Sign in with
Sign up | Sign in
Your question

PC using 1000 Watts

Tags:
Last response: in CPUs
Share
June 26, 2010 11:21:41 AM

I was wondering if some one could help me please
I have just upgraded from a socket 939 3500+
to a Phenom ii x4 940 on a Ga-ma74gm-s2.
With a (just one)8600gt 256meg.
When I first put this together it ran fine and I rang MS to reactivate Win7
then I ran a old nvidia demo (dawn) and my PC shut down and would not
power on again, so I have brought a new 650 Watt PSU.
All is fine not BUT using a Wattage meter the sort that clamps to the electric meter cables I am drawing as much as 1.25k Watts when running Planesoft Western railway screen saver.
But averages out at about 975 Watts so I am expecting this PSU to fail any time soon
in Windows I am using an average of 440 Watts peaking at around 675 Watts when accessing my drives.
I had no idea I was going to dragging so much power??
is this normal?
If this is the case I wish I had not transfer d my windows key.
I brought the moby and chip from a local computer fair.

Thanks in advance.
Andrew

More about : 1000 watts

a c 203 à CPUs
June 26, 2010 11:39:11 AM

Something is wrong with that wattage meter.
Phenom II x4 940 and 8600GT will be using less than 300watts of power, probably under 200watts at idle.
a c 203 à CPUs
June 26, 2010 11:48:24 AM

Taking a closer look using a THG article: AMD Phenom II X4: 45nm Benchmarked
X4 940 in 790GX motherboard with GeForce GTX 280 used about 115watts at idle.
Related resources
a b à CPUs
June 26, 2010 4:00:39 PM

Thats impossible. You should measure your current and voltage also. Post those up. It is very unlikely that you are pulling that much power especially with your hardware.
June 26, 2010 4:09:28 PM

Use a clamp Ammeter to measure the current. You can measure the current on the PC AC input. That will provide you with the over all consumption of the PC.

If you use a DC Clamp Ammeter you can measure the actual current draw for each rail (12V, 5V, and 3V). With that you can calculate the power consumption of the load.

a b à CPUs
June 26, 2010 8:25:22 PM

leon2006
Using the clamp-on Amp meter is great Idea, unfortunately not very practical. Not enough room to put a clamp around say the +5 (or +12 V) line going into the 20/24 Pin MB connector.

Blackhawk
Measuring Current and voltage will give reactive power, but not true power unless the Power factor is known, or equals 1. I did make a interconnect power cord where I cut the +120 line so that I can use a DVM for current, or a clamp-on meter probe (Goes to my O’scope).

Concur – Reading is not correct, or being performed incorrect. I would prefer to get one of the cheap Watt Meters, plug it into the wall and the computer into the wattmeter.
a b à CPUs
June 26, 2010 8:42:30 PM

i heard somewhere that some power supplies have to draw much more than teh system needs to provide the necessary current, eg to power a pc drawing 400watts in dc , the psu will pull 850watts at the outlet, this is where 80+ gold/silver/bronze etc is applied
a b à CPUs
June 26, 2010 8:53:38 PM

850 W with 400 Out would = 47% eff Not seen one that bad yet at least that lasted more than 30 Sec. (PS some of the OLD Series Regulated PSUs were down around 50 %, But provided better power regulation and less ripple)

Even a LOW end 70 % would be only be 570 Watts to supply 400 Watts.

Typical is 80% being 500 Watts to provide 400 Watts, With the better ones even less.

a b à CPUs
June 26, 2010 10:38:07 PM

^Hey Mr. Smart, thanks for lesson...xaira was using it as an example not real life also.

"Measuring Current and voltage will give reactive power, but not true power unless the Power factor is known, or equals 1. I did make a interconnect power cord where I cut the +120 line so that I can use a DVM for current, or a clamp-on meter probe (Goes to my O’scope). "

^Never knew that...thanks for the amazing lesson.
June 27, 2010 10:26:02 AM

If you are using a meter that measure the current by reading the magnetic field (i.e. the ones that clamp over), then it only measures the current and not the voltage, although we know roughly what the voltage is so can work it out from that.

So either the meter thinks you are on ~240V when really you are ~120V, which would mess up the readings. Or you have clamped it over both the neutral and live, when you should only clamp it over one or the other.

Basically sounds like you're not using the meter correctly.
a c 87 à CPUs
June 27, 2010 10:31:38 AM

I must be missing something...

Quote:
using a Wattage meter the sort that clamps to the electric meter cables I am drawing as much as 1.25k Watts when running Planesoft Western railway screen saver. But averages out at about 975 Watts


So let me get this straight. Your measuring your computers power draw by measuring from your main line? Assuming nothing else kicked in, fridge or AC, then even then its not that bad. 1.25kW - 975 = 275W? Theres no way your system draws 1kW, so I'm probably not understanding something. 300W might be possible. Please rewrite whats going on if what I wrote is wrong. I do know there isn't really a way for a 650W PSU to pull 1.25kW from the wall. Even if the OCP wasn't working I'm sure it would fry if it tried.
June 28, 2010 5:13:16 PM

Thanks for the answers
yea I am as miffed about as you all are, the meter was supplied by npower (Smart power)its the sort that clamps onto the live cable from the electric meter and the consumer fusebox 220~240 volts
I have tested the watt meter using an electric fire with 3x 400Watt halogen tubes and shows accurate

I have today swapped out the 650watt PSU for my wifes 450watt while she was out and although its a little less it still gets to 1.1kwatts while running the julia Cpu benchmark in Everest with all the cores running at 100%.
the PSU is not getting over heated.
Idle the pc drops to around 495 Watts and while typing this into open office its running at 820watts
taking off the 125watts my Monitor and modem + router is burning.
And 650watts at post then drops when windows loads.
Nothing else is switched on, fridge was turned off during the tests.
My old system using the same 8600GT ran flat out at about 450 but mostly about 395Watts
and around 185 while surfing the net.
I really am hoping its a glitch on the wattage meter or I will be reluctant to use my pc.
July 14, 2010 1:50:25 PM

Thanks every one for you answers.
i have the answer and a further question i would like to ask.
Firstly the answer, i contacted Gigabyte about my problem and after they looked at my specs told me that the Phenom II 940 was not supported by my revision of board.
also on another part of Toms hardware i read that i could take the cpu core voltage down to 1.20 volts.
which i have done and i have dropped the clock speed from 3.0gig to 2.2.
effectivly dropped my CPU consumption from 125watts to 95Watts i am told.
On test my Pc is now drawing 280watts idle and up to 550watts flat out.
thats half what it was before.
Now for my question, in your opinions could i expect a lower power consumption with a AM2+ or AM3 motherboard?
Thanks in advance
Andrew

a b à CPUs
July 14, 2010 2:45:15 PM

boldpatch said:

On test my Pc is now drawing 280watts idle and up to 550watts flat out.
thats half what it was before.

You would have to have an insane amount of peripherals connected, that you aren't telling us about, in order to come close to those numbers.
Check my config, http://www.tomshardware.com/forum/configuration.php?con...
Only draws 155w DC idle and 450 watts DC "flat out" according to the P-Tuner software that came with my psu, roughly 190w AC idle and 540w AC under load according to my power meter, the components you have listed won't come anywhere near that.
July 14, 2010 2:59:45 PM

I know its getting me down to be honest, no i only have a two hard drives three fans a 8600Gt.
the wattage meter i use is attached to the live wire from my electric to the consumer box.
i have taken it off and modified the power lead to my PC and connected it so it only reads the PC base unit.
i know it sounds insane.
i have tested the wattage meter on other things like our halogen heater with 400 watt tubes.
nothing is getting hotter than it did with my older 3500+ chip and motherboard.
friends laugh until they see it for them selfs.
before i down the core volate and using Everest ultimate edition runing the Julia Floating point cpu test it took it upto 1.2K watts from the wall, but now the same test is 550watt from the wall.
July 20, 2010 9:50:27 AM

Hi everyone
This is embarrassing to say the least, the wall socket type energy monitor came this morning from ebay, I quickly set it up between my Pc base unit and switched on.
Its reads 81 watts at post and fluctuates between 66 ~ 75 watts idle and 81 ~ 86 watts flat out.
I had dropped my CPU core voltage down to 1.20volts in a bid to get the old meter reading down.
But even though this meter shows the old reading to be wrong I am going to leave it at 1.20 volts unless I get problems.
Funny testing every other appliance shows the old meter to be accurate apart from my Pc and a just one table lamp with an energy saving bulb which also showed a false reading on the old meter 65 watts. So its a far cry from the old reading of 1k watts.
Sorry every one but I am glad it was a false alarm I can enjoy my computer now thanks again for your help.
!