@dnewton3
Since I have a background in stats, I can appreciate all the x-bar R and SPC stuff and confidence intervals and Z testing and such. I've done the R90C90 work and can sort of intuit that stuff now. I've analyzed data from MATLAB to JMP to MINITAB to EXCEL and back-- everything from reliability/warranty data to engine test cell data to field test data. I'm talking data files where the farthest column is XIV and tens of thousands of rows.
And I'll back you up-- if you want to publish a paper in a lubrication journal declaring one oil better than another, you need a massive sample size and a good gauge R&R study to accompany the lab data. NIST traceability, calibration certs, all that.
But I think you're missing the mark in insisting that you can't learn *anything* to distinguish oils short of such a scientific study. You don't just magically transition from "completely inconclusive" to "Nobel committee called me" on a step function basis. There is a spectrum. And there is useful information to be gleaned. If I shot a rifle one time and hit a 12" target at 100y, how accurate/precise is that rifle? I can't say with certainty. But I know there's a good probability I can hit a 12" target with hit because my first shot did just that, and-- almost tautologically so-- the first shot represents the most likely outcome. So my confidence in hitting it a second time is much higher than if I'd never shot it at all-- because it's already done so once. So with a sample size of two, I've gleaned useful but anecdotal information.
If I move a light switch up and down, the light goes from off to on and it doesn't take very long to become convinced that moving the the switch causes the light to change and that moving it one direction vs the other always causes the light to move the same direction.
Now what if instead of the light going off and on it just got the tiniest bit dimmer or brighter? Twice in a row it got brighter when you moved it up and got dimmer when you moved it down. When do you conclude that one position of the switch is brighter than the other? Do you need a NIST-traceable light meter with precise quantification of luminous flux? Or can you trust your eyeballs enough to have a reasonable confidence that one position is brighter than the other?
So what are reasonable statement that can be made about the UOA shown in the OP?
-- The change in aluminum levels is nonexistent and essentially zero. There' no difference between these oils.
-- There's no discernible effect on iron levels. Changing back to PUP had no effect on iron vs SS iron levels. There's no difference between these oils.
-- The copper difference is *absolutely* discernible and real. Two runs of SS trended copper much higher and kept going, only for the PUP to revert to the same prior copper level, which was half of the SS value.
What's misleading in the UOA report is labeling copper a "wear metal" when it's very likely to be leaching vs wearing. Copper is catalytic and quite reactive. And how exactly does one "wear" copper into the oil without also wearing any aluminum, tin, zinc, or iron?
To my knowledge, pure copper does not exist in most oil-wetted surfaces in a modern engine. It only exists as a brass or bronze alloy. I know that's true of Cummins engines at least because our engineering standards forbid pure copper on oil wetted (and fuel wetted) surfaces. The injector combustion seals are one of the only places you'll see pure copper on a Cummins engine.
Which means any wear particle generated that contain copper will
also contain the other alloying elements-- tin or zinc especially, but perhaps even lead.
The only way copper can spike up without taking any other element with it is if it is, in fact, leaching and not wear.
So at the end of the day, I interpret this UOA as saying there's no discernible wear performance difference between the oils, and that the high ester content of the Amsoil is leaching copper from a bushing alloy somewhere (presumably valvetrain), and this will come back down after the surface depletion and a bit of passivation.
If a third run of SS showed the copper staying higher at 10ppm/1k miles, I'd abandon SS for this engine because chronic high copper levels are an accelerant of piston deposits. And since Amsoil's stout ester dose is presumably rooted in cleanliness desires, that would be counterproductive.
Based on copper alone, I'd say this is a "win" for the PUP and there's no need to pull 30 samples to reasonably conclude that. The copper went up and the SS went in and it came back down when the SS came out. Cause and effect.
There's no way to translate these results into engine life or failure probability. But that's not the point. The point is upvote vs downvote. To me, the copper leaching of Amsoil gets the downvote here and I'd recommend the PUP here confidently and not need 30 samples to have that confidence.
