This
program measures process capability in terms of fraction non-conforming using
a unique modeling approach that estimates process capability accurately even
when the fraction non-conforming is small (i,e, p < 1000 ppm). The program
is powerful enough to handle virtually any process distribution. It
automatically checks the assumptions of process stability and normality, providing
a graphic display of the process control chart and normal probability plot.
It displays a histogram of the data, with an over-laid best-fitting
Johnson curve and cumulative probability graph. It computes confidence limits
for the fraction non-conforming, thus indicating the risk with limited data.
It measures how sensitive the process is to small changes in the mean or variation,
thus indicating the robustness of the process. The
program has a process optimization procedure to find the minimum fraction
non-conforming, or minimum net sensitivity, or maximum profitability.
This
program also provides all the standard capability indices for normal and non-normal
data: Cp, Cpk, Cpu, Cpl, Cpm, and Cpmk using both long-term and
short-term estimates of variation along with standard errors of the indices.
Lower 95% confidence limits for Cp, Cpk, and Cpm are provided for dealing
with small sample sizes.
