V
Hi,
I work in an electrical component manufacturing industry. We are working on a coil winding application where a coil is first wound and then glued to a LCP bobbin with an epoxy. We are supposed to run a pull test on the parts for coil to header adhesion to make sure that the parts build meet the customer spec. We normally grab a couple of coils being built in the assembly line and run pull-test on them. However, I was wondering if there is any statistically valid method that could be used to come up with a number of samples of parts that should be tested for pull-test?
Another scenario where we need to come up with a statistically valid sampling plan is during the final electrical/mechanical inspection of the parts built at the plant. We have traditionally been using a 32 piece sample data no matter what the production lot size is for a PO placed by the customer (PO for more than 32 parts, offcourse). However, we don't have any valid rationale behind the 32 pcs sample data. We have just been doing this for years and years until now.
There is no specific request from customer yet. But we are worried that the customer will someday ask us about the statistical rationale behind the 2 parts pull-test we have been carrying out (or the 32 parts final inspection plan) till now and we won't have any thing to present to the customer.
I found this paper in the internet (attached along with this post). But I am not sure if this would be applicable for our particular situation or not?? The sample sizes seem too big!
Any help will be greatly appreciated.
Thanks in advance.
I work in an electrical component manufacturing industry. We are working on a coil winding application where a coil is first wound and then glued to a LCP bobbin with an epoxy. We are supposed to run a pull test on the parts for coil to header adhesion to make sure that the parts build meet the customer spec. We normally grab a couple of coils being built in the assembly line and run pull-test on them. However, I was wondering if there is any statistically valid method that could be used to come up with a number of samples of parts that should be tested for pull-test?
Another scenario where we need to come up with a statistically valid sampling plan is during the final electrical/mechanical inspection of the parts built at the plant. We have traditionally been using a 32 piece sample data no matter what the production lot size is for a PO placed by the customer (PO for more than 32 parts, offcourse). However, we don't have any valid rationale behind the 32 pcs sample data. We have just been doing this for years and years until now.
There is no specific request from customer yet. But we are worried that the customer will someday ask us about the statistical rationale behind the 2 parts pull-test we have been carrying out (or the 32 parts final inspection plan) till now and we won't have any thing to present to the customer.
I found this paper in the internet (attached along with this post). But I am not sure if this would be applicable for our particular situation or not?? The sample sizes seem too big!
Any help will be greatly appreciated.
Thanks in advance.
Attachments
-
342.1 KB Views: 660