Overclock.net › Forums › Cooling › Water Cooling › Can you have only radiators as both intake and exhaust?
New Posts  All Forums:Forum Nav:

Can you have only radiators as both intake and exhaust?

post #1 of 8
Thread Starter 
So I've scoured the internet, and only found a few posts where people have small conflicting opinions on this topic as part of answering other questions. What I'm trying to determine is the following: If all my components are liquid cooled, do I need pure airflow fans at all, or can everything have a radiator attached?

My current setup is a custom loop inside a Lian Li PC-O8 with an i7-6700k and a Titan XP on a liquid cooled Asus Formula VIII. For anyone familiar with the case, I've modded the back of the case to mount an additional fan below the gpu, and modded the back of the case to mount fans without the harddrive rack (I don't have any mechanical drives). I've also removed the LED lighting kit that came with the case (there is a control box and a ton of nobs and wires that you can remove near where the PSU sits). Here is a diagram of what it currently looks like using a stock photo:

http://imgur.com/p7Z1oMt

It's not drawn perfectly to scale, as there is no actual overlap between the top radiator and the exhaust fan on the mobo side of the case, and there is a bit more room between the PSU and the rad/fans on the PSU side of the case; but you get the idea.
My temps are alright, and are somewhere between 25-30C idle for both CPU and GPU. and the CPU rises to anywhere between 45-65C under sustained load (GTAV for like 5 hours for example) and the GPU rises to about 45-55C.

I would be inclined to think that the GPU traditionally gets hotter than the CPU, and the fact that the CPU, GPU, and mobo are in the same loop causes heat from the GPU to increase CPU temps and the CPU to reduce temps in the GPU (not reduce, but, decrease the highest temps the GPU would reach on it's own with proportional cooling equipment).

What I'm considering doing is going SLI with another Titan XP, and my actual question and the reason for this post is the following: If all of my components are liquid cooled, do I need any pure airflow fans, or can I just put radiators at literally every intake and exhaust point? The thing I'm especially curious about is, if I have a radiator for intake, and then the fans push the hot air off, then another radiator as exhaust, the hot air from the intake rad would be pushed through the exhaust rad. I'm not 100% sure this is a problem, as the fans on the exhaust rad would still be pushing hot air out, which comes from the water in the rad, not the air outside the rad, and I don't think the air outside the rad would heat the water inside the rad too much. This is the argument though: Will the additional cooling from an additional radiator offset the reduction in cooling from pushing hot air through one of the radiators in the loop? No one seems to have a definitive answer, so I'd love to hear what you guys think. Here is my new setup idea:

http://imgur.com/Reya5ki

I would like to point out that where there red circles are on the overlap between top and rear rads/fans is there to remind everyone that there is no actual overlap, it's just bad drawing on my part. It is a pretty tight fit, however....

Please keep in mind that I will likely split the current loop into two loops, one with mobo and cpu, and the other with the 2 titan xps. I currently have an ek d5 pump, with ek mobo block, titan block, petg tubing and the radiators are a top mounted 240mm and a front mounted 360mm. If you aren't familiar with the PC-O8 from Lian Li, please keep in mind that the front fans intake are from a vent of the side of the case, and also exhaust air into a separate part of the case from the main components--so it's not you're standard air in/air out through the same chamber set up.

Please ask any additional questions you need, as this is quite a bit of a ramble, but I've just put off asking for too long, and don't really have an pc enthusiast friends.

Kind Regards.
post #2 of 8
No one knows because it isn't a one size fits all scenario. It is going to come down to two factors:

1) How strong is your airflow / how frequently is the air in your case completely cycled?

2) What is the delta between ambient room temp and the temp of your coolant?

With a great enough temp delta and a high enough flow rate, your idea will still be inefficient, but it will work. Otherwise you may break even or even come out negative in cases with poor airflow and/or very small deltas.

If you have an IR thermometer, and since you already have the system in question, you can find out very easily without all the equations required for a hypothetical system.

Read the temp of the air coming out of the case (with the radiators you currently have dumping heat into the case). If that temp is lower than the temp of the coolant (might need to get creative to read that, depending on your setup), then adding another radiator that is fed by your "used" air will help. If the same nothing will happen. If the other way around it will hurt temps.

That said you still don't know HOW MUCH it will hurt our help. The delta between the two temps will give you an idea, but nothing exact without also calculating the heat transfer between your red and air, and then calculating your airflow thought said rad.
Arcane
(33 items)
 
Sanctum
(9 items)
 
Fukurou
(15 items)
 
CPUMotherboardGraphicsRAM
Ryzen 7 1700 MSI B350M Mortar Arctic RX Vega 64 G.Skill TridentZ  
Hard DriveHard DriveOSPower
Samsung EVO 960 Consatellation Windows 10 Pro Corsair HX 1050 
Case
Raijintek Aeneas White Window 
CPUMotherboardRAMHard Drive
Core i5-661 GA-H55-USB3 12GB DDR3 1333 WD Caviar Black AALS 640GB 
Hard DriveHard DriveOptical DriveCooling
WD Caviar Green EARS 2TB Vertex Turbo 60GB (SSD) Samsung Blu-ray, Samsung DVD Burner H50 (With push/pull nocturas) 
OSMonitorKeyboardPower
Windows 10 Professional 64 LG 47inch LED LCD Saitek Eclipse Lite-touch Wireless Seasonic X750 Gold 
CaseAudio
CM690 II Passive bookshelfspeakers + amp. 
  hide details  
Reply
Arcane
(33 items)
 
Sanctum
(9 items)
 
Fukurou
(15 items)
 
CPUMotherboardGraphicsRAM
Ryzen 7 1700 MSI B350M Mortar Arctic RX Vega 64 G.Skill TridentZ  
Hard DriveHard DriveOSPower
Samsung EVO 960 Consatellation Windows 10 Pro Corsair HX 1050 
Case
Raijintek Aeneas White Window 
CPUMotherboardRAMHard Drive
Core i5-661 GA-H55-USB3 12GB DDR3 1333 WD Caviar Black AALS 640GB 
Hard DriveHard DriveOptical DriveCooling
WD Caviar Green EARS 2TB Vertex Turbo 60GB (SSD) Samsung Blu-ray, Samsung DVD Burner H50 (With push/pull nocturas) 
OSMonitorKeyboardPower
Windows 10 Professional 64 LG 47inch LED LCD Saitek Eclipse Lite-touch Wireless Seasonic X750 Gold 
CaseAudio
CM690 II Passive bookshelfspeakers + amp. 
  hide details  
Reply
post #3 of 8
Thread Starter 
Quote:
Originally Posted by Zero4549 View Post

No one knows because it isn't a one size fits all scenario. It is going to come down to two factors:

1) How strong is your airflow / how frequently is the air in your case completely cycled?

2) What is the delta between ambient room temp and the temp of your coolant?

1. I think that this question is better suited to more standard cases, where there are front/top intake fans over the main components, and then a rear exhaust. In my case, the case design has compartments, and the main source of heat is the 360mm rad with push/pull fans that blows air nowhere near the motherboard or gpu. So as far as "completely cycled", that would be difficult to determine, as there would be a "cycle" for compartment A and a "cycle" for compartment B, and the amount of "complete cycling" would have varying effects on component temps.

2. I don't have a thermometer in my loop (I read that it slows down flowrate and it doesn't really matter because your component heat is what counts) but I suppose I could add on in if I wanted to do my own experiments. I was hoping to draw on the collective experiments of this community though, before I resorted to tons of my own testing (at this point). The difference between ambient room temp and my component temps at load are typically somewhere in the 10C range and 30C range, respectively.

The biggest question I'd like to address is, based on the technology of radiators (which i'm certainly no expert in, but I'm somewhat well read in) does external heat have an equivalent impact as internal heat. This is to say, at an equilibrium point of hot air moving through the case, will additional hot air blowing out of the radiator from within the loop release more heat than hot air from inside the same loop being introduced to the radiator from outside of the loop?

For example, if I had a rear 240mm rad immediately after the GPU, then after that component, there was another rad, the front intake rad, blowing it's heat into the 240mm rear rad, would the 240mm rad have already rid the loop of enough heat to the point where the heat that remains in the loop would blow off the 360mm rad into the 240mm rad, but the heat from the air would not increase the temperatures by as much as the 240mm rad would have already expelled?

It's super complicated to try and word what I'm asking, so I hope I'm getting my point across
post #4 of 8
If your case has multiple compartments, but all the radiators exist in the same compartment, in theory you can just look at that compartment alone and treat it as if it were the whole case.

In practice however this will never be accurate because no case has perfectly sealed compartments, and if anything is generating heat in another compartment, that heat must go somewhere, and unless that compartment has active cooling, it will go wherever there is least resistance, which more often than not means right back into the compartment that does have active cooling.

Thus, I would still treat the whole case as a single unit and forget the compartments. This still isn't accurate, but it will err on the side of caution, which is certainly better than the alternative.

As for the heat exchange of radiators, it is in fact NOT directional. The only reason radiators typically release heat instead of absorbing heat is because the liquid flowing through them is typically warmer than the air flowing through them. Reverse the equation and you get a reversed result. More radiators / better radiators don't change the formula. All a radiator does is facilitate equilibrium between two liquids/gasses.
Arcane
(33 items)
 
Sanctum
(9 items)
 
Fukurou
(15 items)
 
CPUMotherboardGraphicsRAM
Ryzen 7 1700 MSI B350M Mortar Arctic RX Vega 64 G.Skill TridentZ  
Hard DriveHard DriveOSPower
Samsung EVO 960 Consatellation Windows 10 Pro Corsair HX 1050 
Case
Raijintek Aeneas White Window 
CPUMotherboardRAMHard Drive
Core i5-661 GA-H55-USB3 12GB DDR3 1333 WD Caviar Black AALS 640GB 
Hard DriveHard DriveOptical DriveCooling
WD Caviar Green EARS 2TB Vertex Turbo 60GB (SSD) Samsung Blu-ray, Samsung DVD Burner H50 (With push/pull nocturas) 
OSMonitorKeyboardPower
Windows 10 Professional 64 LG 47inch LED LCD Saitek Eclipse Lite-touch Wireless Seasonic X750 Gold 
CaseAudio
CM690 II Passive bookshelfspeakers + amp. 
  hide details  
Reply
Arcane
(33 items)
 
Sanctum
(9 items)
 
Fukurou
(15 items)
 
CPUMotherboardGraphicsRAM
Ryzen 7 1700 MSI B350M Mortar Arctic RX Vega 64 G.Skill TridentZ  
Hard DriveHard DriveOSPower
Samsung EVO 960 Consatellation Windows 10 Pro Corsair HX 1050 
Case
Raijintek Aeneas White Window 
CPUMotherboardRAMHard Drive
Core i5-661 GA-H55-USB3 12GB DDR3 1333 WD Caviar Black AALS 640GB 
Hard DriveHard DriveOptical DriveCooling
WD Caviar Green EARS 2TB Vertex Turbo 60GB (SSD) Samsung Blu-ray, Samsung DVD Burner H50 (With push/pull nocturas) 
OSMonitorKeyboardPower
Windows 10 Professional 64 LG 47inch LED LCD Saitek Eclipse Lite-touch Wireless Seasonic X750 Gold 
CaseAudio
CM690 II Passive bookshelfspeakers + amp. 
  hide details  
Reply
post #5 of 8
Quote:
Originally Posted by Pryze View Post

If all my components are liquid cooled, do I need pure airflow fans at all, or can everything have a radiator attached?

Generally that's not a good idea. There are motherboard components (the non water cooled ones) and computer components (drives and add in cards) that depend on fresh cool air for cooling.

I would be inclined to think that the GPU traditionally gets hotter than the CPU, and the fact that the CPU, GPU, and mobo are in the same loop causes heat from the GPU to increase CPU temps and the CPU to reduce temps in the GPU (not reduce, but, decrease the highest temps the GPU would reach on it's own with proportional cooling equipment).

Whether a dual loop works better (than a single loop) depends more on the water flow difference than on the amount of heat each component produces. And it doesn't matter what temperature the components run at, only the amount of heat (watts) they produce matters, and they will produce the same amount of heat irrespective of the loop configuration. Then there's also the additional heat dump into the loop from a second pump to consider. I'm not aware of any testing comparing a single loop to a double loop with the same (total) water flow in each, thereby measuring only the differences between the two configurations, not the effect of additional water flow from a second pump.


If all of my components are liquid cooled, do I need any pure airflow fans, or can I just put radiators at literally every intake and exhaust point?

As above, there are motherboard components (the non water cooled ones) and computer components (drives and add in cards) that depend on fresh cool air for cooling.

Will the additional cooling from an additional radiator offset the reduction in cooling from pushing hot air through one of the radiators in the loop? No one seems to have a definitive answer, so I'd love to hear what you guys think. Here is my new setup idea:

See Picture below.


Edited by billbartuska - 10/10/16 at 8:37am
My System
(15 items)
 
  
CPUMotherboardGraphicsRAM
FX6300 Black M5A99X EVO R2.0 Nvidia GTS450 Team Vulcan PC3 12800 
Hard DriveOptical DriveCoolingOS
Samsung 840 PRO Asus DRW-1608P (x2) Custom Water Cooling Win7 (Ult), Win 8.1 & Win Server 2012 R2 
MonitorKeyboardPowerCase
2 X Samsung 915N Ducky Shine III, Blue Cherry/Blue LEDs PCP&C 1kw Lian Li PC-71 (W/Window) 
MouseAudio
Logiteck G400s none 
  hide details  
Reply
My System
(15 items)
 
  
CPUMotherboardGraphicsRAM
FX6300 Black M5A99X EVO R2.0 Nvidia GTS450 Team Vulcan PC3 12800 
Hard DriveOptical DriveCoolingOS
Samsung 840 PRO Asus DRW-1608P (x2) Custom Water Cooling Win7 (Ult), Win 8.1 & Win Server 2012 R2 
MonitorKeyboardPowerCase
2 X Samsung 915N Ducky Shine III, Blue Cherry/Blue LEDs PCP&C 1kw Lian Li PC-71 (W/Window) 
MouseAudio
Logiteck G400s none 
  hide details  
Reply
post #6 of 8
Thread Starter 
This is extremely helpful. My one question would be, what are the components that aren't watercooled? All I can think of is my PSU (don't want to die so I probably wont' try to water cool that) and my SSD, which I don;t think really generates much heat. Oh, and my ram doesn't have a block, but , it's 32gb of 3200mhz ram, it never gets more than like, 15% utilization from what I can see.

Other than that, my formula VIII motherboard has built in blocks, my cpu has a block, my gpu has a block.
post #7 of 8
Thread Starter 
Quote:
Originally Posted by billbartuska View Post


This chart is great, but the question I'm asking is, will a combination of those two scenarios get rid of more heat, or will they eventual equalize, having the cpu/gpu at load for long enough consecutive time. For example, in the right side scenario, if you were to take those two rads on top of one another, but instead of a pure air exhaust on the right, you have a radiator on the exhuast (push/pull for max airflow) would that expell more air than the reintroduction of heat from the first set of rads would pump into the second/exhaust rad on the right of the diagram?

The question of whether you should have 1 rad pump into another rad vs 2 rads pumping into exhaust fans is one that has been answered 1,000 times over, but adding the third rad is sort of the question I'm asking.
post #8 of 8
Quote:
Originally Posted by Pryze View Post

Quote:
Originally Posted by billbartuska View Post


...but the question I'm asking is, will a combination of those two scenarios get rid of more heat, or will they eventual equalize, having the cpu/gpu at load for long enough consecutive time.

Before that happens, there will always be more cooling than at equilibrium.
Any consideration of heat removal should assume that everything has already reached an equilibrium temperature.
If you have enough cooling for the first 5-10 minutes of maximum stress, and then you system overheats - well, you don't have enough cooling.


For example, in the right side scenario, if you were to take those two rads on top of one another, but instead of a pure air exhaust on the right, you have a radiator on the exhuast (push/pull for max airflow) would that expell more air than the reintroduction of heat from the first set of rads would pump into the second/exhaust rad on the right of the diagram?

I kind of don't understand what you're asking.



I doesn't matter what you put on the other side, the warm air from the first set of rads still goes into the case.
And adding the fans you would add to the new, third (exhaust) rad to the existing intake rads (instead of adding a rad and fans) would have the same effect on temperature of the air entering the case, and the same, or very close to, the total amount of cooling - and you'd save the cost of a new rad.

What will that warm air heat up?

Look at you motherboard, see everything that isn't watercooled, all those ICs? They all will run hotter, as will everything else in your case.

You could:
Run all the rads as exhaust (push, pull, or both) and add enough intake fans to equal (or better 10% more than) the amount of air exhausted by the rads.
Mount the rads externally so they intake and exhaust from / to the outside of the case.
Edited by billbartuska - 10/10/16 at 10:18am
My System
(15 items)
 
  
CPUMotherboardGraphicsRAM
FX6300 Black M5A99X EVO R2.0 Nvidia GTS450 Team Vulcan PC3 12800 
Hard DriveOptical DriveCoolingOS
Samsung 840 PRO Asus DRW-1608P (x2) Custom Water Cooling Win7 (Ult), Win 8.1 & Win Server 2012 R2 
MonitorKeyboardPowerCase
2 X Samsung 915N Ducky Shine III, Blue Cherry/Blue LEDs PCP&C 1kw Lian Li PC-71 (W/Window) 
MouseAudio
Logiteck G400s none 
  hide details  
Reply
My System
(15 items)
 
  
CPUMotherboardGraphicsRAM
FX6300 Black M5A99X EVO R2.0 Nvidia GTS450 Team Vulcan PC3 12800 
Hard DriveOptical DriveCoolingOS
Samsung 840 PRO Asus DRW-1608P (x2) Custom Water Cooling Win7 (Ult), Win 8.1 & Win Server 2012 R2 
MonitorKeyboardPowerCase
2 X Samsung 915N Ducky Shine III, Blue Cherry/Blue LEDs PCP&C 1kw Lian Li PC-71 (W/Window) 
MouseAudio
Logiteck G400s none 
  hide details  
Reply
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: Water Cooling
Overclock.net › Forums › Cooling › Water Cooling › Can you have only radiators as both intake and exhaust?