earlier that common practice is to use a 6-sigma range on control chart+ (3-sigma on either side of the mean). It would follow, then, that we will need to calculate a standard deviation in order to establish the limits. This is technically correct but,again, thanks to others we have tables to turn to which make our calculations easy. (On the back of the chart in the appendix.) Since we are working with averages we would need to find the standard deviation of the average values. Statisticians call this the "standard error of the mean." Statistical theory has determined that there are relationships which exist between and among the various averages, ranges, and standard deviations. This has permitted the calcualation of constants which we can use to calculate the various limits. The following formulas show how control limits are calculated Upper Control Limit for Averages, UCL(Xbar) = Xbar + A(2) R
Lower Control Limit for Averages, LCL(Xbar) = Xbar - A(2) R
A(2), read A sub-2, is a factor, derived theoretically, for calculating the limits which permit proper interpretation. The value is dependent on the sample size and is obtained from the table on the reverse side of the Xbar & R
charts located in the back of this document. It will be noted that the two formulas are identical except for the + and - signs. Thus, we can multiply A(2) times R
bar and then add it to and subtract it from Xdouble bar to calculate the upper and lower control limits for the averages. The UCL(Xbar) and LCL(Xbar) are then shown as dotted lines extending through the length of the plottings. Some companies use color codes to designate the center lines and the control limits. Such color schemes have merit in helping persons to understand the charts. SUMMAR
Y
Following is a outline form of the steps we have discussed for the construction of an Xbar & R
Chart. 1. Preliminary a. Determine the characteristic to be studied. b. Determine the frequency of checks c. Plan a procedure for obtaining random samples. d. Select an appropriate sample size. e. R
eview the gauging or testing equipment to insure accuracy and to eliminate auxiliary effects. f. Fill out the headings on the data sheets and the control chart to completely describe the part, operation, and inspection procedures. 2. As samples are obtained:
a. R
ecord the individual measurements on the data sheet. b. Calculate the Xbar and R
for each sample. c. Plot Xbar and R
on the control chart. 3. After 20 or 30 plottings:
a. Calculate R
bar = sumR
/k. b. Plot R
bar on the control chart. c. Calculate the UCL(R
) and LCL(R
) for ranges. UCL(R
) = D(4) R
bar LCL(R
) = D(3) R
bar d. Plot the UCL(R
) and LCL(R
) on the control chart. e. R
emove data outside control limits which can be identified as assignable causes. f. R
ecalculate R
bar. g. Calculate Xdouble bar = sumXbar/k. h. Plot Xdouble bar on the control chart. i. Calculate the control limits for averages. UCLXbar = Xdouble bar + A(2) R
bar LCLXbar = Xdouble bar - A(2) R
bar j. Plot UCL(Xbar) and LCL(Xbar) on the control chart.