cancel
Showing results for 
Search instead for 
Did you mean: 

Ppk chart

vineeth_varghese
Active Participant
0 Kudos

Hi,

Could you please tell me how to calculate pPK values through SPC charts. Now we have cPk and standard values claculated at time of SPC chart creation. How can we get pPK values. Did anyone worked on this, please let me know.

Thanks

Vineeth

Accepted Solutions (0)

Answers (1)

Answers (1)

Former Member
0 Kudos

The difference between Cp and Pp lies in the method for calculating s, and whether or not the existence of rational subgroups is considered. Calculating Ppk presents similarities with the calculation of Cpk. The capability rate for Ppk is calculated using the formula:

,

PPk=min { USL-mue /[3 X std deviation]) ,(Mue-LSL/[3 X std deviation] }

Mue is mean value

Std deviation is direct sigma value from SPC chart

There is no std way by SAP you can derive ,You have to customize the same.

vineeth_varghese
Active Participant
0 Kudos

Sujit,

Can you also please mention wat is the formula for cPk calculation used in standard SAP.

So technicall speaking, how can I build this into SAP, is it through some function module or by some other way? Please suggest.

Not in all cases client want pPK to be seen, only on a need based requirement. How to make it possible.

Thanks

Vineeth

Former Member
0 Kudos

The formula used by sap for cp, cpk..(from sap help)

The following formulas are used in the mean value tracks for the calculation:

cp = (USL - LSL) / 6*s

cpk = min((USL - x), ( x - LSL)) / 3*s

Here, the USL and the LSL are the tolerance range limits; x-bar and s are estimated

values for the expected value and the standard deviation of the original value distribution.

S is calculated according to the process model of the control chart, either as the inner

dispersion or total dispersion of the measured values. X is the overall mean value of the

measured values.

To calculate the process capability index cpk in the qualitative tracks, the following formula is

used:

cpk = u(1 - p) / 3

Where p is the estimated share of nonconforming units and u is the quantile function of

the normal distribution.

This formula typically produces the same value for cpk as with a normally distributed

characteristic with the same fraction of nonconforming units (single-sided).

But, I dont think you have an option to do ppk in SAP. Check for any exit / ehancement point avaialble - to add an additional button for PPK and write your formula to get the values. I have never done this, talk to a good developer to check for the development options to get your ppk.

The other option is if down load the values into a XL and write a macro and put the formula to get PPK. And also you can look at the qm-sti interface, if any legacy spc tool is used. call that spc too thru the interface and get the ppk.

I know I am not giving exact sol. to your problem, hope it helps.

Thanks,

Ram

vineeth_varghese
Active Participant
0 Kudos

Hi,

From above, what I understood is formula is same for both, but "standard deviation"/ "s" calculation is different.

PPk=min { USL-mue /3 X std deviation) ,(Mue-LSL/3 X std deviation }

cp = (USL - LSL) / 6*s

cpk = min((USL - x), ( x - LSL)) / 3*s

Can I make a formula for standard deviation which is different for pPK. Once standard deviation is identified, I will use this new standard deviation value in cpk formula and get pPK. Can anyone suggest what is the best way to go further, Function module or User exit. Please guide in technical terms to maximum extend. I will try for development.

Thanks

Vineeth

Former Member
0 Kudos

Please go through some non SAP links which really has beautiful explanation of your problem

http://elsmar.com/Forums/showthread.php?t=1975

http://www.isixsigma.com/index.php?option=com_k2&view=item&id=1275:&Itemid=155

http://www.qualityadvisor.com/sqc/capability_compare.php

I have observed on thing.The way of calculating the slandered deviation is different in statistic & SAP .

Not much but difference is there if you are using it for Auto Companies as they follows the TS16949 std ( old QS9000)

Best way to deal with such a problem is ask your Client how they want the calculation & then design the same.This will reduce the problem to a greater extend.

I hope this will add values