1. This forum section is a read-only archive which contains old newsgroup posts. If you wish to post a query, please do so in one of our main forum sections (here). This way you will get a faster, better response from the members on Motherboard Point.

idle power of 8500 GT

Discussion in 'Nvidia' started by bruce56, Aug 26, 2013.

  1. bruce56

    bruce56 Guest

    I got a PC which had a Sapphire HD4870 graphics card.
    I measured the idle power draw of the PC, and it was 249 W. I thought
    the HD4870 might have something to do with this, as it has two extra
    power connectors. I have another PC containing this card, and it can
    get hot. I do not need a powerful graphics card, for this particular PC.
    Well HD4870 not considered powerful in 2013, but that is still more grunt
    than I require, so I put in a GV-NX85T512 based on Geforce 8500 GT.
    I believe that was a year older than the HD4870.
    The card has no fan, so I assumed it would use less power. I again
    measured the idle draw of the PC, and it is now 255 W, up 6 W!
    bruce56, Aug 26, 2013
    1. Advertisements

  2. bruce56

    Paul Guest

    Even if grossly wrong, these two power estimates should
    tell you the cards are in entirely different classes. The
    best numbers (for me at least), come from the Xbitlabs
    measurements of the cards. When that data is unavailable,
    I come here. These are "second best" guesstimates.


    HD 4870 150W
    8500 GT PCI-E 40W

    I would suspect a problem with the power measurement technique.

    The best device for the job, would be a Kill-O-Watt meter,
    as it is specifically designed for power measurements.
    And uses the same measurement techniques as digital power
    meters on the side of a house. Different versions of this
    are available, with different power plugs on them. More than
    one company makes them.


    The Kill-O-Watt is for doing power from the AC side of the problem.

    On the DC side, you can use a DC clamp-on ammeter (one with
    a Hall probe), and measure each similarly colored cable
    bundle on the internal power wiring. And that can give a good
    idea of power consumption. The info collected that way, is
    ignoring the inefficiency of the power supply. And certainly,
    a power supply fault could account for a high power draw.

    If you attempt to measure a video card that way (the way that
    Xbitlabs used to do it), measuring the power connector on the
    end of the card, only catches a portion of the power. Both
    12V and 3.3V rails are provided in the video card slot. The 12V
    rail comes from the two yellow wires on the 24 pin main connector.
    And the only other loads on that are fans. So "measure all fans",
    "measure two yellow wires", "subtract fans", gives 12V DC slot current.
    The 3.3V video slot power, is normally in the few watt range,
    and ignoring that is the best policy. To do that one properly,
    requires hacking in a shunt. And considering the power in that
    leg is small, it doesn't throw off power estimation that much,
    to ignore it. If you measure 3.3V entering the motherboard,
    you get the couple watts for the video card, plus 20-30W for the
    chipset. Or maybe the RAM power converter. What loads are on
    3.3V, varies with motherboard design. My old P4, powers
    the RAM converter off 3.3V. More modern systems would run
    it off 12V.

    And finally, use "thermal common sense". If the exhaust air
    from the 255W PC is not warm, then there can't be 255W
    being dissipated. I use common sense as a double-check
    for instrument readings. If the heat coming out, doesn't
    match your measurement, then the measurement is wrong.
    The heat has to go somewhere, and the level of heat
    should match the power reading.

    Paul, Aug 26, 2013
    1. Advertisements

  3. bruce56

    bruce56 Guest

    I was measuring power at the wall socket. The measuring device had been
    sitting on the shelf for about 6 months. I have since checked it by
    testing other items for which I measured the power consumption before.
    It is now giving rather higher readings, so I conclude this power
    meter has gone bad (and it is 3 weeks out of warranty).
    bruce56, Aug 29, 2013
    1. Advertisements

Ask a Question

Want to reply to this thread or ask your own question?

You'll need to choose a username for the site, which only take a couple of moments (here). After that, you can post your question and our members will help you out.