying is that a 90% eff filter at 20um and a 98% eff filter at 20um cannot be distinguished in terms of wear.
Here is the distinction that is CRITICAL to understand:
- Some folks are going to claim that UOAs are unable to see the difference in wear-rates due to efficiency differences. But, that makes a HUGE assumption that a difference actually exists! It PRESUMES that there is an effect.
- Says who? Does it not occur to anyone that the reason there's not statistically distinct wear-data trends between filters is because IT SIMPLY DOES NOT EXIST IN THE FIRST PLACE???
Don't jump to an illogical conclusion of presumed presence. That's a bad mistake to make.