Leave computers on?

Discussion in 'NZ Computing' started by Gib Bogle, Nov 23, 2003.

  1. Gib Bogle

    Gib Bogle Guest

    For years I've been leaving my computers on all the time. A throwaway
    comment about the power used by a P4 has prompted me to revisit this
    policy. My reasoning always was that the startup and shutdown cycle is
    what has most effect on the lifetime of many components. I wonder if
    there is any definitive information on this. The main issue, of course,
    is the survival of my disk drives.

    Gib
     
    Gib Bogle, Nov 23, 2003
    #1
    1. Advertising

  2. On Mon, 24 Nov 2003 09:49:18 +1300, Gib Bogle
    <> wrote:

    >For years I've been leaving my computers on all the time. A throwaway
    >comment about the power used by a P4 has prompted me to revisit this
    >policy. My reasoning always was that the startup and shutdown cycle is
    >what has most effect on the lifetime of many components. I wonder if
    >there is any definitive information on this. The main issue, of course,
    >is the survival of my disk drives.


    Different parts of your computer are affected in different ways by
    their operating cycles.

    One of the real killers is thermal cycling. Each time you turn your
    PC on, all the bits inside warm up, and in doing so, they physically
    expand slightly.

    This thermal expansion (and the contraction that occurs as things cool
    down after it's turned off) produce significant mechanical stresses
    inside some components and those stresses can produce failures after
    many cycle.s

    At the moment of turn-on, some parts of your system (especially the
    power-supply and CRT monitor) are subjected to surge currents and
    voltages that may be significantly higher than those experienced when
    the system is left running. A well-designed system should have no
    problem coping with these surges -- but each one represnts the chance
    for a highly-stressed and slightly sub-spec'd component to fail.

    On the other hand however -- a constant exposure to high temperatures
    (such as the core of a CPU left running continuously) produces a
    gradual decay of the semiconductors to the point where they will
    eventually fail..

    Every time you increase the temperature, you hasten this decay -- and
    it's far from a linear relationship -- which is why a cool computer is
    a happy computer.

    Components that have moving parts (such as fans, hard drives, etc)
    tend to wear out and the more hours they're run, the more worn (and
    likely to fail) they become.

    But to answer your question -- is it better to turn off or leave
    running... well the answer dependson your specific circumstances.

    With the price of electricity now hitting over 13-cents a unit in some
    parts of the country and the average PC chewing about 100W even when
    the monitor is turned off, a machine left on permanently will consume
    its own value in power in 10,000 hours of operation. Surprisingly
    htat's less than 18 months of full-time running -- that's less than
    the warranty period of most machines!

    Economically speaking it makes sense therefore to turn your machine
    off and on as required -- allowing the warranty to pay for any bits
    that might fail due to the increased amount of thermal cycling this
    produces. Sure, leaving it on full-time may result in a reduced
    likelihood of failure occuring -- but the power you'll save by turning
    it off and on will have paid for two new machines by the time the
    warranty expires anyway.

    Of course, having said that, I leave my CPUs on 24/7 -- mainly because
    they get used abuot 12-18 hours a day anyway.

    --
    you can contact me via http://aardvark.co.nz/contact/
     
    Bruce Simpson, Nov 23, 2003
    #2
    1. Advertising

  3. Gib Bogle

    harry Guest

    Bruce Simpson wrote:
    > On Mon, 24 Nov 2003 09:49:18 +1300, Gib Bogle
    > <> wrote:
    >
    >> For years I've been leaving my computers on all the time. A
    >> throwaway comment about the power used by a P4 has prompted me to
    >> revisit this policy. My reasoning always was that the startup and
    >> shutdown cycle is what has most effect on the lifetime of many
    >> components. I wonder if there is any definitive information on
    >> this. The main issue, of course, is the survival of my disk drives.

    >
    > Different parts of your computer are affected in different ways by
    > their operating cycles.
    >
    > One of the real killers is thermal cycling. Each time you turn your
    > PC on, all the bits inside warm up, and in doing so, they physically
    > expand slightly.
    >
    > This thermal expansion (and the contraction that occurs as things cool
    > down after it's turned off) produce significant mechanical stresses
    > inside some components and those stresses can produce failures after
    > many cycle.s
    >
    > At the moment of turn-on, some parts of your system (especially the
    > power-supply and CRT monitor) are subjected to surge currents and
    > voltages that may be significantly higher than those experienced when
    > the system is left running. A well-designed system should have no
    > problem coping with these surges -- but each one represnts the chance
    > for a highly-stressed and slightly sub-spec'd component to fail.
    >
    > On the other hand however -- a constant exposure to high temperatures
    > (such as the core of a CPU left running continuously) produces a
    > gradual decay of the semiconductors to the point where they will
    > eventually fail..


    If you stay within the safe operating area of a semiconductor, there is no
    process of depletion or decay.
    Components such as power supplies also have startup voltage and current
    limited by design.
    There aren't really any parts apart from the cooling fans that will
    experience significant attrition through off/on cycles
    CRT monitor phosphors have a fixed attrition, as do the electron gun
    assemblies due to the filament evaporation, so they should be turned off.
    Disk drives deteriorate over time through evaporation of grease on the
    spindle shaft which is accellerated by temperature
    The mtbf for modern drives is ~300,000 hrs or more, whats that, about 35 yrs
    LOL
    The start stops is specified as 50,000 so it might only last 12 years if you
    turn it on and off 10 times per day

    I'd have to agree with Bruce that its more economical to power off or let
    the power management features take care of powering off as much as possible
    as manufacturers seem to have engineered out any potential increased wear
    and tear due to power cycling to the point that they are insignificant.
     
    harry, Nov 23, 2003
    #3
  4. Gib Bogle

    Craig Shore Guest

    On Mon, 24 Nov 2003 10:37:22 +1300, Bruce Simpson
    <> wrote:

    >With the price of electricity now hitting over 13-cents a unit in some
    >parts of the country and the average PC chewing about 100W even when
    >the monitor is turned off, a machine left on permanently will consume
    >its own value in power in 10,000 hours of operation. Surprisingly
    >htat's less than 18 months of full-time running -- that's less than
    >the warranty period of most machines!


    I think you're out on the maths a bit there. You've considered it at
    1000w haven't you? :)

    at 13c/kwh, 100w/hr = .013c/hr

    10,000 hours x .013 = $130

    10,000 hours is 1 year and 52 days
     
    Craig Shore, Nov 23, 2003
    #4
  5. Gib Bogle

    steve Guest

    Gib Bogle allegedly said:

    > For years I've been leaving my computers on all the time. A throwaway
    > comment about the power used by a P4 has prompted me to revisit this
    > policy. My reasoning always was that the startup and shutdown cycle is
    > what has most effect on the lifetime of many components. I wonder if
    > there is any definitive information on this. The main issue, of course,
    > is the survival of my disk drives.
    >
    > Gib


    I've been told it is the changes in temperature that can wreck the PC. If
    you turn it off and on and it is in a rom that is more of less stable in
    temperature throughout the year, that's probably Ok.

    Many people consign the PC to the "study".....an often small room, typically
    not heated if no one is in it....and the temperatures can vary hugely
    through out the year.

    The resulting expansion and contraction can - over time - create
    micro-stresses in the traces that electricity flows through.

    I leave mine on all the time, too. I also replace parts when they get
    old....not when they break. As a result, I have few system failures and
    have lost only one hard disk in active use in 14 years.

    When a hard drive is 3 years old, I buy a new one...and downgrade the old
    one to a "who cares" box....which I seem to always give away to some
    computerless person before any parts fail.

    --
    Best Regards,
    Steve Withers
    defenestrate: The act of throwing Windows out the window and replacing it on
    your PC with some other operating system.
     
    steve, Nov 24, 2003
    #5
  6. Gib Bogle

    steve Guest

    steve allegedly said:

    > I've been told it is the changes in temperature that can wreck the PC. If
    > you turn it off and on and it is in a rom that is more of less stable in
    > temperature throughout the year, that's probably Ok.


    I should have added: 'and the room temperature is rarely below 20 degrees".

    > Many people consign the PC to the "study".....an often small room,
    > typically not heated if no one is in it....and the temperatures can vary
    > hugely through out the year.


    I should have made clear: between being off and being warmed up and turned
    on. On a Winter day, the PC might go from 5 degrees to 45 degrees - and
    back - several times if used intermittently.

    > The resulting expansion and contraction can - over time - create
    > micro-stresses in the traces that electricity flows through.




    --
    Best Regards,
    Steve Withers
    defenestrate: The act of throwing Windows out the window and replacing it on
    your PC with some other operating system.
     
    steve, Nov 24, 2003
    #6
  7. Gib Bogle

    Gordon Guest

    On Sun, 23 Nov 2003 23:25:15 +0000, Craig Shore wrote:

    > On Mon, 24 Nov 2003 10:37:22 +1300, Bruce Simpson
    > <> wrote:
    >
    >>With the price of electricity now hitting over 13-cents a unit in some
    >>parts of the country and the average PC chewing about 100W even when
    >>the monitor is turned off, a machine left on permanently will consume
    >>its own value in power in 10,000 hours of operation. Surprisingly
    >>htat's less than 18 months of full-time running -- that's less than
    >>the warranty period of most machines!

    >
    > I think you're out on the maths a bit there. You've considered it at
    > 1000w haven't you? :)
    >
    > at 13c/kwh, 100w/hr = .013c/hr


    Lets try 13c/Kwh, at 100W we hace 0.013$/hr ;-)

    >
    > 10,000 hours x .013 = $130


    Okay here ;-)

    >
    > 10,000 hours is 1 year and 52 days


    Leap years? ;-)



    --
    Fairy stories exist so children get used to real life
     
    Gordon, Nov 24, 2003
    #7
  8. Gib Bogle

    T.N.O. Guest

    Steve Robertson wrote:
    > Perhaps some dry joints in monitors?? but who ever leaves monitors on??


    Well, mine stays on for four hours, then goes into standby... it remains
    on, and is warm in the morning, so I assume doesnt cool down.
     
    T.N.O., Nov 24, 2003
    #8
  9. In article <bpr6h3$p83$>, Gib Bogle <> wrote:
    >For years I've been leaving my computers on all the time. A throwaway
    >comment about the power used by a P4 has prompted me to revisit this
    >policy. My reasoning always was that the startup and shutdown cycle is


    Simple answer: do you leave your car running all night
    Do you leave your TV on all night.
    Why not, the same arguements apply to them.

    Heat & bearing failure is what can kill PC's parts during their useful life (and
    caps that arent up to the job)
     
    Steve Robertson, Nov 24, 2003
    #9
  10. In article <wRawb.9267$>, "harry" <> wrote:

    >If you stay within the safe operating area of a semiconductor, there is no
    >process of depletion or decay.
    >Components such as power supplies also have startup voltage and current
    >limited by design.
    >There aren't really any parts apart from the cooling fans that will
    >experience significant attrition through off/on cycles
    >CRT monitor phosphors have a fixed attrition, as do the electron gun
    >assemblies due to the filament evaporation, so they should be turned off.
    >Disk drives deteriorate over time through evaporation of grease on the
    >spindle shaft which is accellerated by temperature
    >The mtbf for modern drives is ~300,000 hrs or more, whats that, about 35 yrs
    >LOL
    >The start stops is specified as 50,000 so it might only last 12 years if you
    >turn it on and off 10 times per day
    >


    At last some facts rather than old wives tales
     
    Steve Robertson, Nov 24, 2003
    #10

  11. >
    >Many people consign the PC to the "study".....an often small room, typically
    >not heated if no one is in it....and the temperatures can vary hugely
    >through out the year.
    >

    I never considered this . You will get big temp drops at night , especially in winter.

    Computors used to have there own climate controlled air conditioned
    rooms. Perhaps this is where we can trace this myth about leaving PC's on back to.

    How many PC failures have ACTUALLY be caused by components being heated/cooled
    or turned off/turned on.Perhaps some dry joints in monitors?? but who ever leaves monitors on??
     
    Steve Robertson, Nov 24, 2003
    #11
  12. Gib Bogle

    Craig Shore Guest

    On Mon, 24 Nov 2003 17:35:44 +1300, "Gordon" <>
    wrote:

    >On Sun, 23 Nov 2003 23:25:15 +0000, Craig Shore wrote:
    >
    >> On Mon, 24 Nov 2003 10:37:22 +1300, Bruce Simpson
    >> <> wrote:
    >>
    >>>With the price of electricity now hitting over 13-cents a unit in some
    >>>parts of the country and the average PC chewing about 100W even when
    >>>the monitor is turned off, a machine left on permanently will consume
    >>>its own value in power in 10,000 hours of operation. Surprisingly
    >>>htat's less than 18 months of full-time running -- that's less than
    >>>the warranty period of most machines!

    >>
    >> I think you're out on the maths a bit there. You've considered it at
    >> 1000w haven't you? :)
    >>
    >> at 13c/kwh, 100w/hr = .013c/hr

    >
    >Lets try 13c/Kwh, at 100W we hace 0.013$/hr ;-)


    Yeah dollars/hr, that's what I meant :)

    >> 10,000 hours x .013 = $130

    >
    >Okay here ;-)
    >
    >>
    >> 10,000 hours is 1 year and 52 days

    >
    >Leap years? ;-)


    1 year and 51 days. Actually it was 51.something days for normal
    years.
     
    Craig Shore, Nov 24, 2003
    #12
  13. Gib Bogle

    harry Guest

    steve wrote:
    > Gib Bogle allegedly said:
    >
    >> For years I've been leaving my computers on all the time. A
    >> throwaway comment about the power used by a P4 has prompted me to
    >> revisit this policy. My reasoning always was that the startup and
    >> shutdown cycle is what has most effect on the lifetime of many
    >> components. I wonder if there is any definitive information on
    >> this. The main issue, of course, is the survival of my disk drives.
    >>
    >> Gib

    >
    > I've been told it is the changes in temperature that can wreck the
    > PC. If you turn it off and on and it is in a rom that is more of less
    > stable in temperature throughout the year, that's probably Ok.
    >
    > Many people consign the PC to the "study".....an often small room,
    > typically not heated if no one is in it....and the temperatures can
    > vary hugely through out the year.
    >
    > The resulting expansion and contraction can - over time - create
    > micro-stresses in the traces that electricity flows through.
    >


    Extreme freezing is bad for electromechanical components that depend on
    lubricants, and cycling of temperatures and humidity that causes
    condensation to form is naturally not recommended. Its a big issue for VCRs,
    so it probably would be for HDs too.
    I have never heard of home PC failures due to "micro stresses in traces", do
    you have some more info ?
     
    harry, Nov 24, 2003
    #13
  14. On 24 Nov 2003, Craig Shore wrote

    >>> 10,000 hours is 1 year and 52 days

    >>
    >>Leap years? ;-)

    >
    > 1 year and 51 days. Actually it was 51.something days for normal
    > years.
    >
    >


    1 year 51 days and 16 hours, give or take the odd millionth of a second.
    (assuming a year to be a calendar year of 365.0 days).


    --
    Nicolaas.

    - "Bother!", said Pooh as he stuffed Piglet's corpse into a trash bag
     
    Nicolaas Hawkins, Nov 24, 2003
    #14
  15. Gib Bogle

    steve Guest

    harry allegedly said:

    > I have never heard of home PC failures due to "micro stresses in traces",
    > do you have some more info ?


    I definitely have.

    You can tell if you come across a pC that willwork OK on a warm day....but
    won't even start on a cold day.

    With some others......it won't start on the first several attempts - maybe
    freezing soon after starting to boot.....but once it is warmed up, it will
    work fine....until you turn it off and it gets cold.

    I've seen several systems like this over the years. Usually older ones, as
    you would expect.

    The symptoms will vary quite a bit - depending on what the function of the
    trace is and when it is used. It might be on a video card and not on the
    mobo at all.

    --
    Best Regards,
    Steve Withers
    defenestrate: The act of throwing Windows out the window and replacing it on
    your PC with some other operating system.
     
    steve, Nov 24, 2003
    #15
  16. Gib Bogle

    w_tom Guest

    Those electronics failed because they were defective when
    manufactured. To find those defects, we thermal cycle on the
    order of 100 degrees just to find those defects before high
    rel equipment is shipped. We thermal cycle everything just to
    find those defects. Those failures in PC traces were created
    in manufacturing - not in normal operation. A new computer is
    best first used at all temperature extremes - so that those
    manufacturing defects are detected when the computer is still
    under warranty.

    If those "micro stresses" caused a failure, they had already
    existed in the manufacturing process. And those "micro
    stresses" were going to appear as defects anyway. But then
    steve does not post what is necessary for valid science
    reasoning. He posts personal speculation - that those micro
    stresses were from thermal cycling only because he feels they
    must have been. He does not even know why those "micro
    stresses" were created.

    Reality, a power cycled computer does not change temperature
    enough to see any significant thermal stress. steve simply
    posts symptoms as if they were proof - a violation even of
    principles taught in secondary school science. He has no idea
    what created those "micro stresses" and simply assumes they
    must be from thermal cycling. And he never posts numbers.


    steve wrote:
    > harry allegedly said:


    >> I have never heard of home PC failures due to "micro stresses
    >> in traces", do you have some more info ?

    >
    > I definitely have.
    >
    > You can tell if you come across a pC that willwork OK on a warm
    > day....but won't even start on a cold day.
    >
    > With some others......it won't start on the first several attempts -
    > maybe freezing soon after starting to boot.....but once it is
    > warmed up, it will work fine....until you turn it off and it gets
    > cold.
    >
    > I've seen several systems like this over the years. Usually older
    > ones, as you would expect.
    >
    > The symptoms will vary quite a bit - depending on what the function
    > of the trace is and when it is used. It might be on a video card
    > and not on the mobo at all.
     
    w_tom, Nov 24, 2003
    #16
  17. Gib Bogle

    w_tom Guest

    Power cycling does affect life expectancy. For example,
    one component with a lowest number was an IBM disk drive -
    40,000 power cycles. That means power cycling seven times
    every day for ... 15 years. Yes, power cycling does affect
    life expectancy ... until we apply the perspective called
    numbers.

    Not one poster who recommends leaving it on will post
    numbers from manufacturer data sheets or other reliable
    source. Instead they wildly speculate (personal feelings not
    even tempered by experience) that thermal cycling is
    destructive. But again, a lie is exposed by applying
    numbers. Semiconductors are manufactured by thermal cycling
    over hundreds of degrees. 500+ degrees and no thermal
    stress. Are the trivial tens of degrees in power cycling
    going to affect anything? Of course not. More numbers they
    fear to first learn.

    Those who say leaving it on extends life expectancy are
    simply promoting junk science. Junk scientists are obvious.
    They must avoid the perspective of numbers. Numbers provided
    above demonstrate perspective - that power cycling is
    irrelevant. Run the computer as you would a TV.

    Then it gets more interesting. Normal computer operation
    involves massive power cycling. For example, that Intel CPU
    requires a motherboard power supply that can go from 1 amp to
    tens of amps ... not in millisecond ... but in microseconds!
    That is the most severe power cycle ... that is also called
    normal operation. IOW leaving a computer running is constant
    power cycling. That is what digital logic does - power
    cycle. Those power cycles called normal operation cause
    uneven and constantly changing heating everywhere on the
    computer die.

    Turn it off or put it to sleep to obtain best life
    expectancy. Power up does not create massive surge that myth
    purveyors claim. And those myth purveyors don't even know
    about the inrush current limiter - that was standard on
    electronics even in 1950 TVs. Just another reason why no
    massive power surge when electronics is powered on.

    Leaving it on to preserve life expectancy is based upon no
    manufacturer data sheets, no understanding of what is
    considered hot to electronic components, and not even
    knowledge of inrush current limiters. They don't bother to
    first learn how thing work before making wild and speculative
    declarations. It is called junk science reasoning. IOW they
    must avoid numbers to promote their 'feelings' - facts be
    damned.

    Turn it off to preserve its life. Ignore junk scientist who
    have no numbers.

    Steve Robertson wrote:
    > At last some facts rather than old wives tales
     
    w_tom, Nov 24, 2003
    #17
  18. Gib Bogle

    T.N.O. Guest

    harry wrote:
    > I have never heard of home PC failures due to "micro stresses in traces", do
    > you have some more info ?


    Even if it was the cause, would the price of a new part warrant an
    investigation? I doubt it... easier to just replace the part.
     
    T.N.O., Nov 24, 2003
    #18
  19. Gib Bogle

    harry Guest

    T.N.O. wrote:
    > harry wrote:
    >> I have never heard of home PC failures due to "micro stresses in
    >> traces", do you have some more info ?

    >
    > Even if it was the cause, would the price of a new part warrant an
    > investigation? I doubt it... easier to just replace the part.


    Not really
    From the point of view of a manufacturer, susceptability to such a fault
    would result in an unsustainable number of warranty claims from resellers.
     
    harry, Nov 24, 2003
    #19
  20. Gib Bogle

    harry Guest

    steve wrote:
    > harry allegedly said:
    >
    >> I have never heard of home PC failures due to "micro stresses in
    >> traces", do you have some more info ?

    >
    > I definitely have.
    >
    > You can tell if you come across a pC that willwork OK on a warm
    > day....but won't even start on a cold day.
    >
    > With some others......it won't start on the first several attempts -
    > maybe freezing soon after starting to boot.....but once it is warmed
    > up, it will work fine....until you turn it off and it gets cold.
    >
    > I've seen several systems like this over the years. Usually older
    > ones, as you would expect.
    >
    > The symptoms will vary quite a bit - depending on what the function
    > of the trace is and when it is used. It might be on a video card and
    > not on the mobo at all.


    Stress won't cause a conductor path to cease to pass current, when the
    stress causes breakage the damage would be permanent.
    I'm willing to stick my video card in the freezer and then de-frost it.
    It is highly improbable that the coefficient of expansion of circuit board
    copper, or its ductility would cause fracturing between operating
    temperature and low temperate climate temperatures.
    Startup problems in cold weather are more likely to be bearings sticking on
    CPU coolers with sensing, or on HDs, or condensation on cold metal parts
    causing switching power supplies to fail to start.
     
    harry, Nov 24, 2003
    #20
    1. Advertising

Want to reply to this thread or ask your own question?

It takes just 2 minutes to sign up (and it's free!). Just click the sign up button to choose a username and then you can ask your own questions on the forum.
Similar Threads
  1. =?Utf-8?B?bWFobW91ZCBvc2FtYQ==?=

    remove winXp & leave win2000

    =?Utf-8?B?bWFobW91ZCBvc2FtYQ==?=, Aug 15, 2004, in forum: Microsoft Certification
    Replies:
    3
    Views:
    948
    Guest
    Aug 24, 2004
  2. LnkWizard

    OT: Leave it to the French

    LnkWizard, Jan 10, 2006, in forum: MCSE
    Replies:
    15
    Views:
    634
  3. Roseb441702

    Question: Do you leave your computer on or turn it off?

    Roseb441702, Aug 9, 2003, in forum: Computer Support
    Replies:
    20
    Views:
    972
    Monsignor Larville Jones MD
    Aug 13, 2003
  4. Replies:
    0
    Views:
    541
  5. kimiraikkonen
    Replies:
    5
    Views:
    1,323
    Walter Mautner
    Feb 7, 2007
Loading...

Share This Page