Component Qualification

[Milas]

Hi

Does anyone know of guidelines which can be usedn when determining the sample size for qualifying components which will be used in medical devices ?

Thank you in advance

 

[Bev D]

Are these ‘off the shelf’ parts or custom designed?

 

[Milas]

Hi, it a combination of both, of the shelf and custom parts

 

[Bev D]

Are you in the early stages or final stages of the development process? In other words are you selecting the parts or trying to validate that the parts selected will work in the final product?

In the meantime you will need to think about testing the components at the extreme limits of the allowable tolerances, not (just) at the nominal.

 

[Milas]

Thanks Bev

We are at the manufacturing stage

 

[Bev D]

I take that to mean that you are post validation?

In any case, I do not recommend using the standard random AQL or Confidence/reliability approach as they are intended for final inspection of a stream of lots…They are not intended for qualification or validation. They are especially weak if you only test product that is air near the nominal and not at the extremes. If you exclude extreme values (at the min and max of the allowable tolerances) you cannot get any information about those parts performance and you will surely get those parts in real manufacturing.

I always select parts at the extremes and use them in the final product (randomly selected across the manufacturing spectrum) and if applicable test the final product at the worst case conditions of consumer use…you don’t need a lot of parts. Maybe 5-10 because this is about directed testing and not probability or variation. Any failure is a failure. The specifications were probably set incorrectly…

 

[Bev D]

If you must use the formulas for estimating the percent defective (or mean and SD).