Cpk calculation for short runs - Manufacturing FFC (flexible flat cable)



Anybody can advise

In our production we are manufacturing FFC (flexible flat cable) The product life cycle for a model is very short due a lot that contain 2000-4000pcs can be finished within 1 hour. everyday a prodcution line (total of 4) are schedule to run minimun 5 to 6 models, each model may only run a day only and may need to wait a week or a month later to run again. Hence the production is almost running dilferent model from day to day.

due to customer request only Cpk data the management are also only concern about it.

our current method are randomly select a model/line in a week and get 30 sample/lot to calculate the Cpk data. and we will use it as a reference for that week Cpk reference for all model.

this method not working very well because the Cpk data keep fluctuating from week to week.

pls give some advise

Al Dyer

Cpk values will fluctuate no matter the sample size. Try increasing your sample size and see of the fluctuation between the samples is consistant.

I think the main concern would be the trends in the Cpk values. Possibly set a target Cpk and react when the trend indicates you are running close to or below the target.

Also consider the variables when reviewing Cpk:



Dave W 2005


Al is right,

There are 'confidence limits' for cpk. The larger the number of samples, the more confident you can be that the cpk value is actually correct (as you have captured more variation).
30 is actually quite a small number. For example, to guarantee that your cpk is actually 1.33 or above, you need to have calculated a cpk of 1.62 for your batch of 30. With 100 samples, this number drops to 1.47.
You should not be surprised if your calculated cpk values vary with 30 samples.

The other question is: are you confident that the same process variation affects all of your product groups in the same way?


Ravi Khare

Thomas Pyzdek has described methods of Short Run SPC in his Six Sigma Handbook...just may prove useful.

He advises to focus on the process than the small/ short batch that gets over before you can start looking for samples. The process focus could consolidate together batches of different products passing through the same process.
Top Bottom