Final Inspection - Specific lot control required?

[bigkahunakal]

Final Inspection

I need some help! We are a continuous run high volume production operation- manufacturing parts for the automotive industry. We are not required by our customers to provide specific lot traceability. All of the products we manufacture are customer supplied- our registrar is not happy with our final inspection process. I understand the requirement- review the preceeding inspection records and complete the physical inspection- but does this process require specific lot control? Like to hear some suggestions from similar industries.

 

[Jim Biz]

The answer may be in what you are defining as a "LOT" ? We contend that a "lot" is all of those products produced under a given Work Order release 1 box - 2 boxes-5 boxes
dosent matter - they were all produced using the same inspection freq's - same paperwork etc.

It can get extreamly complicated if you define a "lot" as each individual box-package etc.

This might not help you much - but if anyone has an "official" definition of a "LOT" - i'd be interested in seeing it

Regards
Jim

 

[Jerry Eldred]

That's not my specialty. I can comment based on a couple of year 'aside' I took doing quality support in our QS9000 subcontractors.

I think you have to have the capability to trace a particular product back to when it was made. If this is a QS9000 compliant part (I hate that word 'compliant'), then in the case of a field failure or latent defect, If the automotive customer traced it back to your part (perish the thought). How would you be able to do containment (as in 8D)? If there is not some sort of lot control or equivalent that would enable you to trace back to the date range that may be affected by the defect, how could a problem be contained?

Let's say there is you make a screw out and it is plated with Smith coating material. It is learned that some screws with Smith coating are failing prematurely due to Smith coating not working properly. You trace back into your product history and find that Screws from Date X to Date Y were produced with Smith Coating Lot Number B. You learn that Smith Coating Lot Number B was defective. You do a little further checking and find that Smith Coating Lot Numbers A and C are all good (due to your product samples, but further inspection of your product samples coated with Smith Lot B determines that Lot B isn't performing correctly. In your Root Cause Analysis, it is determined that the root cause of the screws manufactured between dates X and Y is that Smith Coating Lot B was bad. You would need to go back to Smith Coatings and get their root cause analysis as to why they shipped you bad coating materials to put into your 8D. Your corrective action is to get a better grip on that supplier of coating material (perhaps even change coating suppliers if they don't cooperate), improve sampling methodology for your inspections of the coating, increase incoming inspection frequency or tighten up incoming criteria on the coating material, improve final inspection method to be sure the defective coating problem can't escape. Etc..

After the long-winded example, it brings me back to lot traceability. I'm not familiar with your scenario, and I don't claim to be an expert in when lot traceability is or is not required in a QS9000 context. But it would seem that if there are any reliability requirements on your products, then if there is ever a reliability issue (which even in the best companies can occasionally occur), you have no way of tracking down and containing such issues without some at least minimal control of lots or datecodes or date ranges. I work in the semiconductor industry, and have done subcontractor support in those areas, and it gets downright nightmarish sometimes. But it is an expected matter of course that you have to be able to identify and cordone off production segments with some sort of lot/date of manufacture control.

------------------

 

[Jerry Eldred]

>if anyone has an "official" definition of a
> "LOT" - i'd be interested in seeing it

I'm not sure there is an official definition of a lot. The ability to identify product is the issue versus how a specific lot is defined. Depending on production volumes and the value and complexity of the product, I have seen evreything from serial number to work week code, and everything in between. I had one subcontractor that used shift code, another that used date code, another that used (this one's confusing) lot traceability code of the module. In that last one, they made an electronic module assembly. The bare circuit board had a trace code on it. They used that as their lot number. But it had lets say a dozen different components on it. The computer in their automated production line tracked each component and its individual trace code with the lot trace code of the bare circuit board (they referred to it as the 'substrate'). So everything was tracked against the substrate.

I have seen in the case of large volume single components, a week code or day code or shift code. That was internal tracking. Then when a shipment went out, however many items were shipped together was considered an outgoing lot.

All that confusing rambling to say, I don't believe exactly how you accomplish lot/component traceability is quite as important as just the fact that you need some adequate method to identify product. For example in a screw making factory that only makes one kind of screw, part number 001. If you had a stack of boxes of this part number 001 that were awaiting shipping and somehow accidentally was set aside. Sometime a few weeks later, the stack of boxes is found in the shipping area. How do you know how long they have been sitting there. And how do you know if they are acceptable or rejected product. If there was a reject tag and it fell off, by having a lot trace code (going back again to the original question), you would then have a method to track down their status.

Okay, enough rambling.

------------------

 

[CarolX]

How is this for an "official" definition of "lot" from MIL-SID-105E ----

Lot or Batch - The term lot or batch shall mean "inspection lot" or "inspection batch", i.e. a collection of units of product from which a sample is to be drawn and inspected and may differ from a collection of units designated as a lot or batch for other purposes (e.g.: production, shipment, etc.)

Does that help???

 

[Jim Biz]

Well for our situation it helps me to believe we have done it right all along for whatever reason. Even though the MIL definition is government issue it's the best I've seen.

Maybe Bigkahunakal can use it to devise a plan of tracking by run date and call each selected date-period a "lot"?

Thanks for the input
Jim

 

[Sam]

Toyota describes a lot (generic)as:
"A group of items (whether material,parts,components or products) that are produced under the same manufacturing conditions -- such as the same date,operator, job site,manufacturing equipment,operating conditions,and raw material class, size or composition, etc. as appropriate."

The definition changes from customer to customer and from specification to specification. I have seen two different interpretations from the same customer for different parts; 1- amount of product produced in a standard 8 hour shift and 2- amount of product produced in a standard 40 hour workweek.

Best source for determining the lot size;
the customer and the customer specifications.