Crazy Cooling Concept (oooh alliteration)

Status
Not open for further replies.
Tedster said:
yes but oil is not as good a thermal conductor as heat, plus you would need a stronger pump because oil is more viscuous.

I like the idea of a water/antifreeze system.
Super Peltier Junction
These devices were used in picnic coolers; as they consume 80.6 Watts they should be used with a large heatsink (like maybe 4x6"). We're offering them here at a bargain price but we don't want folks to smoke 'em either. Here's the data: measures 40x44x3.3mm, deltaT=79deg, Imax=8.1A, Vmax=16.1V, Thot=50deg C, Qmax=80.6W Made in USA by Tellurex under their P/N CZ1-1.4-127-1.14. Large quantity available, substantial discount for cash customers buying lots of 1000 or more.
DATA SHEET Hint: 23F004 heatsink fan assembly may be used for cooling the hotside of these.
04U003 $14.95
 
Tedster said:
Water is already conductive. In a closed loop system it wouldn't matter.
we have already established that water itself is not conductive. only when substances dissolve in the water and create ions does the water begin to conduct electricity.
and it would matter because a system cooled with conductive water is far more likely to be destroyed in the event of a leak than if the system had contained pure(er) water.
 
Water/oil or heatsink?

Tedster said:
yes but oil is not as good a thermal conductor as heat, plus you would need a stronger pump because oil is more viscuous.
Sorry mate but oil is a better thermal conductor then water is, but it is indeed the pump you'll need and the power it will use that's going to be the issue in this possibility.
Also when using oil you must calculate in that its viscuousuty changes over its temprange, plus that some synthetic oils can be pretty corrosive too.
Zephead said:
we have already established that water itself is not conductive. only when substances dissolve in the water and create ions does the water begin to conduct electricity.
and it would matter because a system cooled with conductive water is far more likely to be destroyed in the event of a leak than if the system had contained pure(er) water.
Indeed only pure water is not conductive, but that you'll only find in labs.
In short, even distilled water is conductive, less then tabwater I agree but still.
On the other hand, what are the charcteristics of the currently used coolingwater/liquid?
1000cc said:
Super Peltier Junction
These devices were used in picnic coolers; as they consume 80.6 Watts they should be used with a large heatsink (like maybe 4x6"). We're offering them here at a bargain price but we don't want folks to smoke 'em either. Here's the data: measures 40x44x3.3mm, deltaT=79deg, Imax=8.1A, Vmax=16.1V, Thot=50deg C, Qmax=80.6W Made in USA by Tellurex under their P/N CZ1-1.4-127-1.14. Large quantity available, substantial discount for cash customers buying lots of 1000 or more.
DATA SHEET Hint: 23F004 heatsink fan assembly may be used for cooling the hotside of these.
04U003 $14.95
I've visited the site and am still working on the how-it-works.
Perhaps you could write down a pretty article about it.
Explaining this equipment in general with a short "how-does-this-work" weaven in.
It has got potential, that I agree upon, but I'm not just fully through with understanding it.
 
ßeetlejuice said:
I've visited the site and am still working on the how-it-works.
Perhaps you could write down a pretty article about it.
Explaining this equipment in general with a short "how-does-this-work" weaven in.
It has got potential, that I agree upon, but I'm not just fully through with understanding it.
http://www.tropicool.co.nz/ start here.
 
ßeetlejuice said:
Sorry mate but oil is a better thermal conductor then water is, but it is indeed the pump you'll need and the power it will use that's going to be the issue in this possibility.
Also when using oil you must calculate in that its viscuousuty changes over its temprange, plus that some synthetic oils can be pretty corrosive too.

.
i dont agree here The reason Suzuki FAIL wining races in the mid 80s on street legal sportbike was becouse oil cooling on racing bikes was not better than water cooled ..Yamaha and Honda & Kawi went water Suzuki abandon the idea. corrosion is not a problem i have torn dozens and dozens of motorcycle engines thru many years. also the most exotic racing bikes the 500cc 2cycle Grand Prix bikes are all watercooled SINCE GP racing started. This are million dollar bikes.
 
We're not dealing with racing gear here

Although I'd like the idea of racing technology used in a rig.
On the other hand Porsche had all it's best models racing oilcooled and they've ever been the most stable and secure racingcars around the world.
But this is a complete other usage of oil as cooling, it's also a different kind of oil, complete other characteristics.
Also high-power transformers are oil-cooled.
This is a very special oil, which is quite corrosive too, but for that application it's still the best way to have them running "cool".
The transformers themselves are in an oilfilled tub in this.
It's a completely closed system and the oil is also used as an insulator between the poles (end-connectors).
I know this 'cause and old-colleague and close friend of mine works as a design/development engineer for Pauwels trafo (a worldwide known Belgian company building exclusively high output-transformers)
I'm talking transformers you'll find in powerplants and downwards 'til street-level, all are oil-cooled.
I went true the Super Peltier Junction idea, it's a cool "dry" idea and they use it already for CPU cooling I saw.
I just wonder why it hasn't been more commonly known on the market?
Is it to "heavy" for the power-supply?
Too expensive? I Didn't think so.
Not powerfull enough for todays CPU-temps?
I mean you need a bit more cooling for a CPU running aircooled at around 50°C then you need for a fridge (closed airbox) coming from the environmental 30°C max (Belgium).
And you need to do this extra cooling on a surface of about 2.25cm² (taking 1.5cm sides on your cpu (hot-spot) as they're designed today.
Indeed the square bump in the centre is the surface that gives way all heat.
 
power wise they dont need much. heat transfer depends on chip size . peltier are more efficient making heat but they do cool.. the hot side gets very hot needs good heatsink and fans . case ventilation must be good since the peltier creates lots of heat. the ideal mobo would have the CPU on the other side "this way the hot air does not get dump inside the case.http://www.tetech.com/
 
Produce heat to cool, going Peltier

So, if I understand this wright, the problem is the dissipated heat from the Peltier cooler.
To get enough cooling you'd be producing so much heat one wood almost need to have it's heatsink placed outside the case and that heatsink should be huge and very well fan-cooled.
But then we're back to square 1, 'cause wouldn't we be better of then with traditional aircooling?
I mean, no huge heatsink, no power needed for the element and we're already dealing with lots of fanning at this moment, which we wanted to prevent and get rid of in the first place?
Or am I to critical here?
It looks like a very nice concept though.
I'm asking myself the Peltier junction cools one side and heats the other, so what would happen if you mount let's say, for example, 3 junction above each other, or is the heated side that hot that the cool side can't have that.
Of course the top-side would then almost be an electrical grill, I think. :confused:
Mmmh, someone ever experimented with these? :rolleyes:
 
you got the right idea . being inside is not big roblem as long you got enough fans blowing air in and out . i have a game box 3800+ that i must build so i cant mess with peltiers right now .i have seen a CPU one but was very simple using just 1 pelt . They have a watercooled one for watercooled system this really gets cold. and you must insulated so it wont create ice and water around the CPU. they also have a water chiller peltier block that you could run the water thru and lower the water temp.
 
read this since i kind agree . is from another forum. =Watercooling was fine... just wasn't cutting the cooling I need, I don't trust peltiers cooling my components that much after it KILLED MY FX5900!

I can't believe you actually bought that, how much did it cost you? a good 300US at least? You could've saved so much money if you just paid a visit to a junkyard for a radiator, and went to Wal-Mart for a pump. You could make your own waterblocks, but I don't recommend it, so you buy the waterblocks, CPU/VGA, and HDD if you have Raptor or 15000rpm+ SCSI drives. I really don't recommend chipset blocks, since a Vantec Iceberq4 does the job really well.

So yeah just my opinion if you're going to spend that much money on a watercooling setup, the triple radiator... lol i'll bet ANY car radiator will beat that thing's performance, well i guess car rads are huge and weighs a crapload. Anyways, if you're going to spend that much money on watercooling setup, might as well save a tiny bit more for a prometeia mach2 ls.

But still that kit's great, you should buy another XP120 with a Silverstone 120mm fan, and a 226watt peltier. Epoxy the peltier on the radiator, I said I don't trust peltiers on components(GPU/CPU) But I trust it on radiators since it'll just make the water cooler. Then use the XP120 to cool the peltier.
 
He mention car radiator - just so you dont get confuse is the very small radiator under the car dash board used as a heater to warm the car in the winter. My idea would be a peltier helping a Heat sink but not being sandwich between the heat sink and the CPU this way if the peltier fail the heat sink would take up the slack until you replace the peltier without any fear of damage. using peltier with water cooled is a bit less work. "but whole point was to avoid water by using peltier" A good size midtower case is a must or even a full tower.
 
the only reason i mention a mother board with a cpu on the other side is a "you could using a new case system with large heat sink on the left side. another option would be if the case was wide enough now you would have enough in bewtween wall room . the left side would act like a wind tunnel too cool the cpu and peltier and the right side would look like a regular mother board minus CPU . "here i think im ahead 10 years in design" prob some scumbag might read this and get rich. OF course ideal would be for both CPU and graphic card to be on the left side. and cool the same way . but at the moment i dont think MSI, ASUS ,DFI ETC or ATI and NVIDIA are going to kidnap me. or comply with my wish............MAYBE chaintech LOL.
 
I live in scottsdale AZ

I live in scottsdale AZ todays temp is +112°f! :hotbounce
My main reason for adding to this post is
there is a company here who used to produce cryonically sealed
server towers (they also freeze ppl) they also have some baseball
players head frozen there :eek: .
I am not sure how or why they did it or why they seemed to stop.
But the idea 5 years ago cost about $300usd to seal it.
I'm pretty sure they we're try'in to reach super conductivity.
They also anealed a cpu (amd 586 133) freezing it down to -250° f
It supposedly ran faster 586 333 or sum thin like that.
Just a thoughtful thought :approve:
 
112 degree is ok if you case are super flown with at least 3 to 4 120mm fans with a flow of 50 to 80 CFM each. im sure you also use a A.C a good HS like a XP120 AND a 120mm on it . you can built a chiller from a stores that sells used equipment for restaurants like juice chillers they have a very small freon compressor inside.
 
That would be kewl...er

1000cc
112 degree is ok if you case are super flown with at least 3 to 4 120mm fans with a flow of 50 to 80 CFM each. im sure you also use a A.C a good HS like a XP120 AND a 120mm on it . you can built a chiller from a stores that sells used equipment for restaurants like juice chillers they have a very small freon compressor inside.


Check this thread
You can bet I'm keep'in it chill.
I have considered chiller system but it's out of my realm mentally & $$$'s :haha:
 
SOcRatEs said:
Check this thread
You can bet I'm keep'in it chill.
I have considered chiller system but it's out of my realm mentally & $$$'s :haha:
if you can find a aquarium chiller used also you could use it.. this were on ebay new for $289.00
 

Attachments

  • 1-e8_1.jpg
    1-e8_1.jpg
    16.6 KB · Views: 5
Brilliant!!

yes that would work with the new fluid systems.
Now you got me think'n,
freon chiller out side to super cool
h20 (w/ dessolved ions)
pumped into pc box through sealed hose to all the heat sinks
would definately work!
I have a very good aquarium center that has surpplus....Hmmmm? ;) :
 
i would use a half liter to a liter insulated resevoir tank run the chiller way ahead get the water temp down then let it balance it self. should run nice a cold 20 or 15 celcius with ease.
 
This is south florida it gets in the 90s but very humid of course the A.C can keept it at 72- 75 or 80 most of the time, some days it strugles to keep it at low 80s. right now is 76 degree at 9.40 pm but i can crank the ac to maybe 70 at night.
 
"I went true the Super Peltier Junction idea, it's a cool "dry" idea and they use it already for CPU cooling I saw.
I just wonder why it hasn't been more commonly known on the market?'

Here's why:
1. Peltiers take up a crapload of Watts. Most PSU's don't have that amount of extra Watts. If you don't give it enough juice, you don't transfer enough heat off the CPU. When this happens, your CPU will fry because the peltier is a very bad conductor of heat, making it worse off. When that happens, you'll do better with the CPU alone without any heatsink/fan. Also, peltiers with not enough wattage will not transfer enough heat across, causing the same problem as mentioned above.

2. Peltiers that do move enough energy across will create a bigger need for a bigger heatsink on the other side. If you don't cool down the hot side fast enough, the cold side will also start heating up. And since peltiers create more heat on the hot side than absorbing heat on the cold side, you 'd probably need something like a water cooling system on the hot side.

3. They cost money, as some other posts have mentioned.

4. You get condensation problems. But that is specifically for high powered peltier systems. Its not that hard to control the peltier, but a normal thermostat will not bode well for your peltier, so you'd need to sink in more money for a proper temperature controller for you peltier (something to do with the way peltiers operate).

5. Peltiers are not very power efficient at the moment. Not much money has been put into improving these peltiers.

But if you can overcome all these problems, I'm sure peltiers will be used more often. In fact, I was just researching these peltiers, I found that someone is coming out with a peltier operated CPU cooler. Full set will cost quite abit, and it addresses most of the problems mentioned. Comes will a complete set of instructions on how to prevent condensation (doesn't use thermostats). Check the link on the other thread I started.
 
Status
Not open for further replies.
Back