PDA

View Full Version : Case Fan - suck or blow?



Tony
20-01-2006, 07:36 PM
The title says it all - should my case fan be blowing air in or sucking air out?

tweak'e
20-01-2006, 07:38 PM
depends on where it is placed and what part of your cooling design it is in.

Tony
20-01-2006, 07:54 PM
It is a vertical mid-tower case, power supply at the top. The case fan is on the back panel, between the power supply and the cards. The component probably most in line is the Northbridge heat sink.

I ask because I took it out to clean the dust off, and I'm not sure if I put it back in the same way round. :blush:

Cptn Hotshot
20-01-2006, 08:03 PM
I think we have been down this track before on something similar :D

http://pressf1.pcworld.co.nz/showthread.php?t=47317

tweak'e
20-01-2006, 08:30 PM
while thats a good read.......to save tony some trouble, rear case fan should blow out.

bob_doe_nz
20-01-2006, 08:43 PM
Whatever floats your boat :D :lol:

Greven
20-01-2006, 08:55 PM
it probably wouldn't make much difference if you had the rear blowing & the front sucking except you would get a lot more dust (or if you actually dust/vacume behind your computer desk, no difference.)

Tony
20-01-2006, 09:07 PM
Well on a 50/50 chance, it looks like I got it right - the rear fan is sucking air out of the case, as is the power supply (I looked at the other thread). I haven't got a front fan.

Thanks everyone - as usual you can count on PressF1 to leap into action with great advice. :thumbs:

apparition
20-01-2006, 09:10 PM
Yes the front is for sucking and the rear is for blowing

this is a good one, it even sucks the dust out
http://www.dse.co.nz/cgi-bin/dse.storefront/43d09a6f0534fccc273fc0a87f990742/Product/View/XH5088

Tony
20-01-2006, 09:33 PM
Yes the front is for sucking and the rear is for blowingWe're talking about fans here, right? :D

robsonde
20-01-2006, 09:51 PM
best design I have seen is to have sucking in the front and blow it out the back..

more air out then in results in negitve case presure, dust and crud gets sucked in all the cracks in the case. this give lots of dust in the case.

more air in than out results in positive case presure, air is forced out all the cracks and dust can only get in throught the sucking fan. put a filter on the one fan and never have dust in the case again :-)

easy way to make a filter is just to strech a bit of pantyhose over then fan grill.

tweak'e
20-01-2006, 10:04 PM
to small things....

the fans that are commanly used do not work well with any sort of pressure so you don't really end up with postive or negative presure...well any decent amount that is.

pantyhose works fine........for about 3 days then you have pull your case apart to clean it ! i found it much easier to let the dust just get blown straight through and out again ;)

Trev
20-01-2006, 11:33 PM
That reminds me of 2 400w PSU I bought from DSE. The 2 fans on them were both sucking in air. Took em back and got a refund.

Trevor :)

Agent_24
21-01-2006, 01:19 AM
I reversed the rear fan on my case once to see if it would provide better cooling, but not - my cpu temp increased 10 degrees. you definently want the fan sucking out.

SurferJoe46
21-01-2006, 08:11 AM
There WAS a thread on this a while back, but it needs to be reiterated here.

You should have laminair flow through the case by allowing air to be inserted (blowing in) through the front of the case over the hdds, and then another fan educting ( blowing out) of the case from the rear.

Any other fans on cards, the CPU or CHA will draw from that airstream going past them from the front inductor. I have also added a sidecase fan that blows directly onto or in the proximity of the vid card.

Keep in mind that if you have more eduction (out) than induction (in), you might end up starving your PSU and it's fan(s) from sufficient air flow. If the case pressure values go negative, then there may not be sufficient airflow in the case to allow the PSU to cool.

I ran a Baumanometer in the case to test the pressure and found that I had a slightly positive (+) case pressure which then helps push more air into the power supply to help it cool.

The dust issue be damned...you are operating a machine that will naturally require some maintenance and cleaning it of dust is a small price to pay.

Filters are a joke and they clog up so fast here in the dry California semi-deserts that they are totally useless. Placing your tower on or very near the floor will also dirt it up a lot faster than having the unit about 18 to 24 inches off the floor.

Greg
21-01-2006, 08:43 AM
this is a good one, it even sucks the dust out
http://www.dse.co.nz/cgi-bin/dse.storefront/43d09a6f0534fccc273fc0a87f990742/Product/View/XH5088
Can you link to the product itself? It sounds interesting.

Tony
21-01-2006, 09:08 AM
Can you link to the product itself? It sounds interesting.
Look for XH5088

Greg
21-01-2006, 09:11 AM
Look for XH5088
Ta.

I woulda thought that an intake fan (a sukka) would be more useful to have anti-dust capability.

The_End_Of_Reality
21-01-2006, 09:21 AM
This thread has got me thinking, are my fans ok?

4 blowing in from the front (added after built)
1 blowing in from the side window (standard)
2 sucking out from the back (standard)
2 PSU sucking out (standard)
1 CPU (side window blows directly on to it) (standard)
1 GPU (standard), most likely going to add a dual fan PCI card to blow more onto GPU (add after build)
1 Northbridge fan (standard)

A total of 12 fans :lol:

How does all that look?

Thanks all :), Thanks Tony for letting me use your thread :)

Nyuuji
21-01-2006, 09:54 AM
Using the latest scientific analysis techniques, "End of reality"...

I'd say you blow 25% more than you suck. ....lol

Sorry, couldn't help myself.

Tony
21-01-2006, 11:33 AM
This thread has got me thinking, are my fans ok?I would have thought your main problem would be making sure everything was properly bolted down to stop it blowing away in the gale. :)

The_End_Of_Reality
21-01-2006, 11:44 AM
I would have thought your main problem would be making sure everything was properly bolted down to stop it blowing away in the gale. :):lol: I made sure everyting was screwed in tight enough when I built it :)

Agent_24
21-01-2006, 12:25 PM
Have you considered a top fan between the PSU and cdrom drives that sucks up and out?? might as well get as many fans as you can! :lol:

how did you get 4 fans in the front??? I sure would like to know what case you used

Jester
21-01-2006, 12:34 PM
edited by me - have PM'd :cool:

Graham L
21-01-2006, 01:15 PM
That reminds me of 2 400w PSU I bought from DSE. The 2 fans on them were both sucking in air. Took em back and got a refund.

Trevor :)

But that's how the manufacturer designed it. They make power supplies, not computers. They naturally arrange the air flow so the PSU gets the first go at the cool air from outside the case. Why should it bother with air that's been past a few KW of CPO and video cards? :D

Terry Porritt
21-01-2006, 01:40 PM
You should have laminair flow through the case by allowing air to be inserted (blowing in) through the front of the case over the hdds, and then another fan educting ( blowing out) of the case from the rear.


I'm going to do a Graham and be pedantic.

Laminar flow is the last thing you want or will get inside a case.

The air flow will be turbulent, and the convective heat transfer coefficient in turbulent flow is higher than for laminar flow (streamline viscous flow)

SurferJoe46
21-01-2006, 06:44 PM
I'm going to do a Graham and be pedantic.

Laminar flow is the last thing you want or will get inside a case.

The air flow will be turbulent, and the convective heat transfer coefficient in turbulent flow is higher than for laminar flow (streamline viscous flow)

Uh....not!

Laminaire flow is actually just air flow that is zoned. It makes little sense to send heated air all over the case after it has picked up btu's on the HDDS. Turbulence is good if true molecular impact and expansive values are ignored, but it impedes the heat removal if the air arriving at...say.....the CHA is already much above ambient because it took on the heat from a hot device like the HDDS.

Enermax worked out a really fine case with a tunnel effect that started at the front intake fans, sent the air over/thru the HDDS and then proceeded to keep the flow straight out the back of the case via exhaust fans.

The idea was simply intelligent. The rest of the case got it's air from a side fan that sent non-preheated air to the motherboard and the residual of that went out the PSU's eductors. They advertised that the cooling effect was at least 20%-25% better than just allowing the air to flop around inside the case and get battered about striking ribbon cables, expansion cards and other objects thereby generating dead air pockets and slow flow over critical components.

I tend to believe that those numbers were pretty close to truth...we had to also make large gen-sets run in very cramped quarters in a boiler room in which I worked. At first they used just "generally indirect" turbulated air to cool the very close areas. The manifolds and the fields and generator cases ran very hot and would blister the dielectrics all the time. We replaced many sets of windings yearly. After switching to a more "tunnel-ram" effect and keeping the manifolds airflow seperated from the general cooling airflow, the gen-sets never had any dielectric failures again.

I vote for laminaire flow...I've seen it work much better than sloppy and non-directed air.

True, with the smallness of the internals of the towers and the obstructions inside, laminaire flow is difficult to achieve. Any attempts to keep pre-heated air from being used to cool another hot object is a good idea.

Heated air has less effectiveness in removing heat from an object on a parabolic scale as the latent heat level rises; much more air (higher flow-rates) will become less effective as the temperature increases exponentially.

At some point, no matter how fast the airflow is, it will not be able to remove enough heat to assist in cooling at all, turbulated or laminaire. Stratification of the airmass will increase to where the impact and transfer of energy via molecular conduction will cease as the heat load coefficient is reached. We could go into thermodynamics of superheating here, but that is too OT for a computer tower in someone's home/office.

Terry Porritt
21-01-2006, 06:52 PM
Sorry, but uh.... yes.

Laminar flow means streamline flow, but the air from a typical 120mm fan is turbulent, the Reynolds number based on average air speed is about 30,000, that is, about 15 times that for the onset of turbulence.

If the flow was laminar there would be hardly any flow at all. It's basic physics.

Terry Porritt
21-01-2006, 09:57 PM
Before I go to bed, I'll just add I fully agree with you Joe about directing the air flow intelligently as per Enermax example, but being pedantic as I said originally, that flow will be turbulent not laminar :)

You can have a turbulent directed flow, as for example turbulent flow in a duct.

dolby digital
23-01-2006, 09:25 AM
This thread has got me thinking, are my fans ok?

4 blowing in from the front (added after built)
1 blowing in from the side window (standard)
2 sucking out from the back (standard)
2 PSU sucking out (standard)
1 CPU (side window blows directly on to it) (standard)
1 GPU (standard), most likely going to add a dual fan PCI card to blow more onto GPU (add after build)
1 Northbridge fan (standard)

A total of 12 fans :lol:

How does all that look?

Thanks all :), Thanks Tony for letting me use your thread :)

I wondered why my lights have dimmed ;)

The_End_Of_Reality
24-01-2006, 04:52 PM
how did you get 4 fans in the front??? I sure would like to know what case you used Easy the case alows for 4 80mm fans in the front here (http://raidmax.com/product.php?pid=12&xcSID=9bc0a9fae84097d07ce87eb7cdfa18de) is the link to the case website, the case is black :)

BTW it is in my signature :p


I wondered why my lights have dimmed :lol: Yeah there is a rather large power surge when it gets turned on, the light in the house dim until it has spun everything up :p :D

SurferJoe46
26-01-2006, 07:51 AM
Sorry to re-open this post, but I just found some of the info I didn't have when I spoke/posted before.

Of special interest to me (and I think another poster who shall remain nameless) is the Peltier device. See the following....



Chiller Module – the Chiller Module includes a dual thermal plate heatsink, 6-TECS, two laminar flow FHEs, and a low speed fan. The Chiller keeps the coolant chilled and transfers heat outside the PC.

Chiller / Pump Module – the Chiller along with the compact reservoir and pump are integrated into a single mountable unit.

FHE (Fluid Heat Exchanger) – the FHE is a multi-channel block which transfers heat to the coolant. The CPU FHE acts as a low resistance medium to transfer heat from the CPU to the coolant.

Heatsink – a heatsink is a metal body including at least one thermal plate and a plurality of fins which conducts and transfers heat away from another body. CoolIT Systems dual thermal plate laminar flow design is used to remove heat from the TEC’s.

Laminar Flow - the motion of fluid or air when it is a smooth, even, regular or nonturbulant flow.

Peltier – see TEC.

TCM (Thermal Control Module) - the TCM uses an IC (integrated circuit) sensor to accurately detect and control the coolant temperature, by regulating the amount of power that is applied to the TECs and exhaust fan. The TCM will automatically adjust the cooling output of the system to conform to ambient temperature and CPU load. If desired, the TCM can be manually adjusted to suit individual user preferences.

TEC (Thermoelectric Cooler) – a charged ceramic substrate (also called a peltier) that pumps heat from one side to another.

Graham L
26-01-2006, 12:46 PM
The problem with a Peltier device as a "cooling device" is that it actually adds to the amount of heat which has to be removed from the system. It takes power to move the heat from its cold side to its hot side. It's a heat pump.

I wouldn't called it a "charged ceramic substrate". It may be fabricated in/on a ceramic substrate, but that is an insulating mechanical support. The business is done by a stack of thermocouple junctions. The only "charge" is the fairly high price of the things. ;D

The Peltier effect is just the converse of the Seebeck effect, which (as everyone knows ;) ) is how thermocouples work. You can actually use a Peltier device to generate electricity by heating one side and cooling the other. I'm not sure how efficient it is.

More terminology: A "heatsink" [b[isn't[/b] a sink. It's a heat exchanger. Sinks are what you wash the dishes in.

kingdragonfly
26-01-2006, 04:42 PM
If I might digress, if anyone knows someone who sells water-cooled PSU's in NZ, please drop me a line. I know there's several mod sites, but I value my life enough *not* to do DIY plumbing + electricity.

If you want to replace the CPU, VGA and NorthBridge fans with dual 1300 rpm fans:

Thermaltake Tai Chi case
http://www.thermaltake.com/xaserCase/taichi/vb5001sna/vb5001sna.htm

For those interested in eliminating most or all fans:

ThermalTake Rocket
http://www.a1-electronics.net/Heatsinks/2005/6th/Thermaltake_Rocket_Aug.shtml

Zalman Reservator
http://www.zalman.co.kr/eng/product/view.asp?idx=160&code=021

QuietPC New Zealand branch
http://www.quietpc.com/nz/products.html

VIA EPIA-TC6000E Mini ITX
http://www.via.com.tw/en/products/mainboards/mini_itx/epia_sp/

mCubed HFX ATX Home Theatre Case with HPC Heatpipe CPU Cooler & HGC Heatpipe
http://www.mcubed-tech.com/eng/produkte.htm

Zalman TNN 500AF ATX Full Tower Case
http://zalman.co.kr/eng/product/view.asp?idx=64&code=020