Calculating the RFT using the Cpk value

[lee01]

Help needed:

I understand that if your Cpk value is 1.33 your looking at a RFT value of 99.9937%. and if you have a Cpk of 0.50 your looking at a RFT value of 86.64%.

This is on most documentation out there but, is there a formula that I can use to create the actaul RFT for all my measured processes? I do not want to guess what a RFT value is simply by comparing the documentation and guessing what RFT value I have if I have a Cpk of 0.43 etc. . .


Best Regards

Lee01

 

[Darius]

My recomendation is not to use any relation RTF vs Cpk, as Don Wheeler said, the Cpk take in account "normality" (Gausian distribution, I hate the term "Normal", because most of the people tink, is what it should be, if no special causes happen), and as more the distributions departure of "normality" happen, the estimated values get garbage.

There are may tables of such relationship on the net (and in some books, like "Inovative Control Charting" by Stephen A. Wise and Douglas C. Fair).

If you still want to do it, use a interpolation of the table shown of the file, I once obtain an aproximate formula

dpm (1 tail ) = Exp( - 5.356 * Cpk^1.859 + 12.581))

in the two tails case, multiply the result by 2

and RFT= 100 - dpm / 10000

beware: Deming (1986):“no process,except in artificial demonstrations by use of random numbers,is steady and
unwavering".

 

[howste]

Doesn't that number come straight from a "standard normal" table? It should be easy to come up with the RFT number using the table and an understanding of Cpk. Like Darius implies though, you need to know if the distribution is normal and if it is centered in the tolerance or not. To calculate the real value, you would need to know not just Cpk, but also the mean, standard deviation, and the specification limits.

 

[Al Dyer]

Aren't the two mutually exclusive since there is no direct relationship?

Just enquiring?

Al...