Steve Walker wrote:
On Mon, 14 Jun 2010 19:49:22 +0100, geoff wrote:

In message , Dave Osborne
writes
geoff wrote:
In message o.uk,
Dave Liquorice writes
If stuff survives the first few days or weeks of regular use it will
essentially last forever. Always assuming that it has been designed
well, no components right up at the top of their power rating with
inadequate methods of removing the heat resulting in scorched circuit
boards and weakend joints after a year or three of use or simple
component failure due to being under too much stress.

Lovely sweeping statements based on invalid assumptions there

Not at all.

http://en.wikipedia.org/wiki/Bathtub_curve
Yes, at all

Your assumptions concerning good design, components not being stressed
and heat transit are only ones that can be made in an ideal world

they are certainly not encountered in real consumer electronics

And I'm not sure how you equate "essentially last forever" with what
parts 2 and 3 of the curve - random and wear out failures

Tin whisker growth in germanium transistors?

Actually chips can and do fail after a period of working perfectly, as the
stresses of thermal cycling can break the welds on the fine wires
connecting the actual IC to its external pins.


No, that's unusual, especially in 'proper' packaging, because they are
not 'tight'

Normal CHIP (as opposed to package) failure mode is dopant migration,
causing the semiconductor to become less of a semiconductor: that's
happens in a few seconds at 200C, a few hoursr at 199C..a few days at
198C..you get the picture. That and crackled seals leading to chemical
poisoning are the two main causes of ageing.

I've never had a bondwire actually come off.





SteveW