[WEB4LIB] Chip creep and power surges

[Walt_Crawford at notes.rlg.org]

>May I ask David (or anyone else) to explain
>what "chip creep" means?  This is a new one for me.  Thanks!

David may have a subtler answer, but (theoretically) thermal changes in the
PC's internal environment could cause chips to creep out of their sockets.
It's true enough that the largest thermal change in a PC's environment, in
most office buildings/libraries at least, is probably when it's warming up.

Let's just say that I've never seen one of the major PC magazines report on
the Dangers of Chip Creep: possible, but not a major issue. Again, maybe
over 15-25 years, or if you're going through power cycles several times a
day, but...

As to Debra's comment about the power required to start up a PC: as far as
I can tell, that's an urban legend.

Yes, there's probably a small surge on power-up, mostly for the power
supply's own fan. No, there's no evidence (that I've ever seen) suggesting
that it's even a fraction of an hour's worth of computer use. It's wildly
improbable given the sheer numbers involved. Let's say that a typical PC
draws 40 watts (ignoring the display for the moment): a guess, but probably
within 50% of the truth in one direction or the other. That means it draws
40 watt-hours in an hour. The power-up "surge" lasts considerably less than
half a second (more like 1/10th second, but...). For that surge to draw as
much power as one hour of normal operation (40 watt-hours), it would have
to draw 40*3600*2: 288,000 watts. Which would instantly trip the circuit
breaker for that circuit: at U.S. voltage, that's 2500 amps. More
realistically, if the power surge even equalled one *minute's* power usage,
and you turned on a power strip with 10 PCs attached, the power strip would
self-destruct since you'd be drawing 720 times normal wattage. (Amazing
what a little calculation can do...)

Motors do have significant power-on surges--maybe double the normal power.
But in terms of overall power use, the power-on surge is simply irrelevant:
it's not even up to the "noise" level.