«

Jean-Pierre Gauthier wrote:
> On the first day of school, my grand-daughter's 9th grade computer
> teacher told the class that it was better not to turn off a computer.
>
> Sounds odd to this semi-literate computer whizz.....
>
> What would be the benefit?/drawbacks?
>
> Thank you,
>
> Jean-Pierre Gauthier

I've assembled a couple of hundred computers over the past
several years, and they're *all* turned on and off
regularly, some of them several times a day. I know most of
the owners personally, most of them first-time owners. I've
kept track of nearly every computer, most of which are used
and abused by multiple users. About half of them do not use
UPSes even though power fluctuations, brown-outs, sudden
and gradual outages are common here. Failure statistics are -

About 10 PSUs - I had to use cheap ones for some time as it
used to be impossible to get decent ones in my area. Some
died during use and some at turn-on. None of the better
ones have failed yet.

One hard disk - the owner admitted that his nephew had been
constantly tinkering with the innards of his computer. (I
use Seagate exclusively).

Four motherboards - one of them was the one with the failed
HDD. Another was struck by lightning via the LAN port. Two
from that infamous batch of bad Taiwanese capacitors.

One CPU, an early Athlon XP - the heatsink fins were choked
with dust and must have been overheating regularly.

Four or five optical drives - a mix of normal wear and tear
and abuse.

One monitor - the result of heavy power surge that also
destroyed the PSU. The main filter capacitors had vented
and were leaking badly.

Three RAM sticks - all of them no-name products that I was
forced to use at the time.

Sorry about the long-winded narrative. My assessment is
that while power-on surge contributes something to the
failure rate, especially with low-grade components, it
doesn't make a big statistical difference with good quality
stuff. You can draw your own conclusion. I don't have
comparable data for 24/7 machines.

While I don't have numerical statistics, my impression over
the years corroborates another poster's that machines that
are normally left on all the time somehow tend to fail when
they're turned off and then on again, more so than those
that regularly go through the on-off cycle.

My advice to customers is to leave the computer on for
short periods of unuse such as at dinner time, and turn it
off for longer idle periods such as when they're retiring
for the night or going off to work.

Mitch <> wrote:

>And just WHAT is a power-up surge?
>Power is regulated -- this is NOT analogous to a water pipe with too
>much pressure.

When does a light bulb burn out? Approx 99 and 44/100% of the time it
is when you turn it on, not while it is lighted. Why? Because of
the initial power surge and the thermal stress of being heated.

When does the motor in a refigerator or a washing machine fail? Again
in the vast majority of instances it is when they are turned on (e.g.
when the refrigerator thermostat switches on the electric motor that
runs the compressor? Why? Because all electric motors are under the
greatest stress, and therefore most likely to fail, when they are
first started up.

Electronic components, including the CPU and RAM, are also subject to
thermal stress when power is first applied, just like a light bulb.
And hard drives use an electric motor, and they are subject to same
initial startup stresses as any other electric motor.

Ron Martell Duncan B.C. Canada
--
Microsoft MVP (1997 - 2006)
On-Line Help Computer Service
http://onlinehelp.bc.ca
Syberfix Remote Computer Repair

"Anyone who thinks that they are too small to make a difference
has never been in bed with a mosquito."

Mitch mumbled incoherently in 24hoursupport.helpdesk:

snip

> And a light bulb doesn't have a sophisticated power-management
> component.

Ah, but those "sophisticated power management components" are themselves
subject to thermal stress when power is initially applied.

--
The Old Sourdough
I pretend to work. They pretend to pay me.

In article <Xns983334B35CF8C1258ggbeerisgood@216.196.97.131 >, The Old
Sourdough <> wrote:

> Ah, but those "sophisticated power management components" are themselves
> subject to thermal stress when power is initially applied.

Thermal, sure -- but the suggestion was that simply having power
running through the device was a problem. As one poster wrote, the
'jolt' of electricity is a damaging factor.

In article <>, Ron Martell
<> wrote:

> When does a light bulb burn out? Approx 99 and 44/100% of the time it
> is when you turn it on, not while it is lighted. Why? Because of
> the initial power surge and the thermal stress of being heated.
>
> When does the motor in a refigerator or a washing machine fail? Again
> in the vast majority of instances it is when they are turned on (e.g.
> when the refrigerator thermostat switches on the electric motor that
> runs the compressor? Why? Because all electric motors are under the
> greatest stress, and therefore most likely to fail, when they are
> first started up.
>
> Electronic components, including the CPU and RAM, are also subject to
> thermal stress when power is first applied, just like a light bulb.
> And hard drives use an electric motor, and they are subject to same
> initial startup stresses as any other electric motor.


That's nicely put -- but does it really have to mean that it is that
time which causes the wear?
It would be clear from that info that the change of state from inactive
to active is a time of enough stress to make failure more common. Is
that necessarily also the time when the wear happens?

I'm suggesting it isn't. I'm suggesting that long hours running makes a
bigger difference. And I'm suggesting that when you aren't getting use
from the computer, there is no reason to leave it on.

I have to wonder why people would argue against turning it off -- the
most natural and normal thing to do. Is it laziness, do you think, or
some desire to have machinery at your service?
To them -- why not just turn it off?

This will be the reason for conflict between Ron Martell and me. Ron
assumes facts only from observations. He violates basic science
principles as taught in Junior High School science. He knows without
first learning the underlying theory and therefore has no logical
hypothesis. He then selectively samples data - his observations. From
only watching a light bulb when it is turned on and by only using
observations; somehow he knows. IOW Ron is promoting classic junk
science reasoning.

Learn from experts on light bulbs. Power cycling light bulbs is not
destructive. If it were, then the yellow traffic light bulb that
flashes all night long would be the first bulb to fail. It hardly ever
fails. This because of facts well documented by light bulb
manufacturers even in the IES Handbook - an industry bible. Notice
that well respected sources - not just observations - are also cited.
Light bulb life expectancy is a function of hours of operation. So
well defined that life expectancy is even defined in math formulas.

Reality takes longer to explain. Junk science reasoning cannot be
bothered to learn so much before somehow knowing. Therefore Ron says:
> When does a light bulb burn out? ... Because of the initial
> power surge and the thermal stress of being heated.

That is a lie. His conclusion is so incorrect and so based only on
selective observation as to be a lie.

Leaving computers on does not [prolong life expectancy. This 'leave
it on' nonsense comes from people like Ron who do not first learn
science. His reasoning - microprocessors thermal cycling. Yes they
change tens of degrees. But when manufactured, they are thermal cycled
500 and 900 degrees - and that is not destructive. Why then are tens
of degrees so destructive? Junk Science reasoning. Ron simply
speculates; then converts that speculation into facts without doing
what was even taught in Junior High science.

Turn the computer off at the end of the day. Power cycle like a
radio, TV, or other electronic appliance. Manufacturer specifications
do not support this nonsense damage due to power cycling.

Unlike Ron who somehow knows, I come from an industry where we
located failed parts, replaced it, and discovered why that part failed.
Just did another household appliance that failed when powered. Of
course. Due to long hours of operation, that appliance reset circuit
had failed. Reset circuit only does something useful during power on.
But the part is under greatest stress only during normal operation.
The part failed due to too many hours of operation. Therefore when
the appliance restarted, that reset circuit was not functional; would
not restart the appliance. Ron would instead blame power cycling
because he only observes; does not first learn what has failed and why.

In another case, a pullup resistor in a computer power supply only
does something useful when power supply restarts. That resistor also
failed due to too many hours of operation. And that failure only
appeared during its only useful function - restart. Ron again would
have blamed power cycling when too many hours of operation causes the
part failure.

In each case, only using observation, Ron would declare power up as
destructive. And yet in each case by first learning facts, we discover
hours of operation caused failure. IOW we learn by using what was
taught even in school science.

Light bulbs fail on power on due to too many hours of operation.
Just like those other examples, Ron used only observation and junk
science reasoning to proclaim power on as destructive to light bulbs.
Junk science reasoning is alive and well when others only 'assume'
power cycling is destructive. They don't know. They have observation
that even violates what was taught in junior high science. They know
and need not bother to get a technical education, read manufacturer
datasheets, repair only to learn why failures happen, learn the
underlying theory of why components degrade, etc. Leaving computers on
- the mythical power on surge - is classic junk science promoted by the
same reasoning that promoted Geritol and smoking cigarettes for better
health. Power cycling damage to household appliances is a classic junk
science myth. Turn it off when done.

Ron Martell wrote:
> When does a light bulb burn out? Approx 99 and 44/100% of the time it
> is when you turn it on, not while it is lighted. Why? Because of
> the initial power surge and the thermal stress of being heated.
>
> When does the motor in a refigerator or a washing machine fail? Again
> in the vast majority of instances it is when they are turned on (e.g.
> when the refrigerator thermostat switches on the electric motor that
> runs the compressor? Why? Because all electric motors are under the
> greatest stress, and therefore most likely to fail, when they are
> first started up.
>
> Electronic components, including the CPU and RAM, are also subject to
> thermal stress when power is first applied, just like a light bulb.
> And hard drives use an electric motor, and they are subject to same
> initial startup stresses as any other electric motor.

"Mitch" <> wrote in message
news:030920060313043252%...
> In article <>, Ron Martell
> <> wrote:
>
Snipped (what all of it ?.) <g>.

Just done a quick calc. for everyone Mitch.
Say someone uses their machine for 12 hrs a day.
Say they don`t use it for 12 hrs a day, but leave it
switched on.
My maths tells me that 12x365 =4380 hours.

So their machine runs for a wasted 4380 hours a year.
I`d rather have my electric bill <g>.

"w_tom" <> wrote in message
news: ups.com...
> This will be the reason for conflict between Ron Martell and me. Ron
> assumes facts only from observations. He violates basic science
> principles as taught in Junior High School science. He knows without
> first learning the underlying theory and therefore has no logical
> hypothesis. He then selectively samples data - his observations. From
> only watching a light bulb when it is turned on and by only using
> observations; somehow he knows. IOW Ron is promoting classic junk
> science reasoning.
>
> Learn from experts on light bulbs. Power cycling light bulbs is not
> destructive. If it were, then the yellow traffic light bulb that
> flashes all night long would be the first bulb to fail. It hardly ever
> fails. This because of facts well documented by light bulb
> manufacturers even in the IES Handbook - an industry bible. Notice
> that well respected sources - not just observations - are also cited.
> Light bulb life expectancy is a function of hours of operation. So
> well defined that life expectancy is even defined in math formulas.


Some years ago I designed a simple lamp flasher for a shop display and was
forced to redesign to keep a residual current through the filament -
precisely because power cycling IS destructive to lamp filaments! By passing
a small current to keep the filament partly heated thermal shock was
avoided. Also filaments are coiled and often "coiled coil" construction, as
the filament wire has a pronounced positive temperature coefficient the turn
on current is somewhere around 12x its normal running current - and because
the filament wire is coiled the turn on surge also causes a burst of
alternating magnetic flux which causes the filament to contract in sympathy,
adding mechanical wear to the equation.

Go back and tell your so called lighting 'experts' that!!!!!

Mitch <> wrote in
news:030920060313043252%:

> I have to wonder why people would argue against turning it
> off -- the most natural and normal thing to do. Is it
> laziness, do you think, or some desire to have machinery at
> your service? To them -- why not just turn it off?

So why did you start this insane thread?
Shut it off and leave us the **** alone.

"Mitch" <> wrote in message
news:030920060301130597%...
> In article <Xns983334B35CF8C1258ggbeerisgood@216.196.97.131 >, The Old
> Sourdough <> wrote:
>
>> Ah, but those "sophisticated power management components" are themselves
>> subject to thermal stress when power is initially applied.
>
> Thermal, sure -- but the suggestion was that simply having power
> running through the device was a problem. As one poster wrote, the
> 'jolt' of electricity is a damaging factor.

Most switch mode power supplies of the type used in PCs & monitors have a
NTC thermistor in series with the internal fuse. At room temp the NTC has a
resistance often as high as a few thousand Ohms, as 230V is applied to such
a high resistance very little current flows - but enough to heat the
thermistor, being NTC the thermistor's resistance then falls and as less
voltage is dropped across it nearly the full 230V is applied to the SMPSU.

The degauss coil for a CRT display is exactly the opposite - using a PTC
thermistor, here the thermistor is in series with a degauss coil across the
mains input. This produces a large burst of alternating magnetic field
around the CRT which decays slowly to demagnetise it.

The most common argument seems to be that as long as the hard drive is
spinning the heads never touch (and cause wear to) the disk surface.