Arduino vs 555 reliability

On Apr 21, 2012, at 10:34 PM, Guy Sotomayor wrote: I think you're arguing about different things. Today's PCs are largely junk and generally less reliable than those of years ago, but I'd say that's more due to poor engineering on the part of the designers of the actual systems. They take shortcuts, don't do proper thermal engineering (which is a lot more important in today's PCs than it was 20 years ago), and generally do a poor enough job that you can't expect an assembled system to last more than about 5 years with any degree of confidence. BGAs, in particular, do a great job of pulling their pins off the board when subjected to thermal cycles; if you don't design your thermal relief well enough, it's one of the best ways to hasten your system's doom. Note that it's not generally the ICs failing in these systems, at least not on their own. It's usually their containers that cause the various failures. Today's high-end systems, though, are much more robust and designed with durability in mind (well, a lot of them, anyway). I'd say you're about as likely to find an IBM Z-series mainframe alive in 30 years as you are to find a VAX alive (not dead due to natural causes, in any case) today. I will cop to MOS being somewhat less resilient to actual abuse than bipolar logic, and things with tinier gates do exacerbate that; beyond that, I seem to recall that smaller geometries are a bit more susceptible to electromigration, and that tends to get worse with higher temperatures. Of course, the actual MTBF data doesn't always bear that out, so I could have been told wrong. - Dave (amusing side note about MTBF: we had a customer who developed a really awful board, but claimed it had an MTBF of over 100 years; it turns out the way they "calculated" this was by adding the MTBFs of the individual components)