John Needleman

John Needleman

3/15/2017

Manf 3000

Capability analysis

The data given in figures 1 and 2 are measurements taken off or a micrometer. The desired measurement is .258±0.005”. The parts measures came from two separate machines as represented in the figures to the right. To compare the two machines the upper and lower tolerances and frequency of measurement are found. With this information and an expected normal curve overlaid on the frequency distribution we can compare the machines. The frequency of distribution can be seen in the graphs below.

As can be seen in the graphs, both machines are producing parts out of specification in an unacceptable fashion. Machine N-7 is preferable to machine N-5 as the there is a significant portion of parts within the normal curve, where machine N-5 looks to be operating generally high and out of tolerance. This will cause more parts to be out of specification.

The process capability can be found through grouped data analysis. The process capability and other key values are shown in the chart to the left, corresponding to each machine.With a relatively high standard deviation compared to the tolerances given in specification (.003 vs .005 system is completely unacceptable and must be revised or totally scrapped.

Through statistical process control and the use of x-bar and r-bar charts we can further evaluate each machines performance. These can be seen in the graphs below.

The N-5 graphs show that the machine being examined is not very consistent. The parts vary widely and reflect a higher range than machine N-7. The range within each subgroup of machine N-5 shows that there is almost no repeatability in the process, while machine N-7 shows a more of a trend. Each N-7 grouping was more tightly packed. While this is promising the machine still reflects some error as the parts are consistently shifting larger and larger. This shift along with the consistency of each grouping leads to more parts being out of spec in one grouping as opposed to random parts throughout as in machine N-5. Neither process is ideal. N-5 is hopeless and needs to be given an in-depth analysis to find the cause of error as the error is simply random. Machine N-7 however has some order to it. The consistency of parts getting larger can suggest that tool wear or even a shifting stop could be factors in the trend to larger parts. If this trend can be corrected through a simple fix the machine can show some good promise and likely remain in operation. These conclusions are reflected by the process capability using the range of the machines. The process capability of each machine based on the ranges is .98(N-5) and 1.88(N-7) reflecting the idea that N-7 can be corrected.