I find this topic interesting because it's part of the ongoing debate bewteen viscosity and wear verses EPA requirements.
The two methods used are Resistive and Capacitive. In the first method, the oil film is considered to be a high resistance resistor in a circuit. In the second method (Capacitive), the oil in the clearance is considered as the capacitor's dielectric. The capacitive method is the most used since it has been found to be less variable (more repeatable), and more accurate overall.
Here is the test set-up:
Four different viscosity straightweight (monograde Newtonian) oils of 5, 20, 30, and 40 weights were tested in an operating GM 3.8L V6. A plastic shim was placed on the outside of the bearing shell (for insulation) and an AC signal of greater than 100 kHz and about 500 mV was applied to the journal and the bearing shell. So essentially, the inside of the bearing (the metallic bearing shell), the clearance, and the journal form a capacitor, with the oil film acting as the capacitor's dielectric. The dielectric of the oil is measured and the film thickness is then plotted as a function of capacitance. The measurement is called MOFT, for "Minimum Oil Film Thickness."
Oil inlet temperature to the bearing was maintained at a median temp of 120 C. The engine is brought up to temp for one hour, idled for a few minutes, and then brought up to an rpm of from 1600 to 2500 rpm. The bearing/journal clearance was 20 um and the test started with 40 weight (monograde) oil. The oil film varied from a high of 10 um at a crank angle of 270 degrees to a minimum of 3.8 um at 450 degrees crank angle, all at 1600 rpm.
For every ratchet of viscosity grade downward, the minimum oil film is reduced by an average of 1.25 um.
So by about a 5W, the hydrodynamic film is near rupture at a crank angle of 450 degrees, and the lubrication is now mixed hydrodynamic/boundary. This is where boundary protection additives come into play.
The higher the rpm, the thicker the film for any of the mongrade oils. The highest film thicknesses came at 2500 rpm and 50% torque.
In a another phase of the test, multigrades (Non-newtonian) oils with VII's were used. Overall (a generalization), multiviscosity oils showed slightly greater average film thicknesses, lower bearing temperatures, but higher shear rates than did the monograde's. [to be expected].
From my reading of the four papers on this subject, it appears that the minimum weight of oil that can be used in modern engines is SAE 5 weight, and the maximum weight appears to be 40 weight. I believe this is why 5W30's and 10W30's have been the most popular oils for the last three decades; they appear to to be the median weights for all around service for oil temperatures at approximately 250 F bearing inlet temps. This may be another reason why 5W40 oils are becoming increasingly popular as well.
As an added note, a number of companies were involved in the ASTM task force test:
Exxon/Mobil, Pennzoil, GM, Shell, Texaco, Ethyl, Lubrizol, Rohm and Hass, Southwest research Institute (SWRI), Cannon, and Paramins.
[ August 20, 2003, 03:35 PM: Message edited by: MolaKule ]