Sign in with
Sign up | Sign in
Your question

My 25 watt system consumes 50 watts - why?

Last response: in Systems
Share
June 2, 2010 10:14:58 PM

I have built a system *very* similar to the one described in the article http://www.tomshardware.com/reviews/25w-performance-pc,... but my system's power consumption is consistently around 50 watts when idle. That's *twice* as much as the article claims for their system.

My system uses an H55M-ED55 motherboard, not exactly the same as that in the article but same chipset and everything and same manufacturer, just a few 'goodies' less than the one in the article. I've used a Core I3 processor rather than Core I5 one but that should reduce the power consumption by a little. I am using exactly the specified power supply. I've tried a couple of different disk drives but they're not really significant power users anyway and only vary the power by a watt or two.

My system consumes very close to 50 watts when idle, it makes little difference whether I run Windows or Linux on it, nothing seems to reduce the power.

So why can't I get anywhere near what the article claims?
June 2, 2010 10:22:34 PM

Which parts exactly....?

It's hard to tell why your system is drawing more power without knowing what parts exactly you used?
June 3, 2010 12:08:10 AM

^ Also, how did you measure the power? KillAWatt meter?
Related resources
June 3, 2010 9:23:16 AM

fastx21 said:
Which parts exactly....?

It's hard to tell why your system is drawing more power without knowing what parts exactly you used?


I thought I had said fairly fully:-

MSI H55M-ED55 motherboard - very like the one used in the test but fewer goodies, same chipset etc.
Intel Core I3 530 processor - slightly lower power than the one in the test
Generic DVD drive - makes a trivial difference if removed anyway
500Gb WD 'Green' disk drive - using different drives makes little difference to the power consumption
FSP220-60LE power supply - Exactly the power supply used in the test
2 x 2Gb Kingston DDR3 1333Mhz memory - Same as in the test

There's nothing else


June 3, 2010 9:27:42 AM

Shadow703793 said:
^ Also, how did you measure the power? KillAWatt meter?


With a Brennenstal power meter, I've checked that it measures pretty accurately with other loads including ones with poor power factor and strange load waveforms. For example it measures the consumption of a 14 watt CFL (compact fluorescent lamp) as 14.7 watts and the consumption of a WD Mybook II NAS is almost exactly the claimed 6 watts. I've yet to find anything where the power meter doesn't give the right power in cases where I know the power consumption by other means.
June 3, 2010 3:39:06 PM

Question: How does your power, under Prime95 load, compare.

Your memory is diff - yours, 2 x 2Gb Kingston DDR3 1333Mhz: Test bed, Corsair CM3X2G1600C9DHX (1.8 V) No much difference

They do not provide (didn't catch) their definition of Idle and/or what power saving options might be turned on.

Did you verify your Voltages matched the artical's voltages.

I always take these evals with a grain of salt, as they normally do not get "Off the Self" parts. The get most of their parts straight from the Manuf, Which may, or not, be identical to retail items
June 3, 2010 4:10:16 PM

RetiredChief said:
Question: How does your power, under Prime95 load, compare.

> Your memory is diff - yours, 2 x 2Gb Kingston DDR3 1333Mhz: Test bed, Corsair CM3X2G1600C9DHX (1.8 V) No much difference

I think I did try running with only one 2Gb chip and found it made virtually no difference.

> They do not provide (didn't catch) their definition of Idle and/or what power saving options might be turned on.

There's not much to play with, they did try reducing some voltages and found it made no worthwhile difference. I'm investigating playing with some of the processor options.

> Did you verify your Voltages matched the artical's voltages.

Pretty much so, all standard.

> I always take these evals with a grain of salt, as they normally do not get "Off the Self" parts. The get most of their parts straight from the Manuf, Which > may, or not, be identical to retail items

June 3, 2010 7:11:05 PM

In Intel chips, stock voltage differs from part to part.

In the article they undervolted their CPU to 0.888V.

Did you do this with your setup? You won't get their numbers if you don't undervolt and use stock voltages.
June 3, 2010 8:13:48 PM

>
banthracis said:
In Intel chips, stock voltage differs from part to part.

> In the article they undervolted their CPU to 0.888V.

> Did you do this with your setup? You won't get their numbers if you don't undervolt and use stock voltages.


It says in the article after they decreased the CPU core voltage:-
"We measured a decrease in system power at the reduced settings, but the impact was negligible at idle. At peak load, we had a decrease from 80W to 76W total system power."

So I don't think reducing the CPU voltage has much effect, especially at idle which is what I'm interested in.
October 12, 2010 7:00:47 AM

Noting wrong with your building. The article shows 50 watts at efficieny run (which of them ...). In other review it's the same value (see system power consumption at http://www.tomshardware.com/reviews/d510mo-intel-atom,2...).

The article's title is wrong and bad written and mess it up the watt-hour's rate with the idle state comsuptiom (which I think you aim). I agreed that watt-hour is the most technical, but conclusion miss it.
December 31, 2010 11:13:00 AM

Hello all,

Did you get to lower your readings chrisisbd ?
I too, have got this exact same MB (MSI H55M-ED55 ) along with a Core i3 540 and 4Gb of Kingston Memory and an Intel SSD (no floppy or DVD). I am using 47W at iddle from the wall when I was expecting 31W as per all reviews. My PSU is PicoPSU 150W (close to 99% efficiency); but got roughly the same readings with a standard ATV PSU.

Any help appreciated; I bought this system for a 24/7 server and would love to have 25/31W rather than 47W

Cheers
!