Attribute control charts limit calculations

All attribute control charts follow the same three sigma control lim. it away from the centerline methodology of the variable control charts:

Control limits for attribute charts = centerline 土 3 s (3.8)

For constant samples (C or nP charts)

For Poisson distribution:

For changing sample sizes (U or P charts) 

For Poisson distribution:

Centerline = Poisson average number of defects in a sample u

C and U Charts are considered as a special form of control charts in which the possibility of defects is much larger, and the probability of getting a defect at any specific point, place, or time is much smaller.

The relationship of attribute charts to the six sigma concept is through the defects implied in the charts. The centerline represents the defect rate. These defect rates can be translated into an implied Cpk, as shown in the previous chapter.

Several assumptions have to be made in the case of the attribute chart connections to six sigma:

1. There is one or a complex set of specifications that are not readily discernible that govern the manufacturing process for the parts.

2. These specifications are either one- or two-sided, resulting in one- or two-sided defects (defects < LSL and defects > USL).

3. The manufacturing process is assumed to be normally distributed.

4. There is a relationship between the process average and the specification nominal. In some definitions of six sigma, an assumption is made that there is a 1.5 a shift from process average to specification nominal.

The control limits of the attribute charts are not related to the population distribution. Therefore, the method of finding the population standard deviation a is quite different from that used in variable control charts, as shown in the examples below.