Mil-Std-1916 Inspection Sampling question

D

diptankd

Hello everyone,
I have a question I hope you could help me with. If I pull a sample using mil-std-1916 say 192/100,000 on over time and DCMA is not avaliable to do the inspection concurently and I inspect the parts and everything is acceptable in my 192 parts so I accept them and green tag them and I have all my documentaion there so the parts are ready to be released. In the morning DCMA comes to work and pulls there own sample to there requirements and finds a defect, I did not check the parts 100% the 192 that I checked were good. I realize that the parts need to be screened for any more defects but, can I be written up for passing the parts even if my sample was good, could they hold me acoutable for passing defective parts? #2. If they are not doing the over time my manager is requiring me to do can DCMA tell me that I can or cannot check certain parts. unless they are present. Let me know what you think.
diptankd
 
O

osmqc

I have a different question: In Mil-Std-1916, what is the meaning of "single-sided specification" and "double-sided specification"?
 

David-D

Involved In Discussions
I have a different question: In Mil-Std-1916, what is the meaning of "single-sided specification" and "double-sided specification"?

MIL-STD-1916 utilizes single and double sided specifications in reference to variable sampling. Single sided means that there is only one end of the specification (minimum or maximum) she double sided specifications have both upper and lower limits. For example, "3.00" min work be single sided specification while 3+/- 0.005" is double sided.

This comes into account because the variable sampling of 1916 uses one parameter "k" for both single and double sided specs. It is analogous to a Cpk, assessing how close to the spec limit(s) the process/product is. For two sided specifications, 1916 additionally assessed the "F" parameter which is similar to the Cp assessing the potential capability of the process. Both of these are in addition to the requirement that all the individual samples conform to requirements. I'd suggest you look at paragraph 5.2.2.2.3 and table III of the spec
as well as the examples of 30.2 and 30.3 in the appendix. You may also find the guidance handbook MIL--HDBK-1916 useful as well.

If you don't have it, MIL-STD-1916 is available at:

http://quicksearch.dla.mil/qsDocDetails.aspx?ident_number=120287

David
 

Steve Prevette

Deming Disciple
Leader
Super Moderator
Hello everyone,
I have a question I hope you could help me with. If I pull a sample using mil-std-1916 say 192/100,000 on over time and DCMA is not avaliable to do the inspection concurently and I inspect the parts and everything is acceptable in my 192 parts so I accept them and green tag them and I have all my documentaion there so the parts are ready to be released. In the morning DCMA comes to work and pulls there own sample to there requirements and finds a defect, I did not check the parts 100% the 192 that I checked were good. I realize that the parts need to be screened for any more defects but, can I be written up for passing the parts even if my sample was good, could they hold me acoutable for passing defective parts?

In ANY random sample there is always the possibility that the parts you HAPPENED to choose were GOOD, but all the rest of the production run might be BAD. Highly unlikely (that is the whole point of random sampling) but could happen on occasion. You can always have a "false negative" now and then.

It should also be pointed out that you may not have been unlucky, and got a "representative" sample of the population and justifiably "Passed" it. I can just about guarantee that if you made 100,000 items, and the AQL is say 1%, that there is at least ONE bad item in the 99,808 items, and most likely about 998 of them!

What level of AQL and sampling did the contract specify? That would be the real question of accountability for passing defective parts.
 
Top Bottom