hetsinker2, "Well, my three reasons are all real possibilities. If you have been in the engineering profession for a number of years, 25 years for me, you would know this. Companies do use different safety margins and some companies really don't understand reliability very well. If you dispute this, well you have not been around the electronics industry that much."
It is true that I have not been in _this industry_ for 25
years, but currently I am, and I was mainly objecting
the AMD "ignorance" part of your statement.
"Oddly enough, most electronic product don't have this constraint. Usually thermal decisions are made based on a projected lifetime, not whether or not the box stops working at a particular temperature. That's why I forgot to include this possibility. But microprocessors are different."
There is one set of rules when you are dealing with a
single FET or few dozen of them. If you have 40,000,000
intertwined FETs, rules are slightly different, you
are right.
95 deg.C is not something unheard of in the semiconductor
industry, but 60 deg.C is simply ridiculous, it indicates
that no margins at all.
Also, your "Intel designs with a greater margin of safety"
does not stand a reality check, just remember the
release of 1.13GHz P-III.
- Ali |
