Searching \ for '[TECH] Leaving PC on overnight reduces lifespan?' in subject line. ()
Make payments with PayPal - it's fast, free and secure! Help us get a faster server
FAQ page: www.piclist.com/techref/index.htm?key=leaving+overnight
Search entire site for: 'Leaving PC on overnight reduces lifespan?'.

Exact match. Not showing close matches.
PICList Thread
'[TECH] Leaving PC on overnight reduces lifespan?'
2008\12\08@104448 by Virchanza

picon face

All good things must come to an end, and every computer claps out
eventually.

Let's say you get up in the morning, and about 2 hours later you turn
your PC on. It remains on for the day and then you turn it off before
you go to bed. You sleep for about 9 hours. So altogether, out of a
24-hour period, your PC is on for 13 hours.

I'm wondering, if you leave your PC on overnight, effectively having it
on 24 hours out of 24 hours, then will it only last half as long... ?
Will it clap out after 8 years instead of after 16 years?

(One thing that crossed my mind is that maybe the PC would even last a
little longer if it's left on because you wouldn't have the damage done
from heating up and cooling down.)

The reason this came to mind is that I lost my mobile phone today and so
tonight I'm using my laptop as an alarm clock (I'm using "crontab" under
Linux). I'd use it for this purpose all the time but I don't want to
curtail its lifespan.

2008\12\08@110901 by olin piclist

face picon face
Virchanza wrote:
> I'm wondering, if you leave your PC on overnight, effectively having
> it on 24 hours out of 24 hours, then will it only last half as
> long... ? Will it clap out after 8 years instead of after 16 years?

My guess is the electronics is actually better off left running.  Running by
itself doesn't wear it out, but thermal cycling does.  Any moving parts do
wear out with on time though.  If you have it configured to power down the
disk drive when not in use, then that shouldn't be much of a difference to
it.  The case and CPU fans however will wear out sooner.  Even in low power
standby mode, some fan is usually still running, even if at a lower speed.

Then there is the issue of wasting power.  Definitely turn off the monitor,
speakers, and anything else external.  I keep all that stuff at home on a
separate outlet strip so that I can shut it off when I keep the computer on
to crunch something when I'm not there for extended time.  Still, even a
"off" computer takes a surprising amount of power.  Many keep the USB ports
powered.  Shut down a computer but leave it plugged in and feel how warm the
power supply is after a hour or so.  I use my computer at home infrequently,
so I have it on the master outlet strip to completely turn it off when not
in use, which can be days or even a week or two.  The peripheral outlet
strip is fed from this one, so there is one master switch to shut off
everything.

> The reason this came to mind is that I lost my mobile phone today and
> so tonight I'm using my laptop as an alarm clock (I'm using "crontab"
> under Linux).

Get a real alarm clock.


********************************************************************
Embed Inc, Littleton Massachusetts, http://www.embedinc.com/products
(978) 742-9014.  Gold level PIC consultants since 2000.

2008\12\08@112745 by Herbert Graf

flavicon
face
On Mon, 2008-12-08 at 22:44 +0700, Virchanza wrote:
> I'm wondering, if you leave your PC on overnight, effectively having it
> on 24 hours out of 24 hours, then will it only last half as long... ?
> Will it clap out after 8 years instead of after 16 years?

Perhaps a little less helpful, but honestly, does it matter? How useful
will your PC be in even 8 years?

FWIW I've had consumer class machines running 24/7, and others being
shut down every night; either way, by the time something fails (usually
a PS or HD, I ignore optical drive failures, they fail way too often)
they are so obsolete it doesn't really matter any more.

As a barely related aside: one of my older machines (being used as my
main server) was crashing every few hours (I thought it was done). It
got worse when under load.

Started swapping pretty much every component to determine what the
problem was. In the end I was down to the MB/CPU, since I didn't have
those to swap. On a lark I loaded that machine up to cause it to crash,
then quickly had a look at the CPU temp (all fans were fine, heatsink
looked fine). The CPU was hovering at idle at about 71C! My rule is if
an electronic component is too hot to touch comfortably it's running too
hot, and 71C is definitely too hot.

I was stumped, the heatsink was fine, the fans were running, Vcore was
correct. While looking at stuff I noticed allot of dust had accumulated
on top of the HS. Closer inspection revealed that yes, the fan was
running, but no air was actually passing the HS fins! A quick vacuum and
the machine's been up for days now with zero issues. Been using
computers for many years, but I'm still capable of such a basic rookie
mistake! :)

TTYL

2008\12\08@120131 by Picbits Sales

flavicon
face
> I'm wondering, if you leave your PC on overnight, effectively having it
> on 24 hours out of 24 hours, then will it only last half as long... ?
> Will it clap out after 8 years instead of after 16 years?
>
> (One thing that crossed my mind is that maybe the PC would even last a
> little longer if it's left on because you wouldn't have the damage done
> from heating up and cooling down.)
>

Tricky one this.

The fans for the GPU/CPU/PSU will wear out sooner and the hard drive may or
may not last as long as its being cycled but then there is the consideration
of the finite lifespan of the hard drive bearings.

I run my own servers from home and they are on 24/7. I have them in the
attic and can hear them rumbling away at night. Its half reassuring and half
annoying that I can hear them but I've got used to them (and my wife is half
deaf anyway so she doesn't hear them). I've yet to have a hardware failure
in 3 years of having them constantly on.

A few years when I worked as an IT manager, we ran a Novell network. One
machine was dedicated as a printer server, had no HDD, a slow processor and
had a floppy drive to boot up from. Whenever we had a power cut, it would
just turn off and when the power came back on it was ready for work again.
Nobody remembered who installed it, how long it had been there for or really
what was on the "magic floppy" that made it work. I just left it alone and
it was only replaced when we needed more printers at that end of the
building (we went for a dedicated print server box). That was over 10 years
ago and it would still probably be sitting there today without a problem if
that extra printer hadn't come along......

The server itself died about a year after I joined the company so I built a
mega super dooper new one with all the latest Raid controllers and loads of
storage/memory. I always remember the day the MD called me in and pointed
out the up time was over a year.

When we did have a catastrophic power failure that the UPS's finally gave up
on, everything shut down as it should have done but when the power came back
on all hell broke loose. The CPU fan had seized and didn't start back up
(machine complained loudly about that as it was all monitored) and 3 of the
16 case/hdd cooling fans had seized along with a couple of other fans that
were noisy. It was only when they were stopped that they actually gave up.
Typically I was on holiday that day and was called in to sort it all out.

My Linux servers will run quite happily without rebooting but I'm always
adding new bits and bobs to them which requires a (heartstopping) reboot
occasionally. My Windoze machines run for a maximum of a month before
something drains all the resources or system processes misbehave.

Dom

2008\12\08@121109 by William \Chops\ Westfield

face picon face

On Dec 8, 2008, at 7:44 AM, Virchanza wrote:

> I'm wondering, if you leave your PC on overnight, effectively having  
> it
> on 24 hours out of 24 hours, then will it only last half as long... ?

In my experience, leaving a PC on 24h/day does not "effectively"  
reduce its useful lifespan.  (this is based on perhaps hundreds or  
thousands of PCs in labs at work, mostly left on ALL the time.  PCs DO  
end up in the "eWaste" recylcing bins when they break, and I  
occasionally look through these bins.  I don't know whether they're  
broken or just thrown away because they're so old, but the CURRENT PCs  
showing up in the bins are pentium 2s and pentium 3s...

(And we just had a report from a customer of a router that has be UP  
for over ten years now...)

Leaving it on all the time may shorten the life; even cut it in half,  
but it seems that the fundamental lifetime is much longer than the  
"useful" lifetime anyway.

Your fundamental enemy is probably dust...

BillW

2008\12\08@125103 by Alan B. Pearce

face picon face
>When we did have a catastrophic power failure that the UPS's finally
>gave up on, everything shut down as it should have done but when the
>power came back on all hell broke loose. The CPU fan had seized and
>didn't start back up (machine complained loudly about that as it was
>all monitored) and 3 of the 16 case/hdd cooling fans had seized along
>with a couple of other fans that were noisy. It was only when they were
>stopped that they actually gave up. Typically I was on holiday that
>day and was called in to sort it all out.

I know that feeling. We had a batch of machines that had a particular
Micropolis full height 5 1/4" HD in them, came in two flavours, 45MB and
75MB. If the machines were left on they would run 'forever', but after some
period of use, if turned off the HD wouldn't start again. It would spin up,
and then fail because it couldn't find the servo track.

The problem turned out to be a design problem. The flexible cable that went
to the heads inside the 'clean area' had a sharp bend at the fixed end, and
so in operation cracks appeared in the conductors at this point. Provided
the machine wasn't turned off, the conductors kept making contact, and so
the drive kept working. But, when turned off and the heads moved to the
landing zone, the cable go pulled just bit tighter around the clamp post,
and this pulled the conductors enough for them to part company. Then when
powered up, it would spin up, attempt to load the heads, complain there was
no servo track, unload the heads, and spin down again.

Sometimes it was possible to convince them to run again, and once the heads
were loaded the unit would run until powered down again - but it was always
a red flag to get that drive changed NOW.

2008\12\08@125508 by Neil Cherry

flavicon
face
I leave my computers on 7x24x365 and things do die. I think it's
mostly caused by poor power quality in my area. I've had some
bad drives, memory has gone bad (tough to figure out), a mother-
board go bad and power supplies die. Usually it's the the things
with moving parts (drives, fans, etc). I find that laptops aren't
really designed for 7x24x365. None of the stuff I have is high
quality (except the laptops), no real servers just desktop systems.
I use these for my home automation/development systems/network
servers. Currently I'm testing out a 500MHz system and I'll have
my hands on a 1.6G Atom within the week. For most of what I'm running
I don't need the extra MHz but I'll be trying to move in a PVR
into the HA system in the coming months and that requires some
horse power and storage. I need to figure out whether it's two
systems or one system. I really like small narrow purpose systems
but sometimes it's good to move to a single system, sometimes it's
not.

--
Linux Home Automation         Neil Cherry       spam_OUTncherryTakeThisOuTspamlinuxha.com
http://www.linuxha.com/                         Main site
http://linuxha.blogspot.com/                    My HA Blog
Author of:            Linux Smart Homes For Dummies

2008\12\08@233918 by M. Adam Davis

face picon face
On Mon, Dec 8, 2008 at 10:44 AM, Virchanza <.....virtualKILLspamspam@spam@lavabit.com> wrote:
>
> All good things must come to an end, and every computer claps out
> eventually.

Aside from the mechanical aspects of the computer, the electrolytic
capacitors (used on the motherboard switching power supply for the
CPU/memory/northbridge voltage) are generally the ones that go first.

Electrolytic caps have a limited lifetime that's worsened by heat.
Good caps have an MTBF of 2,000 hours, but that cuts in half at the
temperatures that exist in a PC case near the CPU.

I used to have a stack of CPU and motherboards that were replaced due
to capacitor bulging/leaking and a few bags of replacements.  When I
needed another workhorse machine I'd grab the soldering gun (the
negative is usually soldered to a huge ground plane - no thermal
relief because the additional resistance/inductance wasn't good for
the switching supply) and replace all the electrolytics.  Get another
few years out of a motherboard for $20 for the nicer (higher heat,
longer lasting) electrolytics.

Nice thing about these caps is that when they failed it was obvious.
A lot of electrolytics fail without visibly changing, and there are a
lot of 'good' motherboards out there replaced due to normal component
aging.

Of course, there are other aging related failures, but modern chip
making and PCB manufacturing has limited things like copper migration
on the PCB, or semiconductor migration (and whiskers) in the chips.

Thermal cycling is a big deal with the elctrolytics, though, and often
they fail in a way that allows them to keep operating if already on
(with increased ripple, which can cause its own problems in the logic)
but they will be enough out of spec that the switcher won't start
reliably if turned off.

So... it's a toss up.  Regardless, give your PC as much cooling as
physically possible - heat is the greatest enemy for electrolytics and
mechanical components.

-Adam

--
Please rate and vote for my contest entry:
mypic32.com/web/guest/profiles?profileID=50331

2008\12\09@002913 by apptech

face
flavicon
face
> Electrolytic caps have a limited lifetime that's worsened by heat.
> Good caps have an MTBF of 2,000 hours, but that cuts in half at the
> temperatures that exist in a PC case near the CPU.

Sort of.

Electrolytic capacitor life is at  arated temperature - typically either 85C
or 105C.
Rated lifetimes are usually 1000 or 2000 hours.
Capacitor life approximately double er 10C drop in operating voltage so eg a
1000 hour cap run at 45C average can be expected to have a lifetime of ABOUT

       ~~~   1000 x 2^(85-45)/10 = 16,000 hours or about 2 years
continuous.

A 105C cap in the same environment SHOULD give an extra factor of
2^(105-85)/10 = 4x or 8 years continuous operation.

The mechanism for this is mainly electrolyte dryout. This applied to "wet"
aluminium electrolytic capacitors but not to eg solid Aluminium or solid
tantalum (should anyone be daring enough to use the latter in a power
circuit).

Interestingly, the above is for applications with rated voltage applied. Run
them at a given temperature with no voltage applied and the life time is
much shorter than with rated voltage applied.


  Russell McMahon

2008\12\09@044755 by Vitaliy

flavicon
face
Sorry for jumping in so late, I just wanted to answer the question in the
subject line.

Leaving PC on overnight definitely reduces lifespan. It contributes to an
unhealthy lifestyle (staying up late, working too much, drinking too much
coffee). Turn off your PC, and live longer.

Vitaliy

More... (looser matching)
- Last day of these posts
- In 2008 , 2009 only
- Today
- New search...