The CCT (Certified Calibration Technician) being developed has passed the first hurdle: developing the body of knowledge (BOK). 12 Metrologists (including me) met and worked out the contents of the BOK. This is a comprehensive yet general BOK aimed at the average bench tech with 3-5 years of experience. Let me encourage all the Metrology labs and Quality Managers that look into this forum to watch for more updates here.......and once the BOK becomes publicly available to begin preparing technicians for the certification exam.
pauljackson 3/9/2005
There have been a number of analytical attempts to include the bonus tolerance in the capability equation. Most of them transform the total available tolerance (amount@MMC)+(bonus) relative to the consumed tolerance (position deviation) into a surrogate variable that can be compared to a constant limit so that the capability equations can be used.
The ones that I have studied are "Percent of Tolerance" recorded by Marty Ambrose and first developed at GM fuel/handling division, "Adjusted True Position" authored by Glen Gruner, "Residual Tolerance" by me. All of those methods change the underlying variation of the data by manipulating the variables and then using the variation of the surrogate variable. Todd Minnick offered three methods to address this problem one of which is similar to the method that I think is best. Todd created a surrogate variable in his Method 2 but he kept and applied the variation from the individual sources separate in the equation. The equation that I think works best is Cpu = (MEAN VARIABLE TOLERANCE [amount@MMC+bonus] - MEAN POSITION DEVIATION)/ 3*SQRT(sigma^2[size]+ sigma^2[position]. This formula can be used to generate the process capability prediction when both of the distributions size and position deviation are normal and are "in-control" subject only to random cause variation. Unfortunately the distributions for the position deviation are often non-normal so the prediction may over or under estimate the true capability of the process to the "VARIABLE TOLERANCE."
pauljackson 3/9/2005
The only proper way to do this currently is by estimating the PPM defective from attribute gaging either with A hard gage built to the virtual conditions pass/fail or with variables data that figures in the bonus and compares the deviation to it pass/fail. Continuous data capability typically ignores the bonus tolerance because the formulas for the process capability indices Cp-Cpk and Pp-Ppk work only with constant values for the USL and LSL.