Overclock.net › Forums › Cooling › Water Cooling › The importance of the "order" of your loop?
New Posts  All Forums:Forum Nav:

The importance of the "order" of your loop? - Page 3

post #21 of 110
gpu temps are 26c idle 42c full system load while folding at home, also 46c-47c temps with furmark.

and cpu 28c idle 64c full system load

proof:



Edited by pgmoney - 8/21/10 at 6:30pm
    
CPUMotherboardGraphicsRAM
core i7 2600K ASUS P8P67 Deluxe EVGA gtx 580 SC 2x4 Corsair Vengence 1600mhz 
Hard DriveOptical DriveOSMonitor
G.SKILL Phoenix Pro 120GB ssd 2x1TB Seagate sony, philips Winodws 7 x64 PRO Hannspree 25" HF259H 
KeyboardPowerCaseMouse
Logitech G19 Enermax evo galaxy 1250 Coolermaster HAF-X CM storm sentineal game 
Mouse Pad
sub zero gamers pad 
  hide details  
Reply
    
CPUMotherboardGraphicsRAM
core i7 2600K ASUS P8P67 Deluxe EVGA gtx 580 SC 2x4 Corsair Vengence 1600mhz 
Hard DriveOptical DriveOSMonitor
G.SKILL Phoenix Pro 120GB ssd 2x1TB Seagate sony, philips Winodws 7 x64 PRO Hannspree 25" HF259H 
KeyboardPowerCaseMouse
Logitech G19 Enermax evo galaxy 1250 Coolermaster HAF-X CM storm sentineal game 
Mouse Pad
sub zero gamers pad 
  hide details  
Reply
post #22 of 110
Quote:
Originally Posted by Zenophobe View Post
same question as above... why didn't you splint your gpu to its own rads since you have 2.. and go Res - pump- 360 - cpu - 360 - gpu gpu gpu -res
the system evens out eventually and the water is cooled to it's max before entering cpu. even in single loop i get right at 2.5GPM flow rate and everything runs the same if i split the rads or not i tried it trust me
    
CPUMotherboardGraphicsRAM
core i7 2600K ASUS P8P67 Deluxe EVGA gtx 580 SC 2x4 Corsair Vengence 1600mhz 
Hard DriveOptical DriveOSMonitor
G.SKILL Phoenix Pro 120GB ssd 2x1TB Seagate sony, philips Winodws 7 x64 PRO Hannspree 25" HF259H 
KeyboardPowerCaseMouse
Logitech G19 Enermax evo galaxy 1250 Coolermaster HAF-X CM storm sentineal game 
Mouse Pad
sub zero gamers pad 
  hide details  
Reply
    
CPUMotherboardGraphicsRAM
core i7 2600K ASUS P8P67 Deluxe EVGA gtx 580 SC 2x4 Corsair Vengence 1600mhz 
Hard DriveOptical DriveOSMonitor
G.SKILL Phoenix Pro 120GB ssd 2x1TB Seagate sony, philips Winodws 7 x64 PRO Hannspree 25" HF259H 
KeyboardPowerCaseMouse
Logitech G19 Enermax evo galaxy 1250 Coolermaster HAF-X CM storm sentineal game 
Mouse Pad
sub zero gamers pad 
  hide details  
Reply
post #23 of 110
I don't think there can be any argument. I don't think it's possible for a loop to equalize in temp if you are removing heat at one point and introducing it at another. It may reach a point where the overall temp of the loop stops rising but that doesn't mean the temp has equalized across the loop. There has to be some type of heat build up and then heat dump (at which point the water in the loop would be the coldest) and right after that heat dump, you want to have whatever you are trying to cool. It's simple scientific logic. You want to cool something with the coldest available coolant you have, which is right after a radiator in your loop.

If you have two radiators, separating them will be beneficial, especially if they are larger radiators and have the capability to dump more heat than is introduced into the line per block.

For instance. If you have 3 CPU block and 2 radiators. And 1 radiator is capable of dumping the heat from 2 CPU blocks, then hooking the loop up with both radiators in a series and then the 3 blocks in a series is less efficient than having it rad > cpu > cpu> rad > cpu. The second radiator will dump all heat introduced into the loop and allow for lower temps on the third CPU whereas if you had it as rad > rad > cpu > cpu > cpu. The water is constantly building up heat before it hits that third CPU and there is no available heat dump before it does, thusly making overall temps higher across the board. The second radiator is made useless because there is no available heat left in the water to dump.
 
Project Al
(69 photos)
 
CPUMotherboardGraphicsRAM
Intel 2700k Gigabyte UD4 MSI 580GTX Lightning Xtreme Edition 3GB 2x4GB G.Skill @ 1600mhz CL6 
Hard DriveHard DriveOptical DriveOS
OCZ Vertex II SSD OCZ Vertex II SSD LG Blu-Ray/HD-DVD Combo Drive Win 7 Ult 
MonitorMonitorKeyboardPower
Achieva Shimian Acer G235H XArmor U9BL Corsair HX1000W 
CaseMouseMouse PadAudio
Fractal Define R3 Silver Arrow Razer Naga Ultimate Razer Goliathus Speed Edition Polk Audio RTi-A1 
AudioAudio
Dayton Audio DTA-100a Asus Xonar STX 
  hide details  
Reply
 
Project Al
(69 photos)
 
CPUMotherboardGraphicsRAM
Intel 2700k Gigabyte UD4 MSI 580GTX Lightning Xtreme Edition 3GB 2x4GB G.Skill @ 1600mhz CL6 
Hard DriveHard DriveOptical DriveOS
OCZ Vertex II SSD OCZ Vertex II SSD LG Blu-Ray/HD-DVD Combo Drive Win 7 Ult 
MonitorMonitorKeyboardPower
Achieva Shimian Acer G235H XArmor U9BL Corsair HX1000W 
CaseMouseMouse PadAudio
Fractal Define R3 Silver Arrow Razer Naga Ultimate Razer Goliathus Speed Edition Polk Audio RTi-A1 
AudioAudio
Dayton Audio DTA-100a Asus Xonar STX 
  hide details  
Reply
post #24 of 110
Quote:
Originally Posted by Shrimpykins View Post
I don't think there can be any argument. I don't think it's possible for a loop to equalize in temp if you are removing heat at one point and introducing it at another. It may reach a point where the overall temp of the loop stops rising but that doesn't mean the temp has equalized across the loop. There has to be some type of heat build up and then heat dump (at which point the water in the loop would be the coldest) and right after that heat dump, you want to have whatever you are trying to cool. It's simple scientific logic. You want to cool something with the coldest available coolant you have, which is right after a radiator in your loop.

If you have two radiators, separating them will be beneficial, especially if they are larger radiators and have the capability to dump more heat than is introduced into the line per block.

For instance. If you have 3 CPU block and 2 radiators. And 1 radiator is capable of dumping the heat from 2 CPU blocks, then hooking the loop up with both radiators in a series and then the 3 blocks in a series is less efficient than having it rad > cpu > cpu> rad > cpu. The second radiator will dump all heat introduced into the loop and allow for lower temps on the third CPU whereas if you had it as rad > rad > cpu > cpu > cpu. The water is constantly building up heat before it hits that third CPU and there is no available heat dump before it does, thusly making overall temps higher across the board. The second radiator is made useless because there is no available heat left in the water to dump.
WELL I tested both and same results i have rebuild my loops many times over with the same thinking but the temps stay the same.

not to menton my electrical engineering degree from UM was put to the test in many configurations, only to prove a system over a long operating time, say gaming sesion or folding all day does balance out regardless of rad placement. there is no logic in spliting the rads if I get the same performance and i am satisfied with the performance i get, compared to what others have done with their loops.

are you saying my temps are bad?


besides you just answered your logic vs. fact in your own post:
"I don't think it's possible for a loop to equalize in temp if you are removing heat at one point and introducing it at another."

you are introducing heat to the loop regardless of where you choose to cool it.
Edited by pgmoney - 8/21/10 at 6:46pm
    
CPUMotherboardGraphicsRAM
core i7 2600K ASUS P8P67 Deluxe EVGA gtx 580 SC 2x4 Corsair Vengence 1600mhz 
Hard DriveOptical DriveOSMonitor
G.SKILL Phoenix Pro 120GB ssd 2x1TB Seagate sony, philips Winodws 7 x64 PRO Hannspree 25" HF259H 
KeyboardPowerCaseMouse
Logitech G19 Enermax evo galaxy 1250 Coolermaster HAF-X CM storm sentineal game 
Mouse Pad
sub zero gamers pad 
  hide details  
Reply
    
CPUMotherboardGraphicsRAM
core i7 2600K ASUS P8P67 Deluxe EVGA gtx 580 SC 2x4 Corsair Vengence 1600mhz 
Hard DriveOptical DriveOSMonitor
G.SKILL Phoenix Pro 120GB ssd 2x1TB Seagate sony, philips Winodws 7 x64 PRO Hannspree 25" HF259H 
KeyboardPowerCaseMouse
Logitech G19 Enermax evo galaxy 1250 Coolermaster HAF-X CM storm sentineal game 
Mouse Pad
sub zero gamers pad 
  hide details  
Reply
post #25 of 110
Quote:
Originally Posted by Shrimpykins View Post
I don't think there can be any argument. I don't think it's possible for a loop to equalize in temp if you are removing heat at one point and introducing it at another. It may reach a point where the overall temp of the loop stops rising but that doesn't mean the temp has equalized across the loop. There has to be some type of heat build up and then heat dump (at which point the water in the loop would be the coldest) and right after that heat dump, you want to have whatever you are trying to cool. It's simple scientific logic. You want to cool something with the coldest available coolant you have, which is right after a radiator in your loop.

If you have two radiators, separating them will be beneficial, especially if they are larger radiators and have the capability to dump more heat than is introduced into the line per block.

For instance. If you have 3 CPU block and 2 radiators. And 1 radiator is capable of dumping the heat from 2 CPU blocks, then hooking the loop up with both radiators in a series and then the 3 blocks in a series is less efficient than having it rad > cpu > cpu> rad > cpu. The second radiator will dump all heat introduced into the loop and allow for lower temps on the third CPU whereas if you had it as rad > rad > cpu > cpu > cpu. The water is constantly building up heat before it hits that third CPU and there is no available heat dump before it does, thusly making overall temps higher across the board. The second radiator is made useless because there is no available heat left in the water to dump.
This!

While I do not have water cooling, yet, it would be impossible any other way.
Ra
(17 items)
 
  
CPUMotherboardGraphicsRAM
i7 930 D0 @4.2 Asus P6T PNY GTX 470 SLI 6gb OCZ Gold 1600 
Hard DriveHard DriveOSMonitor
90GB Kingston SSD 1Tb WD Caviar Black Windows 7 (64 bit) PCBank 3view PB2700 
MonitorMonitorKeyboardPower
acer x243w ASUS VE247h Logitech G15 (version 1), Logitech G13 PCPC Silencer 950w 
CaseMouseMouse PadAudio
HAF X Logitech G700 steelseries X-Large audio-technica ATH-A700 
  hide details  
Reply
Ra
(17 items)
 
  
CPUMotherboardGraphicsRAM
i7 930 D0 @4.2 Asus P6T PNY GTX 470 SLI 6gb OCZ Gold 1600 
Hard DriveHard DriveOSMonitor
90GB Kingston SSD 1Tb WD Caviar Black Windows 7 (64 bit) PCBank 3view PB2700 
MonitorMonitorKeyboardPower
acer x243w ASUS VE247h Logitech G15 (version 1), Logitech G13 PCPC Silencer 950w 
CaseMouseMouse PadAudio
HAF X Logitech G700 steelseries X-Large audio-technica ATH-A700 
  hide details  
Reply
post #26 of 110
Quote:
Originally Posted by pgmoney View Post
WELL I tested both and same results i have rebuild my loops many times over with the same thinking but the temps stay the same.

not to menton my electrical engineering degree from UM was put to the test in many configurations, only to prove a system over a long operating time, say gaming sesion or folding all day does balance out regardless of rad placement. there is no logic in spliting the rads if I get the same performance and i am satisfied with the performance i get, compared to what others have done with their loops.

are you saying my temps are bad?


besides you just answered your logic vs. fact in your own post:
"I don't think it's possible for a loop to equalize in temp if you are removing heat at one point and introducing it at another."

you are introducing heat to the loop regardless of where you choose to cool it.
I never said your temps were bad. You may not be able to measure the difference but I can assure you there is a difference.

And that's BS that it balances. You are telling me if you put your CPU then board block then 3 gpus in a series you are getting the same temps than if you put a radiator right before that 3rd GPU. I bet you see temp drops in that 3rd GPU if you do, thusly proving my theory.

There is physical heat dump there that can be removed from the coolant before trying to cool something with it. I don't care if it's a .01C increase, it's still an increase in temps that could be removed.

And yes, you are introducing and removing heat of the same amount regardless of where the radiator or blocks are in a loop, if the heat can get to that radiator to be cooled.

Like I said.

If you have 3 CPU block and 2 radiators. And 1 radiator is capable of dumping the heat from 2 CPU blocks, then hooking the loop up with both radiators in a series and then the 3 blocks in a series is less efficient than having it rad > cpu > cpu> rad > cpu. The second radiator will dump all heat introduced into the loop and allow for lower temps on the third CPU whereas if you had it as rad > rad > cpu > cpu > cpu. The water is constantly building up heat before it hits that third CPU and there is no available heat dump before it does, thusly making overall temps higher across the board. The second radiator is made useless because there is no available heat left in the water to dump, thusly making the overall efficieny of the loop less than it possibly could be.

It's like taking 100C water and trying to cool a 150C plate, and then taking 99C water and cooling the 150C plate. The 99C water will absorb more heat than the 100C water. It may be absolutely miniscule but it's still there.

Simple fact: The coldest point of a loop is DIRECTLY after it leaves the radiator and/or somewhere within the radiator itself.
Edited by Shrimpykins - 8/21/10 at 7:18pm
 
Project Al
(69 photos)
 
CPUMotherboardGraphicsRAM
Intel 2700k Gigabyte UD4 MSI 580GTX Lightning Xtreme Edition 3GB 2x4GB G.Skill @ 1600mhz CL6 
Hard DriveHard DriveOptical DriveOS
OCZ Vertex II SSD OCZ Vertex II SSD LG Blu-Ray/HD-DVD Combo Drive Win 7 Ult 
MonitorMonitorKeyboardPower
Achieva Shimian Acer G235H XArmor U9BL Corsair HX1000W 
CaseMouseMouse PadAudio
Fractal Define R3 Silver Arrow Razer Naga Ultimate Razer Goliathus Speed Edition Polk Audio RTi-A1 
AudioAudio
Dayton Audio DTA-100a Asus Xonar STX 
  hide details  
Reply
 
Project Al
(69 photos)
 
CPUMotherboardGraphicsRAM
Intel 2700k Gigabyte UD4 MSI 580GTX Lightning Xtreme Edition 3GB 2x4GB G.Skill @ 1600mhz CL6 
Hard DriveHard DriveOptical DriveOS
OCZ Vertex II SSD OCZ Vertex II SSD LG Blu-Ray/HD-DVD Combo Drive Win 7 Ult 
MonitorMonitorKeyboardPower
Achieva Shimian Acer G235H XArmor U9BL Corsair HX1000W 
CaseMouseMouse PadAudio
Fractal Define R3 Silver Arrow Razer Naga Ultimate Razer Goliathus Speed Edition Polk Audio RTi-A1 
AudioAudio
Dayton Audio DTA-100a Asus Xonar STX 
  hide details  
Reply
post #27 of 110
Quote:
Originally Posted by Shrimpykins View Post
I never said your temps were bad. You may not be able to measure the difference but I can assure you there is a difference.
PROVE IT! Please. also if my temps are not bad what is the problem?

Quote:
And that's BS that it balances. You are telling me if you put your CPU then board block then 3 gpus in a series you are getting the same temps than if you put a radiator right before that 3rd GPU. I bet you see temp drops in that 3rd GPU if you do, thusly proving my theory.
I am telling you this, not to mention who hooks a 2nd rad to just the 3rd card? pics please

again prove me wrong please! pics or it didn't happen!


Quote:
Simple fact: The coldest point of a loop is DIRECTLY after it leaves the radiator and/or somewhere within the radiator itself.
NO need to prove as everyone knows this, but the temps of your components will not change.

exactly and that is why the water that leaves the second rad is at the coolest temp possible (ambient) before hitting the blocks, flow rates differ and if you got good flow rate by the time the heat gets back to the rads it has absorbed as much heat as possible.

when you break up the rads the temps never reach ambient
Edited by pgmoney - 8/21/10 at 7:35pm
    
CPUMotherboardGraphicsRAM
core i7 2600K ASUS P8P67 Deluxe EVGA gtx 580 SC 2x4 Corsair Vengence 1600mhz 
Hard DriveOptical DriveOSMonitor
G.SKILL Phoenix Pro 120GB ssd 2x1TB Seagate sony, philips Winodws 7 x64 PRO Hannspree 25" HF259H 
KeyboardPowerCaseMouse
Logitech G19 Enermax evo galaxy 1250 Coolermaster HAF-X CM storm sentineal game 
Mouse Pad
sub zero gamers pad 
  hide details  
Reply
    
CPUMotherboardGraphicsRAM
core i7 2600K ASUS P8P67 Deluxe EVGA gtx 580 SC 2x4 Corsair Vengence 1600mhz 
Hard DriveOptical DriveOSMonitor
G.SKILL Phoenix Pro 120GB ssd 2x1TB Seagate sony, philips Winodws 7 x64 PRO Hannspree 25" HF259H 
KeyboardPowerCaseMouse
Logitech G19 Enermax evo galaxy 1250 Coolermaster HAF-X CM storm sentineal game 
Mouse Pad
sub zero gamers pad 
  hide details  
Reply
post #28 of 110
I've changed my position on this.

The reason pgmoney gets the temps he does with daisy chained rads is because water temp never reaches the same temp (or higher) as his devices. If your water temp is lower than your device temp, heat energy will transfer to the water.

I don't know much about the rate at which that energy transfers (slows or speeds up) as the two mediums become closer and closer to the same temperature... but I'm sure there is some curve involved there.

If your rads cool the water to near ambient even though they probably don't (lets say 23c, which would be an average air conditioned house at 73f)... then as the water goes through his loop it isn't being heated enough to effect its ability to cool devices that are way higher temp... (53c and 42c)

A system of temp probes would be necessary to verify... or you could just stick a thermometer inside your reservoir and compare the water temperature to the temperature of GPU3.

Here is a total BS conceptual visual of what water temps might be throughout his loop.
Rad: 23c
CPU: 25c (fine for cooling a 53c CPU)
GPU1: 28c (fine for cooling a 42c GPU)
GPU2: 29c (fine for cooling a 42c GPU)
GPU3: 30c (fine for cooling a 42c GPU)

The temperatures in the GPU's do not gradually rise because they are all doing the same work and the water is more than capable of handling all of it.

So even though there are peaks and valleys in the water temperature as it passes through those devices, because of the difference in the temperature of the water and the temperature of the devices and because of the high rate of flow (not allowing any particular gallon of water to absorb much heat), the loop doesn't lose its ability to wick that heat energy away... and wouldn't until there were many more devices in the loop.

Example:
Loop of X number of CPUs and 1 Radiator.
The radiator would be able to cool X number of CPUs where X represents the number of CPUs it takes to produce enough heat energy to raise the temperature of the water to about the same temperature as the CPUs. Where the amount of heat going into the loop is as high as possible without exceeding the amount of heat going out of the loop.
Edited by Lucretius - 8/21/10 at 7:45pm
post #29 of 110
Quote:
Originally Posted by Lucretius View Post
I've changed my position on this.

The reason pgmoney gets the temps he does with daisy chained rads is because water temp never reaches the same temp (or higher) as his devices. If your water temp is lower than your device temp, heat energy will transfer to the water.

I don't know much about the rate at which that energy transfers (slows or speeds up) as the two mediums become closer and closer to the same temperature... but I'm sure there is some curve involved there.

If your rads cool the water to near ambient even though they probably don't (lets say 23c, which would be an average air conditioned house at 73f)... then as the water goes through his loop it isn't being heated enough to effect its ability to cool devices that are way higher temp... (53c and 42c)

A system of temp probes would be necessary to verify... or you could just stick a thermometer inside your reservoir and compare the water temperature to the temperature of GPU3.

Here is a total BS conceptual visual of what water temps might be throughout his loop.
Rad: 23c
CPU: 25c (fine for cooling a 53c CPU)
GPU1: 28c (fine for cooling a 42c GPU)
GPU2: 29c (fine for cooling a 42c GPU)
GPU3: 30c (fine for cooling a 42c GPU)

The temperatures in the GPU's do not gradually rise because they are all doing the same work and the water is more than capable of handling all of it.

So even though there are peaks and valleys in the water temperature as it passes through those devices, because of the difference in the temperature of the water and the temperature of the devices, the loop doesn't lose its ability to wick that heat energy away... and wouldn't until there were many more devices in the loop.

Example:
Loop of X number of CPUs and 1 Radiator.
The radiator would be able to cool X number of CPUs where X represents the number of CPUs it takes to produce enough heat energy to raise the temperature of the water to about the same temperature as the CPUs. Where the amount of heat going into the loop is as high as possible without exceeding the amount of heat going out of the loop.
exactly, thank you. plus i got a full board block between the cpu and the gpu's
but chipset does not put off alot of heat
    
CPUMotherboardGraphicsRAM
core i7 2600K ASUS P8P67 Deluxe EVGA gtx 580 SC 2x4 Corsair Vengence 1600mhz 
Hard DriveOptical DriveOSMonitor
G.SKILL Phoenix Pro 120GB ssd 2x1TB Seagate sony, philips Winodws 7 x64 PRO Hannspree 25" HF259H 
KeyboardPowerCaseMouse
Logitech G19 Enermax evo galaxy 1250 Coolermaster HAF-X CM storm sentineal game 
Mouse Pad
sub zero gamers pad 
  hide details  
Reply
    
CPUMotherboardGraphicsRAM
core i7 2600K ASUS P8P67 Deluxe EVGA gtx 580 SC 2x4 Corsair Vengence 1600mhz 
Hard DriveOptical DriveOSMonitor
G.SKILL Phoenix Pro 120GB ssd 2x1TB Seagate sony, philips Winodws 7 x64 PRO Hannspree 25" HF259H 
KeyboardPowerCaseMouse
Logitech G19 Enermax evo galaxy 1250 Coolermaster HAF-X CM storm sentineal game 
Mouse Pad
sub zero gamers pad 
  hide details  
Reply
post #30 of 110
I'm still curious about what the temperature of the water is at all points on your radiators.

1) Entry Rad 1
2) Exit Rad 1
3) Entry Rad 2
4) Exit Rad 2

It could be that your 2nd rad isn't doing as much as it could, even though it doesn't matter.
Edited by Lucretius - 8/21/10 at 8:51pm
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: Water Cooling
Overclock.net › Forums › Cooling › Water Cooling › The importance of the "order" of your loop?