Alternative cooling methods

R

RJA

Weaksauce
Joined
Mar 4, 2012
Messages
89
I've switched between water and air in the past but ever since my first WC setup back in 2006,
I've been stuck on the idea of combining a sub ambient temperature cooling method with a
conventional one in a practical manner. Back when I set WC up (Zalman Reserator cooling CPU & GPU),
I happened to switch to DSL from Charter. That winter I modded the hole for the cable wire
and fed some of the tubing outdoors to shave of a few degrees.
I crave a practical helper for air cooling, one that wouldn't complex and expensive, otherwise it's just logical to go with a full WC setup.

I've been toying with the idea of using a portable fridge to feed, no matter how minimal,
a sub ambient temp cooling element to the case. I've ordered a HAF X and since I've been
stuck on finding a way to use the bottom air intake to feed cooler air.
Right now, I'm debating using a copper pipe sitting halfway in the mini fridge
with a fan mounted inside to cool via bottom air intake
or using a plastic duct to feed into the bottom of the case.

I can't exactly call up myth busters and have it resolved once and for all.
This idea just keep coming back every once in a while, it would be nice
to have the curiosity finally satisfied.

I wouldn't mind seeing the high end cases offer air cooling helpers soon,
in one shape or another. An extra tray in the bottom of the case housing
a cooling helper, too far out?
 
You will probably just kill the mini-fridge if the compressor is kept running. Pretty sure that was the explanation I have seen elsewhere (they aren't designed for continuous duty, or dissipating that much heat). There were cases that had a tec unit mounted in the front at some point, but I am not sure how well they worked.
 
And not exactly energy efficient. But as others said, the refrigerator would need to be pretty large to absorb that kind of heat output - they are designed to cool things down once and then keep them cool, not cool down something continuously generating heat. If you wanted to go that way, just mount the radiator in front of the vent on a window A/C unit.
 
Ive been thinking alot of WCing and designing my own setup.

The fridge idea was on of them.

How I would work the fridge idea would be to put a larger res. and the radiator physically inside the fridge with the pump being inside the res. as well. Then make sure to have the fridge closed and set to just above freezing point. I would say set it to 40 degrees F.

With enough fluid in the res and the radiator in the freezer and that part of the system closed then the freezer wouldn't have to work very hard to keep the whole thing cold.
 
That'll help in the short term, if you are just running the computer for a little while, because the cooled water in the radiator will act like a heat soak, but ultimately you are still putting the same amount of heat into the fridge no matter what the water temp. You are taking 100W of heat away from the CPU and that heat is going somewhere, in this case into the fridge.
 
I was under the impression that if you run subambient temperatures you have to insulate your motherboard to prevent condensation damage.
 
USMCGrunt said:
I was under the impression that if you run subambient temperatures you have to insulate your motherboard to prevent condensation damage.
Click to expand...

Under the presumption your hardware and/or coolant will be cooled below the dewpoint of the air the computer is in...then yes. If the computer is inside a sealed refrigerator, it should be safe once it equalizes to the non-humid air of the fridge....but....

1) OP, fridges don't really work. People have tried them on a whim, along with rad chiller buckets of ice for external rads. It doesn't work that well for a sustainable long-term cooling solution.

b) If you want to go sub ambient options are in order of extremity for long term: water chiller(and water loop)->phase change->TEC (Peltier)&water-loop. Any of those will not be that quiet, and get pricey quick. There are a few other options like sub-ground outside-the-house radiator loops that you need a trench digger and trash your lawn to set up....all of these should have the motherboard/hoses/cooled hardware treated against condensation as a precaution.
 
Thanks guys. I've finally come to terms with reality. It wold probably work in a huge lab with unlimited funds at disposal but for a end user who cares about noise, $, aesthetics, clutter free setup and low maintenance factor, it's finally been debunked.
 
If you run a chiller unit that isn't meant for constant operation (read: any refrigerator), eventually the compressor will burn out. They simply aren't meant to handle objects that put out a constant amount of heat. They chill the food down when food is placed inside them then the food stays cold.

For a true air conditioner-style system you can either build your own (requires a lot of knowledge) or buy an expensive Vapochill unit. Know that they are expensive up front, use a ton of power, and are quite noisy.

Basically go with air for simplicity, effectiveness, and lowest cost and possibly sacrifice noise or maximum performance or you can go with water for maximum non sub-ambient performance, quiet (if you build with the right parts), but sacrifice low cost.
 
dealmaster said:
If you run a chiller unit that isn't meant for constant operation (read: any refrigerator), eventually the compressor will burn out. They simply aren't meant to handle objects that put out a constant amount of heat. They chill the food down when food is placed inside them then the food stays cold.

For a true air conditioner-style system you can either build your own (requires a lot of knowledge) or buy an expensive Vapochill unit. Know that they are expensive up front, use a ton of power, and are quite noisy.

Basically go with air for simplicity, effectiveness, and lowest cost and possibly sacrifice noise or maximum performance or you can go with water for maximum non sub-ambient performance, quiet (if you build with the right parts), but sacrifice low cost.
Click to expand...

Nice-ish chillers have min/max coolant temp settings to allow the chiller to cycle off.
 
Ironhand said:
Ive been thinking alot of WCing and designing my own setup.

The fridge idea was on of them.

How I would work the fridge idea would be to put a larger res. and the radiator physically inside the fridge with the pump being inside the res. as well. Then make sure to have the fridge closed and set to just above freezing point. I would say set it to 40 degrees F.

With enough fluid in the res and the radiator in the freezer and that part of the system closed then the freezer wouldn't have to work very hard to keep the whole thing cold.
Click to expand...

physics fail

the volume of your reservoir is irrelevant, the fridge compressor has to pump all the heat generated by the system regardless
 
jojo69 said:
physics fail

the volume of your reservoir is irrelevant, the fridge compressor has to pump all the heat generated by the system regardless
Click to expand...

Not really irrelevant. The volume of water in a water loop increases the heat capacity of the loop. More water in a loop means it takes longer for the coolant to increase in temp the same amount under the same thermal load. Means a compressor wouldn't have to work as often, I'd hazard a guess.
 
I do sometimes forget that some computer systems do not run at 100% load throughout their lifetimes.

I suppose if you only load the thing for a couple hours every night, a gaming session or whatever, the increased heat capacity of the loop might extend the life of the compressor.

That is not how I use my systems.
 
Magnus said:
No love for peltier combined with wc?
Click to expand...

Peltiers are incredibly inefficient and require massive amounts of power to cool to the temperatures they achieve. Also, chips put out much more heat these days than they used to when pelts were used more, so they struggle with modern chips.
 
In the past, I would see something for sale that would get my imagination going. Few years ago, I could have bought a water fountain cooler and was debating using that under mild settings to allow the compressor to take longer breaks but it just didn't seem it would be much more effective than a WC setup already is. Then last year, I read about a small company in Philadelphia that makes the world's smallest AC unit and I started brainstorming again.

The reason I started a thread is that up until now, I was convinced that a "helper" could be effective enough to make a difference yet operate without becoming "drained" and that's where I was wrong.

I even considered options to draw the heat outside the case right away, passively and expel it there but that would be a very expensive customized setup that would become useless with any components upgrades.


This is the closest I came to a clean setup: I found these on youtube and was considering running one them either next to the case or elsewhere depending on noise. I couldn't find a price and the BTU was rather low so I gave up.
http://www.youtube.com/watch?v=pURYLzJGUOU
 
There was a thread where a guy ran his tubing into the crawlspace under his house and had rads and fans there.

I think that worked out pretty good and all it costs is a drill bit, some extra tubing and some wire.
 
my thoughts on this is to take a old airconditioner apart and drop the cold side of the heat exchanger into antifreeze then drop a copper coil into the liquid and connect the copper tube into your water loop. basicly creating your own very cheap very powerful water chiller.
 
Cupidsrage said:
my thoughts on this is to take a old airconditioner apart and drop the cold side of the heat exchanger into antifreeze then drop a copper coil into the liquid and connect the copper tube into your water loop. basicly creating your own very cheap very powerful water chiller.
Click to expand...

and watch the condensation drip off your motherboard
 
Do you remember where? I would like to see what he did.
So I did some more digging about IQ1000MM and used their BTU calculator.
http://www.iceqube.com/pdf/IceQube_AC_CALC.pdf


Within desired parameters, temperature spread is 7 degrees, so if unit is on at 80F, it will be off at 73F.

This IQ1000MM is intriguing. According to the calculator, for a full tower case and a heat output of 550W,
I would need a 2200 BTU unit but the calculator doesn't account for existing cooling and the target temperature at 24C (75F)
is way beyond my needs, I would be happy with anything below 40C (104F).
Here are the smallest ones I found: http://www.iceqube.com/enclosureairconditioners.asp
It looks compact and should be durable but I'm curious how loud it is and how much it costs.

They also have compact TEC's to be placed inside enclosures.
http://www.iceqube.com/thermalelectric.asp

ISO certified. I wish someone would review that stuff or if one is using it to post impressions, otherwise I'll never get info if it's feasible or not.
 
Last edited:
that's true you would have to either regulate the temp very closely or you would have to Insulate very well.
 
jojo69 said:
physics fail

the volume of your reservoir is irrelevant, the fridge compressor has to pump all the heat generated by the system regardless
Click to expand...

The volume does matter as it takes longer for a larger volume of water to become heat soaked.

This has been learned in actual practice in cars with air to liquid intercoolers where temps can reach over 300 F. So whatever someones little computer system can generate for heat energy can be dispensed with quite effectively if done properly.
 
I'm curious about the "fridge isn't design to handle heat" statement that is being repeated over and over. Little "dorm" fridge maybe not, but think average home fridge should work.
Can anyone supply some links where it was actually used and failed?

Most CPUs really aren't huge heaters.. 175 watt is good guess and that's about 500 btu/hour... about the same as average home fridge does... and that's keeping fridge/coolant at 5c.

Small "dorm" fridges are in the 150 btu/hour range.. or about 61watt/hour constant. Don't know what that converts to compared to 500 btu being 5c.
 
You must log in or register to reply here.
Back
Top