1. This site uses cookies. By continuing to use this site, you are agreeing to our use of cookies. Learn More.

Cooling Delta Temperatures

Discussion in 'Hardware' started by GrandeP, 6 Aug 2012.

  1. GrandeP

    GrandeP What's a Dremel?

    Joined:
    13 Sep 2011
    Posts:
    35
    Likes Received:
    0
    Call me a noob but what exactly is the purpose of delta temperatures? I know that you take the ambient temperature away from various cores/graphics cards.

    It is something I have never understood. :duh:
     
  2. Margo Baggins

    Margo Baggins I'm good at Soldering Super Moderator

    Joined:
    28 May 2010
    Posts:
    5,650
    Likes Received:
    268
    the delta temperature is the temperature difference between how hot something is and the ambient temperature.

    Like, my room ambient temperature is 20 degrees, my cpu is idling at 28 degrees, which is 8 degrees delta. Its a useful way of quantising cooling performance.
     
  3. theshadow2001

    theshadow2001 [DELETE] means [DELETE]

    Joined:
    3 May 2012
    Posts:
    5,226
    Likes Received:
    162
    By stating the delta temperature, you ignore the effect that the ambient temperature has on the cooling and so removing that variable from the benchmark. If you just stated temperatures as they are read from the system the results of a test would vary wildly depending on whether the room was hot or cold, done in summer or winter or even the time of day.
     
  4. doyll

    doyll Minimodder

    Joined:
    1 Jul 2011
    Posts:
    730
    Likes Received:
    16
    I think of it as the opposite.

    Ignoring the ambient temp gives your temps that mean nothing where as subtracting ambient temp gives you the actual rise in temp.
     
  5. asura

    asura jack of all trades

    Joined:
    22 Apr 2009
    Posts:
    1,747
    Likes Received:
    78
    Ignoring wasn't best choice of words; you're both saying the same thing, remove the ambient temperature from the CPU load temperature to produce a figure which is easier to use in comparisons, imagine the graph with the ambient and load figures instead of just the delta T figures, they become harder to read.

    Some people go the extra mile and produce C/W figures, these are even easier to read and use from chip to chip.
     
  6. Shirty

    Shirty W*nker! Super Moderator

    Joined:
    18 Apr 1982
    Posts:
    12,571
    Likes Received:
    1,759
    Exactly, comparing absolute temperatures is arbitrary considering that one test could be performed in a room at 18°C and the same test in another room at 28°C. Clearly the absolute temperatures recorded in the second room will be 10°C hotter.

    We normalise this by removing the ambient temperature from the equation. Doing so will allow you to calculate the absolute temperature you would achieve using the same hardware in the room of your choice, simply by adding your room temperature on to the delta temps.
     
  7. doyll

    doyll Minimodder

    Joined:
    1 Jul 2011
    Posts:
    730
    Likes Received:
    16
    CPU temp - ambient / watts disipated (how much ur cpu outputs) = C/W. So have to use some software to determine CPU wattage? No thanks.

    I'll stick to using CPU temp - ambient = delta. I don't know of anyone who can tell me how many watts my CPU is using ;)

    Something I never see is what the humidity is during testing. Moist air cools better than dry air because the moisture absorbs heat. ;)
     
    Last edited: 7 Aug 2012
  8. Shirty

    Shirty W*nker! Super Moderator

    Joined:
    18 Apr 1982
    Posts:
    12,571
    Likes Received:
    1,759
    Of course, no two chips are exactly the same, so it's always worth bearing in mind YMMV and taking the published figures as a guideline rather than gospel ;)
     
  9. asura

    asura jack of all trades

    Joined:
    22 Apr 2009
    Posts:
    1,747
    Likes Received:
    78
    Not difficult to get C/W at all! Simply build a small reservoir around your exposed chip add 1g of water, measure the time (t) it takes to raise the temperature of the water by 1C, divide 4.186 by t and you have your answer! :lol:
     

Share This Page