what calculations for caps

[Athanatos]

so i'm looking and looking and all i can find are length conversion tables but no calcs, but i did find that dancap program, which is nice and all but it's asking for some values which haven't got a clue how to calculate what temps for the evap. temp/cond. temp/return gas temp

now before asking me technical stuff, i will save you the issue and tell you that i have yet picked up any supplies other than the computer parts that i need to cool (2x gpu G80 gtx, 1x cpu amd x2 5600)

regards

Ted

[The MG Pony]

Caps what? Capciters? Cap tube? Pipe caps?

Details go a loooong way my friend!

[lana]

so i'm looking and looking and all i can find are length conversion tables but no calcs, but i did find that dancap program, which is nice and all but it's asking for some values which haven't got a clue how to calculate what temps for the evap. temp/cond. temp/return gas temp

now before asking me technical stuff, i will save you the issue and tell you that i have yet picked up any supplies other than the computer parts that i need to cool (2x gpu G80 gtx, 1x cpu amd x2 5600)

regards

Ted

Hi Ted,
The software you have (DanCap) is good. The data you requested is obtained from the operating conditions of the cooling system. They don't need to be "calculated". You have to specify your system more clearly.
Best Regards,
LANA

[Gary]

Cap tube sizing for computer overclocking is a whole different world, and as far as I know there are no programs/charts designed specifically for that purpose.

Perhaps this will help:

http://www.xtremesystems.org/forums/showthread.php?t=29694&highlight=cap+tube+rule

BTW, this thread should be in the CPU Overclockers forum.

[Athanatos]

i was worried about the length, after reading the document about length type conversions and looking at the flow vs length graph i wanted to make sure that i can get an optimum efficiency of the cap tube

[Gary]

The lower the refrigerant flow, the lower the evaporating temperature. But if there isn't enough flow to handle the heat load, performance drops off the cliff. The trick is to get just enough flow for the heat load in order to get the best possible performance at the lowest possible temperature.

My intent in the linked thread was to establish a ballpark starting point. You will probably need to experiment a bit to get the flow just right.

Think of yourself as a pioneer. :)

[Gary]

To give the fridgies here an idea of what overclocking is all about. consider the following:

Let's imagine that you have a conventional cap tube freezer. What happens if the cap tube is restricted? There is a small area at the beginning of the evaporator that is heavily frosted. The suction pressure is running low, so this small area is very cold. That's exactly what we want, i.e. a very small very cold area. So, the cap tube needs to be far more restrictive than is the case on a conventional system.

When cooling a computer CPU, the area to be cooled is a flat surface about the size of a postage stamp and we want to make it as cold as possible.

That postage stamp sized surface area is in fact the CPU heat spreader and beneath that thin spreader in the very center is a 1/4 inch by 1/4 inch hot spot. That hot spot is where the action is. The colder you can make that spot, the faster the CPU does it's calculations. And to make this even more of a challenge, the CPU is surrounded by moisture sensitive components.

There is an ongoing worldwide competition to have the fastest computer on Earth and there are people out there with cascade systems, using exotic refrigerants, freezing that tiny hot spot to impressively low temperatures.

I think of overclocking as drag racing for computers. You might want to check it out. It's fascinating what people are doing with these computers.

[The MG Pony]

They need you to go back to Xs and kick around some of the members lol.

It is entertaining to see how small they can scale down the cooling systems to a computer level, coldest one guy has goten to thus far is around -160/-170C under load

[nh3simman]

To give the fridgies here an idea of what overclocking is all about.

There is an ongoing worldwide competition to have the fastest computer on Earth and there are people out there with cascade systems, using exotic refrigerants, freezing that tiny hot spot to impressively low temperatures.

I think of overclocking as drag racing for computers. You might want to check it out. It's fascinating what people are doing with these computers.

Interesting, I have seen overclocking mentioned on this site but not realized the extent. Are computer geeks actually using these exotic refrigeration systems? I suppose there is no choice since you can't just buy a chip cooler over the counter.

[star882]

With CPUs becoming more complex nowadays (*especially* the multi cores), heat has become more of a problem. I have actually engineered a cooling system for desktop computers that uses standard refrigeration hardware. Many think that will be the future of computer cooling. (Those who have dealt with Prescotts will know what I mean. And as for Britney...)
Of course, DIYers mostly use Peltiers since they're cheaper. But companies will probably go with compressor-based cooling since it is actually cheaper in volume. (However, Dell seems to be using Peltiers even for quad cores. I don't know why, but I think they would not want to design refrigeration hardware for the first few prototypes and the first generation of the system. But Peltiers sized for a quad core would be more expensive than compressor-based cooling and material costs dominate in mass production so I don't know what's going on. http://www.dailytech.com/article.aspx?newsid=5663 )

[Sledge]

Hi

Do you have any info on what has been done already? This is interesting, but dont want to go re-inventing the wheel.

This must require some fairly tiny equipment. Any details on it.

[star882]

For a modern CPU, 1/4HP is the accepted minimum. Small, but not tiny. With multiple CPUs, the compressor must be larger. An 8 way Britney typically uses a 2HP compressor.