Abstract
This brief paper is based upon a full report of certain experiments conducted for the ASME Research Committee on Lubrication. The experimental work comprised friction measurements under imperfect lubrication conditions up to loads of 10,000 psi or more. Eight lubricants were compared whose viscosity-pressure characteristics were known. A Kingsbury oil-testing machine was used, with bearing surfaces designed to eliminate hydrodynamic action so far as possible.
A condition of mixed lubrication prevailed, in which rapid wear of brass on steel occurred simultaneously with hydrodynamic effects. All tests were run well to the left of the familiar minimum point on the coefficient-of-friction diagram. Three methods of plotting the data were tried: (a) The direct plot of friction per unit area against load per unit area, holding the speed and viscosity constant; (b) a conventional plot of the coefficient against Z1N/P, where Z1 is the viscosity at atmospheric pressure, N the speed, and P the load per unit of projected area; and (c) a modified diagram in which each of the foregoing curves is replaced by a family of curves with different values of the pressure-viscosity parameter. This parameter, the product of load per unit area by the pressure coefficient of viscosity, helps to explain earlier discrepancies and to set up a better correlation.
The usual contrast between friction of fatty and mineral oils was confirmed, and it was concluded that boundary lubrication and wear are accompanied by hydrodynamic action on a microscopic scale.
