P
I am starting to look at acceptance plans and am running into some questions. Help would be appreciated:
1. Why are there c=0 sampling plans as well as sampling plans that allow for the detection of errors, c>0? I am under the impression that either will get you to the same AQL in the end.
2. It seems after reading different posts that Zero Acceptance Number Sampling Plans by Squeglia is quite recommended. However, I am under the impression that this is for attribute data. My company doesn't use much attribute, but mostly variable data. What method should we use?
3. I have also been reading about MIL STD 105E (something that our company does use, barely- I am trying to change that) and that seems to be obsolete, again, what should my company use instead?
4. (this might be the long one) When using either a c=0 sampling plan or a c>0 plan I have come up with the question, what lot size should I use? We make small injection molded plastic parts with a fast cycle time. I can make 550+ parts per hour. When I look at the sample size for a lot, should I look at per hour, per 8 hour shift, per 24 hour day (we have lights out manufacturing), per month, or per year? We typically inspect parts every 8 hours as an in process inspection. As the lot size increases the amount we need to sample for every 8 hours goes down. I.E. for every 8 hours at 1%AQL I need 50 parts per 8 hours, but if I look at 24 hours of run I need 60 parts across 24 hours, or 20 parts for each inspection. This is HUGE for our small company. If I carry it out further and we estimate we make 1e6 parts per year, then we only need to inspect ~6 parts per inspection period. Where do I set my sample size?
Thank you for your input, I have appreciated the advice found on other threads on this website and am impressed with the knowledge found here on this site.
1. Why are there c=0 sampling plans as well as sampling plans that allow for the detection of errors, c>0? I am under the impression that either will get you to the same AQL in the end.
2. It seems after reading different posts that Zero Acceptance Number Sampling Plans by Squeglia is quite recommended. However, I am under the impression that this is for attribute data. My company doesn't use much attribute, but mostly variable data. What method should we use?
3. I have also been reading about MIL STD 105E (something that our company does use, barely- I am trying to change that) and that seems to be obsolete, again, what should my company use instead?
4. (this might be the long one) When using either a c=0 sampling plan or a c>0 plan I have come up with the question, what lot size should I use? We make small injection molded plastic parts with a fast cycle time. I can make 550+ parts per hour. When I look at the sample size for a lot, should I look at per hour, per 8 hour shift, per 24 hour day (we have lights out manufacturing), per month, or per year? We typically inspect parts every 8 hours as an in process inspection. As the lot size increases the amount we need to sample for every 8 hours goes down. I.E. for every 8 hours at 1%AQL I need 50 parts per 8 hours, but if I look at 24 hours of run I need 60 parts across 24 hours, or 20 parts for each inspection. This is HUGE for our small company. If I carry it out further and we estimate we make 1e6 parts per year, then we only need to inspect ~6 parts per inspection period. Where do I set my sample size?
Thank you for your input, I have appreciated the advice found on other threads on this website and am impressed with the knowledge found here on this site.