CPK CPU CP is equal to zero, what can happen from?Re: CPK CPU CP is equal to zero, what can happen from?funny data, bad limits, bug in JMP, it is the correct results???????? You need to provide more specific information. Sample data would really help.Re: CPK CPU CP is equal to zero, what can happen from?Please see my data as below
| No. | Length (mm) |
| 1 | 302 |
| 2 | 300 |
| 3 | 300 |
| 4 | 305 |
| 5 | 303 |
| 6 | 305 |
| 7 | 302 |
| 8 | 303 |
| 9 | 305 |
| 10 | 304 |
| 11 | 305 |
| 12 | 305 |
| 13 | 305 |
| 14 | 304 |
| 15 | 305 |
| 16 | 304 |
| 17 | 302 |
| 18 | 303 |
| 19 | 304 |
| 20 | 303 |
| 21 | 303 |
| 22 | 304 |
| 23 | 302 |
| 24 | 303 |
| *Specification | 295 - 325 |
Re: CPK CPU CP is equal to zero, what can happen from?I apparently are not understanding your question. Here are my results of Cp and Cpk for your data
Re: CPK CPU CP is equal to zero, what can happen from?Hi txnelson,
Thank you for the reccommendation but normally I use Cpk from "Quantile sigma" , it is shown like below.
Can I ask more question, between Cpk from "Long term sigma" and "Quantile sigma" , in which case should we use it?
Thank you
Re: CPK CPU CP is equal to zero, what can happen from?Oh and by the way you are using a SHASH distribution!
Re: CPK CPU CP is equal to zero, what can happen from?I'm not sure why the SHASH distribution gives zero as the result but I've not had much success using SHASH for process capability. In your case your data only has 5 distinct levels, and the distribution requires 5 parameter estimates so you really are running out of degrees of freedom and the SHASH is probably a total over-fit of your data.
