Size Distributions of Particles Extracted from Disk Drive Parts: Comparison with the MIL-STD-1246 Distribution

Abstract

MIL-STD-1246B states that naturally occurring particulate contamination on surfaces follows a lognormal distribution with a geometric mean of 1 micron. Thus, when the logarithm of cumulative particle count is plotted against the square of the logarithm of particle size, an approximate straight line should result. However, deviations from this "natural" behavior have been observed in many manufacturing environments involving specialized machining or cleaning. Complex disk drive components, such as machined castings or ground surfaces, must be inspected by first extracting contaminants into a fluid by some means (e.g., ultrasonically, or with a pressurized spray), and then monitoring the particle population in the fluid. The cleaning method and the extraction method, in particular, can cause a significant departure from the lognormal approximation. The substrate material also influences the extent of this departure, as does its cleanliness level. These aspects are brought out by the plotting of actual particle size distributions obtained by extracting particles from substrates that are cleaned using a variety of methods. These are compared against the MIL-STD-1246 distribution. Conclusions are drawn regarding the effects of the cleaning process and the measurement procedure on the applicability of the MIL-STD distribution for various substrates. Implications of these conclusions for the derivation of particulate cleanliness specifications and for modifying MIL-STD-1246B are pointed out.