A process is capable if it can meet the customer specification limit.
It doesn’t need to meet it in every output to be capable.
So a process is acceptable if he meets the customer specification limit in most cases.
Under this concept, we have the definition of center and spread.
- Spread: a process have low spread when all the output are one near to the other;
- Center: a process have high centering when all the output are near to the mean (the target defined by the customer);
The processing capability measures the spread of the data but doesn’t say if all the outputs are near the center. This means only that you can meet the target sometimes.
The process is acceptable if you’re near the center.
Example: you are shooting with a gun. If all the shots are near each other, but not all near the target, the process is capable but not acceptable. If the process is all near the target the process is acceptable.
For measuring spread and centering, you can use this measure:
|New Process not under statistical control||Process alredy under statisticall control|
You can calculate this index with this formula:
- CP = (Upper spec limit – Lower spec limit) / 6 * SD
- CPk = minimum value between (mean – Lower spec limit)/ 3 * Standard Deviation and (upper spec limit – mean)/ 3 * SD
The PP and PPk are similar, but instead of using the real standard deviation, you use a sample’s standard deviation.
Looking at the formula, you can easily see that CP doesn’t talk about mean but just the standard deviation. So you only say that yes, the Standard Deviation can fit in the limits, but maybe you have the wrong outcome because you’re not centered on the mean. For this reason, the CPk introduces the mean in the formula.