So recently I ordered a ‘Kill-a-watt’, a gadget that lets you plug it in between a device and the power supply that tells you how many watts, volts, amps yada yada it uses. In these times of insane energy bills (mine is now £100 a month combined gas and electricity) I think it’s a smart move to work out where all the power goes. Sure, some companies quote figures, but can you trust the seller of some kit to say how much it costs to run?

The Kill-a-watt cost me £19.95 from here: http://www.reuk.co.uk/buy-KILL-A-WATT.htm and arrived this morning. Here is what the little beastie looks like:

So how do all my bits and pieces of hardware stack up? First lets try that nice flat screen monitor:

The Monitor


The monitor in question is an IIyama 19 inch ProLite H481S. It’s pretty lovely, but looking a bit small these days. this gets auto-turned off by vista when I’m away from the PC, and I also turn it off each evening. It’s guzzling 30 watts, which isn’t to be sneezed at, that’s the equivalent of 3 energy saving light bulbs.

The official power consumption for this model is 40W, so it ooks like they actually overestimated its usage. I’m running it at the native res, but the (very poor) speakers were not in use.

Edit : It seems that I can push it 1 watt higher by playing ‘painkiller’ by Judas Priest through the inbuilt speakers. I think i generalize this to playback of any heavy metal.

The Printer

The printer is an HP deskjet 960C, which I often leave switched on when not in use. Looks like it only uses 2 watts, probably for that little green LED. No great worry there.

The Router

I use a very flaky Linksys wireless-G ADSL gateway router. Obviously it’s always on, and because it gets upset when turned off (sometimes this piece of crap won’t turn on until its been left off for 30 minutes…) I leave it on all night too. Looks like all those flashing lights only use up 8 watts, which is not irrelevant, but not a disaster either

I can’t find any official information on what power this unit is supposed to draw, so you heard it here first!

The PC

Holy cow, the main PC base unit really does guzzle the power. This reading (160 watts) was taken after windows had booted and settled down, its an Intel dual core 6600 2.40GHZ CPU  (2 gig RAM) vista machine with a Geforce 8800 GTS video card. Interestingly, I tried some experiments with heavy processor use (rebuilding my current game project, the sequel to kudos), and some mega disk thrashing (copying huge amounts of files) and the power consumption didn’t really change much at all. Make of that what you will, maybe the disk is always drawing a lot of power anyway, or maybe neither processor core is idle, or steps back its power drain when idle.

The Game

I tried starting a 2 vs 2 AI vs Me skirmish game of Company of Heroes, set to 1280 1024 res with everything set high, to see if maybe the video card could push the power consumption higher. A noticeable change, but nothing dramatic to report. maybe those people trying to sell 300 and 400 or even 500 watt power supplies are overstating their case?

Sleep Mode

ZZZZZZZZZZ. I set vista to go to sleep mode if left for 15 minutes or more, and am very happy with how it springs back to life damn fast. I still make the effort to completely power down the beast overnight, and a BIOS timer wakes it when my alarm clock goes off. is this worth the hassle? In sleep mode it looks like its drawing a measly 3 watts, less than the flipping router.  Interesting stuff :D

So there you have it. later today I might check out what the TV, the Wii, the other gadgets and gizmos draw to see if theres anything stupidly hungry that I’m leaving on (although generally I’m obsessed with turning all this stuff off).

48 Responses to “Computer stuff and power requirements”

  1. Drealmer says:

    I have always been told that battery chargers, even when not in use, consume some power. And indeed, they tend to get hot even when the device they are supposed to charge in not plugged. Could you try with your cellphone charger or something like that?

  2. cliffski says:

    my cellphone charger seems to be using 2 watts when plugged in and charging, and zero watts when I disconnect the phone. Thats a motorola charger.

  3. Trails says:

    Try running a disk drive (i.e. DVD), in both read and write?

    I suspect this draws substantial power.

  4. Trails says:

    Oops, also meant to mention: interesting article.

  5. DrLang says:

    The power draw of a charger when it is plugged in vs not plugged in probably depends on the type of power supply it is. Switching power supplies tend to be extremely efficient, and you should expect your cel phone charger to draw very little extra. Now that big heavy wall wort AC adapter is another story.

  6. EvanED says:

    Happen to have an old CTR around for comparison with the LCD by any chance?

  7. twitter says:

    Check out the InfoWorld results for CPU use: http://weblog.infoworld.com/sentinel/archives/2008/05/vista_pcs_worki.html

    Vista uses a lot, even when idle, and that costs power. 165 Watts is not bad for a gaming rig.

  8. Jeff Schmidt says:

    Interesting article. Gives me something to compare my Dell ‘gaming’ laptop too. Note that I did NOT by one of those insanely priced Alienware/Dell gaming laptops. I bought an Inspiron with a 2 Ghz Core2Duo mobile processor, an Nvidia GeForce Go 7900GS and 7800rpm disk drive. More expensive than a typical home user laptop, but less than 2 grand (in Jan 2007, to give you a timeframe in which to put that price in context).

    Anyhow, I believe the power brick that came with the laptop is rated at 90 Watts, so that tells me that my laptop is drawing approx 1/2 the power that a desktop would. Granted, it’s not running as fast, but it’s fast enough for most games, and for doing stuff like watching fullscreen video, compiling software, and using Blender 3D to render scenes (granted, I haven’t done any highly complex scenes, or used a photo-realistic ray tracer, yet).

    All in all, I think the laptop gives me an excellent power/performance ratio, it sounds like. I get probably 75% ro 85% of the performance of a Desktop, at about 50% power usage.

  9. cliffski says:

    Some more readings:
    The kettle pulls 2814 watts! the electric fan heater pulls over 2,000 watts (as expected). it brings it home to you when you visualise switching on 90 monitors whenever I toggle the heating on in this room…
    My energy efficient lamp is indeed only pulling 14 watts.

  10. Adam Piggott says:

    Had one of those dudies for a year or so now, I still like going around the house on occasion to see what’s drawing what. Kettle, as you stated, is a good eyebrow-raiser – try the washing machine as well! Oh wait, we bought these things for the hell of it, we don’t even know how to use a washing machine ;-)

    My UPS is a bit of a greedy, around 30W when not powering anything. Not so great, may invest in a new one if I can find out if it’ll be worth it.

    My quad-core system is around 230W idle and over 350W when using Microsoft’s consume.exe, and even more if consume and ATI’s GPU stresser is used (only a wee Radeon HD 2600 Pro). These figures are from memory so might be a bit off, but they shouldn’t be far off.

    I find it interesting how your wattage doesn’t vary much between idle and load whereas with the same tester my system does. Maybe if you give consume a whirl you’ll find otherwise – this does literally max out the CPU.

  11. mgysgthath says:

    High wattage PSUs are not just window dressing (unless they are no-name brands). An 8800 GTX requires 30 Amps on the +12volt rail (combined) to operate properly. The idle wattage estimate is around with a 9800 Pro uses under full load, somewhere in the neighborhood of 180 watts. Whether this is at the wall, or from the DC transformer in the PSU, I’m not sure.

  12. Peter Green says:

    Nice article!
    I wonder if they make PSU’s economical at all?
    I remember a long time ago I had a PSU that powered down it’s fan when not needed, nice and quiet not to mention power saving.
    Maybe there are more economical PSU’s to be had?

  13. Phil says:

    If you don’t want to fully shut your computer down, but still reap the power saving benefits then use Vista’s (And now XP SP3’s) hibernation mode.

    Last thing I do at night is:

    start -> cmd -> shutdown /h

    when I turn on in the morning it takes slightly longer than from sleep (about 15-20 seconds) but all my programs are open as if the computer had just been to sleep.

  14. Slick Denis says:

    Ah! Heh heh! DENIS! It’s all goin’ cockers now, lads. Heh heh heh. Ah, oh boy. I wish I knew how much power my PC used, but I don’t have one of these power meter things.

  15. FokeyJoe says:

    Thanks for posting your results, mine are mostly comparable. Though, you should check out whether the power goes down if you are more aggressive with your power management, e.g. stopping the disks after a short time.
    Regarding PSU ratings, the docs with PSU’s say that when they’re 80-85% efficient, it’s not when running at max capacity, but at average capacity, which is closer to 60-70% of the total wattage it’s capable of.
    I found my HP Photosmart 2610’s brick pulls a massive 15W without the printer being plugged in! As with all devices, it’s best to unplug or switch them off at source.
    I use my PC’s to heat my room, with aluminium cases and heatsinks I don’t need my central heating as much. Fan heaters are devil’s work!! :)

  16. Jeff Schmidt says:

    mgysgthath,

    Are you sure about those numbers? 30A is a *lot* of current. watts = amps * volts, so 30A @ 12V = 360 Watts. That sounds ridiculous, even for a high-end video card.

  17. bleh says:

    This has been one of the useless articles I have read today. People should know everything draws power even if it pluged in or in hibernation or if the side of the PSU says it only draws 100watt’s when idle. This is why most global warming weenies always say unplug everything not in use like I do. If i’m not using the PC why have it one?

  18. bleh says:

    This has been one of the useless articles I have read today. People should know everything draws power even if it pluged in or in hibernation or if the side of the PSU says it only draws 100watt’s when idle. This is why most global warming weenies always say unplug everything not in use like I do. If i’m not using the PC why have it on?

  19. Mr.PSU says:

    I also bought a wattage-meter within the same price-range. I thought it gave good estimates, but my new Quad-core intel computer draws less than 90W (whole system with 100% cpu load (all cores) with seti@home) full load. Maybe it’s my 600W zalman PSU that’s very efficient (worth the money to buy good PSU if efficiency is 10-30% better than cheap psu) but I frankly think the wattage meter is not telling the whole story. Electricity is a more complex phenomenon than you might think at first.

    I’m not looking to be “green” but electricity does cost money and I use about 10-20% of my income on electricity, so it makes sense to ditch old crt’s for LCD’s that use about 25-50% of the electricity and replace old computers with new ones that consume less and have more computing power.

  20. Tony P says:

    We need to start building intelligent power supplies for laptops too. My 90W power supply actually uses 225W. When the computer sleeps it still uses 225W. There should be better designs so that the power system on the computer tells the brick to only come on every now and then.

  21. […] Cliffski’s Blog » Computer stuff and power requirements […]

  22. Dbug says:

    Getting readings of 90 Watts on some desktops is not a problem with the wattmeter.

    My system runs 82 Watts, mainly because I was paying attention when I built it up. Since it isn’t for gaming, I went with a motherboard having GMA-950 graphics. That still works great for HDTV. I’m running an E6300 Core 2 Duo overclocked from 1.86 to 2.25 MHz (pushed fsb speed a bit), stock fan. There’s a 1 terebyte SATA drive, DVD burner, and USB Eye-TV Hybrid (HDTV tuner) being powered.
    So that’s 82 Watts for everything but the monitor, and that’s while doing video compression with both cores very active.

    Hard drives run around 10 Watts apiece. I figure each 10 Watts continuous consumed costs me roughly $1 month U.S. at .13/kwh. (power is expensive where I am). Instead of keeping show archives in more drives in the computer, I use externals which I can turn off. eSATA cases give the same speed as internal (at least triple what I get with USB 2).

    I’m curious if any gaming GPUs have low power consumption when not doing heavy work. When I saw the big heatsinks/fans I knew I’d be best off avoiding them.

    Oh, I used a cheapie power supply with no efficiency rating given for $15 with free shipping! It was rated at 600 Watts. I wouldn’t trust it under a really heavy load, but at under 100 Watts it’s so cool I expect it to be reliable.
    The rating on a power supply is a maximum total output. A higher number means it can handle more, but does not imply that it’ll use more at a particular power level. Supplies do have to be rated higher than what you use to handle surges when drives spin up etc.

  23. Dbug says:

    Opps, my CPU speed was GHz, not MHz.

    Note meter should be set to show Watts, not Volt-Amps for actual consumption.
    Voltage and current are out of phase with some loads (like unloaded AC adapter transformers, motors etc) which would give a misleading high reading

  24. Inwards says:

    I’m typing this on my Dell XPS 720 (QX6800, two NV 8800 GTX’s – ships with a 1000 Watt power supply) with a 24 inch HP w2408. I’ve connected the both to my UPS which displays my total power draw on a nice little LCD. Right now, it’s show 344 Watts. Crysis runs around 440. I haven’t run a torture test yet, but I suspect that it’ll only get worse…

  25. shoez says:

    Battery chargers use power when turned off because of the way they work. The wall power you have is alternating current(A voltage sine wave) and your sensitive device wants direct current(A constant voltage.) Your charger(wall wart) does this using a transformer and a recitifier. The transformer uses two sets of wires wrapped around a metal core to step the input voltage up or down to the secondary voltage. The rectifier takes that amplitude scaled AC signal, and flattens it out at a positive DC offset. When there is no load on the DC side, there is still current flowing through the rectifier, because it is just a bunch of wires wrapped around a hunk of metal. This power consumption is called “copper loss” because it is simply due to the current flowing through the copper wires fighting the resistance of the wires themselves. This power is expressed as heat.

    Lots of other devices continue to use power for other reasons. Many modern devices never truly turn off, but instead minimize the activity of the most power-consuming parts. Your DVR is a good example. It needs to keep track of what’s going on even when you don’t want to watch TV, so it turns off the output rendering equipment, and continues to use its tuner and hard drive. Your computer and video game consoles consume power to run their internal clocks, so that they know what time it is when they are turned back on. The Xbox360 and PS3 also maintain their USB voltages when powered down so that they can charge controllers.

    As power gets more expensive, hopefully consumers will demand and inspect information regarding power consumption. Then it will be useful for designers to be vigilant in power saving, and we will see a reduction in “on” and “off” power consumption. We have also seen the invention of more efficient electrical and electronic mechanisms for accomplishing the same goals, like fluorescent light bulbs, and liquid crystal displays.

  26. >>maybe those people trying to sell 300 and 400 or even 500 watt power supplies are overstating their case?

    Every computer is different. One computer might be a very low end processor with hardly any software, but another computer might have a bunch of hard drives, a high-end video card and processor. Every computer has different power usage…duh!!!

    Most cheap power supplies are NOT capable of delivering the maximum rated power…they are junk. To be on the safe side with power supplies, it is best to buy one with much higher power rating than is required, because they will regulate better and run cooler. Never buy the cheapest power supply for expensive computer equipment…one of the stupidest mistakes that people make…but also don’t go crazy and spend way too much either.

  27. mikepikefl says:

    I’ve wanted one of those thingies for a while! Very cool info. I’m curious about the readings of that same LCD when it’s in “standby”, “suspend”, and “off” DPMS modes.

  28. wicko says:

    You were surprised at how many watts your PC consumes? If anything, I would have been more surprised at how *little* it consumes. We always hear about how you need huge power supplies to run new video cards, when its not even close to the truth. I did the same thing with my system a few months back, it was 180 at idle and 260 while playing crysis. Yet the recommendation was a 500 watt power supply or greater. you could easily run this system off a 350-400 watt power supply. Maybe they are concerned about efficiency, but even so, I think the average power supply is the most efficient at 80% output.

  29. 2Paranoid says:

    I have a Kill-a-Watt meter also and did some recent checking.

    Below is the following info:
    – Machine, Idle watts, 100% CPU watts using Folding@Home to max the CPU(s)

    – 2.4GHz E6600 Quad, 97w, 188w
    – 3.0GHz, ?w, 181w
    – PS3, 185w, 208w

    Sorry, forgot to write down the idle watts for the 30.GHz PC. Both PC’s have nvidia 8800 video cards. If I remember correctly the watts go up another 10 to 15 when both the CPU and video cards are max’ed.

    I have a 52″ LCD TV and it runs around 200w when the screen is black and 300w when it is white, and everything in between depending on how bright the picture is.

    A heating pad toggled between 0w and 16w. I expected more.

  30. Mike L says:

    I was wondering… has anyone has tried testing how much power an external hard drive consumes?

  31. […] amps yada yada it uses. In these times of insane energy bills mine is now ??100 a month combined gashttp://positech.co.uk/cliffsblog/?p=21How many watts can a 15 amp circuit support?Oct 25, 2007 … how many watts Will A 10 amp Power […]

  32. Some context... says:

    I too bought one of these meters and measured my PCs (and everything else).

    Wall chargers and PC power use is generally lost in the noise when you start to measure lighting, heating and hot water. For example: an electric shower is typically 9kW. With five people in the house having a daily shower that runs for best part of an hour. Everything (several PCs, STB, chargers, TVs) in standby would have to left on for a week to get close to one hour of shower.

    Kettles, Irons, Washing Machines all take orders of magnitude more power than electronic goods.

    But if you really want some context, go and look at a business park at night. Big 500W lights illuminating car parks. Buildsing left illuminated… Domestic power use is irrelevant in comparison.

  33. GrimRC says:

    Give Linux (decent livedvd?) a go, and see if it’s more efficient.

  34. re: wicko says:

    as the poster above said a power supply should be capable of giving more power than needed to run at full load to handle surges such as when the computer turns on/hdd’s spin up etc.

  35. Brandon Hunt says:

    I’m interested in seeing the power consumption of some other devices like the Wii, and also interested in seeing comparisons of consumption between game consoles, and maybe consumption of devices like a television, DVD player, paper shredder, etc.

    If you decide to check the consumption of any more devices, please contact me, because I’m interested in seeing them.

    -Brandon Hunt

  36. John says:

    LED’s use no more than 20mW, not 2W

    > maybe those people trying to sell 300 and 400 or even 500
    Average watt usage and peak watt usage are entirely different ratings ;)

    Anyway, help save the environment and buy a laptop. The energy savings will pay for higher price tag. =P

  37. zocker666 says:

    Crazy shit indeed. Who needs a PC that draws 160W in Idle and 190W while gaming? And in the comments theres systems using even more than 190W.

    I have a server running at 85W in idle mode (4 HDDs, ~2.3TB storage, Athlon 64 X2 Dual Core @2.2GHz). The same server with 4HDDs, ~750GB storage and an Athlon 1GHz ran at 80W idle before that. When using the CPU(s) for video encoding it jumps to about 120 (both for the old and new system). I use a highly efficient PSU that was not really expensive. The HW upgrade (CPU, MB, RAM) was no really expensive, ~300EUR (currently ~$470). While the server does not have a GFX card, I do all my gaming with my laptop and it’s absolutely no problem doing that.

    The Laptop is an Intel Core 2 Duo @2GHz with a GeForce 8600M GS and gaming is not a problem at all. Sure, I can’t use all the highest settings etc. but who really needs to? Gaming is about fun not all the latest fancy stuff you won’t notice during normal gameplay anyway. The Laptop uses ~35W when idling in Windows and ~65W when gaming.

    But then again, I am forgetting about the ability of all those high end, money sucking gear to really enlarge your XXX size … :)

  38. Dan says:

    If you want to know what draws the power in your house, put that thing on the fridge

  39. grayjamn says:

    Saw this and thought it would be relevant to a question asked on whether or not PSU manufacturers are looking at making more energy efficient devices.

    Apparently they are!

    http://www.antec.com/us/productDetails.php?ProdID=27430#

    Thanks for posting your experiment results!

  40. JimT says:

    Iteresting results, thanks for taking the trouble.

    Now THAT’S a knife….

  41. freddo says:

    Just want to make a comment in reply to that made above about phone chargers/transformers etc. Every phone charger I’ve had in recent years (as are many other types of DC adapters these days) are switchmode. These don’t have transformers, and do the same job with switching circuitry. I don’t have concrete figures, but they should be using very little power with no load attached. It would be nice to confirm this though.

    You can tell the difference between transformer and switchmode dc adapters because switchmode adapters are much smaller and lighter (ie no big chunk of metal inside).

  42. […] usage of some devices, I always thought the monitor would eat more than that!!  furthur reading: cliffski Posted in headlines | Print Post | Email […]

  43. bill smart says:

    […] amps yada yada it uses. In these times of insane energy bills mine is now ??100 a month combined gashttp://positech.co.uk/cliffsblog/?p=21Bill Smart – LinkedInView bill Smart’s professional profile on LinkedIn. … Executive Consultant at […]

  44. […] Computer stuff and power requirements I can’t find any official information on what power this unit is supposed to draw, so you heard it here first! […]

  45. […] have kind of always wanted to do what was done in this article, Computer stuff and power requirements ,i.e. buy something like a Kill A Watt and measure my household electrical […]

  46. Graeme says:

    Gotta say that I found this article, and many of the comments very good reading. I’ve just placed my order literally an hour ago after reading all this, reason being I’ve built my own server and I want to price up how much it’s going to cost to have it turned on 24 hours a day, or whether it’ll work out cheaper paying for a service provider.

    Big thanks for the article, it’s really made me think although I’m going to annoy the hell out of the wife with the measuring tool when it gets delivered and I start turning appliances off whilst she’s using em!

  47. feargoanota says:

    Wonderful page. will come back.