Xref: utzoo comp.misc:2059 comp.terminals:612 Path: utzoo!mnetor!uunet!lll-winken!lll-lcc!pyramid!csg From: csg@pyramid.pyramid.com (Carl S. Gutekunst) Newsgroups: comp.misc,comp.terminals Subject: Re: Does Turning off PC's everyday do any real damage? Message-ID: <16473@pyramid.pyramid.com> Date: 7 Mar 88 18:17:23 GMT References: <1727@ssc-vax.UUCP> Organization: Pyramid Technology Corp., Mountain View, CA Lines: 42 Keywords: I hate Keywords In article <1727@ssc-vax.UUCP> dmg@ssc-vax.UUCP (David Geary) writes: >I am interested to know, from a TECHNICAL viewpoint, whether >or not it causes any damage to turn a PC off everyday. This is a religious war that comes up anually.... Here's one engineer's view: In general, electrical equipment functions most reliably when it's power and temperature are constant. There is a laundry list of components and subsystems that can be damaged by changes: stresses from thermal expansion and contrac- tion (and resulting misalignment of critical parts, like disk drive heads); surge voltages into the MOS components; surge currents into the power-supply components and CRT filament; wear on bearings of rotating parts (fans). Of course, leaving a CRT on does no good if you don't crank the screen intensity down whenever you have a constant display.... All of which suggests that, yes, you should leave the equipment on all the time, if you have a stable and spike-free line power. That's a critical "if." At home, I *do* turn my computers and other electronic equipment off after use, because of the poor quality of incoming electrical power. Better to have the clean, controlled power-cycle of the "off" switch than the erratic surges coming in over the power lines. The exception: I leave my terminals on, with their temperature-sensitive CRTs. >However, here at Boing, we are supposed to turn off all >pc's every night to save electricity. We had a similar policy at Burroughs. It was overturned when an engineer in the terminals division proved that the increased maintenance cost of a power- cycled terminal exceeded the electricity cost by a factor of 10. This was in 1979, when CRT terminals cost twice as much as they do now, and electricity cost half as much. But consider this: a typical terminal only costs $65 per year to run continuously, versus $15 per year to run during business hours. That means that if only one out of every ten terminals fails during the year because of power cycling, you've wiped out your cost saving. I would estimate that, for terminals, the number is much higher than one in ten per year. Our lab has about 20 minicomputers in it. All have conventional Wyse, Ampex, AT&T, and Espirit console terminals. Some (on the production systems) stay on all the time. Others (on test systems) get power cycled a lot. The difference is striking and aggravating -- aggravating since we're constantly having to find replacement terminals for the test systems.