Mini Air Conditioner

[wrack]

Hi,

I have a tiny project in mind. I got 3 full tower at home and they put up some nasty noise and not to mention the heat so I was thinking something about out of box solution rather than installing cooling in all of them.

I was wondering if I make a nice wooden cage which fits them all and by my guess the inner volume wont be more than 1 cubic meter. So if I do that and put all 3 cases inside the cage and then put an extremely small Air Conditioner then would the whole setup work sufficiently with boxed coolers which came with the CPUs? All CPU are Intel 3.2 GHz LGA 775 (Preseler core).

If it can work then how big/powerful air conditioner would I need.? I am pretty handy with making things so attaching it with the cage wont be a problem.

In long run I want to make that cage a server rack (but with normal boxes).

Any help would be appreciated.

Cheers

 

[BonzaiDuck]

Well, the project doesn't seem "tiny." By the time you pay for the small air-conditioner, the hardwood and/or lexan, do all the cabinet-work, and so forth, the expense would exceed the price of three very efficient heatpipe coolers with low thermal resistances. Figure you could spend $50 each on three heatpipe coolers with thermal resistance of 0.11 C/W, and your room ambient averages maybe 30C or a little less. The mini-air-conditioner is going to cost you maybe $100 alone. The project will probably employ some sort of accordian ducting hose, and there's the cost of materials and hardware to build the box.

I can't estimate whether or not there would be any condensation problems, but you need to consider the possibility. I'd think you're going to blow cool air in, and exhaust warm air from the AC.

I'm just not enthused about something like this, myself.

Why not just put the air-conditioner in a window, improve the seal around doors and windows in the room, and raise your electric bill by keeping the room itself at a constant 65F or around 20C degrees?

Then buy a coat-rack at an estate sale, and keep a parka handy . . .

 

[wrack]

Thanks for the quick reply.

Well I live in a rent house and not allowed to put a proper window air con system and can't perform any modding to the room either. The whole reason of this is to reduce the noise and I could care less about the heat at the moment but both will be good.

So the only choice was a room within a room but I am ready to give a go and buy 1 three heatpipe coolers with thermal resistance of 0.11 C/W (no idea which one is good) and turn other machines off and see the difference (I have a pretty good idea on what the sound levels are like with only 1 machine when it gets hot).

So the question is:

Can you suggest me a good 3/6 heat piper cooler with good performace and low noise? Would be nice if the store can deliver it to Australia as well

Cheers

 

[BonzaiDuck]

There must be as many resellers like this one in Australia as there are Roo-s.

PC Case Gear

These folks sell Zalman CNPS-9500-LED, and the whole gamut of ThermalRight coolers as well as the Scythe ("Ninja" etc.) You'll have to shop around to get the best prices.

Here's Joe Citarella's database of heatsink ratings. The lower the thermal resistance (C/W, or difference in centigrade per thermal watt), the better:

Air-Cooled Heatsink Roundup at OverClockers.com

I use the SI-120, having replaced an XP-120 with it. Some people here swear by the Zalman CNPS-9500. I'm stunned that my own cooler's rating is 0.14 C/W, because I thought it was lower. The Ultra-120 looks like a good bet, and the XP-90C, but there are the noise factors. The larger the fan, the more airflow with less noise.

Citarella's ratings show how the thermal resistance varies with airflow, and we've long since concluded that there is some threshold beyond which pushing more air (CFMs) through the heatsink fins doesn't change the rate of heat removal. So the important thing to remember here is that some of these coolers are very effective when the fan speed is as low as 1,000 rpm.

 

[BonzaiDuck]

I said $50 per unit, but USD's seem to buy a little bit more right now. Prices at that reseller are $10 to $20 higher in Australian bucks. But I still think -- in cost/benefit terms -- you're just as well to forego the cabinet-work, and the mini-air-conditioner probably costs proportionately more in Australian dollars, so "it's a wash."

I just picked that reseller at random off the web. I note that here in the States, you can find some outlets with rock-bottom prices, and other small resellers who are really trying to get rich on each and every sale. So -- shop around. My remarks about the exchange rate may not be so significant.

 

[wrack]

Thanks a lot BonzaiDuck.

I will shop around for a cooler with 120 MM and lowest C/W possible.

Cheers

 

[BonzaiDuck]

I had all sorts of Rube-Goldberg ideas per cooling these last few years, myself.

For instance, there are several variations for an evaporative cooling arrangement for water cooling, that employs 4-inch-dia. PVC plumbing pipe and fittings to make a "nuclear-cooling-tower." Another guy just buried copper plumbing pipe -- maybe 25-feet "going" and 25-feet returning -- in his back-yard, with a pump sufficient to move the water through the pipe and through all his computer-cooling water-blocks. He'd just water his lawn, and on a decently warm day, the evaporation would refrigerate the contents of the pipe. Both of these solutions promised a drop in CPU temperature to an equilibrium that was more than 10C below room temperature, but the cooling tower works less effectively on cool, humid days. Neither idea makes for "computer portability." But they are cheap cooling solutions, so effective that one absolutely must prepare the motherboard and CPU with insulation against condensing moisture.

But with phase-change, water-cooling, chilled-water-cooling, TEC-chilled-water-cooling and so on, there are tradeoffs. Water cooling may consume less power and generate less noise, but it retreats from the KISS principle -- meaning "Keep it simple, stupid." More levels of complexity mean more opportunities for parts to fail, and WC has always carried a very minor risk of leakage. Further, you have to service a water-cooling setup for cleaning and fluid replacement every several months. TEC requires more power, and there is always another component that might eventually fail.

Three laws of thermodynamics generally explain the idea that in order to remove energy from one place, it must go somewhere else. There is also the cost of energy removal, so with TEC, you spend more and more electrical energy to displace the heat you want to go somewhere else.

With heatpipe solutions, you can't cool below room ambient, and you can only approach the effectiveness of regular water-cooling. But the expected lifespan or MTBF for heatpipes is about a million years.

 

[AgentJean]

Originally posted by: BonzaiDuck
I had all sorts of Rube-Goldberg ideas per cooling these last few years, myself.

Another guy just buried copper plumbing pipe -- maybe 25-feet "going" and 25-feet returning -- in his back-yard, with a pump sufficient to move the water through the pipe and through all his computer-cooling water-blocks. He'd just water his lawn, and on a decently warm day, the evaporation would refrigerate the contents of the pipe. Both of these solutions promised a drop in CPU temperature to an equilibrium that was more than 10C below room temperature, but the cooling tower works less effectively on cool, humid days. Neither idea makes for "computer portability." But they are cheap cooling solutions, so effective that one absolutely must prepare the motherboard and CPU with insulation against condensing moisture.

If he'd burry the pipes several feet under ground he'd be using a geo-thermal cooling solution. I wounder how hard it would be to tap off an existing geo-thermal system to cool your computer.

 

[BonzaiDuck]

Well, my knowledge is casual, pop-science. I understand that earth temperature some few feet below the surface is a constant 55F. This supposedly allows houses which are built into hillsides to reduce AC costs during the summer. The project I saw, however, buried the pipe about six to eight inches below the surface, or a few inches below the lawn-sod.

Now I suppose, if you lived near some sort of . . . . volcanic thermal vent, you could use the heat to power some sort of turbine-air-conditioning rig. And that . . . . would be totally "Rube Goldberg," indeed.

I'd think the irony of it, is that in places like Iceland, such opportunities abound, but it's also cold enough there that you wouldn't want to knock yourself out trying to cool down your computer.

You might even want to warm it up a bit . . .

Does anybody know of people who take their laptops into the sauna-bath with them? You wonder whether they have more problems with computer-cooling, or with the LCD screen fogging up.

 

[wrack]

Thanks for the answers and explainations.

I was just searching and came across this one.

VapoChill Micro from Asetek http://www.asetek.com/main/page.asp?sideid=623

what do you think if it's worth considering?

 

[thecoolnessrune]

Most reviews show that the principle is sound, but it simply doesn't live up to expected standards. If you want quiet cooling, Get an Ultra 120. It will be perfectly quiet most of the time.

 

[gamefreakgcb]

The Ultra 120 is a huge heatsink BTW (height) make sure it fits your case though, that said I have mine but am waiting for my proc to arrive.

 

[BonzaiDuck]

Again -- despite what others may say -- I'd pick the heatpipe cooler according to several criteria, but three of them are essential:

1) Lowest thermal resistance
2) Fits the processor, your motherboard, and the case without interfering with other components, and
3) Does not obstruct airflow -- only promoting it.

You can probably always find more than one cooler that resolves criteria #1 equally, so you have a choice to resolve #2 and #3.

The other considerations:

4) minimize noise
5) minimize torque on the motherboard due to the weight of the cooler and its fan, and how much of that weight hangs farthest from the motherboard mounting.

I once looked very hard for thermal resistance data on one of the earlier MicroCool models -- it was "so-so." I think I had seen newer models which boasted big promises. But you can't tell until you find a review site that provides comparison tests with other coolers which have known thermal resistances, or which actually report the thermal resistance for the specific cooler.

Few manufacturers report the thermal resistance measures on their products, even if the product has the lowest C/W value among competitors. So much thanks to Joe Citarella (OverClockers.com) and other reviewers who keep comparison tables that are updated frequently with new models.

 

[gamefreakgcb]

Originally posted by: wrack
Thanks for the answers and explainations.

I was just searching and came across this one.

VapoChill Micro from Asetek http://www.asetek.com/main/page.asp?sideid=623

what do you think if it's worth considering?

The idea behind that cooler was to eliminate one link in the heat transfer process (Heatplate -> Heatpipe) since this one has the heatipe fluid stored in the heatplate, but they screwed up the HSF itself.

 

[wrack]

Sorry to be so ill-informed but what is HSF..?


Edit:
I am actually going to post photos of my case so some one might have a better idea of what can and is needed to be done.

 

[NoStateofMind]

Originally posted by: BonzaiDuck
Again -- despite what others may say -- I'd pick the heatpipe cooler according to several criteria, but three of them are essential:

1) Lowest thermal resistance
2) Fits the processor, your motherboard, and the case without interfering with other components, and
3) Does not obstruct airflow -- only promoting it.

You can probably always find more than one cooler that resolves criteria #1 equally, so you have a choice to resolve #2 and #3.

The other considerations:

4) minimize noise
5) minimize torque on the motherboard due to the weight of the cooler and its fan, and how much of that weight hangs farthest from the motherboard mounting.

I once looked very hard for thermal resistance data on one of the earlier MicroCool models -- it was "so-so." I think I had seen newer models which boasted big promises. But you can't tell until you find a review site that provides comparison tests with other coolers which have known thermal resistances, or which actually report the thermal resistance for the specific cooler.

Few manufacturers report the thermal resistance measures on their products, even if the product has the lowest C/W value among competitors. So much thanks to Joe Citarella (OverClockers.com) and other reviewers who keep comparison tables that are updated frequently with new models.

Ok, give us links to this heatpipe cooling you keep talking about.

 

[wrack]

Originally posted by: gamefreakgcb
The Ultra 120 is a huge heatsink BTW (height) make sure it fits your case though, that said I have mine but am waiting for my proc to arrive.

I take it you are talking about Thermalright Ultra 120..!
http://www.pccasegear.com.au/prod4119.htm

That thing is a monster, not sure if I can fit it in my case.

 

[gamefreakgcb]

Originally posted by: wrack
Sorry to be so ill-informed but what is HSF..?


Edit:
I am actually going to post photos of my case so some one might have a better idea of what can and is needed to be done.

An acronym for HeatSink & Fan. .

 

[wrack]

But ofcourse OH:

TA

 

[BonzaiDuck]

PC Surgeon --

There are thou-sands, and mil-lions, and bil-lions and zil-lions of links on heatpipes. [Carl Sagan, here -- astronomer turned heatpipe hype-promoter.]

All the newer CPU, VGA -- even chipset coolers use heatpipes. There are whitepapers on heatpipes in general, I just don't have the links handy.

I don't remember the name of the company, but I came across a white-paper posted by a long-time NASA contractor. The contractor had actually invented or patented heatpipes, which came into use sometime around the Apollo program.

There are myths about heatpipes, such as "the orientation of the heatpipe must facilitate gravitational influence on the flow of coolant, or it won't be as effective." This is b***s***. The things were designed to work in a weightless environment from the git-go. Certain VGA coolers like the Zalman ZM80D instruct to avoid installing the cooler "upside-down" because the proximity to the CPU heats up the "cool side" of the heatsink -- making it less effective. It has nothing to do with gravity.

If you want links for specific reviews or comparison updates in a database of heatpipe cooler rankings, go to OverClockers.com. Joe Citarella writes and updates those articles and he does all the testing. I can try and hunt down the white-paper I mentioned, but you might find it just running a general web-search on the topic.