• Welcome to Overclockers Forums! Join us to reply in threads, receive reduced ads, and to customize your site experience!

Rad - Push or pull? (Case flow involvement)

Overclockers is supported by our readers. When you click a link to make a purchase, we may earn a commission. Learn More.

Mpegger

Member
Joined
Nov 28, 2001
Normally, since I only use a single set of fans on my rad, I use them in a push setup. However, with my loop out of the case as I await a new cpu block, I've been re-evaluating this setup because of the case and number of fans in it.

My case is a NZXT Switch 810, and I currently have 5 140mm fans in the main fan positions of the front and rear of the case; 4 in the hard drive cage area (2x2 in push-pull config), and 1 in the rear which I have also switched to a intake. So that's 5 140mm fans in total pulling air into the case, with the intent of all the air being exhausted out the top of the case, through the 3x120 rad (Swiftech MCR320). My thinking, is that the 5 fans pulling air in and trying to force the air out through the top, would be almost equivalent to already having fans in a push config on the rad, so using the 3 120mm Gentle Typhoons on the pull side of the rad to exhaust out the case, would be close to a push-pull config, without 3 extra fans. This should help with keeping the water loop closer to ambient without having the fans (during the summer) at full speed all the time. Thoughts?
 
Last edited:
You would be wrong in thinking that having 5 fans pushing air in equals 3 fans push/pulling air through a rad...

The air being pushed in is going out of every crack, screwhole and joint in the case. Have you tested the temps of the drives with ONLY the 2 intake fans running and not havin them in push/pull? I would be surprised if there was a drastic difference in temps. In all likelihood, the Rad alone is restricting enough "ambient" pressure to negate a good portion of that airflow.

I run push pull on my systems, 2 with a CPU and GPU loop through a 3x120 Rad and 1 Corsair AIO. I also run Deltas or similar server grade fans simply because I can. If I throw max OC at them, I don't have to do anything but turn 4 knobs to speed up the fans.

The rig behind me was idling yesterday at ~44-46C in a room temp of 32-33C... 8 Delta AFB1212VHE at full speed isn't putting the rig at ambient, your gentle typhoons whispering at the rad certainly arent going to.
 
Push is always better than pull. Push/Pull is really a waste of time. The rig I'm building now has 9 fans pushing and 4 fans pulling. Unfortunately, my setup is the nature of the beast. I have 2 x 480's in the lower chamber. I don't want both blowing in, so I have one pushing in and one pulling out. I want the fans on the outside. I have a 480 at the top pushing out and a 120 pushing out.
 
Push is always better than pull. Push/Pull is really a waste of time. The rig I'm building now has 9 fans pushing and 4 fans pulling. Unfortunately, my setup is the nature of the beast. I have 2 x 480's in the lower chamber. I don't want both blowing in, so I have one pushing in and one pulling out. I want the fans on the outside. I have a 480 at the top pushing out and a 120 pushing out.

Push is NOT always better as Pull, it depends on a few other factors as well... Fantype, blade, rad thickness, fpi, (build-in) rad shroud, etc, etc
But most fans are *slightly* better at Push vs Pull... although you wont see that translated in a 10°C delta difference :), it has more to do with bearings & noise profile.


Push/Pull is ALWAYS better, it offers upto 30% improvement/flow for less noise.

However it DOES depend on the rad used... on a low fpi, 30mm AIO type rad it IS a waste of time. On a high-fpi, thick rad like some HWlabs or an 80mm AC Monsta , push-pull is almost a must, unless you want to use crazy, loud & expensive fans
 
Push/Pull is ALWAYS better, it offers upto 30% improvement/flow for less noise.
That was why I was thinking of installing the 3 GT fans on the rad on the pull side (exhaust). The 5 case fans all set as intakes should almost act like 3 fans on the push (intake) side of the rad, since the rad will be the exhaust point for the case. So without having to actually have 6 fans on the rad, I can still get a push-pull effect.

Btw, when I said ambient I meant the water temp. :eek:
 
That was why I was thinking of installing the 3 GT fans on the rad on the pull side (exhaust). The 5 case fans all set as intakes should almost act like 3 fans on the push (intake) side of the rad, since the rad will be the exhaust point for the case. So without having to actually have 6 fans on the rad, I can still get a push-pull effect.

Btw, when I said ambient I meant the water temp. :eek:

only if you (hermetically) seal the rest of your case... you are overestimating the "positive pressure" adagio :)

for starters, it is very difficult to create any meaningfull positive pressure with normal axial fans... they are just not very good at it (put your hand in front of one and it will blow backwards)

Second, most cases "leak" like a sieve through a whole range of cracks & joints, not to mention the perforated PCI slot blanking plates that have become the "standard".
As Adragontattoo said

That said.. i *can* be done, but its a numbers game... if your intake fans have filters, then most likely their CFM is only half or less IF they have the SP to overcome the filter restriction in the first place. So, 5 intake equals 3 exhaust is an optimistic guesstimate... you need the performance curves of the fans to work out the (theorethical) optimums

and you still have to seal all the cracks & crevices :)

That said... just try it

PS: in most cases, using a shroud between rad & fan and having the fans in "Pull" will yield better results as having the fans bolted directly to the rad in push OR in pull.
In some cases dhaving a shroud between rad & fan and having the fans in push yields better results.
It has to do with minimising the "hub shadow" and thus having a larger radiator surface exposed to airflow.
 
Last edited:
I have the same rad, and have it setup in pull. It works great, and keeps my gpu and cpu plenty cool. I say go for it.
 
Simple to test on your own...do you have rpm monitoring on one or all of the top fans? Do the top fans slow down slightly when you open the case door?
If so, then they are boosting what's going through the rad. If not, then you're getting nothing from changing.

I personally have always like pusher fans on the rad. The extra turbulence helps "scrub" off heat. Didn't Martin or Skinee prove that once??
 
Simple to test on your own...do you have rpm monitoring on one or all of the top fans? Do the top fans slow down slightly when you open the case door?
If so, then they are boosting what's going through the rad. If not, then you're getting nothing from changing.

I personally have always like pusher fans on the rad. The extra turbulence helps "scrub" off heat. Didn't Martin or Skinee prove that once??

Nope, inconclusive.. to many variables at play... RAD itself, fpi, fan profile, hub size, blade rake & shape, etc etc

More fans are slightly better at push, probably its easier on the bearings compared to pull..

However, as soon as you add shrouds their "character" becomes more pronounced and it becomes easier to choose the right rad, fan, push or pull combo.
 
Nope, inconclusive.. to many variables at play... RAD itself, fpi, fan profile, hub size, blade rake & shape, etc etc .

Ummm, none of those variables change at all from just moving a fan from push to pull.
And it's free to try unless you're billing yourself for your own time.
 
Just a little more info on why I believe this should work:

My previous case was a XClio Windtunnel, a case which was not made for water cooling, but which I cut the top panel to accept the same 3x120 rad I am still currently running with. Initially, I had only, and I do mean only the 2 250mm case fans as the only fans in the entire case. And they were more then enough to keep the water at only a couple F above ambient, in a loop that was cooling both the CPU and GPU. Those 2 250mm fans were not screamers (only around 800rpm iirc), and most certainly did not have the pressure that better fans do, but they definitely moved air. I have no idea on their CFMs, but as slow as they turned, for their size they were more then capable, and probably even better at moving large volumes of air, then even the best 120mm fans on a rad would be. And yet, with as little pressure as they may have had, they were still able to push through the rad and keep the water (and entire system) cool.

Hence why I believe that the 5 intake fans should be almost the same as having 3 fans as push on the rad, and that using the 3 rad fans on the push side would just be a waste, and would be better to use those 3 on the pull side.

I'm still trying to decide if I will delid my 4790K, and then redo the loop again. At that point I would try with the fans on the pull side, though it won't make too much of a difference, as the water already idles at ambient intake temps. (The water temp probe must be off, as it's reading 77F, while the intake air temp is around 82F.)

[edit] Using 2 different IR temp meters, at the external reservoir, both are reading about 86F. The reservoir is after the cpu block in the loop, and the water temp probe is in the pump, which is after the reservoir. I don't know how accurate a IR meter is through acrylic, which is what the reservoir is made of. [/edit]
 
Last edited:
Ummm, none of those variables change at all from just moving a fan from push to pull.
And it's free to try unless you're billing yourself for your own time.

Actually they DO, but not always. Some rads are not "centered" in their housing, which means their "build-in" shroud is a bit "deeper" on one side. Thus bolting the fans on the other side while keeping the same airflow direction, changes TWO parameters. If you would want to change only ONE parameter, one would also have to flip the rad around.

Anyways, lets rephrase my statement into : too many variables at play to be able to post a unifying theory :)

Some fans do slightly better in push, some slightly better in pull, some do better with a rad shroud, some dont, etc

Thus: fan X in pull on rad Y will do better with a shroud vs unshrouded in push or pull; however same fan X will do the same with or wihtout shroud on rad Z. etc etc

But you cannot say any fan will always do on any rad.

Now, when DO means overall loop performance.. when all is said & done might just be +/- 3 °C

There are other factors.. which may or may not be more important as temps... when DO means noise & tonality or "timbre", in that case shrouds can make a much bigger difference


I'm one of those guys that will take the little percentages. Say that using a shroud on the Fan-A / Rad-Z combo may yield 2% temp improvement AND at the SAME time definitely improves the sound/noise profile (eg 1db drop or "smoother" sound ) @ 1350 RPM then i'll take that :)

:chair:
 
Last edited:
Just a little more info on why I believe this should work:

My previous case was a XClio Windtunnel, a case which was not made for water cooling, but which I cut the top panel to accept the same 3x120 rad I am still currently running with. Initially, I had only, and I do mean only the 2 250mm case fans as the only fans in the entire case. And they were more then enough to keep the water at only a couple F above ambient, in a loop that was cooling both the CPU and GPU. Those 2 250mm fans were not screamers (only around 800rpm iirc), and most certainly did not have the pressure that better fans do, but they definitely moved air. I have no idea on their CFMs, but as slow as they turned, for their size they were more then capable, and probably even better at moving large volumes of air, then even the best 120mm fans on a rad would be. And yet, with as little pressure as they may have had, they were still able to push through the rad and keep the water (and entire system) cool.

Hence why I believe that the 5 intake fans should be almost the same as having 3 fans as push on the rad, and that using the 3 rad fans on the push side would just be a waste, and would be better to use those 3 on the pull side.

I'm still trying to decide if I will delid my 4790K, and then redo the loop again. At that point I would try with the fans on the pull side, though it won't make too much of a difference, as the water already idles at ambient intake temps. (The water temp probe must be off, as it's reading 77F, while the intake air temp is around 82F.)

[edit] Using 2 different IR temp meters, at the external reservoir, both are reading about 86F. The reservoir is after the cpu block in the loop, and the water temp probe is in the pump, which is after the reservoir. I don't know how accurate a IR meter is through acrylic, which is what the reservoir is made of. [/edit]

hmm, yeah.. well ok. I doubt that your experience is representative, but by all means, please go on with it :)

However.... on thing thou that everybody will agree upon:

IDLE TEMPS are meaningless


I can switch the fans on my rad OFF and still be at 9-10-11°C over ambient on IDLE thanks to the sidepanel exhaust @ 30% PWM moving enough air to "suck" fresh air through the rad. But at LOAD the temps would become VERY unhealthy very quickly for the CPU if the fans were off. Actually reaching thermal shutdown after a few minutes of Prime, even while the rest of the casefans ramp up to 100% in an attempt to compensate.
 
True, there are a bunch of wonky fans that shouldn't even be mounted on a radiator in the first place. My general rule of thumb (so to speak) is if you can stick a finger through the blades perpendicular to the frame, don't buy it...
Then it is anybody's guess as to which way is best to run it, trying both is still best.

Sanyo D, and Panaflo are usually my only choice, but I did break down and get some gentle typhoon 1450's for this rig, and they're doing okay. Noctuas were too wide open and left a much larger Delta from idle to load.
 
Back