This type of chart plots the number of defects per sampled unit and indicate the lack of control in the process. Let's go to the same example of refurbishing computer monitors company. The monitor is assembled, containing main parts like screen, case, wires, desk support and electronic. Each of these components is checked by quality control personnel. For each monitor, the number of defects found is recorder. One more assumption: the size of each subgroup is 100, for each of them 2 monitors is checked.
Needed computations are: percent of defects of each subgroup (e.g. u1 = 2/100 = 0.02, u2 = 3/100 = 0.03, etc.); mean of the process, as sum of all defects, divided to sum of all sample sizes (e.g. mean = [2+3+1+0+2+0+1+6+5+3]/[100+100+100+100+100+ 100+100+100+100+100] = 0.023; lower control limit, using the formula: LCL = mean – 3 * sqrt(mean/ni), where ni is sample size (e.g. LCL = 0.023 – 3 * sqrt(0.023/100) = - 0.0225. If LCL is less than 0, set it to 0; upper control limit, using the formula UCL = mean + 3 * sqrt(mean/ni), where ni is sample size (e.g. UCL = 0.023 + 3 * sqrt(0.023/100) = 0.0685. plot the data on the chart and observe if there is any data falling outside control limits. Resulting shows that collected data fall within control limits, meaning that the process is under control (the above graph, near the table). In order to point the lack of control, let's assume that for test number 8, total number of found defects is 9 (2 scratched tubes, 3 cases, 1 wire, 3 electronics, 1 desk support). Redoing the computations and plotting the chart (on the right) we observe that for batch number 8, there is evidence for a lack of control (too many defects compared with the average rate of them). Investigations should be undertaken, in order to find the cause of the problems.
Resulting shows that collected data fall within control limits, meaning that the process is under control (the above graph, near the table). In order to point the lack of control, let's assume that for test number 8, total number of found defects is 9 (2 scratched tubes, 3 cases, 1 wire, 3 electronics, 1 desk support). Redoing the computations and plotting the chart (on the right) we observe that for batch number 8, there is evidence for a lack of control (too many defects compared with the average rate of them). Investigations should be undertaken, in order to find the cause of the problems.
In order to point the lack of control, let's assume that for test
number 8, total number of found defects is 9 (2 scratched tubes, 3 cases, 1 wire,
3 electronics, 1 desk support). Redoing the computations and plotting the chart (on the right)
we observe that for batch number 8, there is evidence for a lack of control (too many defects compared with
the average rate of them). Investigations should be undertaken, in order to find the cause of
the problems.