Fans - cold air in or hot air out?

[Magno Boots]

That's the question chaps.

Is it better to blow cold air into the miggy or extract the hot air?
I'm thinking of a 12v 60mm fan.

Some sites suggest taping all vents at the top grill, except for where the fan's located in order to direct the hot air out.

Not sure if it's such a good idea.

Does it make any difference?

Please advise.

[Shoonay]

Fred and Dizzy will love this thread...

IMHO It's best to blow directly on the most heat-sensitive hardware, like CPU's, turbo's and various chips...

Anyways, how about a blow-through? A blow-in from the right side and blow-out from the left side, for example?
Feeling a mod enough for the job?

[DrF]

Sucking or blowing, I always prefer to blow anyhow, just me I guess

[zeke1312]

OK, being in the Main Frame and PC support business, your best method is to *pull air out* of the cabinet/enclosure. In part, this moves ambient temperature air from the input vents and moves this air over the components. Pushing air into the box may result in lower air flow through the enclosure and trap some air in areas that are not prone to good air flow. Notice most if not all PCs that have fans have them at the rear of their enclosure. Main Frame systems use the same approach.

[Zetr0]

@Magno

with any 'active' air solution, its best to review your cooling method.

such that... if you are sucking hot air out.. is there a way for the air in. will the force of the air-draft cool the selected area.
Again the reverse is true, if you blowing cold air in, then will there be an exit for the air, or will it become buffered and stagnant flow causing heat up and no cooling. will the path of the air-flow cool the selected components.

there is arguably a theory for both and respectively there is no difference if you use a duct or such method.

In an a1200 desktop there is a lot of places where air is needed and where it can get easly trapped offering no cooling at all...

Some ideas to consider

For an "out-take" consider modifying one of them PC drum fans that exist sitting in a PCI bay, there sheap and can fit internaly to the rear-top left of the case and provide a much needed chipset - internal cooling.

In sofar as an accelerator card they were manufactured in a peculiar fashion, cooling this is tricky as there is hardly any space for an active "fan" cooling solution.

I have been experimenting with old laptop heat sinks and cooling veins but my mind is undecided as to which to proffer. I would suggest some ducting and a good fan,

you can make ducting out of anything from CD jewel cases or even cardboard and sticky tape

good luck with your venture, and please post pics

[jmmijo]

Generally speaking you should have proper air-flow and ventilation in your chassis/case.

If you have to pick one then exhaust is the preferred single-fan solution.

If you can at all place the exhaust near/at the top of the chassis and exhaust it out then that is the optimum placement as convetion will already help direct the heat to the top of the case.

If the enclosure is small then you don't have any choice really, exhaust it out as quickly as possible.

Think of the external USB HDD enclosures, these almost always have a single exhaust fan and vent holes in the front to get proper air-flow through the enclosure

[Shoonay]

Nice to see you bump into the thread, Z, and what a surprise!



Let's take a hyp-no-ethical situation...


I got 3 empty cube-a-like cases. They all have a fan on each and every side wall, which makes 6 per cube, obviously.


Cube A has all the fans to blow air inside case.

Cube B has all the fans to blow air outside the case.

Cube C has 3 fans to blow air inside, and 3 fans on the opposite sides to blow air outside, so generally let's say it blows air through the case.



So, the Ultimate Question to Life the Universe and Everything is: which cube provides the best cooling?
(I do hope it'll not take 7½ million years to compute and check the answer )

[Charlie]

Originally Posted by Zetr0

...there sheap and can fit internaly to the rear... ...or even cardboard and sticky tape

good luck with your venture, and please post pics

Oooo Zet0 - you perv

To quote Bart Simpson:

'Man, this sucks and blows! I didn't think that was possible'

Do both, but concentrate on the sucking.

[Zetr0]

@Charlie

LMAO

[adonay]

http://www.freewebs.com/adonay-/amigaalbum.htm take a look at the second page i made something pretty fast for cooling down my amga "for bppc and bvision" it works ok i plan to use one of those doubble hd clip on fans and screw it to the trapdoor turn the fans and voila airflow .... I also have a small fan regulator that will lower the fan speed and sound .. The fan walls are made by sissor and a DVD cover

[Jope]

Like said before:

Cold air in at the front/bottom, hot air out at the top/back.

Use convection to your advantage, no point in doing it any other way.

If you have hotspots like chips inside the case, these have been traditionally cooled with a fan blowing onto the chip.

I don't really see the debate here ;-)

[Magno Boots]

Thankyou for the replies guys.

It seems that hot air out is the way to go, but it's limited dependant on the machine's size.

The concerns I have are with my A600 / 3.1 rom / 2gb CF and soon to be internal slimline DVD.

The bottom of the housing gets bloody hot despite 10mm rubber feet already. Heaven knows what it'll be like when I bribe LordV into selling me an internal 8mb ram card

Noise and room is the factor. I've noticed that the smaller the fan the bigger the db.

I suppose it's a case of suck it and see.

Thanks again.

[Paul_s]

Fans? You need mineral oil

[Zetr0]

@Pauls

I am just reveling the irony, of a super-cooled pc in a fishtank, running the screen saver of the fishtank.... hmmmm...

[rare_j]

Originally Posted by Magno Boots

Noise and room is the factor. I've noticed that the smaller the fan the bigger the db.

I suppose it's a case of suck it and see.

Thanks again.

I got a pair of 40mm fans from quitepc.co.uk and also some resistors (12v to 7v)so they run really quiet whilst still moving an acceptable amount of air.
In the end I decided to have one blowing air over the back of the 1230IV board inside the case and and one blowing out the top on the other side behind the keyboard.

[Shoonay]

Somebody could please answer my seriously serious question?
I mean *seriously*, now...

[Merlin]

@ 5h00n4y

I can't right now, I'm still working with Reynold's Numbers and trying to work out if your hypothetical cubes have laminar or turbulent flow, fluid mechanics doesn't take 5 minutes you know

BTW, if the cubes are hypothetical, does that mean the fans don't actually exist? It could save me a LOT of working out.....

LOL

Merlin

[Shoonay]

@Merlin: Thanks for your time...
BTW: The situation is hipothetical, the cubes and fans are real, ok?
Yeah, I think it would be easier if we assumed that those cubes are based in a winter environment with lots of winds blowing and stuff.
What good is a fan that blows hot air inside?

[Charlie]

@5h00n4y:
For a cube with one fan on each face (assuming the base is off the ground) I'd say the best arrangement would be:

2-blow (front & base)
4-suck (sides, top, rear)

Why?
Always have more suck than blow - helps reduce static pockets of warm air.
Airflow should be directed upward - in line with convection.
Horizontal flow over components - to the rear where the (hot) PSU is likely to be. Also who wants hot fan blowing in their face..?
(don't answer that)

Cool Cube

I think it would be better to determine where all the worst heat sources are located first as those components will be the ones most likely to fail or cause a failure of a nearby less temp tolerant component due to heat stress. You can do this with a fast response electronic thermometer of some sort or Infrared (IR) camera to find the hotest components after the test rig in its expected operational orientation has been on for some time

Note - The medical thermometers that go in your mouth or under your arms are not fast enough and probably don't have enough of a thermal measurement range to be useful.

May be the ones you use in your ear would be usable. Ones that use a laser or a small thermocouple to measure temperature may be best thermometer to use depending on their sensitivity. An IR camera will probably be least sensitive.

Then you can work on how to cool down those components efficiently without heating up other nearby components more than neccessary. Just putting thermal transfer paste, a heatsink with a a blowdown fan on a hot chip will cool the chip but at the expense of all the other surounding components due to ther hot air forced onto the board by the blowdown fan.

The newer PC motherboadrs and CPUs manage this optimal heat management problem with passive cooling via highly conductive copper heat pipes with radiator fins on one end, strategically placed fans to cool the heat pipe radiator fins, and fan air ducts to direct the heated air from the fans out of the case in such a way that the hot air does not recycle back into the case again.

Note - even good internal case cooling can be defeated by ab inapproprate physical operating enviroment. An air cooled case needs access to cool air on the approprate sides and good ventilation to immediately take away all the hot air expelled by the case while the equipment is powered up.
EG puting a high powered server in an airconditioned closet with enough cooling capacity is fine as long as the critcal airconditioning is on while the server is powered up and for a sufficient cool off period after the server is shutdown.

The 6 fan cube situation:
For a good test you need to have at least one heat source of specified wattage at the same location in each identical cube. Having no specifically rated heat source means your air temps are ambient room temp with the your fans being the only heat source. Varying each heat source wattage could be done as another series of tests.

You'll also need to test the cubes under controlled conditions:

• same ambient room temp, humidity, air pressure so the heat capacity of the air is as close to a constant as possible
• have each cube at same location relevant to walls, floor or table top, etc..
• use same fan type on all sides of cube under the same voltage conditions which will best reflect normal world operation. You could vary fan voltage as a seperate series of tests.
  The airflow rate through each fan will vary due to the test cube configuration. Fixing the RPM will cause certain fans to use more power causing air from fan to be at a different temp because they will have different air flow resistances to overcome.

Suggest you have a fourth control cube with the fan vents but no active fans so you can see where the heat wants to go naturally.

After the environment has reached stability, you'll also want to measure the air temp and air flow rates and direction at:

1. all the fan vents
2. various locations inside the cube including the surfaces of the heat sources in the cube.

So you can evaluate how your cooling system is working.

Assuming there are other air vents in the cubes I'd expect the all pull or all push cubes will have a big possibly hot dead air zone in the middle of the cube. If there are no other vents other than the fan holes how would you get air flow into the cube in an all push or all pull situation?