1. Welcome to TechPowerUp Forums, Guest! Please check out our forum guidelines for info related to our community.

Watercooling loop order

Discussion in 'Overclocking & Cooling' started by D3M0N-G4M3R, Aug 4, 2009.

  1. D3M0N-G4M3R

    D3M0N-G4M3R New Member

    Joined:
    Jan 4, 2008
    Messages:
    91 (0.04/day)
    Thanks Received:
    6
    Location:
    Yorkshire, UK
    Ok im not too sure, i set up my first watercooling setup the other day and went with my instinct on the order however ive heard this second setup works better.

    The kit i used is:

    -Black Ice 240cm rad
    -Swiftech Apogee GT
    -XSPC 12v 10w pump/res (modded to 18w by tigger)
    -Tygun Tubing and Jubilee clamps


    My order is:

    PUMP/RES > CPU BLOCK > DOUBLE RAD

    What I heard was:

    PUMP/RES > DOUBLE RAD > CPU BLOCK

    Atm my temps are:

    IDLE
    [​IMG]

    LOAD
    [​IMG]

    These temps good or bad?

    Lastly what i want to ask is, ive got a spare single rad to slot in the loop somewhere for when i get a good gpu block for my 4870 non reference 1gb. What would the ideal loop setup be for this?
     
  2. erocker

    erocker Super Moderator Staff Member

    Joined:
    Jul 19, 2006
    Messages:
    39,648 (13.30/day)
    Thanks Received:
    14,039
    Best order is: Pump-->Rad-->CPU Block-->Res.-->Pump

    If you are adding another rad + gpu into the loop: Pump-->Rad-->Rad-->CPU Block-->GPU Block-->Res.-->Pump

    Temps look good. :)
     
    WarEagleAU and D3M0N-G4M3R say thanks.
  3. DanishDevil

    DanishDevil

    Joined:
    Oct 6, 2005
    Messages:
    10,203 (3.12/day)
    Thanks Received:
    2,090
    Location:
    Newport Beach, CA
    Is that because the pump heats the water up a bit?
     
  4. erocker

    erocker Super Moderator Staff Member

    Joined:
    Jul 19, 2006
    Messages:
    39,648 (13.30/day)
    Thanks Received:
    14,039
    It really does. I was leak testing without running fans through my radiator and I couldn't believe how warm things were getting.
     
    WarEagleAU and DanishDevil say thanks.
  5. D3M0N-G4M3R

    D3M0N-G4M3R New Member

    Joined:
    Jan 4, 2008
    Messages:
    91 (0.04/day)
    Thanks Received:
    6
    Location:
    Yorkshire, UK
    Ah ill go with that then when i finally do it, you sure that both rads should be next to each other in the loop, wont the water be hot on its journey from cpu to gpu?
     
  6. Icewind31 New Member

    Joined:
    Feb 11, 2006
    Messages:
    215 (0.07/day)
    Thanks Received:
    28
    Location:
    Boston Area, MA
    Loop order doesn't matter as long as res is before pump, put it in the order that makes the most sense in terms of layout in your case.

    Loop temps reach equilibrium over time. Only time you'll notice is when you first turn on the system and measure.
     
    Sadasius says thanks.
  7. D3M0N-G4M3R

    D3M0N-G4M3R New Member

    Joined:
    Jan 4, 2008
    Messages:
    91 (0.04/day)
    Thanks Received:
    6
    Location:
    Yorkshire, UK
    Ah i see, i think ill still go with erockers plan though and go res/pump rad, this pump does get rather warm so it might be better to dump off that heat and cool it completely through 2 rads.
     
  8. erocker

    erocker Super Moderator Staff Member

    Joined:
    Jul 19, 2006
    Messages:
    39,648 (13.30/day)
    Thanks Received:
    14,039
    Since the radiator cools the water, the water exiting the radiator will always be coolest. You do have a point though, it's most likely not much cooler to make a difference and I usually go with looks, organization of the loop before worrying about one or two degrees.
     
  9. Sadasius

    Sadasius New Member

    Joined:
    Apr 6, 2009
    Messages:
    263 (0.13/day)
    Thanks Received:
    140
    Location:
    Ottawa, Ontario
    YES!! Someone who knows. Res before pump and the rest just use the shortest route possible with less tubing possible. Temps out of a rad are maybe .5 degrees at best and no it makes no difference on the rest of the performance. It will not lower your CPU temps any better then the rest of the water. It's a flow and equilibrium thing...:D
     
  10. Binge

    Binge Overclocking Surrealism

    Joined:
    Sep 15, 2008
    Messages:
    6,982 (3.18/day)
    Thanks Received:
    1,752
    Location:
    PA, USA

    it will matter if your put the rad before the cpu for about 20 minutes until your system reaches equalibrium. I learned that a few months ago, but what does matter is that your cpu block gets a lot of flow so you might see .5C temp differences between the two setups :p
     
  11. D3M0N-G4M3R

    D3M0N-G4M3R New Member

    Joined:
    Jan 4, 2008
    Messages:
    91 (0.04/day)
    Thanks Received:
    6
    Location:
    Yorkshire, UK
    5c less or more?
     
  12. Binge

    Binge Overclocking Surrealism

    Joined:
    Sep 15, 2008
    Messages:
    6,982 (3.18/day)
    Thanks Received:
    1,752
    Location:
    PA, USA
    .5C meaning 1/2 of 1C less OR more.
     
  13. D3M0N-G4M3R

    D3M0N-G4M3R New Member

    Joined:
    Jan 4, 2008
    Messages:
    91 (0.04/day)
    Thanks Received:
    6
    Location:
    Yorkshire, UK
    Oh lol haha that might not even be worth faffing with then lol
     
  14. newtekie1

    newtekie1 Semi-Retired Folder

    Joined:
    Nov 22, 2005
    Messages:
    19,919 (6.19/day)
    Thanks Received:
    6,020
    It doesn't really matter, but the pump does put add a little heat to the system. Maybe enough to raise the temps 1°C, but is that 1°C worth the hassle of re-arranging your setup?

    My Ideal loop is: Res--> Pump --> Big Rad --> CPU --> NB --> Small Rad --> Video Cards --> Res

    Usually, I do a 2x120 Radiator hanging off the back of the case, then a smaller 1x120 inside the font of the case where the front fan would go normally. Obviously, this changes depending on what Radiators I actually use in the loop.
     
    WarEagleAU says thanks.
    Crunching for Team TPU More than 25k PPD
  15. D3M0N-G4M3R

    D3M0N-G4M3R New Member

    Joined:
    Jan 4, 2008
    Messages:
    91 (0.04/day)
    Thanks Received:
    6
    Location:
    Yorkshire, UK
    yeah i was thinking of that setup, my big rad then cpu then small rad then gfx card.
     
  16. newtekie1

    newtekie1 Semi-Retired Folder

    Joined:
    Nov 22, 2005
    Messages:
    19,919 (6.19/day)
    Thanks Received:
    6,020
    And it is certainly not an equalibrium thing, temps do not "reach and equalibrium" in a water cooling loop.

    The overall temperature of the fluid might rise, however, the whole loop will not be the same temperature. There will be hot and cold spots in the loop. Generally, hot spots being right after the blocks, and cold spots being right after the radiators.

    However, you can only cool the fluid down so far. Once it reaches ambient temperature, it will not get any cooler. So, if you have a 2x120 radiator, and it is able to cool the fluid down to room temperature before the fluid exits the radiator, then putting a your 1x120 raditor right next to it in the loop is useless, and makes the 1x120 raditor completely wasted.

    However, if you put the 1x120 radiator between the CPU and GPUs, then the heat from the CPU will be removed from the fluid before it gets to the GPUs. This will give you lower GPU temps. However, if the 2x120 radiator can't remove all the heat from the GPUs before the fluid gets through the radiator, this could lead to higher CPU temps also.
     
    t77snapshot says thanks.
    Crunching for Team TPU More than 25k PPD
  17. D3M0N-G4M3R

    D3M0N-G4M3R New Member

    Joined:
    Jan 4, 2008
    Messages:
    91 (0.04/day)
    Thanks Received:
    6
    Location:
    Yorkshire, UK
    ah i understand, well ive got some ideas for the double rad anyways. im gonna have the fans pulling the air through rather than blowing air through, ive heard that works better, the way my case is modded too allows the rad, res/pump to have there own section that is always filled with cold air, hopefully this will keep the rad below room temp by quite a bit.

    I have to agree with you on the equilibrium thing, i think that is nonsense too, ofcourse there will be hotspots.
     
  18. Sadasius

    Sadasius New Member

    Joined:
    Apr 6, 2009
    Messages:
    263 (0.13/day)
    Thanks Received:
    140
    Location:
    Ottawa, Ontario
    What happened? You were doing so well until this post. The only time it is not an equilibrium thing is when you have the worst flow in the world. That way there is actually time to have these hot spots in the loop. But in real world builds it is not like that and I can prove it on my system as well as I have water temp probes after the blocks and also after my rads. Now I mentioned a .5 degree difference between the blocks and the exit of the rad but that is from a rather large loop....My loop which has 2 pumps, 2 rads, 3870X2 full face GPU bock, SB block, NB block, MOSFET blocks and my CPU block. If you look in the picture of my build you will see wires near my pumps. These temp probes are in place specifically for the purpose of this topic as I wanted to show some people first hand.
     
  19. Binge

    Binge Overclocking Surrealism

    Joined:
    Sep 15, 2008
    Messages:
    6,982 (3.18/day)
    Thanks Received:
    1,752
    Location:
    PA, USA
    I can also prove with relative certainty that it does not matter the loop order unless there is a problem with flow, which I did address in my TERRIBLY WORDED post. Here are the points in bullet:

    • Pumps before the most restrictive block may have some effect on the equalibrium
    • Placing a part before or after the radiator has little bearing on the water temperature
    • Equalibrium is reached and is usually reached best when there are adiquate parts to suit the need.
     
    t77snapshot and rampage say thanks.
  20. Sadasius

    Sadasius New Member

    Joined:
    Apr 6, 2009
    Messages:
    263 (0.13/day)
    Thanks Received:
    140
    Location:
    Ottawa, Ontario
    Binge you know your stuff. Yeah I lost an argument with Vapor on XS awhile back about loop order in which I thought the highest pressure from the pump directly to the CPU would be best and he proved me wrong and that it actually did not matter if it was the rad before the CPU or direct from the pump. It's how everyone learns. It was funny too because I stuck to my guns till the end! :laugh:
     
  21. newtekie1

    newtekie1 Semi-Retired Folder

    Joined:
    Nov 22, 2005
    Messages:
    19,919 (6.19/day)
    Thanks Received:
    6,020
    So the fluid leaving the Radiator is the same temp as the fluid entering the raditor? Then what is the point of the raditor? Oh...it is supposed to cool the fluid...but according to you it doesn't...:laugh:

    Sorry, it doesn't work that way, the radiator cools the fluid as the fluid passes through it, the exception might be when running extremely low velocity fans or no fans at all, or if the radiator/s are overloaded. If you get to the point where you are getting an equalibrium, you have overloaded your loop.
     
    Last edited: Aug 6, 2009
    Crunching for Team TPU More than 25k PPD
  22. Sadasius

    Sadasius New Member

    Joined:
    Apr 6, 2009
    Messages:
    263 (0.13/day)
    Thanks Received:
    140
    Location:
    Ottawa, Ontario
    Don't think you grasped it yet......Water is a great conductor of heat. It picks up very quickly and can also get rid of it quickly in various manners and forms. When in a loop, the water is going relatively quick. It does not have to time to sit inside a rad or a block. It picks up heat and transfers it very quickly. Hence why the lack of a temp difference in a loop. It is only when you have very poor flow or circulation in your loop will you see hot spots and usually by then a hose will blow off or a res will crack. A rad basically cools the whole loop down at the same time because of the way water disperses heat so quickly plus it's flow. You are right it's just the 4th dimensional thinking that is lacking which is time. The fluid might pass through the rad 40 times before it drops a degree when your load goes down.
     
    Last edited: Aug 6, 2009
  23. shevanel

    shevanel New Member

    Joined:
    Jul 27, 2009
    Messages:
    3,479 (1.85/day)
    Thanks Received:
    406
    Location:
    Leesburg, FL
    I want to ask this..

    Right now I have rad-res-pump-cpu-rad

    my block is probably restrictive, an APOGEE GT, would it be best if I switched the two lines on the pump and cause the radiator to be before the CPU?
     
  24. Binge

    Binge Overclocking Surrealism

    Joined:
    Sep 15, 2008
    Messages:
    6,982 (3.18/day)
    Thanks Received:
    1,752
    Location:
    PA, USA
    Shev your CPU isn't very restrictive at all and your loop order is simple and fine. It would do nothing to change the config of your loop.
     
    shevanel says thanks.
  25. shevanel

    shevanel New Member

    Joined:
    Jul 27, 2009
    Messages:
    3,479 (1.85/day)
    Thanks Received:
    406
    Location:
    Leesburg, FL
    Thanks. I might pick up a triple rad though.. i'd like to do better than 50 at idle
     

Currently Active Users Viewing This Thread: 1 (0 members and 1 guest)

Share This Page