Power Usage: What'cha Usin'?

Discussion in 'Benchmarks & Performance' started by Taurus, Jun 27, 2008.

  1. Taurus

    Taurus hardware monkey

    Messages:
    3,206
    Location:
    Sacramento, CA
    been playing with my Kill-A-Watt and i thought i would start a thread where those who have the means to measure their computer's power usage can report their hardware and watts used in these different states:

    Off
    Standby - aka. Sleep or S3
    Idle - desktop, minimal apps, no activity
    Load - 100% CPU, all cores
    Game/Max - 100% CPU, all cores, plus 3d benchmarking

    here's the only one i've measured so far... my htpc:

    Processor: AMD Athlon 64 X2 BE-2400
    idle: 1.0GHz @ 1.0v
    load: 2.3GHz @ 1.2v
    Chipset: Geforce 7050PV
    Graphics: onboard
    Memory: 2 x 1GB PC2-6400
    Drives: 320GB 7200RPM HD
    Power Supply: generic SFF 300w
    Extras: USB wireless KB/mouse receiver
    (here list add-in cards, usb devices, etc that use power)

    Off: 5w
    Standby: 6w
    Idle: 41w
    Load: 77.5w
    Game/Max: 80w

    the results really surprised me. for one, 5 watts when completely shut down is an awful lot. i'm curious to see what numbers you guys get here. also, standby is barely 1w more which dissolves any guilt i had with using it. and then the 41w idle reading is lower than i hoped! i was seriously considering getting a BT NAS so i wouldn't feel bad about leaving the comp on to download, but i would barely be saving any electricity (Netgear ReadyNAS Duo is rated at 35w typical).

    post your results! :)
     
  2. Aprox

    Aprox Moderator Political User

    Messages:
    2,737
    Location:
    California, USA
    Wow, nice power usage you got going there. I would be interested in knowing my systems power usage... but I don't want to buy anything! haha
     
  3. LordOfLA

    LordOfLA Godlike!

    Messages:
    7,027
    Location:
    Maidenhead, Berkshire, UK
    Off is actually "Standby" these days. Its so that your network card can do wake on lan or your keyboard can wake from "off", etc.

    "Standby" is actually sleep and has to keep enough power going for the ram refresh cycles as well as wake on lan, keyboard, etc.
     
  4. Taurus

    Taurus hardware monkey

    Messages:
    3,206
    Location:
    Sacramento, CA
    i thought about this shortly after i posted, i went and turned off WoL and "off" went down to 4w. i also have it set so only the power button turns it on, but that doesn't seem to make a difference. at least with this board.
     
  5. ming

    ming OSNN Advanced

    Messages:
    4,252
    Location:
    UK
    Sorry, but i think those results are totally wrong. I don't think it's possible to have a consumption of only 80W at max load. A typical CPU consumes around 40-120W depending on how many cores it has, motherboard around 35-70W depending on load, and GPU anything between 25-180W.
     
  6. gonaads

    gonaads Beware the G-Man Political User Folding Team

    I tend to destabilize the power grid in my area every so often. It drives P.G.&E. nuts. :p
     
  7. sean.ferguson

    sean.ferguson Moderator Folding Team

    Messages:
    1,693
    Location:
    Fife; Scotland
    I agree, if this was the case everyone would be able to get away with running their systems off a 100W PSU rather than the most commonly used 350W PSU. Is this kill a watt system really reliable?
     
  8. ming

    ming OSNN Advanced

    Messages:
    4,252
    Location:
    UK
    Even my PIII 1.0ghz laptop has a power supply rated 235W, though it's likely to only consume probably around 60-80% of that rating. That's around 140W of power consumption minimum.
     
  9. X-Istence

    X-Istence * Political User

    Messages:
    6,498
    Location:
    USA
    Just for you guys saying that 80 watt is impossible, I am running a Via C7 chip that is rated at 20 watts. Running on a motherboard that is rated similarly. One 80 GB sata hard drive and I am still at maybe 60 watts total. Just to prove that having a system running on a 100 watt power supply is entirely possible.

    PAY ATTENTION:

    However the reason he is seeing those results is that the kill-a-watt measures on the AC side, not on the DC side. For example, on my power supply I have an 18 amp 12 volt rail. That means that devices plugged into that can pull a maximum of 18 amps worth of power (remember, W (watt) = V (volt) I (Amps)) that means (12 volts) * 18 amperes = 216 watts. (So, two rails at 18 amps, is about 500 watt, then 5 amps on the 5 volt line, well do the math. Yes, it is a 600 watt power supply). Yet, when I input AC into it, there is NO WAY that it will be pulling 18 amps for 1 rail, why? Breakers tend to pop at 20 or 10 amps, depending on your home.

    See where I am going with this? There is no direct correlation between the two. AC != DC. So a CPU that runs at 120 watts, is pulling that on the DC side, what it is pulling on the AC side is unknown.
     
  10. ming

    ming OSNN Advanced

    Messages:
    4,252
    Location:
    UK
    My bad, didn't take into consideration the low voltage processors.
    About the correlation or non-correlation of AC/DC consumption, your explaination doesn't actually explain why most systems carry a 400-600W PSU if the AC power consumption for the very large majority of today's systems consume a maximum of 80W.

    Let's take my laptop power supply as an example and use the power equation.
    Input: 100-240V 1.6A
    Output: 19V 3.42A

    Correct me if I'm wrong, but output is likely to be in DC and not AC.


    Power (W)= voltage (V) x Current amps (I)
    Input = 100V ~ 240V x 1.6A = 160W ~ 384W (AC)
    Output = 19V x 3.42A = 64.98W (DC)

    Although this is the theoretical power consumption, it just shows that the actual power consumption for most desktop systems cannot be 80W AC.
     
  11. Taurus

    Taurus hardware monkey

    Messages:
    3,206
    Location:
    Sacramento, CA
    i don't see why 80w is that unbelievable. it's a low-voltage processor with no dedicated graphics or any extra cards, for that matter. no optical drive and only one harddrive (for now). i built it to be energy efficient.

    i measured my TV with the kill-a-watt and it's readings matched what others with the same model posted on avsforum.com, so i trust it's accuracy.

    and power supply ratings--that rated wattage is max at each rail (12v1, 12v2, 5v, 3.3v, etc). to pull that many watts out of your power supply with standard components would be damn near impossible. you would need special equipment to put an exact load on each rail.
     
  12. roirraW "edor" ehT

    roirraW "edor" ehT Builder/Installer

    Messages:
    529
    No offense, but there's something not right here. Temporarily ignoring the AC/DC part, 18 amps at 12 volts would become 1.8 amps at 120 volts, so of course it wouldn't pop the typical 15 or 20 amp circuit breaker for a 120 volt circuit.

    Also, it would have to be taken into account the power supplies' efficiency rating. The random, OLD, cheap power supply I just happened to have handy doesn't tell it's efficiency rating, but they all have one whether they say so or not. If one has a 95% rating, then 5% of the AC wattage is waste energy. So the AC wattage will be higher than the DC wattage.

    I just looked up NewEgg's highest customer-rated power supply, and it's rating is greater than or equal to 65%. Not all the p/s on there list the rating.

    They have a category for "80 PLUS Certified". The highest customer-rated of these says up to 84% efficiency.
     
  13. Taurus

    Taurus hardware monkey

    Messages:
    3,206
    Location:
    Sacramento, CA
    i can't be the only geek with one of these things. if anyone is interested in buying one, the model i have isn't the cheapest. here's a cheaper one that does the same thing.
     
  14. Taurus

    Taurus hardware monkey

    Messages:
    3,206
    Location:
    Sacramento, CA
    yeah, i guess i should note that this doesn't make sense to me, either. a 65w component still pulls 65 watts from the wall. well, add in some percentage for power supply inefficiency, but the point is watts are watts. volts and amps will be changed in the conversion, but ideally they would multiply out to the same amount of watts on either side of the power supply.

    and the bottom line is the kill-a-watt measures amp draw at the outlet which is what matters.
     
  15. ming

    ming OSNN Advanced

    Messages:
    4,252
    Location:
    UK
    I've got a confession to make... All my comments previously were made based only on the results. I've only just looked at the link and realised that the consumptipn was tested with an actual product.
    However, I still stand by my comments because those results above is conveying the idea that my Antec Neo480 and Thermaltake 480W PSUs were soo inefficient poweringmy old Athlon 3700+ (single core) system that they both died when consuming/drawing only around 70 -110W of power AC.
    On that note, if it is really true that a similar spec system only consumes around 80-100W, then why on earth do manufacturers put in a 400W psu? Why would one even need to go as high as 1kW psu for a high-end system if it's only likely to consume around 300W max.
    Basing it totally on results in the first post, it says that my 32" LCD tv consumes almost twice as much power as your system which is highly unlikely - (I'm reading off the label on the tv btw... it actually states 152W AC)
     
  16. Taurus

    Taurus hardware monkey

    Messages:
    3,206
    Location:
    Sacramento, CA
    that rating is the max peak usage it will ever use. probably happens when you first turn it on or something for a split second. general usage would be somewhat lower than that.

    how much lower? get a electricity usage monitor and find out.:) judging by this thread, a lot of people would be surprised at much or how little certain things in their home uses. it's really a handy tool.

    i guess this tread can be moved since it's less of a comparison among systems and more a discussion on electricity usage and the accuracy of my monitor.
     
  17. Taurus

    Taurus hardware monkey

    Messages:
    3,206
    Location:
    Sacramento, CA
    here some other things i measured...

    Samsung 61" DLP w/ LED backlight
    off: 1w
    on: 140-170w

    Onkyo 5.1 Receiver
    off: 0w/unmeasurable
    on: 38w

    PS3
    off: 2w
    on: 160w
    game: up to 200w

    Xbox360
    off: 3w
    on: 148w
    game: 160-165w
     
  18. loaderbull

    loaderbull OSNN Junkie

    Messages:
    52
    Location:
    Midlands, UK
    As regards why do manufacturers put in a 300W PSU if it is only pulling 80W, it comes down to effiency of that part. OEM systems have (in general) cheap PSU's, efficiency isn't a selling point on the PSU - it could probably only push out 200W maybe a tad more but they still have to have stability of the system, so the overhead of the original posters PSU is probably double what the components can actually pull.

    So pulling 80-100W leaves room, add to that another hard drive or external USB powered device and it will eat away at the left over wattage... A CD Rom drive spinning...

    Maybe i'm wrong though!
     
  19. LeeJend

    LeeJend Moderator

    Messages:
    5,291
    Location:
    Fort Worth, TX

    Sigh... Good thing you have a power systems engineer lurking around here.

    Watts are not always watts. Watts = Volts x Amps x power factor.

    Power Factor (abbr PF) is the Cosine of the phase angle between voltage and current waveforms. The PF is usually 0.6 or less for electronics. For good equipment like that required in Europe the PF is required to be very close to 1 (+/-0.95).

    1) In the USA where we waste energy for sport, your power supply may not be power factor corrected. (Thought I'd beat the Brits to that comment ;) ) With a non-corrected power supply (most USA electronics) you will use 40% more than the "Watts"

    2) I suspect the power company is charging you by the VA (volt amp) not by the watt. I need to confirm how the current generation of electric meters work but since the power company has to deliver volt-amps which are higher than watts they'd be stupid charging for just watts.

    3) Electronics usually have a really crappy current waveform caused by modern switch mode power supplies which chop up the nice clean sine waves rather badly. This harmonic distortion also causes an apparent reduction in power factor and draws excessive energy to perform the task.

    For the above reasons your power meter may not be reading Watts or V-A correctly. Good power analyzers run in the kilo dollars.

    You best bet is to measure volts and amps and compare them to the watts reading. I did not see if your meter can measure PF and VA and VAR, but if it can't the only thing it will measure correctly is a space heater or lamp.

    Post back if the meter will do anything beyond W, V, A. I might buy one.

    PS 5W for a computer or TV in standby does seem low.
     
  20. Taurus

    Taurus hardware monkey

    Messages:
    3,206
    Location:
    Sacramento, CA
    funny that you posted about VA and PF, Leejend. i just got done reading about it before checking this thread again. i knew what the "VA" mode on the monitor stood for, but didn't get why it was different the watt rating. and since the power monitor can measure V, A, and W, and PF is just a simple calculation derived from those, then yes, it has a PF reading as well.

    but i still stand by the fact that watts are what matters, being that our power company, and i think most others, charge by the kilowatt hour.

    and even if the power monitor is off by a certain amount, we would still be able to compare one system to another.