Theoretical water cooling idea

  • Thread starter Thread starter Brad
  • Start date Start date
B

Brad

Aside from the obviously large amount of space being taken up...

Could a water cooler, (like the common free standing 4-5ft tall including
huge water jug) be used as a water cooler for a standard home computer ?


Later considerations of saving space whether its set it in the my spare room
and run the tubes through the tv-cable face plate or use physically remove
non-essential parts ( the stand... smaller jug for water etc)

Im aware that there is specifically made water coolers for computers but I
have a project interest in mind.

Hence my question:
Can a water cooler reliably keep the water flow constant and be an
effective, silent cooling mechanism ?
 
Aside from the obviously large amount of space being taken up...

Could a water cooler, (like the common free standing 4-5ft tall including
huge water jug) be used as a water cooler for a standard home computer ?

Yes, but simply having a lot of water is only half the
story- the water needs the heat removed which isn't quite
what a water jug is desiged to do- just the opposite, it's
designed to have (nearer) the least surface area possible to
contain the water. In other words, it'd be more of a
novelty than an average, let alone optimal, solution.

Better would be to simply replace the huge water jug with a
common smaller automobile radiator. That is, if you really
don't mind having that much space taken up. Then again,
moving the water further may require a stronger (potentially
more noisey) pump too, so gains from lower airflow needed to
cool the radiator may be offset by more pump noise.

Later considerations of saving space whether its set it in the my spare room
and run the tubes through the tv-cable face plate or use physically remove
non-essential parts ( the stand... smaller jug for water etc)

You're going to need a pretty beefy (abnormally so relative
to what's typically used in PCs) pump for that. All you
really need is to move that water far enough to get the heat
outside of the case. If you want to do this novelty, you
may well find that the water reservoir is completely
unnecessary (providing your pump(s) aren't an immersed
type), you can simply use copper tubing to radiate the heat.

Even so, using a radiator is the easier solution and works
well... but then only you know if you even need
water-cooling. It's not quieter, it's higher-maintenance,
more expensive, cumbersome. It's of most benefit to those
highly overclocking/overvolting "something".

Im aware that there is specifically made water coolers for computers but I
have a project interest in mind.

you might just come out with what that project is, we can't
very well tailor an answer to a question we aren't
considering.

Hence my question:
Can a water cooler reliably keep the water flow constant and be an
effective, silent cooling mechanism ?

A water cooler is a storage tank... unless you have
something specific in mind that does more than this. If
it's tapped into the main water line to get it's water,
that's fine for it, as it's a closed system, but to keep
water flowing you'd have to dump that water out of the
system, very wasteful and expensive in the long run.

It won't be nearly as effective as some alternate methods,
but the refridgeration can do the job given enough surface
area.

There is no noise reduction from water cooling!
Your plans to cool a part or two with water blocks do not
remove the need for other fans. You still need same flow
rate to cool the drives. The power supply could have
somewhat lower rate due to less heated air, but the
motherboard still needs the flow rate. If you put the
pump in the other room where it's (hopefully) sonically
isolated, at most you only reduce CPU noise- as video cards
are easily, inaudibly cooling with aftermarket heatsinks
costing less than a water-cooler implementation.

So you're left with the reduction of noise from the CPU
'sink. A good CPU 'sink can run very quietly, in fact it's
noise can easily be eclipsed by the fans on the case walls,
which release more noise into the room rather than being
kept in the case. Overall the time, expense, and risks are
a bit excessive unless you're trying to break some
overclocking records, but for that you'd be considering
compressor-based cooling, not merely chilled water.

Don't get me wrong, it could be a fun project and certainly
novel... just not useful except in very limited scenarios.
 
Better would be to simply replace the huge water jug with a
common smaller automobile radiator.

Consider using a motorcycle radiator. They are a lot smaller and
certainly are adequate for cooling a CPU.


--

Map of the Vast Right Wing Conspiracy
http://home.houston.rr.com/rkba/vrwc.html

If you can read this, thank a teacher.
If you are reading it in English, thank an American soldier.
 
Consider using a motorcycle radiator. They are a lot smaller and
certainly are adequate for cooling a CPU.

By "common smaller automobile radiator", what he really meant
was the radiator from the interior heater. They come in nice
small sizes and are inexpensive.

For the OP, a word of warning... I've maintained water cooled
radio systems for decades. Ranging from 5 foot high klystrons
with 4" copper pipe to supply glycol coolant pumped from another
room, to very small units with convection cooling using
distilled water in a small heat exchanger mounted on the side of
the equipment... they *all* had two problems. One is that
somebody has to make sure there is always enough coolant, the
second is that eventually they leak.

There is an engineering principle known as KISS, for "Keep it
Simple, Stupid", and it applies to cpu cooling systems in the
worst way! The more complex, the more trouble. There are only
two reasons to use water cooling on anything. One is when
nothing else will do. That doesn't apply to cpu's in a PC. The
other is for fun. Water cooling cpu's is *great* fun.

The system described would be a lot of fun. You won't have time
for computing though, because the design, installation, and
maintenance of that cooling system would be a great full time
hobby all in itself!
 
Water cooling cpu's is *great* fun.

Could you elaborate on that.

I thought I had pretty much doen everything in life that isworth
doing. Now you come along and tell me that cooling cpu's is an
experience no one should miss.
(Barrow, Alaska)

Oh, now I understand.


--

Map of the Vast Right Wing Conspiracy
http://home.houston.rr.com/rkba/vrwc.html

If you can read this, thank a teacher.
If you are reading it in English, thank an American soldier.
 
By "common smaller automobile radiator", what he really meant
was the radiator from the interior heater. They come in nice
small sizes and are inexpensive.

When one custom designs the solution they can pretty much
pick what they want. A motorcycle radiator may be
sufficient if he's got copper tubing running into another
room, as the tubing itself is radiating more heat. Othewise
the smaller the radiator the more likely a fan is necessary.

I didn't necessarily mean a very small radiator, primarily
because the mention of the "water cooler" tended to suggest
there was ample space for something big. On the other hand
a radiator out of a '723 Impala might be a bit extreme,
there are plenty of those miniature "toy" cars on the road
today such that several are likely to have a nicely
downsized radiator... maybe even for a VERY small car like a
Chevy Metro.

For passive cooling something even more space-out might work
better though, like a freezer coil assembly... but I'd
imagine it tends to be a lot more expensive.

For the OP, a word of warning... I've maintained water cooled
radio systems for decades. Ranging from 5 foot high klystrons
with 4" copper pipe to supply glycol coolant pumped from another
room, to very small units with convection cooling using
distilled water in a small heat exchanger mounted on the side of
the equipment... they *all* had two problems. One is that
somebody has to make sure there is always enough coolant, the
second is that eventually they leak.

There is an engineering principle known as KISS, for "Keep it
Simple, Stupid", and it applies to cpu cooling systems in the
worst way! The more complex, the more trouble. There are only
two reasons to use water cooling on anything. One is when
nothing else will do. That doesn't apply to cpu's in a PC. The
other is for fun. Water cooling cpu's is *great* fun.

Yep, agreed. There's almost nothing as reliable as a
high-quality, low RPM dual ball bearing fan on a 'sink
that's dusted out on a regular (as-needed) interval.

The system described would be a lot of fun. You won't have time
for computing though, because the design, installation, and
maintenance of that cooling system would be a great full time
hobby all in itself!

Either that, or rightabout the time it's finished he decides
to get a Pentium-M based box and finds the water cooling
isn't of benefit at all. Then again, sandwiching a peltier
inbetween that setup could be a bit quieter than a
compressor-based solution, but again the added complexity
requires more attention to potential failure points.
 
kony said:
Either that, or rightabout the time it's finished he decides
to get a Pentium-M based box and finds the water cooling
isn't of benefit at all. Then again, sandwiching a peltier
inbetween that setup could be a bit quieter than a
compressor-based solution, but again the added complexity
requires more attention to potential failure points.

The number of failure points is a real problem, and monitoring
the whole system is the real fun.

One serious flaw in virtually every water cooled cpu system I've
seen (and in more than one commercial variation too!), is the
lack of a "low flow" alarm mechanism that shuts the system down
before damage is done. I don't know of any inexpensive liquid
flow switches...
 
Could you elaborate on that.

I thought I had pretty much doen everything in life that isworth
doing. Now you come along and tell me that cooling cpu's is an
experience no one should miss.

If you gotta ask, the answer won't mean much.
Oh, now I understand.

I doubt that you do, but it is true that someone who lives in
Barrow is almost by definition adventurous, and liquid cooling
cpu's is an adventure. But adventurous people live other places
too.
 
I doubt that you do, but it is true that someone who lives in
Barrow is almost by definition adventurous, and liquid cooling
cpu's is an adventure. But adventurous people live other places
too.

Sounds more like an example of Eskimo Engineering to me.


--

Map of the Vast Right Wing Conspiracy
http://home.houston.rr.com/rkba/vrwc.html

If you can read this, thank a teacher.
If you are reading it in English, thank an American soldier.
 
The number of failure points is a real problem, and monitoring
the whole system is the real fun.

One serious flaw in virtually every water cooled cpu system I've
seen (and in more than one commercial variation too!), is the
lack of a "low flow" alarm mechanism that shuts the system down
before damage is done. I don't know of any inexpensive liquid
flow switches...


... but that's half the fun, dreaming one up.
I dont' know if a flow sensor is necessary though, one could
just just a temp sensor in somewhat close proximity to the
hottest measured part (of the cooling system).
 
Floyd L. Davidson said:
.... snip ...

The system described would be a lot of fun. You won't have time
for computing though, because the design, installation, and
maintenance of that cooling system would be a great full time
hobby all in itself!

Yabutt - you don't need it. All you have to do is avoid computing
from 15 July through 1 August. :-)
 
kony said:
.. but that's half the fun, dreaming one up.

Ain't that the truth!

I've got most of a system ready to go here to cool a pair of
Athlon CPU's on a Tyan motherboard. Someday I'll get around to
doing it. It's got a Crystalfontz CF633 which can control 4
fans and handles up to 32 temperature probes, so the
possibilities for complexity is great.

The idea is to have a 15 gallon aquarium sitting right next to
a window, using the radiator from a car heater, with
temperature controlled fans sucking air through a plenum that
passes all of the air right next to the window. I might even
try to situate the aquarium such that it has thermal contact
with the window too.
I dont' know if a flow sensor is necessary though, one could
just just a temp sensor in somewhat close proximity to the
hottest measured part (of the cooling system).

I'm not sure that is viable. A probe right on the cpu would
work. But any place else and what you get if the coolant leaks
out... is a nice normal temperature.

A temperature sensor would have a rather complex algorithm to
determine a lack of coolant. When it is first started, that
"idle" temperature is the same temperature it would go towards
if a coolant leak occurs. Both a high temp and a low temp would
have meaning, and both would be rather slow in comparison to
the temperature of the cpu.

In good commercial liquid cooled system I've worked with there
was a flow switch that worked as an instant interlock to shut
the system down if there was no flow.

One system which used convection to circulate the liquid did not
have any way to detect a lack of coolant. And don't you know
that at some point in its life whoever was maintaining it had
failed to keep the reservoir full, and it had overheated! There
were black marks from the heat burnt into the finish... :-) That
was a Collins MX-106 microwave, and one of the worst examples of
poor engineering out of the many that Collins/Rockwell produced
over the years. Collins made a lot of really good HF equipment,
and a huge pile of junk for microwave and telecom.
 
CBFalconer said:
Yabutt - you don't need it. All you have to do is avoid computing
from 15 July through 1 August. :-)

Well, the heat exchanger is certainly less complex here!

We do, however, actually get half a dozen days of the year where
the temperature is 65F or higher. And sometimes even a day or
two in the low 70s.

Oh, we suffer then...
 
Eskimos are damned good at engineering!

You could have fooled me. Any group which clubs its supper to death
seems a bit primitive to me.

One thing that I am definitely jealous about Alaska is the incredible
abundance of salmon. I could literally eat salmon every day. I smoke
it to an internal temp of 100F after having brined it for a day. But
it costs $5 a pound in filet form and that's a bit expensive for
everyday eating, especially when I can eat at least 2 pounds in 24
hours.

I bet you get it for $1 a pound in filets, if even that much.

--

Map of the Vast Right Wing Conspiracy
http://home.houston.rr.com/rkba/vrwc.html

If you can read this, thank a teacher.
If you are reading it in English, thank an American soldier.
 
Yabutt - you don't need it. All you have to do is avoid computing
from 15 July through 1 August. :-)

Not where I live. In Houston, summer begins in May and ends in
November. Anytime I am forced to run the a/c, that's summer.

We have 4 seasons too - December, January, February and Summer.


--

Map of the Vast Right Wing Conspiracy
http://home.houston.rr.com/rkba/vrwc.html

If you can read this, thank a teacher.
If you are reading it in English, thank an American soldier.
 
We do, however, actually get half a dozen days of the year where
the temperature is 65F or higher. And sometimes even a day or
two in the low 70s.
Oh, we suffer then...

Yeah, especially when half the year is spent under 10 feet of snow.

Thanks, but I grew up in snow and swore I would never live around it
again. Too many frozen fingers and toes.


--

Map of the Vast Right Wing Conspiracy
http://home.houston.rr.com/rkba/vrwc.html

If you can read this, thank a teacher.
If you are reading it in English, thank an American soldier.
 
You could have fooled me. Any group which clubs its supper to death
seems a bit primitive to me.

Anybody, apparently, could fool you.

What did you have for supper? Was it alive once? If not, you
aren't either. And just in case you don't understand,
*somebody* clubbed your dinner to death, one way or another. If
that makes an Eskimo primative, it makes *you* primative.

Your statement is either abjectly stupid or simply racist.
One thing that I am definitely jealous about Alaska is the incredible
abundance of salmon. I could literally eat salmon every day. I smoke
it to an internal temp of 100F after having brined it for a day. But
it costs $5 a pound in filet form and that's a bit expensive for
everyday eating, especially when I can eat at least 2 pounds in 24
hours.

I bet you get it for $1 a pound in filets, if even that much.

You have a loose grip on reality... :-)

Regardless, I've been living with Eskimos for a year or two now,
or perhaps a bit longer. (My grandchildren, like their parents,
are Yup'ik.) I've never met an Eskimo that clubbed as seal. Just
like caribou, moose, whales and sheep, they *shoot* them to death.

Not that none of them, or me for that matter, would have a problem
with clubbing our meat to death, just that circumstances don't
make that easy to do anywhere that I've lived recently. Now, I do
recall my father using a 20 pound hammer to dispatch cattle before
butchering them...

But you probably think T-bone steak is a manufactured product,
eh?
 
Back
Top