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.

Is liquid cooling any good?

Discussion in 'Overclocking' started by James A. Donald, Dec 7, 2003.

  1. Most modern CPUs have the capability to report their temperature --
    shows up in bios stats.

    Seems to me as measured by CPU self report, liquid cooling, which
    costs over two hundred dollars, does not get you anything markedly
    better than air cooling, which costs about forty dollars.

    Has anyone obtained worthwhile benefits from liquid cooling?

    Seems to me that ninety percent of cooling is getting good thermal
    contact between the heat sink and the CPU, and the rest is merely a
    detail.

    Of course liquid cooling will give you some benefits on the disk
    drives, but you have to space the disk drives, and if you spaced them,
    you would get some decent air cooling without doing anything special.
     
    James A. Donald, Dec 7, 2003
    #1
    1. Advertisements

  2. James A. Donald

    Adam Webb Guest

    quieter
    handles more heat
    more fun
     
    Adam Webb, Dec 8, 2003
    #2
    1. Advertisements

  3. James A. Donald

    Hank Guest

    Yeah good for a extra couple hundred MHz's.

    Hank
     
    Hank, Dec 8, 2003
    #3
  4. James A. Donald

    mrdancer Guest

    You can easily put one together for under $100. Some DIY'ers do it for next
    to nothing.
    The main benefits to watercooling are quietness and better cooling for
    extreme overclocking. However, for the vast majority of systems out there,
    a good air cooler is sufficient.
     
    mrdancer, Dec 8, 2003
    #4
  5. James A. Donald

    Citizen X Guest

    (James A. Donald) wrote in
    I recently embarked on my 1st liquid cooling project, I spent $17.53 USD on
    the system and am (was) very happy with it. (
    http://www.nathanlandry.com/LiquidCool )

    I like the reduced nose level as swell as the overclocking potential it
    opens up.

    N8
     
    Citizen X, Dec 8, 2003
    #5
  6. Remember liquid cooling is the movement of heat from a
    hot object to another object which itself is still cooled by fans.

    Reason being that temperatures are not huge, and so the temp
    delta between ambient & water isn't so great that you could get
    away with a reasonably sized convection stack.

    What it does allow is an improved enclosure for the radiator:
    o No reason not to mount the radiator separately in a ply cabinet
    o Use a simple perforated metal or mesh front/rear grill on it
    o Baffle in front of & behind the fans inside re noise
    o Use several small low-noise 80mm or 92mm fans
    ---- or 120mm - depends on your radiator size
    o Consider using proper oil coolers re size/quality
    ---- Mocal & Earl come in 115mm & 235mm width
    ---- the 235mm will take 3x 80mm fans ok
    ---- you can vary the height from 5-57 rows, matching # fans

    This allows the very-remote location of the unit from the PC/user.
    So whilst water cooling is an extra step between heat / water / fans,
    it allows you to add distance and even utilise the heat more effectively.

    Remember a few things tho:
    o Quality re pump
    o Large bore tubing
    o Anti-bacterial in the water system
    o Simpler is better

    You can get a silent system with fans - it just takes a bit of effort,
    and case makers could do more to help with their designs. For an
    overclocking system, that's a different requirement re exponential
    cooling need to achieve beyond original clock-speed performance.
     
    dorothy.bradbury, Dec 8, 2003
    #6
  7. James A. Donald

    BigBadger Guest

    --
    *****Replace 'NOSPAM' with 'btinternet' in the reply address*****
    In my experience this is not so...I've used both high end air cooling
    (SLK900u with ducted 120mm fan) and high spec water cooling. My temperatures
    are about the same either way as is my overclocking limit.
    Water used to have a definite edge but with the latest air cooled heatsinks
    the difference is little or nothing. Remember water cooling ultimatly uses
    air to cool anyway...the water is just a way to transport the heat from 'A'
    to 'B' (unless you use a total loss system).
    There is also the argument that water cooling is quieter...this may be true
    to some extent but you can still achieve a quiet air cooled system by using
    a large slow spinning fan and a duct.
    As proof of the above, I'm running a Barton 2500+ at 2500MHz with 2.06
    Vcore, aircooled, right now...my temp as I type is 36C and it only rises to
    48C under Prime 95 torture test...Plus its not too noisy. I've yet to see a
    water cooled system do significantly better.
     
    BigBadger, Dec 8, 2003
    #7
  8. --
    Congratulations.

    You did not seem entirely happy with it. For lack of a safety,
    your CPU and motherboard was destroyed.

    Most modern CPUs can be set to automatically shut the entire
    system down for an hour when the CPU reaches a certain
    temperature. It is a bios setting, which for some strange
    reason usually defaults to let-it-fry. I always switch it to
    do-not-let-it-fry.


    --digsig
    James A. Donald
    6YeGpsZR+nOTh/cGwvITnSR3TdzclVpR0+pr3YYQdkG
    urppO5hKjep2sce9Swu0fEk/6YQiikPAtEH1nR/X
    49x+jdNp48OBiQFq/eJyGQbmSshR9YXMLh1TVi9z7
     
    James A. Donald, Dec 9, 2003
    #8
  9. James A. Donald

    Muttly Guest

    I don't think anyone said that with the right addative in your water
    (something that reacts to uv light), a uv lamp and a windowed case water
    cooling looks damn sweet!
    Martin.
     
    Muttly, Dec 9, 2003
    #9
  10. James A. Donald

    Adam Webb Guest

    Yeah good for a extra couple hundred MHz's.
    ok.... two setups different cooling setups, exact same system

    Barton 2500
    A7N8X
    PC3200 ram (cheap stuff at 2-3-3-8)
    Coolers = Alpha heatsink with 80mm fan or Atlantis waterblock and Radiator

    Max OC Water:
    2400 @ 1.8volts
    2300 @ 1.7volts (used 24/7 for 3 months)

    Max OC Air
    2000 @ 1.7volts unstable
    2000 @ 1.8volts unstable
    1900 @ 1.7volts unstable
    1900 @ 1.8volts, and still might have to back down, im still testing it atm.

    This compares well to other experiences with my other setups for instance:

    XP2100 @ 2133mhz with water
    and only 1560 with air (yup, thats right below stock..... granted its
    sitting a 70c with some silent cooler)

    or

    XP1700 @ 1800mhz with water
    and only 1533mhz with air (another Alpha)
    the above Barton 2500 setup had a max temp of 39c at full load while at
    2300mhz with the watercooling, its now sitting at a toasty 49c at 1900mhz.
    Temps are "high" due to using onboard probe and typical Asus high temp
    readings.....

    atm im using my Athlon 64 with its stock cooler plus 4 case fans its nice a
    cool but relatively noisy. Since i dont think im going to get over 210fsb
    which im stuck at now, watercooling wont bring me anything other than
    silence...which is why i cant wait for the latest Swiftech 5002 to reach
    England.....Cant wait for the silence again :)
     
    Adam Webb, Dec 9, 2003
    #10
  11. James A. Donald

    Hank Guest

    If the #s you state are correct and your still able to hear the telephone
    when it rings and not have to search through the smoke to find it, I bow to
    you for your the God of all air coolers!
    There are plenty of water coolers out there that would kill for #s like
    that, especially at Prime 95 stable.

    Nice Going,
    Hank
     
    Hank, Dec 9, 2003
    #11
  12. James A. Donald

    BigBadger Guest

    --
    *****Replace 'NOSPAM' with 'btinternet' in the reply address*****
    Then I guess I must be a god because the numbers I'm quoting are genuine and
    it's Prime95 stable......here's a screenshot:
    http://mysite.freeserve.com/abz_piper/2507.jpg
    Now I don't know any way to capture noise on a screen shot but you can see
    that the fan is only running 3409rpm...so I'm sure you understand that it's
    not too noisy.
    My 'secret' is using a case side mounted fan to draw cold air from outside
    the case and duct it onto the heatsink...A lot simpler and cheaper than
    water cooling and I agree with your final point...There plenty of water
    cooling users who don't get anywhere near as good results, which for me
    proves the point that water isn't all it's cracked up to be.
     
    BigBadger, Dec 9, 2003
    #12
  13. James A. Donald

    BigBadger Guest

    I can only refer to my experience...And that is that I'm running faster
    (2500MGHz), higher voltage (2.06V) and the same peak temps (48C) as your
    water cooling....but I'm on air !!!!. My air cooling cost a fraction of what
    a decent water rig would cost and took a fraction of the time to fit....Also
    other than blowing out dust from time to time it's maintenance free. It's
    also not noisy because I use a 120mm fan at 3000-3500rpm.
    Here's a screen shot: http://mysite.freeserve.com/abz_piper/2507.jpg
     
    BigBadger, Dec 9, 2003
    #13
  14. James A. Donald

    Adam Webb Guest

    I can only refer to my experience...And that is that I'm running faster
    same temp? thats 10c higher........
     
    Adam Webb, Dec 9, 2003
    #14
  15. James A. Donald

    Bigbadger Guest

    I should heve read your post more thoroughly Adam....I see your
    reported temps are indeed lower than mine....however reported temps
    can be misleading the temps temp i get from a probe touching the cpu
    core are always 3 or 4 C cooler than MBM reports..also if I drop down
    to 1.8V / 2400MHz (which I can run no problem)I take the temps down by
    a good 5 or 6 degrees. I'd be interested to see your temps at 2.06V.
    I would still be confident that running the same speed and voltage
    there would be nothing in it as far as temps are concerned...and as
    far as getting an extra 200MHz overclock with water...I've yet to see
    anyone running at 2700MHz with a water cooled XP2500+
     
    Bigbadger, Dec 9, 2003
    #15
  16. James A. Donald

    Arthur Hagen Guest

    48 - 39 = 10? Remind me not to send my kids to your school. :)

    That said, it appears that BigBadger did one worse, and apparently saw his
    own text quoted in your posting, and based his 48C on that... *grin*

    Regards,
     
    Arthur Hagen, Dec 9, 2003
    #16
  17. James A. Donald

    Adam Webb Guest

    far as getting an extra 200MHz overclock with water...I've yet to see
    and your the first aircooled 2500 @ 2500 ive seen :)

    well done :)
     
    Adam Webb, Dec 9, 2003
    #17
  18. Thanks :) Do you think a "Y" good or bad (I've heard differing opinions on
    this ...)? I've got two AQXFA-Y week 34 2500's in my duallie setup (one
    SPMW, one RPMW) and they happily do 2100MHz at stock voltage with a TT
    Silent Boost. I'd go higher and up the voltage, except both sit at (die
    temps) about 60 deg C idle and 70 deg C under load. Room temperature is a
    about 30 deg C and case temp is about 38 (only one case fan, and a 120mm fan
    + a 80mm fan in the PSU, which is pushing it for a duallie ...). However,
    given that the temperature reported by my Soltek 75DRV5 for a single CPU are
    a good 20 deg C lower (also die temps) I've completely lost faith in the
    whole temperature measuring capability of motherboards, and gone for the "if
    it's stable then it's fine" mantra.

    I'm dipping my toes into the watercooling pool with the XP1700 that got
    replaced by the duallie. If this works out, I might try making a couple of
    waterblocks for my duallie, and see how much higer it can go. I'd try
    ducting and the like, except large HSFs don't like my board, and there also
    wouldn't be much left of the side of my case with two 120mm holes cut in it
    :)
     
    Michael Brown, Dec 9, 2003
    #18
  19. You're comparing your socket temps to his die temps. Which means that any
    comparison is completely stupid and irrelevant. That said, it's very
    impressive that you got a Barton 2500 to 2500MHz on air. Many 2500's won't
    do 2500 regardless of what they're cooled with. Do you have the three codes
    (AXDA etc) on you by any chance?

    [...]
     
    Michael Brown, Dec 9, 2003
    #19
  20. James A. Donald

    BigBadger Guest

    I do agree that comparing temps reported by one board against another is
    pointless...in the case of my board (Abit NF7-S) you can't even compare
    temps using the same board but different bios because they keep
    re-calibrating the thermister. That said I'm running the latest d20 bios
    which reports a good 10C higher than the early bios versions...My probe
    which I have touching the CPU reads 3 or so degrees C lower.

    My cpu is a AQXFA wk34..I'm afraid I don't have the other numbers other than
    it's not a 'Y'.

    --
    *****Replace 'NOSPAM' with 'btinternet' in the reply address*****
     
    BigBadger, Dec 9, 2003
    #20
    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.