Ppk vs Cpk - A Good, clear explanation and How Mini-Tab Handles Certain Statistics

[vk.sangisetti]

Re: Ppk vs Cpk - A Good, clear explanation and How Mini-Tab Handles Certain Statistic

Thanks Bev,
In our case we are calculating Cp for individual memebers, I mean to say there is no group. I think in our case sub group size is always 1..is it right?

Sangisetti

 

[Bev D]

Re: Ppk vs Cpk - A Good, clear explanation and How Mini-Tab Handles Certain Statistic

Thanks Bev,
In our case we are calculating Cp for individual memebers, I mean to say there is no group. I think in our case sub group size is always 1..is it right?

Sangisetti

If you are measuring individual parts you shoudl be using an "individuals, Moving Range" chart (or similar depending on the distribution).
If using the I, MR chart your subgroup is 2: range of the first value to teh second, second to the third, etc. d2 for n=2 is 1.128

If you are trying to use only the individual values and calculating the control limits directly there is no need for d2 as there is no subgroup range. (Note: I am not recommending this...)

 

[aaronan2006]

Re: Ppk vs Cpk - A Good, clear explanation and How Mini-Tab Handles Certain Statistic

I have a little study on SPC. CPk is caculated when the process is stable which means there is no special variance in the process( only commen variance exists). PPk does not take care of the condition whether the process is stable or not. PPk is an overall caculation, but CPk is a within the subgroup analysis.
Any question, please call me.
86-13761644301
mail:aaronan2006***********

 

[hobbyxin]

Re: Ppk vs Cpk - A Good, clear explanation and How Mini-Tab Handles Certain Statistic

for Ppk and Cpk, I think these very confusing , but if we think about how we caculate these two indices, it is easy to distinguish these two:

1. if we group data , then we may caculate st. devi. as Rbar/d2, that means we consider how the varience appeard within group, so Cpk just tell us how is our process stability during limited time, this means short period;
2. If we caculate st. Dev. per individule data against true value, we get a lont term varience of our process,so Ppk tell us how is our process stability during all period,this means long term.
3. So, I interpret Ppk as: it indicate potential capability of all process( with one statistical status).Cpk indicate capability of production period.

hope for your comments!

 

[pratishesh]

Re: Ppk vs Cpk - A Good, clear explanation and How Mini-Tab Handles Certain Statistic

Hello sir,

I came in to the forum today mainly to understand the definition and the usage of the capability indices (Cpk and PpK).

Now in my shop floor we have several machining processes including some important finishing process like Grinding, in which I have lot of doubts in calculating Cpk/Ppk. Basically we are piston rings making cmpany where we do finally a Grinding operation to control the axial thickness.Now we are checking the samples for every 25 pcs and based on the output we give the wheel wear compensation and this steps is continued till the end of the shift.

Now in this scenario how to calculate Cpk/Ppk considering the fact that the output is varying due to factors like the wheel wear , coolant temperature fluctuations and so on.

Can you help me pelase.

Shesh

 

[Kales Veggie]

Re: Ppk vs Cpk - A Good, clear explanation and How Mini-Tab Handles Certain Statistic

Investigate if EWMA (exponentially weighted moving average) will work in your case. Some info on it

A company that I worked for in the past used this technique for a similar process.

 

[Bev D]

Re: Ppk vs Cpk - A Good, clear explanation and How Mini-Tab Handles Certain Statistic

Now in this scenario how to calculate Cpk/Ppk considering the fact that the output is varying due to factors like the wheel wear , coolant temperature fluctuations and so on.

Can you help me pelase.

Shesh

the causes of variation and your reaction to them don't matter. The calculation of Cpk/Ppk are simply a quantificationof the actual variation in the output that you actually experience. How and why you have that variation does not obviate the validity of the calculation.

If youare expereinceing systemic drift due to tool wear, etc. you will have more of a uniform distribution than a bell shaped curve. THIS will effect which formula you should use and what your result means. If you search this forum for tool wear, Cpk and uniform distribution you will find ample information on hwo to handle this situation...

 

[Hamed]

Cpk for one side specification

If the specification is in of one side data like length 35.5 +0/-0.3
Are there any change in Cpk calculation.

Regards,

Hamed

 

[Hamed]

Re: Ppk vs Cpk - A Good, clear explanation and How Mini-Tab Handles Certain Statistic

Hello Friends

What about CpK for one side specification like length 125.5 +0/ -0.3 is same way in calculation for two sides data 125.5+/- 0.3

Many Thanks,

Hamed

 

[Bev D]

Re: Ppk vs Cpk - A Good, clear explanation and How Mini-Tab Handles Certain Statistic

Hello Friends

What about CpK for one side specification like length 125.5 +0/ -0.3 is same way in calculation for two sides data 125.5+/- 0.3

Many Thanks,

Hamed

Hamed - one of the reasons your post from October wasn't answered is that it has been answered numerous times before. try a search of the site and see if you dont' find the answer you are looking for.