Volume 21, Issue 1 (January 1993)
Up-and-Down Methodology Applied to Statistically Planned Experiments
Twenty-four modified small sample up-and-down tests were conducted to estimate the median endurance limits of unplated and cadmium-plated ½-13NC 2A machine screws of SAE Grades 2, 5, and 8, each fastened using standard geometry SAE Grade 2 zincplated nuts or Ranged SAE Grade 8 nuts (termed toolmarker's nuts). The test program was a split-plot experimental design with six mainplots (three grades of machine screws, each unplated or cadmium-plated) and two split-plots (nut-type) within each main-plot. This test program was repeated to form two time-blocks, each time-block involving twelve modified small sample up-and-down fatigue test sequences conducted under test conditions that were as uniform as practical. Based on prior test data, the effect of the split-plot factor, nut-type, was of primary interest in this study. Two statistical analyses are discussed herein: the standard (parametrie), analysis of variance and an alternative nonparametric analysis. Recommendations are made on how to conduct similar multifactor test programs involving up-and-down test sequences to estimate the median endurance limits of interest.