Does Win7 SP1 have a power meter?

[Peter Jason]

Is there any way to find the total power
consumption of a computer from within Windows.

I have a small power meter at the wall plug that
reads 186Watt constantly
http://www.radioparts.com.au/images/11900001.jpg
and I wonder how accurate it is.

I have 4 HDDs going all the time with a 1000W
power supply, and 4 large cooling fans. Also a
23" color monitor.

Peter

 

[DevilsPGD]

In the last episode of <[email protected]>,

Is there any way to find the total power
consumption of a computer from within Windows.

I have a small power meter at the wall plug that
reads 186Watt constantly
http://www.radioparts.com.au/images/11900001.jpg
and I wonder how accurate it is.

I have 4 HDDs going all the time with a 1000W
power supply, and 4 large cooling fans. Also a
23" color monitor.

There's really no way for software to identify power consumption to any
reasonable degree of accuracy beyond guesswork. Even if all computer
components reported their power use (they don't), you wouldn't be able
to get useful reasons from fans or the PSU itself, or take PSU or
component efficiency or signal losses into account.

Usually the best way to determine if a meter is accurate is to get two.
If they give you the same results, they're probably both right, if not,
at least one of them is wrong but there's no way to know which.

 

[Paul]

Is there any way to find the total power
consumption of a computer from within Windows.

I have a small power meter at the wall plug that
reads 186Watt constantly
http://www.radioparts.com.au/images/11900001.jpg
and I wonder how accurate it is.

I have 4 HDDs going all the time with a 1000W
power supply, and 4 large cooling fans. Also a
23" color monitor.

Peter

Someone used a Kill-A-Watt type meter, comparing it
to a proper analog power meter, here.

http://www.pupman.com/listarchives/2005/Aug/msg00936.html

This is the analog power meter used.

http://scopeboy.com/tesla/drsstc/experiment/meterundertest.jpg

And this is the Kill-A-Watt style measurement device used.

http://www.maplin.co.uk/images/Full/38343i0.jpg

The tester got 750W on the analog meter for the device
under test, and 746W measured on the Kill-A-Watt.

"Prodigit meter read 746w, 1140VA, 6A RMS, pf=0.62"

So that's good agreement in my book.

A couple other web pages I checked, they didn't get
that good an agreement. But at least the Kill-A-Watt
shouldn't give completely erroneous measurements,
like some of my other meters here have done. It's
especially difficult to measure a non-PFC ATX supply
on a sleeping PC. The current waveform is terrible,
and upsets my ammeter (claims enough amps to run
the PC in the waking state). When we know the
sleeping PC is drawing about 10-20 watts.
The Kill-A-Watt should do a better job of
making such measurements, and get closer
to the true number than I was able to.

Page 8 and 9 here, show what the current draw
waveform looks like on the three "flavors" of ATX
power supplies. Some measurement methods, don't
like this at all. (Need TrueRMS and wide bandwidth
measurement device.) The one with Active PFC, should
be especially easy to measure accurately, even
with my collection of meters. That's because the
current waveform shape, is very close to a sinusoid.
It's only got a couple "bumps" in the waveform.

http://www.xbitlabs.com/articles/cases/display/psu-methodology_8.html

*******

Your power draw has nothing to do with the "1000W" rating
of the supply. The power used is proportional to the
electrical loads inside the PC. Each additional hard drive
might draw 7W to 10W, for example. If you have big stinking
gamer video cards, the older models of those (like 8800GTX),
can gulp down power, even when idle. Some of the newer
video cards, when you aren't gaming, they turn off
lots of stuff, or turn down the clock to quite a low
operating frequency. And that helps with idle power.
The best modern video card to date, draws ~3W at idle.
In ten year old PCs, the video card could be
drawing 35W at idle.

My "best" PC here, would draw about 60W idle (VIA Chipset).
Some of my older equipment, like a ten year old PC,
might idle at 150W, and not really have all that
impressive hardware inside the box (not a gamer
class machine). This is why it's not a good idea, to
use that ten year old PC as a "router box". It's
a power pig.

Lots of equipment, is still designed carelessly with
respect to saving power. A lot of the networking
equipment is bad for that sort of thing. You can
get your PC down to 60W, and have a "tree" of wall
adapters for the networking boxes, wasting a
total of 100W. That's why my "tree" is on a power
strip with a switch, and when I'm done at the end
of the day, the whole tree is switched off.

Paul

 

[John McGaw]

Is there any way to find the total power
consumption of a computer from within Windows.

I have a small power meter at the wall plug that
reads 186Watt constantly
http://www.radioparts.com.au/images/11900001.jpg
and I wonder how accurate it is.

I have 4 HDDs going all the time with a 1000W
power supply, and 4 large cooling fans. Also a
23" color monitor.

Peter

Sometimes things are exactly as they appear to be. 186W doesn't seem to be
'wrong' for a small modern computer which isn't being forced to work hard.
The hard drives and fans aren't consuming enough power to worry about in
the overall scheme of things. If you want to see if the power consumption
changes try running a program like Prime95 to force the CPU into some
serious work. Run almost any modern FPS game to force the GPU to work hard.
If you have a really good system then you might do both of these at the
same time. If you have a 1kW PS then you must have some sort of serious
computing power -- my most powerful computer has only a 450W and it is
certainly no slouch.

The small system I'm typing on right now (i7-920, 6gB, 3tB HD, SSD, two 23"
LCDs) is consuming 207 watts but much of that is being forced by the BOINC
client running 4 CPU and 1 GPU threads. If I tell BOINC to sleep the power
drops to 135W within seconds (and the cooling fan slows noticeably).

 

[Bert]

In DevilsPGD

Usually the best way to determine if a meter is accurate is to get
two. If they give you the same results, they're probably both right,
if not, at least one of them is wrong but there's no way to know
which.

A man with two clocks never knows what time it is.

 

[Jon Danniken]

In DevilsPGD


A man with two clocks never knows what time it is.

If you stop one of them, you will at least have the correct time twice
per day.

Jon

 

[John McGaw]

If you stop one of them, you will at least have the correct time twice
per day.

Jon

But if you stop both of them will the accuracy be twice as good?

 

[Peter Jason]

Someone used a Kill-A-Watt type meter, comparing it
to a proper analog power meter, here.

http://www.pupman.com/listarchives/2005/Aug/msg00936.html

This is the analog power meter used.

http://scopeboy.com/tesla/drsstc/experiment/meterundertest.jpg

And this is the Kill-A-Watt style measurement device used.

http://www.maplin.co.uk/images/Full/38343i0.jpg

The tester got 750W on the analog meter for the device
under test, and 746W measured on the Kill-A-Watt.

"Prodigit meter read 746w, 1140VA, 6A RMS, pf=0.62"

So that's good agreement in my book.

A couple other web pages I checked, they didn't get
that good an agreement. But at least the Kill-A-Watt
shouldn't give completely erroneous measurements,
like some of my other meters here have done. It's
especially difficult to measure a non-PFC ATX supply
on a sleeping PC. The current waveform is terrible,
and upsets my ammeter (claims enough amps to run
the PC in the waking state). When we know the
sleeping PC is drawing about 10-20 watts.
The Kill-A-Watt should do a better job of
making such measurements, and get closer
to the true number than I was able to.

Page 8 and 9 here, show what the current draw
waveform looks like on the three "flavors" of ATX
power supplies. Some measurement methods, don't
like this at all. (Need TrueRMS and wide bandwidth
measurement device.) The one with Active PFC, should
be especially easy to measure accurately, even
with my collection of meters. That's because the
current waveform shape, is very close to a sinusoid.
It's only got a couple "bumps" in the waveform.

http://www.xbitlabs.com/articles/cases/display/psu-methodology_8.html

*******

Your power draw has nothing to do with the "1000W" rating
of the supply. The power used is proportional to the
electrical loads inside the PC. Each additional hard drive
might draw 7W to 10W, for example. If you have big stinking
gamer video cards, the older models of those (like 8800GTX),
can gulp down power, even when idle. Some of the newer
video cards, when you aren't gaming, they turn off
lots of stuff, or turn down the clock to quite a low
operating frequency. And that helps with idle power.
The best modern video card to date, draws ~3W at idle.
In ten year old PCs, the video card could be
drawing 35W at idle.

My "best" PC here, would draw about 60W idle (VIA Chipset).
Some of my older equipment, like a ten year old PC,
might idle at 150W, and not really have all that
impressive hardware inside the box (not a gamer
class machine). This is why it's not a good idea, to
use that ten year old PC as a "router box". It's
a power pig.

Lots of equipment, is still designed carelessly with
respect to saving power. A lot of the networking
equipment is bad for that sort of thing. You can
get your PC down to 60W, and have a "tree" of wall
adapters for the networking boxes, wasting a
total of 100W. That's why my "tree" is on a power
strip with a switch, and when I'm done at the end
of the day, the whole tree is switched off.

Paul

I am thinking of an el-cheapo tablet computer for
off-peak times using an USB as a HDD because this
will cut the use of the main computer to 1/3rd.

 

[miso]

Actually that seems a little low if it includes the power used by the
monitor.

 

[John McGaw]

Actually that seems a little low if it includes the power used by the monitor.

The power figure does include one of the two monitors, the primary, but
they are quite efficient and are run at relatively low intensity because of
sensitive eyes. Consumption is 20W. The computer itself is a Shuttle SX58H7
which is quite economical with power consumption and the graphics card is
an earlier generation ATI Radeon HD 5700 series which is also relatively
low power, at least when compared to current video equipment.

Considering that I have six computers running 24 X 7 here (five supporting
World Community Grid plus one server) they still make quite a contribution
to home heating despite my attempts to keep the power consumption of each
down. I don't even like to _think_ about what damage they do to the air
conditioning bill during the summer.

 

[miso]

I was referring to the original poster's system.

 

[Peter Jason]

I was referring to the original poster's system.

I fear the screen runs at 50W, though it turns
itself off when not used.