Top Radiator .. Intake or Exhaust?

[BrokenArrow]

I'm still not liking the idea of intaking from the top since that goes against my brain waves of "heat rises"

Convection of "heat rising" is so easily overcome by the weakest radiator fans sucking in cold air.

For God's sake people, it only takes a few minutes to change a fan's direction, please experiment both ways and determine what works best for your rig.

As coolmeister said, too many variables and to each his own.

This is suppose to be a hobby and experimenting should be part of the fun instead of posting for advice.

[rge]

I have tried both ways and sucking in cold air from the top always wins. Try it yourself and see.

Ditto. I have 280 rad in front, and 360 rad up top, and I have run mine both ways with 10 accurate dallas temp probes and IR gun, and #1 not only gave me the best cpu and gpu temps, but also gave me the lowest motherboard, mosfet, and ram temps, and the internal case temps were within 0.1C either way.

May not make sense to some, until you try it yourself like I did, and realize preconceptions are sometimes completely wrong.

When I used figure 1, running prime 95, my external ambient temps were 24C, temps inside case after rad were ~26C. But key is those 3 fans in pull at top using figure 1, blow 26C air rapidly over mobo, ram, mosfets, dropping ram temps by 20C, mobo by 3-4C, and mosfets 6-8C. When I used figure 2, those same fans closest to mobo now pull 26C air across mobo (yep same temp regardless of what one thought) and push 26-27C air through rad, increasing cpu temps by 2-3C, and because pulling air across motherboard was far inferior to pushing air across it for cooling it the ram temps were 10-15C higher (than figure 1), mobo 3-4C higher, and mosfets 5-8C higher. Figure 1 was better for everything inside the case.

It is true that you will lose some rad effect from air resistance from having severely imbalanced air intake/exhaust, but the loss will pale in comparison to the benefits. On mine, I made a second exhaust in rear which along with just passive leakage was enough to retain most rad effect. You could exhaust out front like was mentioned, but you just need to mind whether your exhaust out front will heat up your local intake ambients. I have my exhaust all out rear then through hole in desk at back, so whole room has to heat up, before my local intake ambients will.

Now, after measuring things several times, I wouldnt prevent the expelled rad air from blowing over my mobo even if I could, after seeing how well (2C hotter than ambient air) but fast moving air cools mobo, and everything in case.

[PatRaceTin]

interesting

[PiLsY]

I've got floor, rear and top rads as exhaust at the moment. I tried the bottom as intake and the top and rear as exhaust - loop component temps went up by about 1c and case temp went up by almost 2c. Using all rads as exhaust everywhere i measure air temp inside the case is within 0.5c of ambient temp.

My case temp definitely goes up if any of my rads are set as intake. I have only tested with the fans at around 800rpm though. I'd imagine with higher fan speeds the trend would equalise and then reverse.

I can see what you're saying naekuh about those mosfets though. In my case I don't suffer for mosfet temps (under 50c even in linx) and I've got good clearance between the rads and the board - they're near nothing hot tbh.

rge makes a good point too - if mobo cooling as a concern in any way then blowing air across it will give better temps.

Think BrokenArrow said it best though. You should be taking the posts here as opinions - its up to you to figure out which will work best. I think you can see from the posts here though that whatever you try will work, but you'll have to experiment to find the best setup for your particular rig. In the grand scheme of things swapping fans round on rads really is about as easy as you can get for a loop rebuild . A large part of the fun in this hobby is getting the most from your loop and the countless amounts of configuration changes you can experiment with.

[PiLsY]

What?!?! Why would blowing heated case air through a rad give you best temps? This is a ridiculous statement and goes against the basic principles on how rads work.

If your case has good airflow your "heated" case air will be at most 1c over ambient. Please remember you're watercooling and the majority of stuff that makes heat in your case is now in the loop and applies its heat through the rads. Lots of mechanical hard disks can heat your case up but if you're imaginative and place them near an exhaust and not an intake that is not a problem. There is no reason why case air blown into a rad should be warm, let alone hot.

[aerial]

Imo there are other things worth taking into consideration, such as looks and dust. If you put all fans as exhaust, air sucked into the case through all small vent holes (passively) has very low pressure/speed - doesn't take much dust from outside, especially with lower rpm fans (thick rad). In the other hand, if fans are set as intake, dust filters or rad requires cleaning quite often. It is really good way of having dust-free case, without need of having and/or cleaning dust filters.
Another advantage of having fans set as exhaust, is that if you want them exposed. Having case with a window, if you want certain color scheme, you will see all fans via window (at certain angle ofc). If fans are intake, between case panel and radiator, they are pretty much hidden (push only configuration).
So in the end of the they, these things can be pretty important. In practice, for 24/7 low rpm system, these few degrees of difference between fan configuration are less relevant. I would say they are pretty much never relevant with usual overkill rad setup. On paper ofc all intake, cold air from outside will always win in performance contest.

[lilmanmgf]

I may have to less with mine again. It's 30 screws which really don't want to screw and unscrew though.

Sent from my GSBv1.9 ERIS using Tapatalk

[goodhead]

the extra temp diff to me is not worth the extra dust aerial said, absolutely hate dust so much that i have dust filters inside the case on the rad fans, double filtration = 1/2 time cleaning

just went from a 902 w/ 3 in 1 out to a 800d with 1 in 4 out and the diff in dust is night and day. def would not go back for few degrees

[Aaronharmon]

Well, logically speaking... Cold air rises and hot air decends so keep that in mind. Also, you're going to want that water moving through that loop as slow as possible. That way, it spends more time in the radiator. One more thing to keep in mind, you want your loop to be as restrictive as possible.

Here's a pro tip, no fans whatsoever. They generate too much heat.

[Waterlogged]

Well, logically speaking... Cold air rises and hot air decends so keep that in mind. Also, you're going to want that water moving through that loop as slow as possible. That way, it spends more time in the radiator. One more thing to keep in mind, you want your loop to be as restrictive as possible.

Here's a pro tip, no fans whatsoever. They generate too much heat.

So that's what I've been doing wrong all these years.

Seriously, this is the worst kind of misinformation...everything is completely wrong. Really hope this was a failed attempt at sarcasm.

[coolhandluke41]

Well, logically speaking... Cold air rises and hot air decends so keep that in mind. Also, you're going to want that water moving through that loop as slow as possible. That way, it spends more time in the radiator. One more thing to keep in mind, you want your loop to be as restrictive as possible.

Here's a pro tip, no fans whatsoever. They generate too much heat.

wow

[zalbard]

Seriously, this is the worst kind of misinformation...everything is completely wrong. Really hope this was a failed attempt at sarcasm.

That's obviously sarcasm. Your sense of humour failed.

[lilmanmgf]

Looks like sarcasm to me...

[Church]

Aaronharmon: Compare how fast air rises by convection above heating source in your room to airflow of even very slow fan. One can simply ignore convection in PC case, if it provides airflow only 10% from what slow fan can give. Then add to that fact that passive cooling is 4-10 times less effective then active one from same heat exchange area. Yes, you can compensate that with adding big pile of mega big rads. But imagine how monstrous result would look for passive cooling capable to cool hot modern hardware like overclocked 200-300W TDP cpu-s and multi gpu-s with such heat dissipation even at stock clocks? I don't think that PC with 2-4 Mo-Ra rads by it will look very good. +Yes, no fans are quieter then eg. 500rpm (for many almost inaudible) fans, but you still have fan in PSU, noise of hard disk drives .

= So ignore convection in air flow planning. If taking air from outside case is better, even if it goes against convection direction, so be it.

[zalbard]

Yes, no fans are quieter then eg. 500rpm (for many almost inaudible) fans, but you still have fan in PSU, noise of hard disk drives

Can use SSDs and a fanless PSU.
And house water supply to get rid of the pump!

[Church]

zalbard: most fanless PSUs i've seen are something like 400W. Not exactly well fit for heavily overclocked i7 + two 480/580/5870/5970/6970/6990 class gpu-s

[zalbard]

Use a few of them!

[Panthols]

Just get a 500w for the mb and a 400w for each of the video cards. Something like this http://plumbing.hardwarestore.com/51...le-600821.aspx to hook into your loop and you're set. No pumps, no radiator, no fans. Just make sure you keep the faucet set to cold.

[Aaronharmon]

Can use SSDs and a fanless PSU.
And house water supply to get rid of the pump!

No, I just hook the water lines to the eave spouts and wait for rain. Best option, really.

[Church]

Imho better option then passive PSUs is LC PSUs, like those of Koolance, with 1200-1600W output. Pitty they were lagging a bit in other areas, like efficiency, ripple suppression, cabling modularity, looks and so on, otherwise i loved that idea.
What i'd prefer most - some "kit" that could be used for most PSUs, so that one can take whatever PSU he likes and LC it.

[lilmanmgf]

How loud are your PSU fans? The fan in my antec truepower 750 is inaudible.

[Conumdrum]

Well, logically speaking... Cold air rises and hot air decends so keep that in mind. Also, you're going to want that water moving through that loop as slow as possible. That way, it spends more time in the radiator. One more thing to keep in mind, you want your loop to be as restrictive as possible.

Here's a pro tip, no fans whatsoever. They generate too much heat.

Hmm, sarcasm, yep. Appropriate? No.

[goodhead]

you guys are doing it wrong you just need to exhaust out to a roof fan duhh

case fans are so 2010

[Aaronharmon]

you guys are doing it wrong you just need to exhaust out to a roof fan duhh

case fans are so 2010

I have two of those big ones in push/pull

[NaeKuh]

you guys are doing it wrong you just need to exhaust out to a roof fan duhh

case fans are so 2010

*shrug*

Im on Ion Technology:

Can use SSDs and a fanless PSU.
And house water supply to get rid of the pump!

Dont u remember my thread?!?!?!?

My SSD's are probably the hottest part in my entire system!!!!!!!
Dont believe me?

Tell me what in my system minus my PhysX GPU is capable of running hotter then my SSD!!


Do you see now why i tell people i have no room to upgrade in LC?
My PhsyX GPu even has problems running colder them my 580GTX.


Disclaimer.. my DTS probes are ALL WRONG @ calibrated levels my system does.
So these temps are all wacked out.. which is why u never see me doing temp reviews anymore.

I have yet to find a working cpu, retail or ES, which has DTS calibrated down to the ranges i go, so they always report below ambient values.

And yes that 580GTX is retail reporting below Ambient Values.
You see now i call my system the physic buster... everything it touches idles below ambient and i there is nothing i can do to fix it.