top end pc's.....power usage....suprising

Millwall.FC · 31 May 2007 at 23:16

Jokester said:
Ahh I see what he means, if you want to work out the actual power used by the PC just multiply PSU power by it's effiency, typically 70-80%, but it'll vary with power demand.

Jokester

it is.

james.miller · 31 May 2007 at 23:16

Jokester said:
Ahh I see what he means, if you want to work out the actual power used by the PC just multiply PSU power by it's efficiency, typically 70-80%, but it'll vary with power demand.

Jokester

thats right. according to JohnnyGuru the 620w corsair is roughly 82-83% efficient @ with a ~380w draw. for argument's sake, if that 384w load was the pc alone that would mean the components were actually drawing around 318 watts.

of course that isnt the case, and if the draw form the mains is closer to 350w, the draw of the components could be as low as 290w. Again i will test all of this very soon

Millwall.FC · 31 May 2007 at 23:19

james.miller said:
what difference does that make?

nvm

semi-pro waster · 31 May 2007 at 23:19

james.miller said:
thats right. according to JohnnyGuru the 620w corsair is roughly 83% efficient @ with a ~380w draw. for argument's skae, if that 384w load was the pc alone that would mean the components were actually drawing around 318 watts.

Nah, I'm taking this result, right now. It means I was almost bang on the money.

It would be interesting to know what it is without any other peripherals connected though.

james.miller · 31 May 2007 at 23:22

Millwall.FC said:
you said "as I ran superpi on one core and 3dmark's nature tests on the other, the maximum peak was 1.6amps, 384 watts. that's it, thats all it could muster lol. This is only a short test but if nothing else it does prove that recommended ratings and 90% of what people will tell you you *need* on a hardware forum.....is complete tosh"
but you only measured the draw of the psu not the actualy wattage the pc is using.

forgetting what else was plugged in, i measured what the psu was pulling from the wall socket. that IS what the pc is using. All the components, the inefficiency of the psu also...everything

Jokester · 31 May 2007 at 23:22

Millwall.FC said:
but you only measured the draw of the psu not the actualy wattage the pc is using.

Yes, but the actual PC usage will be even less, as the total power used is the PC power + the power wasted in the PSU (where that other ~20% goes).

The current draw that was measured gives the power used by the PC, the power wasted by the PSU and the power used by the monitor, router and modem.

Jokester

Millwall.FC · 31 May 2007 at 23:35

the next test to do is a insulation resistance test

Jokester · 31 May 2007 at 23:37

Yeah I want to see a Megger test done

.

Jokester

Tee Hee Johnson · 1 Jun 2007 at 00:52

james.miller said:
so what do you think then? and well done gt_junkie and Craig_kennedy83 for guessing virtually on the button

woohooo!

I won at somethingfor the first time ever

lol kiddin!!

Thats really suprising due to how high of a power supply that can be recommended most of the time...

Thinkin about it, seems weird to think that a 400w would do that system... :|

Craig!!

Tom|Nbk · 1 Jun 2007 at 01:08

Im buying one of these wall read out displays so will also do some testing in a few days

Cupra · 1 Jun 2007 at 02:47

goes to show its all about the rails

we need to get 600w drawn from a psu and see what happens then

as the meter could be out

Cyber-Mav · 1 Jun 2007 at 03:51

james.miller said:
not as far as i could see. unfortunately the equipment didnt measure in watts so the best accuracy we had was to withing 0.1amps, 24watts basically. even given that, the total draw was still WELL below with would officially be recommended for the pc. given that the 384w (1.6amps, although in reality it could have been +/-12w of that figure) figure was for everything connected to that strip, its not unfiar to say the draw of the pc could have been closer to 360w or less.

basically looks like a psu needs to be able to supply 380w on the 12v rail for most pcs.

this just confirms that something like a corsair hx520 is plenty for a fully loaded rig, possibly even sli.

Energize · 1 Jun 2007 at 04:02

I knew it would be low, it's ridiculous to see people spending £70+ on 600W+ psus for specs even lower than that when even psu calculators say it's double what they need. Thats why I got a 430W seasonic for £40.

Dravic · 1 Jun 2007 at 04:08

Since my last decent one blew I've been running on a £20 no name PSU. :eek:

I've been amazed that it coped, even for a couple of months.

The Halk · 1 Jun 2007 at 23:41

james.miller said:
thats right. according to JohnnyGuru the 620w corsair is roughly 82-83% efficient @ with a ~380w draw. for argument's sake, if that 384w load was the pc alone that would mean the components were actually drawing around 318 watts.

of course that isnt the case, and if the draw form the mains is closer to 350w, the draw of the components could be as low as 290w. Again i will test all of this very soon

It's a little bit more efficient than that. Around 85-87% at that load. 230V gives more efficient results than 110V.

james.miller · 1 Jun 2007 at 23:46

that was tested at 240v. there's no way it was near 90% efficiant, that would make it opne of the most efficient switch mode power supplies in the world lol

The Halk · 3 Jun 2007 at 06:37

james.miller said:
that was tested at 240v. there's no way it was near 90% efficiant, that would make it opne of the most efficient switch mode power supplies in the world lol

http://www.silentpcreview.com/article692-page4.html

First look at the very bottom table, this shows efficiency at 470W output at various voltages. The most pertinant is the 86.2% at 246V.

Then look at the second table on the page. This table shows the 620W Corsair efficiency at various output levels, at 117V. The closest figure to the 470W output is the 519.1W output, which has an efficiency of 81.9%. Also note the efficiency at 407.1W is 83.9%. The efficiency at 117V for 470W would be between these figures, closer to the 519.1W output. Putting it about 83%. Finally look at the 299.8W efficiency, 84.7.

If it's 83% efficient at 470W at 117V, and 86.2% at 246V.
And it's 84.7% efficient at 300W at 117V, then the efficiency at 240V would lie between 85 and 87%.

Tom|Nbk · 4 Jun 2007 at 15:57

Just recieved my power meter, have done a quick test, discovered that my PC tower uses just over 300w when Im stress testing with Orothos. This is fairly suprising, specs in sig and I will do more testing soon.

GUIClaw · 4 Jun 2007 at 17:50

I know it's not entirely in line with desktop pc's but i've got one of these meters too so I decided to test what my Server and Router draw, they're in my garage so they happened to be really easy to test..

The server is a 2500+/2gb ram with an age-old Enermax 400W PSU with 5 harddrives if I remember correctly, which i've got set to turn off with 30mins inactivity for powersaving, and the router is a 2400+ with 512mb ram and a 40gb hdd (running smoothwall if you're interested

) powered from a noname peice of crap (i think it's around a 250w psu)

When booting the router on it's own it spikes at about 0.5amps (115W) and then settles down to about 0.2amps (46W). When booting the server it spikes at about 1.5amps (345W) due to all the harddrives going clunkclunk

but I watched it as the hdds span down and the total load from the router and server stays at about 230W constant

Also I had a netgear 5port gigabit switch plugged in when testing this, but that didn't even register on the meter (it's a meter that I built myself with a current meter, voltmeter and I dunno why but I added a frequency meter too, all housed in the biggest plastic box I could find :rolleyes:

aha but it's interesting to see how much the national grid fluxuates - it was at about 60Hz the other day and i've seen the voltage go from about 220-240 too, but I bet that's all you guys turning on your power vacuums you call computers

)