Databases Reference
In-Depth Information
TABLE 10.2
Defects per Million Opportunities (DPMO)
Defects per Million
Opportunities (DPMO)
Sigma Level
1
690,000
2
308,537
3
66,807
4
6,210
5
233
6
3.4
common-cause variation has been sufficiently reduced to make outputs consistent and predictable
(i.e.,
3.4 defects per million products).
What constitutes tampering as opposed to reducing variation? By definition, if an action taken
does not reduce the variation in a process, it is tampering.
The sample application in this example needs to make two determinations. First, how many
gages meet the specifications? (How many are good?) Second, is the
≤
process
that produces the
gages satisfactory or does it need to be refined?
10.3
DEALING WITH THE RAW DATA
Now the raw data from the manufacturing process must be imported, manipulated, and analyzed
to provide the information desired.
Figure 10.6 shows the format for the raw data file that contains the measurements of the resistors
fabricated on the substrates. The first line contains the substrate number. The next line contains the
column numbers (1-12). The next eight lines contain the row number (labeled A-H for rows 1-8)
in column 1, followed by the resistance measurements for every resistor produced on the substrate.
Looking at Figure 10.6, an 8
12 matrix is formed for each substrate, where each row/column
element is indicative of a resistance. The file repeats in this fashion for as many substrates that
have been produced in the given lot.
×
FIGURE 10.6
Raw data format for strain gage manufacturing process.
