Reject Rates - Lots and split lots

[RudyB]

Good Morning Everyone,

I'm a manufacturing engineer for a small company that doesn't have a formal quality system in-place (owner's decision). We do have an individual that inspects items upon receipt that are fabricated to our drawings.

For the most part, the lots are accepted/rejected base on the results of sample inspections (he uses the tables in M-S-105).

Typically, our reciepts (lots) are either accepted or rejected based on the sample results. We then report supplier preformance based on lots accepted.

Example - 10 lots inspected, 9 accepted equals a 90% acceptance rate.

Occasionally when we really need an item, he inspects the entire lot, sorting out the good & bad, and making it a split lot.

MY QUESTION(s) - What is the appropriate method to report on lots that have been split (or) is there another approach we should take on reporting vendor performance based on the inspection method we use?

TIA,

Rudy

[Al Rosen]

In Reply to Parent Post by RudyB

Good Morning Everyone,

I'm a manufacturing engineer for a small company that doesn't have a formal quality system in-place (owner's decision). We do have an individual that inspects items upon receipt that are fabricated to our drawings.

For the most part, the lots are accepted/rejected base on the results of sample inspections (he uses the tables in M-S-105).

Typically, our reciepts (lots) are either accepted or rejected based on the sample results. We then report supplier preformance based on lots accepted.

Example - 10 lots inspected, 9 accepted equals a 90% acceptance rate.

Occasionally when we really need an item, he inspects the entire lot, sorting out the good & bad, and making it a split lot.

MY QUESTION(s) - What is the appropriate method to report on lots that have been split (or) is there another approach we should take on reporting vendor performance based on the inspection method we use?

TIA,

Rudy

Hi Rudy, welcome to the cove. I believe that the lot fails even if you sort.

If you have lot sizes that vary, you can use the total no. of items to determine % accepted(rejected).

(Total items in all passing lots/Total no of items in all lots)*100 = % Accepted

[CarolX]

Hi Rudy,

Look at it this way...why did you sort? Because it was rejected on initial inspection.

I track first piece approval through my manufacturing process. If my inspector has an out of tolerance condition, it is still considered a reject, even if engineering approved it "as is".

Hope this helps a bit.

[Jim Wynne]

In Reply to Parent Post by Al Rosen

Hi Rudy, welcome to the cove. I believe that the lot fails even if you sort.

Yes--even if a lot is sorted, if it's lots-accepted-or-rejected that is being counted, then the lot that gets sorted has to be counted among the rejected lots.

In Reply to Parent Post by Al Rosen

If you have lot sizes that vary, you can use the total no. of items to determine % accepted(rejected).

(Total items in all passing lots/Total no of items in all lots)*100 = % Accepted

This could be misleading unless 100% inspection is being done, because it's possible that there are defectives in the lots that "pass." It works as a measure of how many lots pass, of course, but I think it would be better to compare the number of pieces received vs. number nonconforming, regardless of whether the defectives are found in receiving inspection, production, or customer returns.