cooling computer with refrigerator

[WindsorFox]

I've been thinking about blowing cold air into a puter case. Water
cooling only takes care of the CPU / GPU, while the other devices: HD /
CD / DVD, other MB chips, etc are left to fans. I know nothing about
cooling compressors, but what's the possibily of building a small
refrigeration unit and blowing the cold air inside the case?

I did a cursory search on google and didn't find much.

You can buy those now off the shelf, but if you're intent on building
it a small fridge is not the way to do it. You need something with
finned coils like an AC unit or dehumidifier to be able to cool air as
it circulates.

--
I used to have abs. Now, I've just got ab.
One big ol' Ab. - BigSkiff www.titanspot.com

Pyongyang sounds more like the sound effect an ACME catapult makes
as it goes off at precisely the wrong moment for Wile E. Coyote. -
Cadbury Moose

 

[WindsorFox]

The reason you didn't find any information on blowing cold air into a
computer's case probably has something to do with a little thing called
"condensation." If you are cooling surfaces that don't need it to
temperatures below the ambient temperature of the room the moisture in the
air will begin to collect on those surfaces as very dangerous beads of
liquid water.

There is also the problem of a rise in static electricity as the
temp and humidity fall.

--
I used to have abs. Now, I've just got ab.
One big ol' Ab. - BigSkiff www.titanspot.com

Pyongyang sounds more like the sound effect an ACME catapult makes
as it goes off at precisely the wrong moment for Wile E. Coyote. -
Cadbury Moose

 

[WindsorFox]

You obviously didn't look hard enough!!

http://www.comet.co.uk/comet/html/cache/43_250902.html

Oh yeah, a riot... Here's a *real* activly cooled case...
http://www.newegg.com/Product/Product.asp?Item=N82E16811221001

--
I used to have abs. Now, I've just got ab.
One big ol' Ab. - BigSkiff www.titanspot.com

Pyongyang sounds more like the sound effect an ACME catapult makes
as it goes off at precisely the wrong moment for Wile E. Coyote. -
Cadbury Moose

 

[WindsorFox]

Heh... Anyone else reminded of the guys who overclocked a old 486 from
33 MHz to 247 MHz by putting it in a freezer?

Project E.U.N.U.C.H.
http://totl.net/Eunuch/index.html

They actually paid 50p for a 486 CPU???!

--
I used to have abs. Now, I've just got ab.
One big ol' Ab. - BigSkiff www.titanspot.com

Pyongyang sounds more like the sound effect an ACME catapult makes
as it goes off at precisely the wrong moment for Wile E. Coyote. -
Cadbury Moose

 

[Don Freeman]

Oh yeah, a riot... Here's a *real* activly cooled case...
http://www.newegg.com/Product/Product.asp?Item=N82E16811221001

Yeah but will it keep your beer cold? I think NOT!

 

[Rod Speed]

There is also the problem of a rise in static electricity as the
temp and humidity fall.

The humidity will increase in that case.

 

[Andrew Smallshaw]

Have seen CPU's dramatically over clocked with liquid nitrogen cooling.
-300 C, makes a huge difference to the conductive properties of a small
strip of gold leaf.

Somehow I doubt this for a number of reasons...

1) While I'll admit I haven't checked up on the ratings for typical
processor, -300C is awfully cold. Among semis in general, even
military grade devices are typically only rated to -40C. Going
past that is into the realm of undefined behaviour just as much
as going above the top limit, which kind of defeats the object.

2) Liquid nitrogen freezes at -210C under standard conditions, so
it's simply impossible to cool anything below this using the
stuff. As an aside, one of the big breakthroughs in
superconductor research was the discovery of materials that would
superconduct at these kind of temperatures, enabling plain old,
relatively cheap, liquid nitrogen to be used to get them
working.

3) -300C can otherwise be expressed as approx -27K... in other words
_below_ absolute zero. Congratulations! You've instantly
redefined the laws of physics! I'm sure your Nobel Prize is
pretty much guaranteed.

Amazingly enough, you're not the first person I've noticed
making breakthroughs of this scale on this newsgroup...

 

[SubKool_Ntrance]

Somehow I doubt this for a number of reasons...

1) While I'll admit I haven't checked up on the ratings for typical
processor, -300C is awfully cold. Among semis in general, even
military grade devices are typically only rated to -40C. Going
past that is into the realm of undefined behaviour just as much
as going above the top limit, which kind of defeats the object.

2) Liquid nitrogen freezes at -210C under standard conditions, so
it's simply impossible to cool anything below this using the
stuff. As an aside, one of the big breakthroughs in
superconductor research was the discovery of materials that would
superconduct at these kind of temperatures, enabling plain old,
relatively cheap, liquid nitrogen to be used to get them
working.

3) -300C can otherwise be expressed as approx -27K... in other words
_below_ absolute zero. Congratulations! You've instantly
redefined the laws of physics! I'm sure your Nobel Prize is
pretty much guaranteed.

Amazingly enough, you're not the first person I've noticed
making breakthroughs of this scale on this newsgroup...

Ok boys, Overclockers.com
Not my idea, Not my website. I dont think i even researched that. but i
saw it on a promo video at a computer store.

David

 

[DaveW]

That would be a very thermally and electrically inefficient way to cool the
computer. Do a search on computer cooling and you will find all sorts of
effective devices sold.

 

[dj_nme]

The humidity will increase in that case.

I'm not so sure that the cold air from an airconditioner has higher
humidity than the surrounding air.
The reason is that the colder air has less energy to maintain the water
as vapour.
Just look at an airconditioner set to cold, there is water dripping out
of the cold side after it has condensed on the coils.
If you wanted to maintain or increase the humidity of the cold air,
you'd have to use a device to add water vapour back into the air.
It's called a humidifier.

 

[Rod Speed]

Rod Speed wrote

I'm not so sure that the cold air from an airconditioner has higher humidity than the
surrounding air.

Corse it does, just like any air that is cooled does.

The reason is that the colder air has less energy to maintain the water as vapour.
Waffle.

Just look at an airconditioner set to cold, there is water dripping out of the cold side
after it has condensed on the coils.

Yep, beause the cold side is below the dew point.

If you wanted to maintain or increase the humidity of the cold air,
you'd have to use a device to add water vapour back into the air.
Nope.

It's called a humidifier.

Not used when its cooling, only when its heating.

 

[kony]

I'm not so sure that the cold air from an airconditioner has higher
humidity than the surrounding air.
The reason is that the colder air has less energy to maintain the water
as vapour.
Just look at an airconditioner set to cold, there is water dripping out
of the cold side after it has condensed on the coils.
If you wanted to maintain or increase the humidity of the cold air,
you'd have to use a device to add water vapour back into the air.
It's called a humidifier.

Cold air from an AC does have less water in it, BUT, the
parts that get cooled by this air may come in contact with
OTHER air besides that from the AC vent, air which is warmer
and has more water in it, that water settling out onto the
cooler parts if the temp is low enough.

 

[No One]

Corse it does, just like any air that is cooled does.


Yep, beause the cold side is below the dew point.


Yup.


Not used when its cooling, only when its heating.

Tell that to all of the Canadians up North that have humidifiers
running in the winter.

 

[Big Bad Bob]

I've been thinking about blowing cold air into a puter case.

what's the possibily of building a small refrigeration unit
and blowing the cold air inside the case?

I did a cursory search on google and didn't find much.

2 words: peltier devices

 

[Rod Speed]

Tell that to all of the Canadians up North
that have humidifiers running in the winter.

That's when its heating, stupid.

 

[Big Bad Bob]

Water is conductive, and not a particularly good thing to have randomly
appearing inside your computer!

some computers actually are water cooled, but they don't apply the water
directly to the electronics. They typically use distilled water or
water + antifreeze that's chilled, run it through heat exchangers, dand
then pass air through those heat exchangers to cool it down. Chilled
water can then be supplied by an external refrigeration unit. Yes I've
seen a computer that did this (it was an older design that used TTL if
you can believe it). Gandalf may have seen one or two of these in his
military career... can you say UYK-7?

so yeah, what you said, with a bit of a caveat.

 

[No One]

Rod Speed wrote:

That's when its heating, stupid.

Nope. They were running whether the heat was on or not when I was
there. They said the humidity is so low it can cause skin problems.
BTW, what's with the name calling? Insecurity on your part?

 

[Big Bad Bob]

Usually moving the air will do the most good. You shouldnt need to cool it
unless it is a really hot room. Then you might as well cool the room.

there are some limits, but here's the basic calculations...

there are 3 basic thermal conductive layers you have to deal with. The
first two are fixed and more or less cannot be changed (the CPU material
and substrate, followed by the heat sink material). The differential
temp across each of these components is directly proportional to the
heat transfer rate, and inversely proportional to its thermal
conductivity. So double heat transfer, double differential temp.

Ideally in a balanced system the heat transfer rate will be more or less
a constant, determined by the power load of the CPU in watts. This
means that the differential temp of the CPU case and substrate and heat
sink are more or less constants. You can probably read them from the
appropriate tech manuals.

Next is the variable part, the air temp. The average air temp
determines the average conductivity so you can look at the differential
temp of the outlet - inlet air, which is proportional to the heat
transfer rate, and inversely proportional to the mass flow rate of the air.

To reduce the temp of the CPU you must increase the mass flow rate such
that the differential temp of the CPU + heat sink plus the average temp
of the flowing air is equal to the temp that you want to maintain in the
CPU.

So if the CPU is at 140 degrees F, and the temp drop is 40 degrees F
(heat sink + CPU case and substrate), then the average air temp needs to
be below 100 degrees F in order to keep the CPU temp at or below 140
degrees F. If the inlet temp is 80 degrees F and the exhaust temp is
120 degrees F, this would work (in a properly designed system).

However, if the inlet temperature were raised to 90 degrees F, you would
have to exhaust the air at 110 degrees F to maintain the desired CPU
temperature. To accomplish this, the flow rate of air would have to
double because the differential temperature is now 20 degrees instead of
40 degrees. In addition, the fan would consume 8 times the power in
order to double the flow rate (an important factor for laptops) and
would generate additional heat, requiring (possibly) even MORE cooling
air to compensate.

Placing a peltier (or other cooling) device on the inlet would, in fact,
offset this by lowering the inlet temperature, which would then allow
for higher differential temperature of the air, and thereby prevent the
need for a larger fan. It is quite possible that the peltier device
would consume less power than the larger/faster fan (though I'm only
guessing at this point), but only in some configurations (in other
configurations, larger fans might be more efficient).

Basic heat transfer/fluid flow stuff, at any rate.

 

[Big Bad Bob]

Heh... Anyone else reminded of the guys who overclocked a old 486 from
33 MHz to 247 MHz by putting it in a freezer?

heh - good one. Them were the days...

 

[Big Bad Bob]

Have seen CPU's dramatically over clocked with liquid nitrogen cooling.

-300 C, makes a huge difference to the conductive properties of a small
strip of gold leaf.

lower entropy, lower resistance