44

Among some computer users I've heard the claim that turning a computer on and off reduces the life of the computer, here's an example of this claim. In this claim, it is believed that turning the computer on and off stresses or damages various components more than leaving the machine running 24/7. Others claim that the components are damaged more by being left running.

So which claim is correct, does turning a computer on and off reduce the time to failure more than leaving it running?

Or does leaving a computer running all the time, wear it out?

Additional Details:

I guess I'm assuming that one would not turn the computer on and off more than a handful of times a day, but usually just once. As opposed to leaving it running 24/7.

[Side question: does sleeping have a similar effect as turning it off?]
[What about when the computer Hibernates?]

Jason Plank
  • 3,180
  • 1
  • 26
  • 17
Mark Rogers
  • 11,324
  • 11
  • 55
  • 71
  • 2
    Sleeping is leaving the computer on at reduced power usage. (A Mac Mini uses very little power while sleeping.) Hibernating, by contrast, is turning the computer off in a way that it can pick back up where it left off. – David Thornley Apr 13 '11 at 02:19
  • 1
    You have to give some numbers. A fast booting PC could be turned on and off about 60 times per hour. If you compare that, to `let it run 24/7`, you might get different results to `power on/off over night` or `5 times a day`. I would expect different components to react differently. The fan is more polluted by constantly running, I guess, and not so much impressed by turning on and off. Other components might suffer more from changing temperatures. – user unknown Apr 13 '11 at 03:21
  • 2
    possible duplicate of http://serverfault.com/questions/258064/should-servers-be-turned-off-at-night – Paul Apr 13 '11 at 03:53
  • 3
    couldn't find any citations in the serverfault answers, though. – Paul Apr 13 '11 at 03:55
  • 2
    The same question on SuperUser: http://superuser.com/questions/2733/is-it-better-to-leave-your-computer-on-all-the-time – sharptooth Apr 13 '11 at 05:07
  • Also when a computer hibernates it saves its state and then shuts off - so it's the same as turning off in terms of electrical circuits. – sharptooth Apr 13 '11 at 05:10
  • 2
    I'm curious to see Skepticism applied to this concept. – Mark Rogers Apr 13 '11 at 13:32
  • The main point would be the hard drive which is party mechanic (for a normal hard drive). Server hard drives are optimized for continuous operation and have a limited spin-on number. Servers should therefore not switched on and off often, which isn't normally the case anyway. Consumer hard drives are optimized for frequent on/off-cycles and will suffer if they run continuously. The rest of the electronic should not be that much affected, but should be less stressed will off and therefore cool. The boo-up process should not strain it to much. – Martin Scharrer Apr 14 '11 at 21:41

3 Answers3

21

A PC has many failure modes, and it would be hard to address all of them. One of the more common is hard disk failures. Google did an extensive study and concluded that "for drives aged up to two years ... there is no significant correlation between failures and high power cycles count. But for drives 3 years and older, higher power cycle counts can increase the absolute failure rate by over 2%. We believe this is due more to our population mix than to aging effects. Moreover, this correlation could be the effect (not the cause) of troubled machines that require many repair iterations and thus many power cycles to be fixed."

I.e. as far as the harddisk is concerned, there's no clear benefit.

MSalters
  • 2,277
  • 17
  • 16
  • maybe, but in the past there may well have been... Harddrive failure because of increased mechanisal stresses during power cycles is the main reason I've heard for not powering down harddrives (next to the productivity gain from not losing time waiting for the machine to boot of course). – jwenting Oct 24 '11 at 06:10
  • Can you please define "high power cycle counts"? – Jase Dec 04 '12 at 02:56
4

Another thing that nobody has talked about is the unstable power supply in poorer countries. In some countries (for example, Pakistan), you have frequent power cuts and/or brownouts.

Such voltage fluctuations can cause computers to fail.

In such a situation, there is a strong incentive to unplug a computer that is not is use, to prevent it from being damaged.

(Yes, you could get a UPS but they are expensive.)

Oddthinking
  • 140,378
  • 46
  • 548
  • 638
Abdussamad
  • 141
  • 2
  • 1
    [Welcome to Skeptics!](http://meta.skeptics.stackexchange.com/questions/1505/welcome-to-new-users) Please [provide some references](http://meta.skeptics.stackexchange.com/q/5) to support your claims. (I know some of the other answers here don't meet this requirement, but that is because they were written before this requirement was formalised.) Also, at its heart, this question is about whether the price of extra electricity outweighs the price of replacement parts and wait time, no matter what the country. – Oddthinking Dec 01 '13 at 11:56
  • References? You want references about the electricity situation in poor countries like mine? Ok I'll add those. – Abdussamad Dec 01 '13 at 12:03
  • "this question is about whether the price of extra electricity outweighs the price of replacement parts and wait time" Wait time maybe but price of replacement parts? You are assuming that shutting down and restarting computers more often leads to a shorter life span. That is debatable. – Abdussamad Dec 01 '13 at 12:08
  • Sorry, I didn't mean to "beg the question" by saying they *would* have a shorter lifespan. That is what was being asked. I guess I meant even in richer countries, electricity prices for running computers are not negligible. If they were, I don't think the question would come up. – Oddthinking Dec 01 '13 at 12:41
4

Obviously if the computer is turned off then electricity is saved (whether it is actually turned off or in standby/hibernate mode). Over time, the savings in electricity could equate to the cost of a new computer. When this intersection is reached it doesn't matter if keeping the computer on makes it last longer or not.

Cost savings references:

http://computer.howstuffworks.com/question328.htm

http://money.usnews.com/money/blogs/fresh-greens/2009/03/25/simple-green-step-shut-down-your-computer-every-night

sosc
  • 73
  • 2
  • 3
    The howstuffworks article assumes that the computer uses 300 watts all the time. This absolutely isn't the case, especially with idle power modes (http://www.youtube.com/watch?v=mJzcJhEj9z8). – Chris Dennett Apr 13 '11 at 08:43
  • 3
    Of course the life-span of the computer matters. If you haven't saved up the cost of a new computer in electricity between the reduced life-span from turning it on and off and the time it would have broken anyway, you lost money. – Lennart Regebro Apr 13 '11 at 13:31
  • 4
    @Lennart: On the general subject of cost, there's the societal cost in developing countries of a) mining all the exotic materials, and b) disposing of them. – Mike Dunlavey Apr 14 '11 at 22:03
  • 1
    The trouble with this answer is that it doesn't at the question of what is good *for the computer*. Maybe lower electricity use is good for the *owner*, but I doubt the computer cares a flipped bit about it. – matt_black Dec 01 '13 at 19:07