Sub Zero Liquid cooling without a chiller & Tower Cooling

[zipdogso]

Based on what -ACID- said it would seem you need a dry atmosphere. Better still a warm dry atmosphere.

If you are only using this for benching how is this for a possible but potentially carpy idea.

Build your bong or what ever you are going to call it, and box it in allround so it is in a sealed container. You will have to use forced air on the bong because you need a air circulation. Then before you seal up for benching you introduce a quantity of Molecular sieves. Molecular sieves work a bit like that silicon stuff in the "Do Not Eat" sachets in boxed electronics... BUT they are far more efficient and will reduce the humidity in the bong box to near as 0% fairly quickly (but you need the air circulation to achieve this.). Since it's a sealed box the heat will rise...unless you use a TEC air chiller arrangement to keep it down to reasonable levels should be fairly easy as your only looking at keeping it down to say 5º-10º ABOVE ambient. At warmer temps you get better evap and air is more thermal conductive.

Molecular sieves are relatively cheap....I enquired here in the UK to one supplier and said I needed enough to reduce to 0% an volume of about 6 cubic feet he calculated that at saturation (i.e. 100% humidity) it would take 25grams of the stuff !! And promptly gave about 125grams free in samples !!
I don't know how much it costs but it must be pretty cheap !! Don't know how much you would need to keep down a constantly increasing humidity, that's why it would probably only work for benching.

Like I said this may turn out to be a very carpy idea....but it's something to go on.

[just a noob]

i bet if you got a heat exchanger of some sort, you might be able to do it, i would imagine if you had an inclosed tower, filled with a high percentage alchohol, or other liquid that evaporates quickly, but i don't know how cold that could get, i would guess that if you made it out of finned aluminum that might help, along with a condenser at the top of the tower, but thats all just a guess from a 15 year old

[Webby]

Ok, here's a tech question the TDP of a quad core is approxamatly 250watts, is this 250 per minute or per hour? I need this info for btu conversion, as 1lb of water can remove 1,000 btu of heat in the tower type system, therefore 1 imperial gallon weighs 8lbs can remove 8,000 btu, so a 500litre /hr pump can move 132.5 gallons of water/hour x8,000=1,060,000 btu/hr.=310.7kw/hr

Okay you do know what a Watt is don't you? 1 Watt is 1 Joule per Second so a processor with a TDP of 250W will be generating 250 Joules of heat energy every second. Simple

A kilowatt hour is the energy used by a 1kW device in an hour so 3600000 Joules.

Hope that helps

[fart_plume]

Okay you do know what a Watt is don't you? 1 Watt is 1 Joule per Second so a processor with a TDP of 250W will be generating 250 Joules of heat energy every second. Simple

A kilowatt hour is the energy used by a 1kW device in an hour so 3600000 Joules.

Hope that helps

Helps alot thank you!! Learn something useful every day!

So a 500 litre /hr pump would be enough to cool a quad back to ambient@ stock speed. so a 1,000-1,500 litre pump would be need to take an oc'd quad below ambient.(estimation)

[Webby]

As already discussed a 250W TDP CPU consumes 250 Joules per Second so that equates to 900000 J per hour (250 J/s * 60 s/min * 60 min/hour) or 900kW/hr so you will require a 1500 litre per hour flowrate to dissipate that much heat (based on your figure of 500l/hr = 310kW/hr dissipated).

1500 l/hr is equivalent to 25 l/min which is pretty swift. Are you sure your calculations are correct?

EDIT: Ah I think I may have seen where you have gone wrong 1060000 BTU is equal to 1118300000 Joules or 1118300 kJ for 500 Litres in an hour so 2236 kJ per L/hr so even with a flowrate of 1L per hour you would have more than enough capacity to remove the heat from a 250W source.

Oh also since you are working in BTU (British thermal units) did you use an english or american pound to calculate your gallons per hour figure (ignore this they are the same)? Also what was your source for the 1lb of water = 8000 BTU ? Is it based on a system similar to what can be implimented in a home built system or is it for the optimum conditions 0% humidity, low pressure, high temperature resulting in 100% evaporation?

[fart_plume]

Well according to the paper I have 1lb of water can dissipate 1,000 btu, water weighs 8lbs to the gallon , therefore 8,000 btu. The paper doesn't give a time varibleI can e-mail the paper to you, if you like. I ciould have very well made an error, being human there is always room for that.

[Webby]

See my edit in my last post for where your calculation error is.

Oh and I answered my own question in my edit the 8000BTU per lb of water is based on evaporating it all (the latent heat of vaporisation of water is 2500 J/g or 2500 kJ/kg where 1 kg = 1 Litre) oh and a pound is only 453.57g so depending on your definition of a gallon (the is an American and British one of those) it will vary the figures a bit again.

[fart_plume]

The pound i'm refering to is American, and I got the btu rate for water from a paper on industrial evaporation coolers, i can send you the file if you like, has a lot of useful info for such a short paper.

[DoGMaN]

I wish I still had picks of my old "bong" cooler. It was about 14' tall double walled to contain as much of the vapour as possible, very similarly to what zipdogso said. It used a four fan forced air setup that blew into a reservoir and up the main pipe against the flow of the water. Now remember this was quite a while ago I think I was using a P4 2.26 then OCed but I forget to what. With the height and added airflow. Best water temps as I recall were 5ish below ambient. Good luck with your build FP.

[[XC] riptide]

As already discussed a 250W TDP CPU consumes 250 Joules per Second so that equates to 900000 J per hour (250 J/s * 60 s/min * 60 min/hour) or 900kW/hr so you will require a 1500 litre per hour flowrate to dissipate that much heat (based on your figure of 500l/hr = 310kW/hr dissipated).

1500 l/hr is equivalent to 25 l/min which is pretty swift. Are you sure your calculations are correct?

EDIT: Ah I think I may have seen where you have gone wrong 1060000 BTU is equal to 1118300000 Joules or 1118300 kJ for 500 Litres in an hour so 2236 kJ per L/hr so even with a flowrate of 1L per hour you would have more than enough capacity to remove the heat from a 250W source.

Oh also since you are working in BTU (British thermal units) did you use an english or american pound to calculate your gallons per hour figure (ignore this they are the same)? Also what was your source for the 1lb of water = 8000 BTU ? Is it based on a system similar to what can be implimented in a home built system or is it for the optimum conditions 0% humidity, low pressure, high temperature resulting in 100% evaporation?

EDIT: Oh wait the rest of the post makes more sense.

[Martinm210]

Hi got invited here, this is one topic i could almost call myself a forum expert on !!!

Cooling towers are fantastic, cheap easy to make and limitless cooling power (they just dump the heat load into vapour)

I,ve made loads for people all shapes an sizes and depending on persons needs how much they need looking after too.

Few questions for you

1. Most important what is the humidity levels were you live as cooling tower's work by dumping water into the air. If you live in a high huimidity location then they become much less effective as the air is already holding a high water percentage.

2. Is this for everyday running, benching crunching ect

I would only use plain water in the tower, chemicals are too risky and you can really mess up you lungs long term.

Downside's to this are you have to strip and clean more often the a normal water cooling loop and you are maknig a ideal breeding ground for alge ect (uv lights slow this down loads tho as it kills it)

They can be quite big and not very movable.

but enought with words lets dig out some picts !!!!!
messing around with salt @ sub cooling chambers


stock


I made loads of these ones and i think this is what your looking for friend only a wee bit bigger that the lexan one i made with the led fans (looked great at night)

That's awesome, and I didn't know UV lights kill algae...that's good to know. I finally have a legit performance reason for the UV...

[aspire.comptech]

Computer UV lights don't have near enough power to kill much of anything. You need the kind used in like tanning beds

[fart_plume]

A silver rod in the res will kill most biologicals...............

[xVeinx]

If you want something that will perform exceptionally well for this type of thing, use acetone. You'd have to make sure that you used all metal blocks and you can find tubing that wouldn't break down from the acetone. Apart from that, make your chiller with glass, etc. You'd have a killer setup

[fart_plume]

If you want something that will perform exceptionally well for this type of thing, use acetone. You'd have to make sure that you used all metal blocks and you can find tubing that wouldn't break down from the acetone. Apart from that, make your chiller with glass, etc. You'd have a killer setup

Killer is right

andyc

That's one of the parameters is too try and find a coolant that "IS NOT VOLITILE!"

I've already have 33% burn scarring and I would prefer not to have more. I have no fear off fire, but I do have a healthy respect for it.............

[NaeKuh]

dude why is all the vendors afraid of touching silver?

I mean common a little derlin T looking thing but inside plasma coated with silver would do the job. Shouldnt be too expensive and would get us needy people off copper safe and other funky additives. :\

[fart_plume]

Probably afraid their security isn't good enough to prevent thefts of the raw material, we had problems with thieves trying to steal full 4 x8 sheets of titanium(at the time worth $50/lb)

[aspire.comptech]

I highly doubt that's the reason....

Silver is expensive for sure but we aren't talking sheets of it.

[fart_plume]

no matter what the form of the raw material security is always a concern, aside from the expense of the material, it could be the expense of getting the parts coated or the cost of the equipment to do so. Al ot of different things may be preventing them from doing this.

[Glyph]

I work with a lot of highly volatile liquids at work. Chloroform, diethyl ether, hexafluorocyclobutane (fun stuff!!), acetone, dichloromethane, hexafluorobutyne, tetrafluoroethane, trichlorotrifluoroethane, liquid nitrogen, liquid argon ($$$), butane, tetrahydrofurane, thiophene, tetrahydrothiophene (stinkiest stuff i know), hexane, pentane, isopropanol, ethanol, methanol, formaldehyde, trifluoroacetone and several others i can't mention.

All of them are useful... but the ones that are cheap are highly flammable, the ones that are not flammable are highly expensive, the ones that are cheap and non-flammable are toxic to varying degrees, the ones that are non-toxic, non-flammable and cheap... are not volatile.

the only ones i can really say that are useful to a computer modder are...(Drumroll please).... Water, trifluoroethane and perfluorohexane. Trifluoroethane is better known as R134a refrigerant.

While you technically could full your setup with R134a, it would boil away so fast that the expense of an open-loop system would be excessive. If you get a recondenser going it would have to be stronger than the evaporation of R134a in order to recondense it to liquid. you'd be better of connecting THAT system to your CPU....

congrats, you've reinvented the phase-change cooler.

The other fluorocarbon liquids (like perfluorohexane) i would reccomend because they are non-conductive and won't damage electronics if they leak. But they are also quite expensive and while they would make excellent evaporative coolants in an open system you'd be spending a crap load of cash just for a few hours of run time. It's actually more expensive to use the flurocarbons than it is to use liquid nitrogen for the same level of cooling power. this is because the fluorocarbons are totally artificial substances that must be made from base elements and purified to remove the toxic by-products of the synthesis.

So in the end, alternative liquids are going to be more expensive than currently available cooling options which are more mature and give better performance.

just my 2 cents.

[fart_plume]

Never said it would be easy or cheap

[kinghong1970]

damn Andy... that's harsh... even for me just reading about it on the internet... can't begin to imagine what it did for your wife...

respects, condolences, and prayers bro...

[NaeKuh]

congrats, you've reinvented the phase-change cooler.

just my 2 cents.

ROFLROFL i was reading your post and was going hey! isnt that phase changing and then you poped this line.

[-Acid-]

The joys of being in a different time zone lol

The black tower is a forced draft design but i worked out the length the upper part should be so that a soild heat exchange happens and little water as posible is lost im my given climate.

Question you recon your quad will be throwing out 250w, what volts ect will you be pushing ect
the reason is i think my quad @ 4.2 is 220ish

[fart_plume]

No problem, I don't consider it a derailment but a stern reminder of what can happen if you get stupid.

It's one reason I started the thread, it's a challenge to come up with a cooling system that will make the phase boys turn their heads and go WTH?
I don't believe the limits of liquidcooling are anywhere near there potential, let's get to it and figure it out.