P
Hi all,
Sorry I have browse a little bit in the forum, but not able to find something relevant...
If You want to calculate the capability (Cpk or PpK) of a batch, and only this, we do know that a representative sample of minimum 32 parts is required to get a correct calculation.
But when you have large batch of 500 000+ parts, what sample size would you take to have a reliable result? is there any clue for large batch size?
Just to explain you what we are doing:
We are manufacturing batch which are between 500 000 and 1 000 000 parts (approx this is done in 3 days).
At each determined frequency we are taking some samples => at the end of the batch we have collected let's say 300 samples.
=> The capability is calculated (continous dimensional datas). According our customer the result must be over 1.00
Is there any rational or clue to write to say that 100 samples or 50 or whatever instead of 300 are enough to make the capability calculation?
Thanks
Fred
Sorry I have browse a little bit in the forum, but not able to find something relevant...
If You want to calculate the capability (Cpk or PpK) of a batch, and only this, we do know that a representative sample of minimum 32 parts is required to get a correct calculation.
But when you have large batch of 500 000+ parts, what sample size would you take to have a reliable result? is there any clue for large batch size?
Just to explain you what we are doing:
We are manufacturing batch which are between 500 000 and 1 000 000 parts (approx this is done in 3 days).
At each determined frequency we are taking some samples => at the end of the batch we have collected let's say 300 samples.
=> The capability is calculated (continous dimensional datas). According our customer the result must be over 1.00
Is there any rational or clue to write to say that 100 samples or 50 or whatever instead of 300 are enough to make the capability calculation?
Thanks
Fred
Last edited by a moderator: