Ppk results shown as asterisk after the transformation of Non-normal data

[PB Alli]

Hi all,

I used Minitab 17 to calculated the Ppk of my 30 samples non-normal data and it's resulted as an asterisk for Ppk results in 1 set of data (The rest of data are include with both normal and non-normal and it's normally shown the Cpk/Ppk result in both normal and non-normal run). So I would like to ask if there any cause that make this data set failed to shown the Ppk results?

BTW my failed data set is as below:
#12
n = 30
Subgroup = 1
LCL = 5176.5
UCL = 5278.0
5240.6
5241.6
5241.8
5242.4
5242.0
5241.4
5240.4
5242.2
5242.2
5241.4
5242.2
5241.4
5241.6
5242.4
5243.4
5242.2
5239.0
5240.4
5241.4
5240.4
5238.2
5242.8
5237.2
5238.8
5239.2
5239.6
5237.4
5234.4
5237.1
5234.6

*FYI. for my failed run I used Weibull base on its p-value has the most fit distribution.

Thanks.

 

[bobdoering]

When you say "failed"...what do you mean? This is the result I got...looks OK to me.

 

[PB Alli]

When you say "failed"...what do you mean? This is the result I got...looks OK to me.

Hi bobdoering,

when I run minitab the ppk result is turn out as below picture no matter how many time I have run:

 

[Bev D]

I'm not a minitab user - maybe Miner can chime in here.
Is your data sequential parts?
the sample size is pretty small for a capability index - is this just short term data collection?
a simple run chart shows you have a substantial decrease in values a the end - that would lead to your non-normal look of the data. this 'instability' makes any transformation inaccurate. perhaps that's why minitab gives you an *?
Bob - not sure how you get a Ppk of 1.96 when the data is so small compared to the tolerance?

 

[Miner]

This is all that Minitab Help has on the subject. Why are some capability analysis results shown as asterisks in the output?

None of these specific examples appear to apply in this case. However, the asterisk appears when the index cannot be calculated. The non-normal capability that Minitab uses calculates a z-score from the Weibull function then converts that to Pp/Ppk. It may be that Minitab's Weibull function does not provide a z-score that large (i.e., the programmer set limits on the calculation).