PC power consumption

Message

#1 Post by **K. Eng** » Sat Mar 10, 2007 3:47 pm

I used a device called kill-a-watt to measure the power consumption on my desktop computer. Basically, you use this device by plugging it into a wall socket, and plugging your computer or other appliance into the device's socket. The kill-a-watt will measure the power flowing through the wall socket.

http://www.p3international.com/products ... 00-CE.html

Specs are:
1.47 GHz Athlon (Barton, 512kb L2, 266 MHz FSB, 1.38Vcore), nForce2 Ultra 400 chipset, 512 MB DDR-SDRAM, 80 GB Seagate 7200.9 HDD, Radeon 9600 AGP graphics (128MB), 802.11g NIC, Logitech X-230 speakers, Dell 1703fp display, Netgear WGT624 router, and Samsung ML-1710 printer.

With the system unit, display, and speakers off, the printer in standby, and the router on, total power consumption was 21 Watts. If I powered down the printer completely, power consumption for the system unit and router dropped to 15 Watts.

This just goes to show that even when a desktop is powered off, it still consumes a nominal amount of power.

With the system unit, display, speakers, and router on, and the printer in standby, total power consumption was 120 Watts. This is with the computer idling or working on light tasks like word processing and IMing.

So I decided to crank up the stress on the computer. Running 3DMark03, max consumption went up to 140 Watts, though the typical use here seemed to be about 135 Watts.

For fun, I measured the power consumption of a Napa platform Centrino notebook with an Intel Core Duo processor, 512 MB of RAM, 14.1" WXGA display, and integrated graphics. 20 Watts at idle.

One thing to consider is the efficiency of the power supply. The system unit of my Athlon box draws 85 Watts by itself, and the power supply is rated 68% worst case (I think typical is probably 75%). So I am guessing that the system components draw about 61-62 watts while 23-24 watts are lost in the conversion from AC to DC.

#2 Post by **rkawakami** » Sat Mar 10, 2007 4:02 pm

Was the laptop running without any battery? If not, some of that power could have been going into charging the battery. Granted, it's still the total amount of power being drawn that counts but I'd be curious to see what just the laptop itself was consuming.

leoblob · #3 Post by **leoblob** » Sat Mar 10, 2007 5:07 pm

Desktop computer power supplies are always on. When you hit the "on" button, you are completing a 5V circuit (that's been running the whole time) that tells the rest of the power supply to essentially wake up (that's known as the 5Vsb signal). The 110VAC power input is always applied and "live."

With older computers (486s), the on/off switch actually controlled the 110VAC line going into the power supply.

One thing about those power meters (I have one)... They may not be so accurate when measuring the power draw from something like a switching power supply in a computer since the phase relationship between the input voltage and current can get all twisted around.

#4 Post by **K. Eng** » Sat Mar 10, 2007 5:38 pm

rkawakami wrote:Was the laptop running without any battery? If not, some of that power could have been going into charging the battery. Granted, it's still the total amount of power being drawn that counts but I'd be curious to see what just the laptop itself was consuming.

The battery was in the laptop, but not charging. 20 Watts max at idle is about right for this machine, taking into account the inefficiency of the AC adapter. Also, in AC mode, some of the notebook's power savings features are disabled and the display panel runs in full power mode.

#5 Post by **K. Eng** » Sat Mar 10, 2007 5:40 pm

leoblob wrote:One thing about those power meters (I have one)... They may not be so accurate when measuring the power draw from something like a switching power supply in a computer since the phase relationship between the input voltage and current can get all twisted around.

Interesting. I wonder how accurate the meter I used is.

#6 Post by **tomh009** » Sat Mar 10, 2007 6:29 pm

K. Eng wrote:With the system unit, display, speakers, and router on, and the printer in standby, total power consumption was 120 Watts. This is with the computer idling or working on light tasks like word processing and IMing. (...)

For fun, I measured the power consumption of a Napa platform Centrino notebook with an Intel Core Duo processor, 512 MB of RAM, 14.1" WXGA display, and integrated graphics. 20 Watts at idle.

Hence my plan to use a ThinkPad as a server, even if I need to resort to an external USB disk!

hausman · #7 Post by **hausman** » Sat Mar 10, 2007 8:39 pm

FWIW I have a Kill-A-Watt too. Here are my measurements from a few months back.

IBM TP A21p
Operating: 20W to 40W
Standby: 10W (no battery installed)

Dell Dimension 3000 desktop
Operating: 140W
Standby: 100W
Dell 2001FP 20" LCD: 45W

And tomh009, Crappy Tire has something similar to the Kill-A-Watt for $25 (item #52-8851-2.) That's less than what I paid for the Kill-A-Watt on eBay.

leoblob · #8 Post by **leoblob** » Sat Mar 10, 2007 10:00 pm

K. Eng wrote:Interesting. I wonder how accurate the meter I used is.

I do think, however, that the relative measurements you made are very useful for looking at how power consumption changes as you change operating states of your computer.

I used one of these watt-meters to look at how much power a window air conditioner was using. I was very surprised to see that the "high" setting used almost exactly the same power as the "low" setting (both measured when the compressor was running). I believe this means that my total power consumption would be lower by using the high setting, but having the compressor run less often. This would have never occurred to me unless I had used that meter.

#9 Post by **tomh009** » Sat Mar 10, 2007 10:27 pm

hausman wrote:And tomh009, Crappy Tire has something similar to the Kill-A-Watt for $25 (item #52-8851-2.) That's less than what I paid for the Kill-A-Watt on eBay.

Thanks, Dorian -- I may just check that out.

And 45W for the 2001FP (which I have at the office)! That's about three times what my X31 uses.

#10 Post by **K. Eng** » Sun Mar 11, 2007 1:10 pm

I took a couple more readings.

The Netgear WGT624 802.11g router by itself consumes 4 watts.

A Sony Dream Machine alarm clock consumes so little power that it registers at 0 Watts.

A tree lamp with 2x 60 Watt bulbs (could have 3, but I only have 2 lights installed) consumed 114 Watts.

PC power consumption

PC power consumption

Re: PC power consumption

Who is online