The CP value is the BEST the Cpk could ever be, IF the point averaged exactly at zero deviaiton, if you want some technical stuff as well.
CP is a measure of HOW MANY TIMES (3 x Sigma) will fit within 1/2 of the tolerance band (yes, 1/2 the total band, it assumes that a zero dev or perfect part is in the exact middle of the tolerance band)
CPK, it is how many times (3 x Sigma) will fit between the AVERAGE and the CLOSEST tolerance limit.
Cp ignores centeredness and is strictly a measure of deviation. Cpk includes not only process spread but it's centeredness on the mean.
Dude, you must not have read one of my other posts, I used the PROFILE dimension analogy, Cp = FORM ONLY and CPK = Form & Location. Which is exactly what you just said.
Dude, you must not have read one of my other posts, I used the PROFILE dimension analogy, Cp = FORM ONLY and CPK = Form & Location. Which is exactly what you just said.
Cp is a 'measure' of repeatability of the point to itself
Not true, that would be standard deviation. Cp includes the tolerance limits in that without the total tolerance spread you can not calculate Cp.
Cpk is a 'measure' of repeatability compared to the tolerance limits.
OR,
Cp = FORM ONLY Cpk = FORM and LOCATION (to use a profile analogy)
Not only tolerance limits but also centeredness on the mean.
Cp - includes tolerance spread only
Cpk - includes centeredness as well as spread.
