CPU cooler fan direction....

I have a CoolerMaster Hyper212+ cooler on my PhenomII 955BE at stock speeds, and my case is a standard tower case with the motherboard mounted vertically. I currently have the cooler mounted so the fan blows from front-to-back (or east-to-west), directly into the rear exhaust fan of the case. I've read that 'tower' type CPU coolers like mine actually work a bit better when the cooling 'tower' is horizontal (mine is currently vertical) and with the fan blowing from bottom-to-top (or south-to-north). This way, the air either goes straight in to the PSU and is exhausted, or it goes in to a top case fan if you have a case with bottom-mount PSU. Supposedly, this horizontal mounting can drop CPU temps as much as 4-5C. They say it has something to do with the effect gravity has on heatpipes and the liquid inside..... any thoughts?

(for a 360 view of my CPU cooler, check this link)

 

I would say that this heatsink is meant to be mounted with the pipes vertical and the vanes horizontal. Since the vanes are parallel to the motherboard this implies a horizontal motherboard.

I say this because the pipes are not complete loops so there is no return path, except by internal wicking.

From what I can see of Halo's postage stamp his is a helicoil which should work equally well in either orientation.

 

:confused I'm not sure what you mean . . . in a regular tower case, with the MB in a vertical position, should I have the cooler set as in the 1st pic below, with the fan blowing south-to-north, or as in the 2nd pic, with the fan blowing east-to-west? I ask because I just got a new case (came in today! :celebrate ) and it has a bottom-mount PSU, so I have exhaust fans on the top of the case.... and I've read that heat-pipe tower-type coolers work better when they don't 'fight' gravity and supposedly the south-north scenario is more 'gravity friendly'....

@ Halo: I see you prefer Noctua fans LOL

(BTW- my new case is the NZXT Beta Evo here)

(BTW2- my temps are OK, I was wondering if there's any truth to the 'rumors' online LOL )

 

BTW3- at Newegg (link above) the NZXT Beta Evo case currently has a $10 discount with a 'promo code', free shipping, and a $20 mail-in-rebate!!! Only $30 total (if they actually mail you the rebate )!!! A good deal!!!

 

OK to amplify my comments.

I'm suggesting that the vapour in the copper pipes in DLB's heatsink will naturally rise in a vertical pipe and fall back when re-condensed to liquid.
In a horizontal pipe there is no preferred direction for either liquid or vapour.

 

I have always been puzzled why the tube style heat sink manufacturers do no duct their fans, a propellers efficiency is greatly improved by ducting. Note all case fans and all the older CPU's had ducted fans, I see these late setups with the pipes and open fans. I am willing to be that putting even a paper strip to shroud between the coils would reduce temperature, doing so would force all flow through the vanes non would be lost off the blade tips. My case has a side port ducted to the processor fan which sits right on the processor, to me it is just common sense putting all the air flow to good use.

 

Interesting discussion once again dbl. I'm interested in the thought that the pipes are full of a liquid, I have never seen this written and always assumed it was just a gas (any old current workshop atmosphere) that was "sealed" within the HS tubing/pipe. If it is a liquid it will indeed have a relationship between orientation and cooling potential. I would've thought that if that was the case then "horizontal" would be the optimised orientation so the liquid would accumulate and "distribute evenly" across the base via gravity. Whereas on a vertically mounted mobo the liquid will "tend towards" the bottom of the case side of the HS and cool "the bottom 1/2 of the HS" (irrespective of orientation) and thus the CPU slightly greater (related to CPU and Ambient Temps will be indicative of liquid to gas ratios and thus "affect mass dispersal" (T~m/v). That is the hotter it is the more the liquid is gas and the more the mass is dispersed affecting localised mass collection potential (m to v) and thus temps (localised).

Sorry I'm waffling.

A gas would see heat distributed a lot more evenly and case design and air flow potential would be the "dictators".

I agree with Halo Video Card RAM and PCB get very hot and may affect HS inflow temps!

 

To support my thinking I had recently read of a Air Cooler that did "specifically" have liquid filled tubes, looked similar to a TRUE120 form factor but had the addition of a chamber (reservoir) "atop" the tubes (HS) and thus gas would accumulate and revert back to liquid and flow back towards the HS base via gravity. Overall results were very average and the cost was "prohibitive".

Sadly the make and model eludes me at this moment.

 

No the pipes are far from full. If they were there would be no room for vapourisation.

In fact heatpipes are even more sophisticated.

They contain a 'wick' . The internal liquid vapourises at the hot end, moves along tthe pipe to the cold end and condenses back to a liquid. The internal wick transports the condensate back to the hot end by capilliary action.

Thus you have vapour passing in one direction and liquid passing in the opposite direction within a single pipe.

Also this is how near horizontal heatpipes in laptops can function.

One point, system efficiency increases with temperature difference between the hot end and the cold end - it will not work at all if there is no difference - so the idea of cooling the whole pipe is actually counter productive.

All quite clever really.

 

Awesome answer and "physically probable" so distribution potential variations are limited at best and negligible but active cooling impairs the process, interesting.

 

If you look at heatpipe type coolers, the hot end is normally a substantial block of metal.

The pipe leads away from it but there are no dissipating fins etc until quite a distance from the block. This is particularly marked on laptop types.

The cold end is a bit like a car radiator and the idea is that the fan plays directly on this. Many are ducted to prevent cooling occuring in the pipes themselves.

One reason the car radiator is separate from the engine block has the same thermodynamic roots.

 

So based on thermal dymaics, surely the best way to have the pipe assembly is with the hottest bit at the bottom (i.e. the processor) and the coldest bit at the top (i.e. the fan/vane area). So surely the best way is to have the motherboard horizontal with the cooler standing on it?

 

Does anyone know what liquid is used? My first thought would be Ammonia as it boils well below most others, as it is heated up tiny bubbles form which would rise to the top cool/condense and drop to repeat the cycle. Water would not be practical the temp needed for it to do its job would be way to high for our machines.
I doubt the cooler orientation has much effect on cooling, though if you think for a minute setup vertically you may actually have better flow as the fluid would circulate rather then rise and fall slowly back down! That should get some of you thinking.

 

Wow this has been a very interesting discussion and I have certainly learnt something new. Indeed the "heatpipe" Air Coolers have an internal liquid (unlike my earlier assumption of just a gas) and the earlier versions solely relied on gravity for liquid distribution, but as clearly pointed out studiot the current coolers have a wick and use capillary action to distribute the liquid:
Scroll down to "Heatpipe"
http://en.wikipedia.org/wiki/Computer_cooling
Awesome pics:
http://encyclopedia2.thefreedictionary.com/Heat-pipe

Okay with that indicated an excellent example of capillary action (defying gravity) is a tree (or any plant) and how it draws moisture up to the leaves from the roots in the ground to provide moisture and cool the growing leaves/tips (I can give you a lot more detailed breakdown on this if you want).

With the cooled liquid being distributed via wicks and thus capillary forces (no gravity) I believe HS orientation will have "no significant" influence on liquid movement and thus HS cooling potential. The case airflow potential and as Halo pointed out the Video Card (RAM, VRM & PCB) can influence HS inflow temp potential and may have an influence on cooling.

I use front to back and as I only have 1 of the 5 1/2" bay used I have another 120mm cable tied in the empty 4 slots blowing straight at my CNPS9500 (my whole front panel is perforated and filtered) which then dumps the air straight at a rear 120mm exhaust. So my airflow has dictated my orientation.

OldPhil the 2nd link indicates that the liquid is water or acetone. You will want the right liquid to ensure that is converts to gas at the right rate.

 

Of interest in the 2nd link is the 2nd pic where the whole side of the case in used as a passive cooler for the HS (both CPU and Video Card) and removes "noisy fans".

 

I think the liquid is a complex proprietary (=secret) mix of organic voltatile compounds. These do not chemically attack copper pipes as ammonia does. I would expect low molecular weight alcohols, glycols and ketones since freon (halocarbon) type substances are now banned as refrigerants.

 

W O W ! ! ! This thread really took off in an unexpected (yet excellent and informative) direction. Very interesting stuff here, I'm a big fan of learning something new.

So I finally got my system rebuilt in to the new chassis (NZXT Beta Evo), and I kept my CPU cooler blowing east-west as it was before. This new case has lowered my temps BIG TIME!!! The CPU idle temp is down by 8c-12c, and my GPU idle temp is down by +/- 14c. I'm about to start gaming, and I'll log the temps and post 'em if anyone cares . . . nahh, nevermind . . . I'll post 'em whether anyone cares or not!!! LOL roflmao

Check out this article which exposes the internals of heatpipe construction, fabrication, and materials. Good stuff!

 

After thinking about this issue more and what I wrote post#10 I was actually thinking about an article I had read on a "Air Cooler with a pump" that used a "liquid metal" with "electromagnetic pump" for the cooling. Sorry for my error but the review on the Cooling system said it was very expensive and did not provide "substantial gains" if any differences at all.

I have been trying to find the article for several days and it has mostly eluded me but here is a blurb on it:

http://www.hitechreview.com/it-prod...-available-liquid-metal-based-cpu-cooler/821/

From what I gathered at this early stage the technology is still in it's infancy but certainly shows a lot of promise. More of an academic article in this area of research which is being designed by the Chinese for Laptop cooling at the moment:

http://www.bioheat.ac.cn/chinese/paper/05/Liquid metal based miniaturized chip-cooling device.pdf

Damn I saw a lot better academic article in this field the other day with complex modelling, and I can't find it.

 

Click this link for an in-depth review of the same "liquid metal" that BoldEagle linked to above... interesting stuff.

 

roflmao It does kind of look like a case with a freekin' H U G E fan on it!! LOL I didn't even think about it 'til you mentioned it! Too funny!

Should have no problem using the beloved Noctua P12 or S12 fan(s)...

 

As I "promised" earlier, I'm now posting the temps from my new case.... the screen shot is from HWMonitor; I left it running while I played Left4Dead2 for a couple hours... I think the temps are pretty freekin' impressive.

 

After a bit more research, it seems that heatpipe coolers cool a bit better when the pipes are parallel to the top edge of the motherboard when mounted in a tower-type case. Some CPU cooler manufacturers (Scythe for one) are actually saying that this is a preferred orientation, meaning that with my particular cooler, the fan should be blowing from south to north, up through the cooler as shown below...
I just ordered a bunch of case fans and I'll be installing 'em in the next couple days. At that time, I'll change the mounting of my CPU cooler and see if I actually get any improvement in temps. From what I've read, I can expect about 3-4c lower readings.... I'll post the results!

 

Yup. You are correct, sir! My cooler is 4 "U" shaped pipes with the base of the "U" on top of the CPU, and the extended research I've done tells me that horizontal (or south-north) mounting is indeed better. Today I installed a bunch of new fans in my new case and remounted the cooler to this horizontal orientation, and it has improved my temps by about 3-5c. (see screen shot below)

Again, all very true. In my new case, I have 2 fan mounts on the top of the case, the PSU mounts at the bottom. So I now have the CPU cooler blowing upward directly into 2 120mm exhaust fans. I also have a rear 120mm exhaust fan.

Side fans should be blowing inward, and most cases with side fans are designed with the side fan(s) blowing on to the GPU. I have a 120mm side fan that is oriented to where it blows about 80% of its input at the GPU, and the rest comes in just under the CPU fan, so it makes for great flow. If using a stock CPU cooler (the type that blows air directly at the CPU) would also work better with the side fans blowing cooler fresh air directly at the CPU/HSF. I also have a front 120mm fan blowing cooler air into the case and across the hard drives.

The screen shot attached is after 3+ hours of playing Left4Dead2 with all in-game settings at high. This is with my CPU cooler horizontal, blowing south-north, with 2 120mm fans in a push/pull configuration. The cooler is the CoolerMaster Hyper 212+ (more info here). The case is the NZXT Beta Evo (more info here) with 3 x 120mm exhaust fans (2 top, 1 rear), 2 x 120mm input fans (1 front, 1 side), with the PSU mounted at the bottom of the case and oriented with the fan upward, meaning that it acts as another exhaust fan for the GPU 'zone'. The case has tons of mesh everywhere; the bottom half of the front is mesh, the PCI slot covers in the back are mesh, the drive bay covers are mesh... the really helps with the air flow also. Since I have more exhaust fans than input fans, it means the case has "negative" pressure; meaning the air pressure in the case is lower than the pressure out side the case which naturally causes air to flow into the mesh in an effort the achieve equilibrium. If/when I get some pics of the new rig, I'll post 'em. The "Min" temps in the screen shot are my new idle temps. When I'm not home, and the heat is turned off, the temps are about 5-6c cooler than the "Min" in the screen shot.

 

Awesome improvements buddy, I have just got back from a business training trip (very tired) but will have a good look at this tomorrow.

 

Yes, it's true, I have no life roflmao I left HWMonitor running while I went to work, and naturally since everyone is at work, the heat in the house is turned off. I attached a screen shot showing the LOOOOOOW idle temps I get during the day (obviously in the "Min" column). I imagine at night the temp is at least 2-3c lower. Before I got the new case and before I changed the CPU cooler orientation, the CPU was idling in the upper end of the mid 30's (like around 37-38c), and the GPU would idle around 45c or so. During a heavy gaming session, the CPU "Max" temp has dropped by about 10c with the new case, and another 2c by changing the CPU cooler mounting from east-west to south-north(I also added a 2nd 120mm fan to the cooler for a push/pull configuration), and the GPU "Max" dropped from the mid 60s to an all-time high of 57c. I haven't seen it go over 57c yet, but I also haven't run FurMark for an extended time at 1680x1050 with the AA/AS cranked up LOL So.... yeah, I'm pretty happy with my new setup!! LOL :dood

 

Let it be known that I have less of a life than previously stated. Just for "fun" I decided to run a CPU stress test with Everest Ultimate. Attached is a screen shot showing the temps from a 20min run of the Everest Ultimate CPU stability test. I included the temps from HWmonitor too.....

(jeeeeez... I need a life... R E A L L Y badly!)

 

Hoping to get a new case soon LanCool K60 and I'm seriously considering the H-50. Just need some time this w/e to generate some "before" baseline runs.

 

I have a GetForce GTX295 and in the year i have had it i have never been able to get it to run consistently below 88c.
It has been in several different cases and with god knows how many fan combination's but it still runs a bit on the warm side.

P.S. it is running graphics (folding) so it is under stress all the time but it seems to do OK at these temps and all the other components are well within the limits.

 

I was looking at that same case and the LanCool PC-K58 too, but after rebate and promo, I got the NZXT Beta Evo for only $29!!!

Since it's under a large load 24/7, I don't think 88 is too high. I have seen some cards (9800GTX, 275GTX) get over 100 running FurMark stress tests and they were OK. While I don't recommend running a GPU over 100 for any amount of time, I think modern high-end GPUs can take the heat much better than cards of a couple years ago. I'm sure you've already tried this (in fact, I think we actually discussed this in a different thread a few months ago), but have you tried using Riva Tuner or maybe eVGA's Precision to lock your GPU fan speed at a higher level? I have mine locked at 75% with Riva Tuner, and it's never hit higher than 68c, and in my new case, it has yet to get over 57c. My previous case was much more silent than this new one (I now have five 120mm fans howling away) and I couldn't hear the GPU fan then....

 

Yes we did discuss it and i sometimes do that but it gets a bit loud so i am always trying something different.

When i put my GTX 260 in the rig it runs at 50 to 55c

 

Have you considered an 80MM fan on top of the card over the chipset?

 

I have tried an 80mm directly over (or under as it is on my rig) the air intake for the GPU but it only dropped by 2 or 3 degrees, I have also set one to suction boost the air coming out of the GPU but still no real success.

At the moment it is at 89c and that is about the best i have achieved on a consistent basis.

 

Did you strip it Bill and then apply a quality TIM?

The 4850s (I have one) came out with a shocking "Fan Profile" and I modded the cards BIOS to change the fan speeds. It used to only be at about 20-30% at 50C and 40-50% at 70-75C (it often idled at 80C) now it runs at 50% for low temps and scales to 100% at 80C. It rarely gets to 80C now.