22 Apr
2012
22 Apr
'12
12:02 p.m.
OK. How many 1,000,000's of parts have you looked at? What long term st=
atistical analysis have you done?
And how many have you _personally_ looked at? I am not talking about
gettign data from elsewhere, which may or may not be reliable?
Extrapolating long tail effects from single samples does not yield accura=
te results.
True enough. But either I have been very unlucky, or there really is the
effect that complex ICs are less reliable than simple ones.I am not
talking about a few more microcontrolelrs failing than 555s. I am talking
aobut often having to repalce a complex IC and very rearle having to
repalce a simple one.
=20
In other words, I'ev; had many thousands of things across the bench by=20
now. And I can assure you that complex ICs, modern ones, old ones,=20
whatever, are mcuh more likely to be dead than simpler ones.=20
Still a small sample set. I would argue that todays computer systems are *much*
more reliable than
the systems that were in general use 10-20 (or more) years ago.
My first thoguht was 'have I jsut telepoted to a another uinverse' and
then I relaised you can't really compare like with like. Thew old
computers that I (and others here) run are oftne very expensive manchiens
(when new), as opposed to the <$500 stuff that most people have today. It
can't be fair to compare the build and reliabiltiy of those.
And I will agree that there are plentyo of (presuambly) modern systems on
the internet that are almost always avaialble (whether becuase each
machine is reliable or becuase there's redunancy or ... I don;t know), so
there must be plenty of reliable modern machines.
Comparing those to a home computer of 10 ro 20 eyars ago is unfair.
Compaing them to a minicomputer or similar is more reasoable, and I've
found such machiens (yes old ones) to be relaivble when properly used.
-tony