Case Fan Guide for Better Temperatures

Discussion in 'Die-hard Overclocking & Case Modifications' started by Psychlone, Sep 3, 2009.

  1. Rotary8

    Rotary8 Master Guru

    Messages:
    226
    Likes Received:
    0
    GPU:
    SLI 680
    i have the reverse airflow with my case. when i installed the corsair H50, i had to rethink how that would work but the corsair 800D really thinks outside the box when it comes to cooling your case. the best i seen :)
     
  2. YuKsS

    YuKsS Maha Guru

    Messages:
    1,418
    Likes Received:
    0
    GPU:
    Asus gtx 1070 strix oc
  3. Psychlone

    Psychlone Ancient Guru

    Messages:
    3,688
    Likes Received:
    0
    GPU:
    Radeon HD5970 Engineering
    ^ I would have liked to see some temps on that.

    A little positive pressure is good, but too much makes the fans fight and reduces their efficiency...this really is rocket science, but is easy to imagine and see in real life.

    Put 2 fans in a closed system with 1 outlet and draw smoke or fog through it - you'll see that the smoke/fog gets chopped up and becomes really turbulent between the 2 fans, where it doesn't get so turbulent if you remove one of the fans.

    Anyway, good video - again, I'd really like to see some before/after temps rather than a "pinwheel test".

    Psychlone
     
  4. Rmmccoy

    Rmmccoy New Member

    Messages:
    1
    Likes Received:
    0
    GPU:
    ATI Radeon HD 4850 1GB
    Has the reversed jpg been moved? The previous jpg comes up fine but even it I try to just insert the url for the second, I get a 404. I guess one picture really is worth a thousand words.
     

  5. stricker

    stricker New Member

    Messages:
    2
    Likes Received:
    0
    GPU:
    nvidia 450 1gb
    fans

    I fined that i get best temps when i have all my fans eahausting the air out. This creats a nagitive air flow; this pulls more hot air out! :)
     
  6. tuco

    tuco Ancient Guru

    Messages:
    3,018
    Likes Received:
    0
    GPU:
    2 460's
    Lol... are you sure thats the best?
     
  7. peg66

    peg66 New Member

    Messages:
    1
    Likes Received:
    0
    GPU:
    ATi HD6850
    This is my first post here so hello all! :grin2:

    I reckon the problem is broader than that - many genuinely don't seem to understand that cooling is in essence about heat transference (i.e. that heat isn't "magicked away" by heatsinks and fans, it's simply shuffled elsewhere) and because of that continue to think only in terms of 'gains' (something got cooler) rather than 'trade-offs' (somewhere else must have got hotter). Unfortunately even those who understand the basic principles involved often treat system cooling as little more than an afterthought (most evident with case selection, where even aesthetics is given a far higher priority than ventilation by many these days) and it's not until the issue is forced upon them, when they hit that point where throwing more money at bigger heatsinks and fancier fans gets them nowhere (generally when they overload their cases capacity to exhaust hot air quick enough,) that they scramble for fixes. That's not to say that many people get things wildly wrong in the end, just that too many seem to get things right more by serendipity than design, which is a pity really as heat management doesn't get all that complicated at the level most of us play at, and we all know there's only one Golden Rule - '"Out from the rear" with Golden Rules!' :rpg:

    I know it's a pedantic point but I think it's worth clarifying: the only imperative for any fan in a given case is that it does it's part reliably and effectively - it doesn't necessarily have to push lots of air to do that, and, as ascl points out, in certain situations shouldn't. Fans should be selected for a purpose, purpose usually dictates the specific type and size of a fan, and cost usually decides the exact specs (RPM, CFM, dB) you end up with (and the fans potential efficiency at a given voltage). It's actual effectiveness of course depends what you do with it - as you and others rightly stress, aim it badly and it's "on paper" efficiency means nothing.

    I'm glad you said usually! I was desperate enough :)flame2: uninsulated house during a scorching Perth heatwave) to make the following changes to my old system (a Centurian 5 ATX case, AM2 mobo/95W X2 w/ Mugen2, HD4850, Zalman wrapped HDD, drive bays removed from the case):

    Code:
      ------------------	------------------	--------------
      =                =	=    HDD HDD >PSU>	=        >PSU>
      =           =PCIe=	|D.V.D.      >PSU>	=        >PSU>
      >                =	=                =	=            =
      >    =GFX==COOLER=	>           =PCIe=	=       =PCIe=
      =    VREG/GPU/RAM=	>                =	>            =
      =         NB     =	>    =GFX==COOLER=	>=GFX==COOLER=
      =     R  <CPU<   <	=    VREG/GPU/RAM=	=VREG/GPU/RAM=
      =     A  <CPU< P <	=         NB     =	=     NB
      =     M  <CPU< C <	=     R  <CPU<   <	= R  <CPU<   <
      |D.V.D.      >PSU>	=     A  <CPU< P <	= A  <CPU< P <
      =    HDD HDD >PSU>	=     M  <CPU< C <	= M  <CPU< C <
      ------------------	------------------	---------------
    
    By flipping the case upside down and reversing both the rear case fan and CPU HSF I got better than expected improvement to all temps in the entire lower section of the motherboard (NB only marginally), and the increase in flow seemed to offset the extra heat convecting into the expansion bay section of the case (enough to help stabilize gfx temps; still +5*C at load but temps had been at melting point before I reversed the lower fans so I took it as a win). After experimenting with extra high-front fans I realised I only needed extra (i.e. quicker) venting high in the case, not extra airflow, and opted to drill out the mobo tray/back panel and made a proper "reverse-ATX" case out of it. That returned the PSU to venting duties (instead of feeding the intake :bugeye:) which made all the difference: the result was a further drop in lower case temps and a remarkable drop in upper case temps (gfx card temps were only slightly better than ATX at that stage but, more importantly, the fan on the cooler was a whole lot less chatty, one of the issues that prompted the exercise in the first place).

    By the time I added a Silenx IXG-80HA2 gfx cooler (sexy in every way; still pimpin on my 6850 I'll add), swapped out the front 80mm intake for a 120mm and tweaked fan speeds a little I admit I was rather impressed with two initially unintentional aspects of my design. First, with the 120mm rear intake being constant speed and the matching (and very close) HSF being variable I lucked onto an arrangement where at idle both NB and MOSFET were pushed air yet at load it seemed heat was instead being pulled from their heatsinks! However it worked the big upshot was that there was very little variance in temps at that end of the board - regardless of load. I also noticed that the heat load making its way up the case made little difference to expansion bay/gfx temps; at idle the convected heat definitely improved airflow around the gfx cooler, at load, while hotter, less of it made it up the case, because as load increased so too did the volume of air pushed out of the front case vents.

    The third illustration shows what I was left with after I attacked the case with a junior hacksaw (offing the entire drive bay section of the case - all in the name of science of course!) and sectioned off the case with a stiff bit of card between the NB and gfx card. For the curious gfx temps rose again by a sizable amount, with the smaller case volume equating to higher expansion bay temps, and moving the front fan much closer to the back of the graphics cooler wasn't anywhere near as effective as I'd imagined it would be. I did manage to claim some of that back (by swapping the 80mm fan back in) but at least now I know the value/effect of the volume of a/that case.

    That's where my experiments with that particular case stalled. My next step is to proof air ducting to a gfx card (the IXG-80HA2 cooler being an ideal test subject due to the orientation of the fins) and I've decided that my trusty Centurian 3 wasn't the best platform to use so am looking at other options. I'll probably create a custom case rather than customise an off the shelf model because I'm also dead keen on testing laminar flow (turbulance is EVIL!!!! :D) but I admit I'm still pipe-dreaming that part atm - I'll be looking forward to help and feedback with that project later down the track.

    Enjoy yourselves, Nick.

    [Edit] These pics are barely better than my crappy ascii art but the combination should better illustrate why reversed airflow was much more efficient than front to back airflow (in the "motherboard" half of the case at least) with the parts I had, and why that held true regardless of case size or motherboard orientation.

    The following shows the system in a state similar to the one described by Psychlone - reversed rear chassis and CPU fans in an ATX case:

    [​IMG]

    The next shows two comparable reverse-ATX variants; the one on the left while still in planning mode, the rightmost dummied up for inspiration just before the system was stripped to onboard w/stock HSF and retired as an regular ATX system (so it's the tray, not the front case fans, that's misaligned in that image):

    [​IMG]
     
    Last edited: Oct 8, 2011
  8. adjago

    adjago New Member

    Messages:
    3
    Likes Received:
    0
    GPU:
    evga gtx285 1gb
    nice thread master i wll do it thanks!!!
     
  9. Edamski

    Edamski Member Guru

    Messages:
    107
    Likes Received:
    0
    GPU:
    Palit 460 GTX 1gb
    Nice little read!
     
  10. calob0903

    calob0903 Member

    Messages:
    18
    Likes Received:
    0
    GPU:
    GTX 770 2gb
    Is it ok to boost my graphics card fan speed to 70% all the time??

    I have a gtx 770
     

  11. ErfanDL

    ErfanDL New Member

    Messages:
    9
    Likes Received:
    0
    GPU:
    ASUS 5850 Direct CU TOP
    nice guide
     
  12. dsbig

    dsbig Ancient Guru

    Messages:
    3,849
    Likes Received:
    3
    GPU:
    SAPPHIRE NITRO 390x
    case with back exhaust dont work too well, when using 2 evga ACX cards in sli.


    with my HAF X case

    the back exhaust fan blows out hot air and the gpu coolers suck it right back in, causing the temps to skyrocket.



    I switched rear exhaust to intake and one of my modded front bay fans to exhaust and also turned the fan around on the cpu cooler.


    and gpu is now around 5-10c cooler when gaming

    and the difference does matter when the top gpu was hitting almost 90c before when playing stressful games.
     
  13. EspHack

    EspHack Ancient Guru

    Messages:
    2,331
    Likes Received:
    0
    GPU:
    ATI/HD5770/1GB
    yea.. but it will make noise and wear out the fan in a few years
     
  14. chiadretti

    chiadretti Member

    Messages:
    15
    Likes Received:
    0
    great read
     

Share This Page