Page 1 of 2 12 LastLast
Results 1 to 15 of 25
  1. #1
    Joined
    Apr 2007
    Location
    chi town
    Age
    24
    Posts
    48

    Loop or Parallel Watercooling?

    Currently I'm cooling only my CPU, but I have a 110 gph pump, and I am thinking about taking out my heatpipes and cooling my MOSFET and NB. I may want to do my GPU's also.

    Would You recommend a huge loop or a y-splitter at the pump and two parallel (NB and CPU together, GPU MOSFET together)?

    THanks
    I like to make werewolf movies!

    The Lord will return soon and destroy this wicked planet with fire and brimstone, then it won't matter what the fastest of anything is.
    SYS 1:
    eVGA 680i A1 | E6600 @ 3.7ghz 1.41v | 2gb OCZ Reaper 4-4-4-15 @ 900mhz
    | SLI eVGA 8800 Ultra's | BFG 1000wt PSU| tt kandalf case | Razer AC-1 | 250gb SATA | G-Flow waterblock |
    WATERCOOLED


    SYS 2:
    Asus P5N-E sli | e6400| 2gb OCZ gold 667mhz|BFG 7900 gs 256mb | AeroCool zeroDBA 620w PSU | Lian Li case | OCZ Vindicator HSF| Soundblaster X-fi

    SYS 3:
    ABIT IP35 Pro| q6600| 2gb Corsair xms2 800mhz|evga 8800 gts 320mb| OCZ gameXstream 600w | custom case | OCZ Vindicator HSF

  2. #2
    Joined
    Dec 2002
    Location
    Blaine, WA
    Age
    60
    Posts
    30,989

    Re: Loop or Parallel Watercooling?

    I would do a loop.
    Main Rig: EVGA 141-BL-E769-A1 LGA 1366 Intel X58 CLASSIFIED/WATERCOOLED ED. cooled by the Monster Water Cooling Setup
    HTPC One BIG Case Asus 8 CORE Monster
    ASUS: 5 * BIOSTAR: 1 * CHAINTECH: 1 * EVGA: 3 * GIGABYTE: 5 * SUPER MICRO: 1 *TYAN: 2
    180+ GHZ total power for PC Perspective Killer Frogs Rosetta @ Home Team as The Uncle B's!!




    Spoiler!

    I'm the Uncle your Aunt won't talk about. Go ahead and pull my finger!

  3. #3
    Joined
    Mar 2004
    Location
    mtl,canada
    Age
    36
    Posts
    37,301

    Re: Loop or Parallel Watercooling?

    is that better for the flow?

    add a couple 120mm fans overhead to drop temps by 10+ celsiusD.S.C-12(2)-disclaimer : whatever u do with your hardware/software is your
    responsibility, which i dont hold if u break anything

    GRAB here OcBible v1.55 and Guidemania v1.21
    scary pooch?pooch#1taking ball from dogpooch pulln on ropeme on board after 2 years
    of no board
    http://gonny.se/ <-cool lady http://www.mediafire.com/?8dmg21ms1vx
    IN MEMORY OF HOW BM WAS TREATED HERE...

  4. #4
    Joined
    Dec 2002
    Location
    Blaine, WA
    Age
    60
    Posts
    30,989

    Re: Loop or Parallel Watercooling?

    Well, it depends, but anytime you use a split, you allow possibilities for flow to be diverted due to restrictions. I use a loop wherever possible. One instance where I haven't is a dual CPU system. It works here because the resistance of each block is identical.

    Main Rig: EVGA 141-BL-E769-A1 LGA 1366 Intel X58 CLASSIFIED/WATERCOOLED ED. cooled by the Monster Water Cooling Setup
    HTPC One BIG Case Asus 8 CORE Monster
    ASUS: 5 * BIOSTAR: 1 * CHAINTECH: 1 * EVGA: 3 * GIGABYTE: 5 * SUPER MICRO: 1 *TYAN: 2
    180+ GHZ total power for PC Perspective Killer Frogs Rosetta @ Home Team as The Uncle B's!!




    Spoiler!

    I'm the Uncle your Aunt won't talk about. Go ahead and pull my finger!

  5. #5
    Joined
    Aug 2004
    Posts
    10,583

    Re: Loop or Parallel Watercooling?

    It all depends.
    NB coolers are usually low resistance and mosfet coolers (at least the ones I've seen) high, like memory cooling.
    In a loop (series) the high resistance pieces would choke the entire system, and if done in parallel the NB cooler along with the high resistance mosfet coolers would get minimal flow, as most would go to the low resistance side.

    Also 110gph (less than 2gal/min with 0 load) with out a very high head # is very weak, and most likely not suitable for multiple blocks. 1-2 gal/min is generally the best performance range. Your blocks will have a lot to say about that.

    Anyway, in a parallel situation the balance is important and you need to carefully consider what the flow will be using the 1/f=1/r1+1/r2... formula which requires acccurate resistance #'s to give a realistic figure. You need to be sure each loop gets enough water to do it's job.

    If the mosfet coolers are low resistance I would say series is the way to go (with a more powerful pump), totally separate loops (different pump) if not.

    A lot depends on what you have, and how much you are willing to spend. Memory heatsinks (air) do a good job on mosfets.


    "The significant problems we face cannot be solved at the same level of thinking we were at when we created them."
    - Albert Einstein (1879-1955)

  6. #6
    Joined
    Feb 2001
    Location
    Santa Cruz, CA
    Age
    45
    Posts
    340

    Re: Loop or Parallel Watercooling?

    Flow and the change in temperature are important things to keep in mind while you are building your loop system..... For example a loop that looks like this

    pump->CPU-->VGA-->MEM-->VR-->radiator-->(back to pump)

    may or may not not cool as well as a parallel loop like this;


    /--->VGA--->MEM\
    Pump----- ---------> (radiator---->(back to pump)
    \--->CPU----->VR/

    For example if you are using a small pump on the first system, your pump has to overcome the head loss of all of the tubing, fittings and water blocks in that loop one after another and in turn will pump at a lower rate (gpm)... If you add enough tubing and fittings, it wont pump at all.

    In contrast, by using the second loop configuration show, the pump has less resistance to go through by only having to pump through 2 blocks on each loop. True, the pump has to split the gpm between the 2 loops, but it will be pumping at a higher rate of speed anyhow, since it is pumping against considerably less head pressure..

    Another very important argument for a parallel system is the change in temperature of the water as it enters the system...

    In the first system (series)he water is coolest after it exits the radiator and passes through the pump.... Therefore the CPU, get s the most cooling since tthe cpu is hot and the water is cool .... But where did the heat go... Its now in the water... How cold is the water when it gets to the VR? At that point the water has been heated by every other device on the loop... How effectively is this water cooling the VR? Obviously, water that is 35 deg C will cool a heck of a lot better than water that is 55 deg C ... In fact none of you would ever think of cooling your cpu with 55 deg water right?

    One thing I have noticed in most water cooled systems are temperature monitors for supply and return temperatures.. This data would give you an effective measurement of how much heat your system is absorbing from your hardware...But for most of us on a budget this may not be an option...

    In general, a parallel system will work fine as long as the system is balanced, where you have the same number of blocks on each loop. Keep in mind that the CPU and video card should be the first items on any loops... Also as some one else mentioned, If you have block with a relatively low resistance (head loss) like a VR block, go ahead and throw a couple or three of them in series (After the primary block)...

    In my system , I have almost all devices run in parallel... I have the CPU on a loop. I have the VGA on a loop, I have 2 Koolance memory coolers in series on a loop, and I have 2 hard drive coolers in series on a loop.. I have excellent flow through each loop and each block/device gets the coolest possible water to ensure it can absorb the most heat form the device...
    ok i have to go to the sharks game.. bbl
    Scott
    LOTS of Koolance water cooling hardware/cases/coolers/ for sale cheap. PM if interested.


  7. #7
    Joined
    Apr 2007
    Location
    chi town
    Age
    24
    Posts
    48

    Re: Loop or Parallel Watercooling?

    Thanks alot, shredder. Thats clarifies a bunch for me. I;m just hesitant on adding more blocks. Heres what i have for a system. I have koolance pump/reservoir combo (1.8gpm pump) which pumps directly into my Innovatek G-flow CPU block, which has very low resistance. From there I go to my triple radiator, then from the radiator to a t-splitter where I have a fillport/fillport resivoir, then that al feeds into the main reservoir/pump.

    All of my tubes are 3/8 in., and as short as possible for my case without kinking.

    I would like to first-off, cool my NB, and maybe later do my GPU's.

    Would I be safe adding the NB block (I'll get one with the lowest resistance I can find) with my pump?

    And if I could add the VR & NB coolers, what kind of loop or parallel sequence would you think might give me the best cooling?
    Last edited by bareftjacky; 04-30-2007 at 10:08 PM. Reason: addition
    I like to make werewolf movies!

    The Lord will return soon and destroy this wicked planet with fire and brimstone, then it won't matter what the fastest of anything is.
    SYS 1:
    eVGA 680i A1 | E6600 @ 3.7ghz 1.41v | 2gb OCZ Reaper 4-4-4-15 @ 900mhz
    | SLI eVGA 8800 Ultra's | BFG 1000wt PSU| tt kandalf case | Razer AC-1 | 250gb SATA | G-Flow waterblock |
    WATERCOOLED


    SYS 2:
    Asus P5N-E sli | e6400| 2gb OCZ gold 667mhz|BFG 7900 gs 256mb | AeroCool zeroDBA 620w PSU | Lian Li case | OCZ Vindicator HSF| Soundblaster X-fi

    SYS 3:
    ABIT IP35 Pro| q6600| 2gb Corsair xms2 800mhz|evga 8800 gts 320mb| OCZ gameXstream 600w | custom case | OCZ Vindicator HSF

  8. #8
    Joined
    Aug 2004
    Posts
    10,583

    Re: Loop or Parallel Watercooling?

    Most likely you can add the nb cooler in series. The Vr block is the unknown, and you will have to figure it's resistance. In my opinion not worth doing.
    It sounds like shredder is using 3 pumps (with one rad?), which eliminates the problem, and is not really parallel. Good solution though, if expensive.
    the Temp rise problem is really insignificant in a series loop. My dual core, vid anf NB blocks raise the temp 1C. The problem is flow and balance.


    "The significant problems we face cannot be solved at the same level of thinking we were at when we created them."
    - Albert Einstein (1879-1955)

  9. #9
    Joined
    Feb 2001
    Location
    Santa Cruz, CA
    Age
    45
    Posts
    340

    Re: Loop or Parallel Watercooling?

    Well, at this point if you only have a CPU and NB cooler go ahead and run them in series.. Glasman has all of his devices on one series loop and in only seeing 1 deg tem rise on those devices.. That is plenty of cooling as long as that temperature difference is measured while under full load (ie gaming)... (Glasman , you can easily turn your pump down a bit till you get a delta t of 5 or 10 so and save some power,, you may even get more life out of your pump..)
    Maybe you will want to cool your VGA's and VR's on another loop when you get more blocks...
    BTW, my system is run with one koolance pump for all of the loops and works great. Water cooling is very forgiving as long as you have flow through your loops..

    If it were my system, i might run one loop cpu-nb-vr-vr-vr
    and another loop vga-vga-nb
    You will know as soon as you fill the system if the loops are flowing well -watch the flow of the air working out of it or get a flow meter...

    Remember , water cooling is very forgiving ...it does not take a lot of flow to cool your system as long as the water going to each item is cool to begin with..

    scott
    LOTS of Koolance water cooling hardware/cases/coolers/ for sale cheap. PM if interested.


  10. #10
    Joined
    Feb 2001
    Location
    Santa Cruz, CA
    Age
    45
    Posts
    340

    Re: Loop or Parallel Watercooling?

    Quote Originally Posted by TheGlasMan View Post
    Most likely you can add the nb cooler in series. The Vr block is the unknown, and you will have to figure it's resistance. In my opinion not worth doing.
    It sounds like shredder is using 3 pumps (with one rad?), which eliminates the problem, and is not really parallel. Good solution though, if expensive.
    the Temp rise problem is really insignificant in a series loop. My dual core, vid anf NB blocks raise the temp 1C. The problem is flow and balance.
    Your system is a good example of a working series circuit. But you need to be careful of what you are saying... Temperature rise is directly related to flow.. Slower flowing water will heat up more than fast flowing water.. If you are only seeing a 1 deg rise in temp of the water.. then you are pumping extra water needlessly through your system. this is not a bad thing, but the water that is going through your loop is hardly absorbing any heat from your hardware before it gets back to your radiator.. for an efficient heat exchange, we look for return temps to be about 5 deg higher than supply temps.. of course if you dont care about the power consumption (or noise) of your pump, then go ahead and run it at full speed.

    You are right about balance though.. I have found by using short parallel loops with not too many devices on each loop, I can take advantage of low head loss, where the pump has the least amount of 'work' to do and run at the highest rpm and yield the highest flow rate/gpm through the pump. This results in more than adequate flow rates through all of my loops. (i turn my pump down very low unless i am in an oc situation).

    Scott
    LOTS of Koolance water cooling hardware/cases/coolers/ for sale cheap. PM if interested.


  11. #11
    Joined
    Aug 2004
    Posts
    10,583

    Re: Loop or Parallel Watercooling?

    My system is performing to expectations, but I think we have different ideas of what we are trying to accomplish. I would be very upset with a 2C temp rise, 5C would seem to be steaming my components (relatively speaking of course ) I wouldn't want to run the fans needed to take 5C out of the loop anyway.

    PS I fold 24/7, even when gaming.


    "The significant problems we face cannot be solved at the same level of thinking we were at when we created them."
    - Albert Einstein (1879-1955)

  12. #12
    Joined
    Mar 2004
    Location
    mtl,canada
    Age
    36
    Posts
    37,301

    Re: Loop or Parallel Watercooling?

    hmm....that puts more than a cat in the bowl

    add a couple 120mm fans overhead to drop temps by 10+ celsiusD.S.C-12(2)-disclaimer : whatever u do with your hardware/software is your
    responsibility, which i dont hold if u break anything

    GRAB here OcBible v1.55 and Guidemania v1.21
    scary pooch?pooch#1taking ball from dogpooch pulln on ropeme on board after 2 years
    of no board
    http://gonny.se/ <-cool lady http://www.mediafire.com/?8dmg21ms1vx
    IN MEMORY OF HOW BM WAS TREATED HERE...

  13. #13
    Joined
    Feb 2001
    Location
    Santa Cruz, CA
    Age
    45
    Posts
    340

    Re: Loop or Parallel Watercooling?

    Quote Originally Posted by TheGlasMan View Post
    My system is performing to expectations, but I think we have different ideas of what we are trying to accomplish. I would be very upset with a 2C temp rise, 5C would seem to be steaming my components (relatively speaking of course ) I wouldn't want to run the fans needed to take 5C out of the loop anyway.

    PS I fold 24/7, even when gaming.
    No, I dont doubt your system preform well.. your 1 deg delta t proves that..but how much lower are your hardware temps than with a OEM sink and fan.. WAY WAY below i bet!! I know when I turn my pump up i can eliminate the 5 degree difference, but there goes my silent system. Sometimes I like it quiet, sometime i like it loud.. But even if my components are 5 degrees warmer than i could have them, they are sill a heck of a lot colder than oem temps...

    For my system, i dont think i always have to keep my core temp 20 degrees below what an oem heat sink and fan will accomplish unless I am oc'd... (which is not much lately... but hopefully it will be soon with all of your help and my new machine)...

    Again, though, about the 5 degrees and your fans: the lower the gpm/flow the more time the water spends in the radiator and therefore would achieve more cooling as a loop with a high gpm/flow ...water that is getting blasted through a radiator does not have as much time to cool as water that is pumped through it slowly, regardless of fan speed. so .. slower pump speeds lead to higher heat absorbtion through the loop of blocks, but they also get proportional increased cooling through the radiator. Again, if you are trying to achieve a 1 or 2 degree change in temp, you can still be assured that the radiator will cool the water using the same fan speed regardless of gpm..

    that was a mouthful...

    dangt,, what about the cat?
    LOTS of Koolance water cooling hardware/cases/coolers/ for sale cheap. PM if interested.


  14. #14
    Joined
    Mar 2004
    Location
    mtl,canada
    Age
    36
    Posts
    37,301

    Re: Loop or Parallel Watercooling?

    hmmm..so a slower pum means a better one? and i tought numbers on paper were everything..now, if you double the rad, will u see a huge decrease i ntemps?

    what neko?

    add a couple 120mm fans overhead to drop temps by 10+ celsiusD.S.C-12(2)-disclaimer : whatever u do with your hardware/software is your
    responsibility, which i dont hold if u break anything

    GRAB here OcBible v1.55 and Guidemania v1.21
    scary pooch?pooch#1taking ball from dogpooch pulln on ropeme on board after 2 years
    of no board
    http://gonny.se/ <-cool lady http://www.mediafire.com/?8dmg21ms1vx
    IN MEMORY OF HOW BM WAS TREATED HERE...

  15. #15
    Joined
    Aug 2004
    Posts
    10,583

    Re: Loop or Parallel Watercooling?

    Shredder, your whole theory is defective. Regardless of flow the same % of water is in the rad at all times. If you did the math you would see that every molecule is in the rad the exact same amount of time, as long as there is flow (this is not an open system) the high flow molecule will just make more passes.

    And like I said, we have different goals. If I wanted a stock silent system I would use air and save a few hundred bucks. I recommend www.silentpcreview.com to you, as these guys sleep with their computers, which is not my goal. Lots of good info you can use. Low temps are one of things you give up to be quiet, but that is not a good thing when your system runs 100% (well 140% if you went by manufacterers spec ) 24/7.
    Last edited by TheGlasMan; 05-02-2007 at 06:58 AM.


    "The significant problems we face cannot be solved at the same level of thinking we were at when we created them."
    - Albert Einstein (1879-1955)

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •