The primary goal of the research was really to see whether or not neutron activation analysis (NAA) was better for quality control purposes than the industry standard, ICP-AES. The uncertainties show that NAA is MUCH more accurate than ICP-AES. I understand that most oil manufacturers don't necessarily have regular access to a nuclear reactor, but ICP-AES is a really lousy way to determine additive concentrations (one guy that we worked with referred to ICP-AES as a "really expensive random number generator").
Which brings me to the 540 Rat blog article...
The guy seems to know what he's doing. As he points out, a "Mechanical Engineer is clearly the most qualified Engineer to test motor oil" (but evidently not the most qualified engineer to know the difference between proper nouns and regular ol' nouns ). The only problems I have with his data:
1. He doesn't provide any uncertainties anywhere. Seriously, numbers without uncertainties are meaningless. If you don't believe me, I'll sell you my 100,000,000 hp 2002 Ram for $100,000 cash (note: +/- 99,999,775 hp). You never know, it could have 199,999,775 hp and be a good deal...
2. The other problem is this statement:
"The ppm (parts per million) quantities of zinc, phosphorus, moly and in some cases titanium, shown in the ranking list below, are taken directly from the Lab Reports that came back from the Professional Lab 'ALS Tribology' in Sparks, Nevada. Some oils have MORE ZINC than phosphorus, while other oils have MORE PHOSPHORUS than zinc. It just depends on the particular oil’s formulation. Either way, the numbers below are correct and are NOT typos."
The ppm data are mass fractions. Because zinc is ONLY added to oils in the form of ZDDPs (while the phosphorus in oils can be from a variety of molecules) and because ZDDPs always have a 2:1 phosphorus atom to zinc atom ratio, the phosphorus to zinc ppm ratios should never be less than [(30.97*2)/65.39]=0.947. It's pretty basic chemistry. If you take a look at his measurements for "5W30 Joe Gibbs Driven LS30 Performance Motor Oil, synthetic," for example, you can see that the ratio is 1496/1610=0.929. What does this mean? That his data (or the data that he got from ALS Tribology) has some really high uncertainties.
That said, there's a lot more to a good motor oil than ZDDP-content. The best way to measure the quality of a motor oil is through mechanical tests, which is exactly what the guy who wrote the 540 Rat blog article does.