environmental stress screening of electronic components 1

[qassure]

presently COTS (commercially off the shelf) components are prefered than MIL approved components, due to cost, availablity and other factors. Environmental steree screening (ESS) is carried out on these components to weed out the infant mortality (i.e. to remove the components which are likely to fail early). In practice ESS is carried out at various levels like component level, card (PCB) level and module level. MIL-STD 2164A gives general idea about the ESS. But the following quiries are not clear. please any one may give some insight.
(a) what is the significance of ESS at these various levels.
(b) What should be the severity of ESS at various levels.

 

[IRstuff]

The whole purpose of ESS is a a cheap screen for gross workmanship flaws, that's it. There's not enough acceleration to get at infant mortality. We used to burn-in 883 ICs for a solid week. A 24-hr or 48-hr ESS at significantly lower temperature is not much of a stressor, even with vibration.

TTFN

 

[qassure]

Thank you Mr. IRstuff for your reply.
Generally, We carry out thermal cycling with 5Deg.C per minute rise / fall in temperature for 10 cycles for a module with pre and post vibration. In case of PCBs thermal shocks used to be given. The severity of temperature at high or low depends on the modules / PCBs. In spite of these, is burn-in necessary for weeding out the infant mortality. If so at what temperature we should carry out burn-in of module / PCB .
Further what is the merit / demerit of HASS (Highly Accelerated Stress Screening) over ESS.

 

[IRstuff]

Unless you have specific data or other manufacturer's information, too much burn-in can be a bad thing, as it eats into the useful life of the parts.

HASS is an alternative approach to simple thermal acceleration of reliability failures. Application of voltage and temperature is supposed to aggravate failures in weak and potentially unreliable parts. Naturally, there is little evidence that such efforts buy real reliability, since there is little motivation to spend scads of money to test to fail for complex systems. So while the parts that do fail during HASS are indeed weak, it's unclear whether such parts would have actually been reliability failures in the field.

TTFN