Most oil analysis labs, including Blackstone, get wear metals counts from ICP spectroscopy. ICP only catches particles up to a certain size, and not all mechanical wear chucks out particles in that size range. Some wear modes produce bigger particles, to which ICP spectroscopy is completely blind. Furthermore, ICP-detectable particles can come from chemical processes that have nothing to do with what most people would consider mechanical wear -- processes like acid attack from overextended oil change intervals, or highly active surface chemistry from aggressive anti-wear additives in the oil. So, if you see big numbers, it's difficult or impossible to say whether they're anything to worry about. And if you see low numbers, maybe everything's fine, or maybe there's a wear mode that ICP spectroscopy can't detect. You can't tell anything for sure just from looking at the report.

The S54 and S65 with the original BMW rod bearings are lucky in that they sometimes produce lead and copper particles in ICP-detectable sizes when their rod bearings wear abnormally, AND they don't seem to produce ICP-detectable lead or copper in significant quantities for any other reason (barring leaded fuel or new oil cooler hardware or something). So, if you see those high numbers from one of those engines, then yeah, it's probably rod bearings. But again, if you don't see those numbers, you have no idea.

As for the updated BMW rod bearings on the S65, or any aftermarket rod bearings, we don't know how or to what extent they will show wear on ICP spectroscopy, if they even do at all. Aftermarket bearing manufacturers sometimes say they went for leaded bearings partly to facilitate oil analysis, but there's been no rigorous testing of that idea and the data from the field is incomplete. What little evidence we have on the updated BMW bearings for the S65 does not suggest that oil analysis is useful for tracking wear in them -- though we don't even know what we're looking for because we don't really know what they're made of (the initially-accepted idea that they were bimetal aluminum/tin bearings turned out to be untrue).

ICP spectroscopy can be useful for tracking early-stage wear when you have a lot of background knowledge about the application in question. You have to know in advance whether and how any potential wear problems will show up on oil analysis, and you have to know about any possible sources of false-positives. Then you have to sample frequently enough to catch the wear while it's still in that early stage, before the wear particles are too big for ICP spectroscopy to detect.

In other words, if you're an end user, running oil with unknown chemistry, in a hand-built car engine with largely-unknown metallurgy, operated in the real world under uncontrolled and constantly varying conditions -- which, if you're reading this, you probably are -- you should assume that trying to use oil analysis to catch wear is mostly futile. And that's the main problem here.

It's generally assumed that frequent and regular sampling, starting from just after a build or repair, can reduce the futility. That's why I'm doing it. But I'm sampling every 5k miles and using a slightly more extensive analysis package (Polaris Labs Advanced Engine Plus), and I'm not taking low numbers to mean a clean bill of health. I'm already planning to revisit my rod bearings within 60k-80k miles after the first change (I have an S65). What I'm looking for are signs that I might have to do them sooner, as well as signs of other kinds of trouble like fuel dilution and other contaminants -- which, by the way, is where oil analysis is truly useful in a car engine. Maybe that's for another post.

In sum:
  • Is a single oil analysis or a few them at irregular intervals better than nothing? Sure.
  • Is it more informative if you have a complete trend? By a mile.
  • Can it provide information on important things, like how your engine is running or how your filters are doing? Absolutely.
  • Is it enough on its own to tell you whether or not there's mechanical wear in your engine? Not at all.