When does containment event not require 100% sort?

[gtr01]

Greetings,

I am a QE that has transitioned from a IATF Tier 1 to a Tier 2 and we make a single components to send to a tier 1 assembly facility. Our process produces an incomplete part every 10000 cycles, and have issues with them escaping to customer. Our process shows satisfactory capability and excellent PPM, however customer still demands 100% sort all parts made when they receive defect. Sorting 50000 parts (typical batch size) for an unfinished part seems unnecessary granted extremely low possibility that defect can occur.

Is it reasonable to push back on these demands for 100% sorts and put more pressure on the customer to provide statistically relevant data to justify such a sort?

Any feedback is appreciated.

Greg

 

[Sidney Vianna]

Is it reasonable to push back on these demands for 100% sorts and put more pressure on the customer to provide statistically relevant data to justify such a sort?

Any feedback is appreciated.

Customers are not always right, but they are always the customer. Can you dialogue with the customer? Can you explain that the sorting process costs are eventually (most likely indirectly, obviously) passed on to them? Can you assess from both organizations' perspectives that it is more cost efficient (for the business relationship) to live with a 100 DPPM rate? Until both organizations agree that the current performance is acceptable from both a quality and cost perspectives, any unilateral attempt for you to use statistics to convince the customer is wrong will likely backfire.

Good luck.

 

[chemohd]

possibility to check with customer on their quality acceptance limit. There must be a reason they are rejecting the lot( batch).

 

[optomist1]

Hi gtr01,

"however customer still demands 100% sort all parts ...", how is or are these demands documented, a performance spec, what does your PO state and what was originally specified (source package or RFQ)...and your firm quoted?

Cheers,
Optomist1

 

[PaulJSmith]

I have seen acceptance of as much as 15% incorrect parts in the aerospace industry, and have also seen automotive customers who see one NC part and want someone to fly to their location for a 100% sort on their dock. Every customer will have different expectations of their suppliers, and it usually depends on the criticality of the part.

It's best to talk with your customer and work out a solution that's reasonable for both parties. If their documented requirements state 100% sort on this particular part, then you're probably stuck with that, and should do it before it leaves your facility.

It's also possible that such a requirement was simply a copy/paste from a previous spec, and may not actually be necessary for that part. An open dialog with them may get that changed to something more reasonable and cost-effective for both.

Talk with your customer.

Also, if you know you're having an escape issue, you need to deal with that internally.

 

[Matt's Quality Handle]

Customers are not always right, but they are always the customer. Can you dialogue with the customer? Can you explain that the sorting process costs are eventually (most likely indirectly, obviously) passed on to them? Can you assess from both organizations' perspectives that it is more cost efficient (for the business relationship) to live with a 100 DPPM rate? Until both organizations agree that the current performance is acceptable from both a quality and cost perspectives, any unilateral attempt for you to use statistics to convince the customer is wrong will likely backfire.

Good luck.

What is the impact of an incomplete part at your customer's facility? Does it create downtime for them? Is it a pass through that they have no opportunity to detect, but could cause functional problems with THEIR customer?

In short, this is the default mode for the customer. Likely to move them off this position will require cooperation, not a statistical proof.

If I were your SQE, some questions that I would ask of you are:

• What causes this? How do you know?
• How do you know the defect rate won't get worse?
• What are you doing to improve the process? When will it take effect?