Xref: utzoo comp.misc:2096 comp.terminals:626 Path: utzoo!mnetor!uunet!yale!cmcl2!phri!roy From: roy@phri.UUCP (Roy Smith) Newsgroups: comp.misc,comp.terminals Subject: Re: Does Turning off PC's everyday do any real damage? Message-ID: <3174@phri.UUCP> Date: 13 Mar 88 03:09:17 GMT References: <1727@ssc-vax.UUCP> Reply-To: roy@phri.UUCP (Roy Smith) Organization: Public Health Research Inst. (NY, NY) Lines: 54 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. Here's some of the things that I've heard are bad about turning computers on and off every day. I've never seen hard experimental evidence that these are true, but they all make a lot of sense from an engineering point of view. 1) Thermal cycling of CRT filaments. When a filament gets hot it expands and when it gets cold again it contracts. The thermal cycling leads to mechanical stresses on the filament which cause it to break. Tungsten (which I assume CRT filaments are made of) has a negative thermal coefficent of resistivity, meaning a filament has a lower resistance when cold than when hot; that translates into large inrush currents when you first turn a CRT on (i.e. before the filament gets hot). Decent CRT designs will have some sort of current limiting, but it's not 100% effective. Over the past few years, there have been extensive discussions of light bulb life on the net; if you can find those old articles somewhere, I'm sure a lot of what was said applies to CRT filaments as well. Thermal cycling, to a lesser extent, will also cause damage to solder and contact connections. 2) Power supplies. When a power supply is turned on, it may take some time for the output to stabilize; before then, it is putting out some strange voltage. CMOS may not mind running off of a variable power supply, but TTL does strange things on 1.5V, especially if you've got several supplies and some have come up to rated voltage before the others do. You may see abnormally high transient currents. Even if the power supply stays strictly within the range from 0 to its rated output, the intermediate voltages in between may do strange things to the logic circuits. 3) Disk drives. When you power down a disk, the heads have to land (yes, I know not all disks have landing heads, but most small winnies do). While landing on a dedicated landing zone is better than landing on some random part of the disk, its better not to land at all. When you power up again, you get big inrush currents in the drive motor until it gets up to speed. On the flip side, every hour your PC is powered off is one less hour it's exposed to potential power-line uglies. With any kind of decent surge protection, though, I'd say that this is of pretty minor significance. If, on the other hand, you know that you routinely get massive surges in the middle of the night, it might be worth considering. If you've got a large number of PCs (like 100), you may be in a good position to do a controlled experiment. Pick half your PCs at random and turn them off every night. Leave the other half on. Keep careful records of repairs. At the end of, say, 6 months, sit down and work up some stats. Then, do the entire net a very big favor and let us all know how things work out. -- Roy Smith, {allegra,cmcl2,philabs}!phri!roy System Administrator, Public Health Research Institute 455 First Avenue, New York, NY 10016