Replacement of obsolete components

- J
- jesper.holm
  
  Contact options for registered users
posted
15 years ago

Thu, Mar 12, 2009 9:05 AM

Hi,

I'm looking for some guidelines regarding replacement/substitution of obsolete components in our products.

Currently the procedure is as follows: Passive components are evaluated by datasheet, and replaced without testing in case of matching data.

Semiconductors are evaluated by datasheet, and tested in 5 pcs. of the product.

The test is a standard functional production test, which basically consists of a verification of the functionality; signal levels, band widths, memory addressing etc. No temperature cycling, EMC test or burn in is performed.

My worry is, that you in case of e.g. a die schrinked SRAM or uP, you might reduce your margin to an unacceptable level, making the product more sensible to EMC, temperature or what ever...

How is this qualification done in the industry? E.g automotive?

Thanks

Jesper

- D
- David L. Jones
  
  Contact options for registered users
Vote on answer
posted
15 years ago

Thu, Mar 12, 2009 10:43 AM

It's entirely up to each individual company, case-by-case, and in many cases up to the whim of individual departments/people within companies. There is no rule of thumb or industry standard qualifications, but different components may require as much trouble as formal product re-verification (e.g. EMC, safety standards, specific industry standards etc)

For most apps you can pretty much get away safely with data sheet comparisons and standard functional tests as you have done. Some components and circuits are obviously more likely to cause issues than others, by orders of magnitude. It really comes down to the design engineers judgment.

Dave.

- T
- Tim Shoppa
  
  Contact options for registered users
Vote on answer
posted
15 years ago

Thu, Mar 12, 2009 1:08 PM

Boilerplate disclaimer on tube testers a half century ago said something like "While measurements done on the tube tester can be useful, performance of the tube in the actual radio being repaired is the final test".

Having gone through some "six sigma" training and also some price analysis, the real question isn't whether the new component matches the old component, but rather whether the new component matches the application. Specsmanship can be used to drive business costs and processes way out of control both money and time-wise. The product didn't work with 1 percent capacitors? God-damn-it, next time we'll spec 0.1 percent capacitors!

Tim.