Overclock.net › Forums › Cooling › Water Cooling › Cold air In or Warm air Out for radiator? Also chipset cooling?
New Posts  All Forums:Forum Nav:

Cold air In or Warm air Out for radiator? Also chipset cooling? - Page 4

Poll Results: Cold air In or Warm air Out for radiator

 
  • 60% (12)
    Cold air In to radiator
  • 25% (5)
    Warm air Out from case through radiator
  • 15% (3)
    Makes no difference
20 Total Votes  
post #31 of 49
Quote:
Originally Posted by Ashcroft View Post


Of course air flow volume makes a difference.

I assume you mean the case internal temp rather than ambient but a radiator putting out 20c warmer air has to be more than 20c over ambient itself to do that. By using cool ambient air to cool the rads they are cooler and put out cooler air.

Even my single rad does not put out air anywhere near that warm and does not over heat the air cooled motherboard components in my small system. A dual rad system will do so even less.

The testing in the link above showed some air cooled component temps lower with the rads intaking so they somehow avoided it too.

The idea that rads heat internal components in a case significantly, and that the order of components and rads in a loop has an effect on performance is a common and completely understandable supposition by people that haven't tested it.
It's natural to think of rads as being hot and the coolant coming out of them being cool again. Water cooling in PC's deals with subtle temp differences rather than large ones like a car cooling system. The very mildly warmer air from a rad can still effectively cool air sinks but that mildly warmer air is enough to make a large difference to another rad because it is also cool. You only need to touch the heat sinks on a MB and a rad to tell how different they are.

That all only applies if you use (noisy) high-flow fans with a very low resistance radiator, which wont cool the water very effectively unless you only use low-power parts. If you use a thick and/or dense radiator then the air will get significantly warmer towards the water's temperature and no amount of airflow will cool the remaining components below the temperature of the warm air.

If a radiator is 50C then any object in its exhaust path will also end up being 50C, regardless of its materials or amount of flow, if said object also needs to dissipate its own heat then it'll definitely end up above 50C (often twice this for passive components).
   
build server
(10 items)
 
CPUMotherboardGraphicsRAM
FX-8150 (stock currently) ASUS Crosshair V Formula AMD R9-290X Corsair Vengeance 16GB @1866MHz 9-10-9-27 
Hard DriveHard DriveHard DriveHard Drive
Samsung 850 pro 2x256GB RAID0 WD Red 2x2TB RAID1 OCZ Vertex 3 120GB Seagate Shingle 8TB 
CoolingOSMonitorKeyboard
EK Supremacy EVO + EK 290X R2 + backplate + EK ... Windows 10 Pro 4k 24" samung PLS freesync Corsair K70 Red 
PowerCaseMouseMouse Pad
Seasonic 1000W Platinum Raven RV01 Corsair M60/M45 Corsair 
AudioOther
HDMI 8ch LPCM 24@192k DAC generic VIA usb3.0 card for an eye tracker that... 
CPUMotherboardGraphicsRAM
AMD AthlonII 640 x4 Gigabyte 880GM-USB3 nvidia GTX 460SE Corsair Dominator 2*2GB 
Hard DriveOptical DriveCoolingOS
Seagate 500GB some sony DVD burner stock, case has a nice side duct though Windows 7 home premium 64bit 
PowerCase
Arctic Cooling 550R Gigabyte mATX 
CPUMotherboardGraphicsRAM
Intel Atom 230 some all-in-one Mini-ITX board Intel GMA 950 some 1GB stick 
Hard DriveOptical DriveCoolingOS
Seagate 320GB LightScribe DVD burner? passive stock Oracle Linux 
PowerCase
integrated in the motherboard, 65W power brick compaque stock Mini-ITX case 
  hide details  
Reply
   
build server
(10 items)
 
CPUMotherboardGraphicsRAM
FX-8150 (stock currently) ASUS Crosshair V Formula AMD R9-290X Corsair Vengeance 16GB @1866MHz 9-10-9-27 
Hard DriveHard DriveHard DriveHard Drive
Samsung 850 pro 2x256GB RAID0 WD Red 2x2TB RAID1 OCZ Vertex 3 120GB Seagate Shingle 8TB 
CoolingOSMonitorKeyboard
EK Supremacy EVO + EK 290X R2 + backplate + EK ... Windows 10 Pro 4k 24" samung PLS freesync Corsair K70 Red 
PowerCaseMouseMouse Pad
Seasonic 1000W Platinum Raven RV01 Corsair M60/M45 Corsair 
AudioOther
HDMI 8ch LPCM 24@192k DAC generic VIA usb3.0 card for an eye tracker that... 
CPUMotherboardGraphicsRAM
AMD AthlonII 640 x4 Gigabyte 880GM-USB3 nvidia GTX 460SE Corsair Dominator 2*2GB 
Hard DriveOptical DriveCoolingOS
Seagate 500GB some sony DVD burner stock, case has a nice side duct though Windows 7 home premium 64bit 
PowerCase
Arctic Cooling 550R Gigabyte mATX 
CPUMotherboardGraphicsRAM
Intel Atom 230 some all-in-one Mini-ITX board Intel GMA 950 some 1GB stick 
Hard DriveOptical DriveCoolingOS
Seagate 320GB LightScribe DVD burner? passive stock Oracle Linux 
PowerCase
integrated in the motherboard, 65W power brick compaque stock Mini-ITX case 
  hide details  
Reply
post #32 of 49
Quote:
Originally Posted by Paul17041993 View Post

If a radiator is 50C then any object in its exhaust path will also end up being 50C, regardless of its materials or amount of flow, if said object also needs to dissipate its own heat then it'll definitely end up above 50C (often twice this for passive components).
This is assuming that the heat of the radiator will transfer all heat to the air. The air will be warmer than ambient, but will not be 50c. If the loop is 15c delta, the air temperature will not increase 15c. It may be a 5-7c increase in air temperature at most.

It will definitely increase the temperature of the air cooler components inside of the case, but not at the drastic level that you mention. I know you are using these numbers to explain the theory, but the examples you show are not representative of real world values.
post #33 of 49
It's a common shorthand to say that loop order doesn't matter, but the truth is it can make a very slight difference--in most cases practical and aesthetic considerations for component placement and tube runs override that.

I'm also a little surprised by people saying the air coming off their rads never gets hot. Either you have massive rad over-capacity, the ambient air temp in your room is really high already, or you never run your rig at load. With an "average" loop following the formula of one 120 rad per major chip + 1, when your chips are under load and your fans are running, most people are going to get a very strong change in air temp coming through the rads.

I also think a few of the above comments don't consider the fact that that air, water, and copper all react to heat very differently. The same temperature in an identical mass of those three substances represent vastly different levels of actual heat energy. And the same amount of heat energy will result in different temperatures. For every degree water changes temp, the same mass of air will change by about four degrees. Comparing by volume, the multiple is in the thousands. So if you see a one degree difference in water temp between the input and output side of a rad, that translates into a big change in the small volume of air blowing through it.
Edited by threephi - 4/10/17 at 5:36pm
Blinky
(16 items)
 
  
CPUMotherboardGraphicsRAM
i7-950 @4.034 GHz Asus Rampage III Formula GeForce GTX 770 6GB Crucial Ballistix 1600MHz 
Hard DriveCoolingCoolingOS
Samsung 840 Pro SSD EK Supreme HF (full nickel) Danger Den Monsoon Dual-Bay Reservoir/Pump Win 7 Pro 64 
MonitorKeyboardPowerCase
Dell U2410 Rosewill RK-9000 Seasonic X-650 Cosmos 1000 
Other
Lamptron FC5-V2 
  hide details  
Reply
Blinky
(16 items)
 
  
CPUMotherboardGraphicsRAM
i7-950 @4.034 GHz Asus Rampage III Formula GeForce GTX 770 6GB Crucial Ballistix 1600MHz 
Hard DriveCoolingCoolingOS
Samsung 840 Pro SSD EK Supreme HF (full nickel) Danger Den Monsoon Dual-Bay Reservoir/Pump Win 7 Pro 64 
MonitorKeyboardPowerCase
Dell U2410 Rosewill RK-9000 Seasonic X-650 Cosmos 1000 
Other
Lamptron FC5-V2 
  hide details  
Reply
post #34 of 49
If you only have 120x120 rad surface for each component... then you are under radiated. That 'formula' may have worked at some point, but it does not take into account the wattage of the component. 120 surface area is good for about 100w dissipation, if you put a 250w gpu in a 240 loop with a cpu , and expect decent thermals you followed the novice formula which is absolutely wrong.

If you have an appropriate level of surface area, the water temp shouldn't get more than 10c over ambient, and if you have an alternate intake, the case temps shouldn't rise more than a degree or two.

I think loop order matters... but there are definitely other variables in designing the loop which may make you stray from the best flow layout.
post #35 of 49
My post was aimed more at some of the comments above yours, but FWIW the so-called formula calls for one rad more than what you described, and IMO it's not a bad approach to start out with. YMMV but I don't think it's fair to call it "absolutely wrong".

The gist of it was aimed more at the idea implied in a few earlier posts that changes in air temp and water temp were on a 1:1 relationship. This would only be true if water and air had the same thermal properties, and they don't.
Blinky
(16 items)
 
  
CPUMotherboardGraphicsRAM
i7-950 @4.034 GHz Asus Rampage III Formula GeForce GTX 770 6GB Crucial Ballistix 1600MHz 
Hard DriveCoolingCoolingOS
Samsung 840 Pro SSD EK Supreme HF (full nickel) Danger Den Monsoon Dual-Bay Reservoir/Pump Win 7 Pro 64 
MonitorKeyboardPowerCase
Dell U2410 Rosewill RK-9000 Seasonic X-650 Cosmos 1000 
Other
Lamptron FC5-V2 
  hide details  
Reply
Blinky
(16 items)
 
  
CPUMotherboardGraphicsRAM
i7-950 @4.034 GHz Asus Rampage III Formula GeForce GTX 770 6GB Crucial Ballistix 1600MHz 
Hard DriveCoolingCoolingOS
Samsung 840 Pro SSD EK Supreme HF (full nickel) Danger Den Monsoon Dual-Bay Reservoir/Pump Win 7 Pro 64 
MonitorKeyboardPowerCase
Dell U2410 Rosewill RK-9000 Seasonic X-650 Cosmos 1000 
Other
Lamptron FC5-V2 
  hide details  
Reply
post #36 of 49
Quote:
Originally Posted by threephi View Post

My post was aimed more at some of the comments above yours, but FWIW the so-called formula calls for one rad more than what you described, and IMO it's not a bad approach to start out with. YMMV but I don't think it's fair to call it "absolutely wrong".

The gist of it was aimed more at the idea implied in a few earlier posts that changes in air temp and water temp were on a 1:1 relationship. This would only be true if water and air had the same thermal properties, and they don't.
maybe I was a little harsh by saying it is absolutely wrong, but I never agreed with that methodology. If you have 2x 1080ti, and a six core cpu on 4x120mm of rad space following the 120xcomponent+120 method, you will have a hot mess on your hands. That would give you 4x120, where even having 6x120 on those components would be on the lower end of the spectrum for that hardware.

You don't need a thermal engineering degree to design a watercooling loop, but you need to have a general idea of the wattage of the parts, and the wattage of dissipation that the cooling elements are capable of.

In general, I feel that radiators work best as intakes, since you get cooler water temperature. With decent airflow in the case, case temperatures are still relatively low.
post #37 of 49
Quote:
Originally Posted by EMUracing View Post

You don't need a thermal engineering degree to design a watercooling loop, but you need to have a general idea of the wattage of the parts, and the wattage of dissipation that the cooling elements are capable of.
^^+1
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 #38 of 49
Quote:
Originally Posted by Ashcroft View Post

That is not what the testing I have seen shows or my own more basic testing. I have never seen any that contradicts it.


http://www.overclockers.com/forums/showthread.php/700744-AirFlow-direction-for-rads?p=7111920#post7111920

This testing shows a slight increase in some internal temps, not all, but a significantly lower air/water delta.

The coolant coming out of blocks should not be hot unless the flow rate is insufficient.
Although intake rads will send mildly warmer air into the case it also sends a greater volume of air flow over those air cooled parts. A very low air flow rate at A lower temp can cool worse than a higher temp but greater air flow.

As for the order of components affecting intake and exhaust decisions, I haven't seen any testing but it seems like more theory than practice. If there was a significant temp difference between the rads I could see it for sure, but as before their should not be a large temp difference. It might be enough to make a measurable difference, but significant I find doubtful.

In a full water cooled loop

The only factor needed to listen to in a custom loop tongue.gif

If you have bad restriction, a low flow block, clogged filters/plates/coolant-particles(no comment), YOUR LOOP DOESN'T WORK RIGHT and you can't comment on anything. If your loop can flow properly and not choke your pump out to die, your coolant temp will be less than 1c difference throughout it.

Now, having a bad flowing loop because of improper setup or whatnot, then the exhaust/intake vise versa before/after can come in to play. In my words for that, WASTED MONEY!! Loop wasn't setup right, or wrong. Spent more money on 2nd rad to fix inexperience. Not to be rude thumb.gif

Rads can only take so much heat out. IF, big IF <- , your loop can flow properly, a double in the front and a double in the back is the same as a quad on top/bottom yessir.gif

If it doesn't flow properly, you can add as many as you want and see 5 different temps across the loop until it can flow right wink.gif

But this info comes from me running temp sensors all over my loops, in and out flow everywhere to understand completely. And custom modifying some blocks to flow better as well. Hence all the "HIGH FLOW EDITION" blocks these days, or the most just make it that way now and don't put in their naming.


In a AIO cooled cpu and air cooled gpu, things get trickier. Blower style gpu cooler will pull almost all heat out the back, not affecting the AIO to a point. New style / non-exhaust coolers will heat up your internal temps and cause higher coolant temps on AIO's if setup as Exhaust. Now setting up an AIO on intake and having non-exhausting gpu cooler will make a hot mess in your case after gaming for a while. 5 minute testing doesn't mean squat. who games for 5 minutes besides 80yr old grandma HA. Your exhaust will eventually affect your ambient temps which will affect your internal temps. It's a no win situation until things you bought work as intended. If using a blower gpu cooler and a AIO on cpu, you will always be better off with intake fans in front, and AIO exhausting through the top as heat rises.
Edited by Hambone07si - 4/11/17 at 3:48pm
Red X9
(20 items)
 
  
CPUMotherboardGraphicsGraphics
4790K @ 5ghz 1.350v 75c max load temps Asus Maximus VI Formula Evga Gtx Titan X @ 1500mhz / 7800mhz +112mv EK Acetal / Nickel / Nickel back plate 
RAMHard DriveHard DriveOptical Drive
Corsair Vengence Pro 32gig 2400mhz c11 red (4x8... 6x 256gig (1.5tb) in Raid 0 , OS / Gaming Drive 2 x Liteon 256gig (512gig) storage drives Asus Dvd-RW 
CoolingCoolingCoolingOS
2x 360mm Alphacool XT45 / 1x 480mm Alphacool Mo... EK X-Res 140 D5 vario Pump/Res combo + EK X-top... EK Supreme HF cpu / EK Acetal / Nickel with Nic... Windows 8.1 64bit 
MonitorMonitorKeyboardPower
Asus ROG Swift 27" 144hz with G-Sync Samsung 65" 7150  Logitech G510s Corsair AX1200 
CaseMouseMouse PadAudio
Thermaltake Core X9 custom painted Red and Black Logitech G602 Wireless Corsair Aluminum Core 2 sided  Bose Companion 5 
  hide details  
Reply
Red X9
(20 items)
 
  
CPUMotherboardGraphicsGraphics
4790K @ 5ghz 1.350v 75c max load temps Asus Maximus VI Formula Evga Gtx Titan X @ 1500mhz / 7800mhz +112mv EK Acetal / Nickel / Nickel back plate 
RAMHard DriveHard DriveOptical Drive
Corsair Vengence Pro 32gig 2400mhz c11 red (4x8... 6x 256gig (1.5tb) in Raid 0 , OS / Gaming Drive 2 x Liteon 256gig (512gig) storage drives Asus Dvd-RW 
CoolingCoolingCoolingOS
2x 360mm Alphacool XT45 / 1x 480mm Alphacool Mo... EK X-Res 140 D5 vario Pump/Res combo + EK X-top... EK Supreme HF cpu / EK Acetal / Nickel with Nic... Windows 8.1 64bit 
MonitorMonitorKeyboardPower
Asus ROG Swift 27" 144hz with G-Sync Samsung 65" 7150  Logitech G510s Corsair AX1200 
CaseMouseMouse PadAudio
Thermaltake Core X9 custom painted Red and Black Logitech G602 Wireless Corsair Aluminum Core 2 sided  Bose Companion 5 
  hide details  
Reply
post #39 of 49
Thread Starter 
OP here with an update:
I had to relocate the H60 radiator to the front due to case space limitations. The updated motherboard has so much plastic on it that the rad no longer fits on the roof.

Now it has a 140mm fan doing the pulling, and the H60 fan was repurposed as a roof exhaust. Water inlet-outlet temperatures over a 10-minute stress run stabilized at 40-38 C and cpu at 71C (RyzenMaster read 91 C with offset). This is better than what I had with air cooler (got to 74.25 C in five minutes so I stopped the cpu stress test).
My question is: looking at these numbers, with water at 40 and cpu at 71 do you think my cooler is mounted correctly? It is seated tight, with a little heat transfer compound.
Power dissipated at cpu ~120W (by subtraction between idle, 1-core and multicore tests), which is a bit more than 95W stock rating, measured at power plug with the Cyberpower UPS.
Just wanted to make sure if these temps look reasonable for a 120-mm rad.
post #40 of 49
Have you mounted it on the outside of the case and are using one of those clear fans in pull config? If so I would change that. Move that clear fan, mount the rad where it was with the original rad fan mounted on it in push config. To reiterate....both the fan that came with the rad and the rad are to be on the inside of the case.

I am kinda assuming that you had those fans like that to have both led effects in the front of the case. Those fans are (from the looks of them) not meant for rads...ie poor sp.

Ofc I could be wrong. I am just not aware of any clear 140mm fans with decent static pressure.
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: Water Cooling
Overclock.net › Forums › Cooling › Water Cooling › Cold air In or Warm air Out for radiator? Also chipset cooling?