I copied this here from a thread on another website I did. I thought the information would be useful for the people here. Enjoy.
[quote]Okay, just trying to get some information all in one place. I want to build a chiller along the size lines of chilly1's Liquid Frostbyte. And as you may gather from the questions as I ask them, I'm pretty new to all of this but I want to learn because this is becoming an obsessive hobby and unless you do this stuff yourself, it's a really damn expensive hobby Question 1:
What are the statistics of the compressors that I would need to achieve the following temperature ranges? ie, What do I need to look for on ebay or other resale areas?
0 to -20C
| Use a low temp compressor with R134a for the nessassary capacity high pack pressure for say a 200W load use an aea4440 tecumseh(Creds to chilly1) |
-20C to -40C
| Use a low backpressure medium temp compressr for 134a with R507 and an oil change to low temp oil 300w load use aea9415 Techumseh (Creds to chilly1) |
-40C to -60C (or as low as possible without a cascade)
| Use a low backpressue Danfoss sc12 m series or 15C series or a 1/2 hp medium temp tecumseh/ (Creds to chilly1) |
Heat exchangers, which ones work best or comments on where the different types excel.
Tube in shell
Tube in tube
Plate Question 3:
For the above heat exchangers, what is the approximate size of a good resevoir? I know more has a larger heat capacity and buffers against temperature changes but at the cost of longer pull down times and an increase of overall temps. So I assume, especially based on some of Gary's posts that there is a real balance as far as resevoir size goes. What are the general rules for this? Question 4:
How much does the block matter? I know that in normal water cooling there is a big difference between say a Storm 4 and a Maze 4 but how much of a difference does it really make in subzero temps?
| As noted same as above zero - a better block gets you lower temps. You should really get the best block possible, but one that's easy to insulate is probably a better choice in reality. (Creds to Tyrou and Butcher) |
What is the "best" tube size to use for the chilled liquid? What is the difference between using 3/8" and 1/2"?
| 1/2" - it's 80% larger which makes for much better flow, especially at low temps due to viscosity increases. Also for parallel, 1/2" is virtually a must to reduce pressure losses from tubing. (Creds to Tyrou and Butcher) |
What are some good books that I could get to learn the basics of working with AC? Are there some "Idiot's Guide to AC" books or similiar?
| Check Gary Lloyd's book out(Creds to chilly1) |
I've seen the posts about pumps on the liquid cooling forum but my question is which survive the cold temps best? I think the Iwaki is the general consensus but I would appreciate some feedback on this.
| LOL...just like every other thread on this forum...depends. It's choice and really not one that is over and above the rest. Take what information you get and make the best choice. Like trying to ask people if it's better to get pentium or AMD...LOL |
When cooling 2 GPUs, North, & CPU, is it best to do all in series or do 2 parallel loops? If in series, what is the best order to have the flow go or does it even matter at these temperatures?
For Question 2:
| General consensus seems to be series due to pressure drop when running in parallel.[/QUOTE] Question 9: Considering the temps involved and the fact that there will probably be dissimiliar metals in the loop, what are the best fluids to use? I don't want slushing, freezing, or any electrolytic effects. |
Question 10: WTF does prommie mean?
| Florinert but it is expensive. |
(I looked into Fluorinert and I think that FC-77 seems to be the best for our uses, having a range from -110C to 97C (Onycho))
To everyone that answers, thanks a whole lot. :toast:
| An endearing name for Prometeia :D
I couldn't really get much on this. By concensus of use, the best seems to be tube in tube. For Question 3:
Most people will use anwhere from 2 L to 5 gallons as a resevoir. The larger resevoir takes longer to pull down to working temperature but it provides a large buffer against heat gains as you push the processor. The temperature loss from having to cool more seems to be minimal compared to the gains from the temperature buffering action.