Home Network Power Consumption



  • I finely got around to measuring the power consumption of my home network.

    • 12.60 Watts - HP EliteBook 2530p Laptop - Core2 Duo SL9600 @ 2.13Ghz - 4 GB Ram -128GB SSD - Atheros Mini PCI-E as Access Point

    • 7.60 Watts  - Cisco SG300 10-port Gigabit Switch

    • 7.10 Watts  - Cisco DPC3008 Cable Modem

    • 1.40 Watts -  Power Bricks

    • 1.00 Watts -  Monster Cable Surge protector

    Total Power Consumption = 29.7 Watts  ;D

    Monitoring power usage it bounces around consistently in the 28.2 to 29.4 Watts range.

    Equipment used in test - P4400 Kill A Watt Power Meter



  • My UPS shows the data room here hovering around 490watts.  :P

    Glad its been a mild couple of months here otherwise the AC would be driving that number up a tad.  ;D


  • Netgate Administrator

    12.6 Watts is impressive.  :)
    I'm not totally convinced my own plug-top style power meter is terribly accurate, especially at low levels. Probably good enough for relative measurements though.

    What is the 'Power Bricks' measurement?

    Steve



  • @stephenw10:

    12.6 Watts is impressive.  :)
    I'm not totally convinced my own plug-top style power meter is terribly accurate, especially at low levels. Probably good enough for relative measurements though.

    What is the 'Power Bricks' measurement?

    Steve

    It's listed above in the OP …. 1.40 Watts

    The Ebay seller of the laptop sent a 120 watt HP Power brick with the laptop  ??? ..... contacted him about the issue that it should be a 35 watt brick. He sends me a 35 watt brick but theirs no ferrite choke so I'm still on the 120 watt until I purchase a new 35 watt with ferrite choke. So I would say the power brick watt use is a bit high because of this.

    Need to get rid of some of my power bricks on ebay..... :(



  • @chpalmer:

    My UPS shows the data room here hovering around 490watts.  :P

    Glad its been a mild couple of months here otherwise the AC would be driving that number up a tad.  ;D

    I know what you mean :o
    Under normal work loads my HP Z800 Workstation not including monitors, licks up around 180 Watts …. Fire up a 3D game like BF3 its like a flaming jalapeno..... 350 - 400 Watts  throw in monitors its hitting close to 600 watts

    Hoping HP in the next 18 months with Intel and Nvidia's help can improve on power efficiency with there future Z830 series but I wouldn't bet on it. That is assuming they continue the Z product line.


  • Netgate Administrator

    Ah, OK. So that 1.4W figure is the off load consumption of the laptop power brick?

    Steve



  • That is correct …. laptop powered off ..... but one thing I did not think of at the time is unplugging the laptop from power brick ..... so the 1.40 watts or part of it is WOL.

    Power brick for switch nor did the cable modem power transformer display a power draw when unplugged from device.... so 1.40 watts from the laptop power brick could very well be the WOL.

    One thing for certain it displayed total network power usage consistently in the 28.2 to 29.4 Watt range. The Kill A Watt PP4400 claims 0.2% Accuracy.


  • Netgate Administrator

    Does the laptop have a battery in it? The few laptops I have looked at in  any detail have charged in a periodic fashion, charging for 30secs - off for 2 mins.

    Anyway, less than 30W total is pretty good going, you have me beat!

    In a previous job I had a number of power meters that I used for various tests. All of them cost a lot. The Kill-a-watt just seems too cheap!  :)

    Steve



  • You can save alot of electricity by using power measuring tools, i have a "Smart Plug" with Wifi that shows the current power draw in my iPhone, with price per day and month. Very handy for measuring optimizing the computers running at home. Only my UPS uses around 70W when in "idle" for example. In total the server room uses 490W at the moment.



  • Yes, battery acting as UPS…


  • Netgate Administrator

    Ah well in that case an instantaneous reading may not be that accurate. It would be interesting to record the Wh over, say, 24hrs and average it.

    Steve



  • @stephenw10:

    Ah well in that case an instantaneous reading may not be that accurate. It would be interesting to record the Wh over, say, 24hrs and average it.

    Steve

    Monitoring for 15 minutes thats long enough… ;)



  • I measured 6 watts for a cisco SG200-08 switch

    My first server with a pfsense is with a Asus Pundit-R (p4 2.4 GHz, 1 GB ram, 2 x intel PCI pro1000 network cards and 8 GB SSD) at around 60 watts

    Seems a bit high, as i run a freebsd based nas with a Xeon 1265Lv2, 32 GB ram and 20 * a 7200 rpm hdds which uses 180 watts.

    I have  spare mainbord with 8 GB memory and could try a xeon 1220Lv2 and a very efficient power supply. Perhaps that consumes less, but is a 400 euro investment.

    This pundit has no issues doing line rate donwload at 120 mbps which is nice, and it does not need to be a much faster machine for me i guess.


  • Netgate Administrator

    60W for a P4 seems about right unfortunately. The Netburst CPUs were notoriously power hungry and I expect your desktop model has no power saving features. My own P4 box with 2.8GHz desktop CPU idled at ~55W but it has no video hardware. I replaced it with a P4-M which idles at around 40W. Still not great.  :( I'd replace it but it has 9 on board Intel NICs and runs rock solid.  :)

    Steve



  • Interesting thread, can contribute some more data:

    Cisco SG300-10:  6 W
    Cisco SG300-28:  10 W
    pfSense machine: 33 W (Atom D-525, CF, 1xWLAN, 4 GiB DDR3, no keyboard, no monitor)
    Desktop management machine: 160 W (Asus P5E3-Deluxe, 8 GiB DDR3, 4xHDD, 2xDVD, 1xWLAN, 2xEthernet, 19" TFT, USB keyboard, not yet measured under heavy game load :))
    Backup pfSense machine: 6 W (Alix.2D13, 1xWLAN, CF)

    Peter



  • Xeon E3-1230v2, 16GB RAM, 1 SSD and 4 Hard Drive, 350W 80+ GOLD: around 50W IDLE (probably I can arrive to 40W if I disable Supermicro IPMI and I slow down/remove some fan)
    Watchguard x750e with Pentium-M, PowerD and DC-DC brick: about 25W
    HP 1800-24G: about 15W, depending on how many ports are connected



  • @stephenw10:

    60W for a P4 seems about right unfortunately. The Netburst CPUs were notoriously power hungry and I expect your desktop model has no power saving features.

    With some traffic i saw 90 watts. So i ordered a new mainbord/psu (msi c847ms-e33)
    According to tests (its brother, the p33 was tested) it uses around 20 watts idle



  • Total as shown by my UPS, I don’t have the breakdown per unit: 27 – 31 Watts

    • PfSense Intel DN2800MT Atom, 4GB PC10600 200 Pin SODIMM , 64 GB mSata

    • Linksys PAP2-T VOIP ATA

    • Netgear GS116E 16 Port switch with 6 ports active

    • Netgear GS108PEv2 8 Port POE switch with the following devices drawing power:
                  Polycom 335 IP VOIP Phones x 2
                  Axis M3007 Network Camera
                  enGenius EAP350 WAP



  • @Downloadski:

    @stephenw10:

    60W for a P4 seems about right unfortunately. The Netburst CPUs were notoriously power hungry and I expect your desktop model has no power saving features.

    With some traffic i saw 90 watts. So i ordered a new mainbord/psu (msi c847ms-e33)
    According to tests (its brother, the p33 was tested) it uses around 20 watts idle

    Got it running on pfsense 2.1, as 2.03 would not boot up on it.

    Msi c847ms-e33 mainbord, 16 GB memory(2x8) seasonic x400 fanless platinum psu in a silverstone sugo09 case (180mm and 120 mm fan running) 64 gb ssd and a pci-e 2 port GE card (hp branded intel)

    Running it consumes 27.8 watt



  • Nice power usage numbers Downloadski  :) …. what device did you use to measure power usage?



  • a plug in the wall outlet type of device (live in Netherlands Europe)
    I have it since years and not a  idea how precise it is. This PC uses half to a third from the test pc i used to check pfsense out.

    It might be a unit made by Cresta.


  • Netgate Administrator

    To quote the Cresta RCE1106 manual:

    Accuracy:

    • Voltage: +/-3% of value measured
    • Current: +/-3% of the value measured +/-0.03A
    • Power: +/-5% of the valve measured +/-10VA
    • kWh: +/-5% of the value measured+/-0.1kWh

    That seems about what I'd expect from a plug top style meter costing <£50. I imagine my own meter is similar. I have a hard time believing the figure giving by Kill-a-Watt.  :-
    This seems to support that: http://forums.anandtech.com/showthread.php?t=137169

    I'm not going to stop using it though. Much better to have some reading for comparison that no data.  :)

    Steve


Log in to reply