![]() If the process is perfectly stable then theoretically these two variations are identical and Cpk = Ppk. Ppk is the comparison between the specs and the total variation (within subgroups + between subgroups). Cpk is the comparison between the specs and the within subgroup variation. If it is not, the calculation as stated breaks down then but don’t worry for 1.67, your Cpk is less than zero.Ĥ- An added point in favor of this method is that, if the process is stable, you really don’t need to calculate Ppk. Moreover, the fact that the process remains stable when assessed with the limits defined several runs ago gives much more confidence on the long-term stability of the process.ģ- In the calculation above, bu loosely using the word “distance”, I am assuming that the process average is within the specification limits. If the process is stable across runs (as it should), then there is no need to calculate new limits for each run. The best practice is to establish the control limits in a long enough run and then leave them there for the following runs. If your run is stable but different than a previous run that was also stable (different process average and/or variation), then the process is NOT stable. However, a stable process is not a stable run. The run may be short and there may be not enough points to asses stability in one particular run. Because this last distance is 3 sigma and 1.67*3sigma (within subgroup) = 5 sigma (within subgroup), this is the same than saying that the distance between the process average and the closest specification limit is at least 5 sigma (within subgroup variation).ġ- Strictly speaking, I would replace ” staying within 3 sigma control limits throughout the run” by “being the process in-control”, which implies not only staying within control limits but also the absence of other out-of-control signals.Ģ- I would also not say “throughout the run”. This means that, for the control limit that is closer to the specification limit (be it the upper or lower control limits), the distance between the process average and that specification limit is at least 1.67 times the distance between the process average and that control limit. IF the +/-3-sigma control limits are such that the Cpk is 1.67 or better*, then yes, it is enough. You ask “Is staying within 3 sigma control limits throughout the run enough to accomplish this?” ![]()
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. Archives
December 2022
Categories |