Overclock.net banner

1 - 9 of 9 Posts

·
Registered
Joined
·
168 Posts
Discussion Starter #1
Hi Guys,
I was hoping on a new designed 62mm direct die water block but it's starting to look like that is not to be. Here is my dilemma; Should I go with a TEC chiller or a chiller box utilizing an A/C unit?

The TEC chiller: I would be able to use my current geothermal loop to cool the hot side of the TECs. I would need to make some copper heat exchangers (a non-issue) as well as insulate the socket (again a non- issue). While this would be, by far, the easiest method of bringing down my temps and realizing a higher clock. I'm not sure if it will take me to the low temperature realm that I wish to be in.

The chiller box: This would be a more complicated build but I do already have the components to complete it. I have a 12,000 BTU A/C unit from a semi tractor APU system. It's a modular setup so hose length and system configuration is easy to set up. I would mount the condenser and compressor outside my house and pipe in the evaporator lines where I currently have my geothermal lines running.
I also have a complete water cooling loop installed on my machine. Granted I would sell my UT60 240mm radiator and buy one that more closely matches the evaporator size. I have drawn up a couple of designs for the box itself. I would need to buy the PVC/acrylic(?) sheets to construct it but I have a CNC router so that won't be a problem.

So hopefully you can see my dilemma. Cost being somewhat equal. More work in doing the chiller box. What would you recommend?
Liam has been most helpful in supplying me with a wealth of information. Cudo's to you Liam!!! I would like to get some more input from the other cooling Gurus here.

Just for information sake. I plan on vacuuming the chiller box and replacing the air with nitrogen (I have a ready supply of it). I also have available a CNC router, plasma cutter, laser cutter (for plastic), a 3d printer, a r-134a A/C evac/charging machine, TIG/MIG/aluminum welders. And pretty much anything else I could think of needing. Being a teacher at a tech school has it's advantages! So be creative in your suggestions.

Also, this is my daily driver. I'm currently running at 5Ghz on a FX-8350 and have 2 R9-290s @ 1150/1475. I hope to gain AT LEAST 500Mhz on the CPU and some on the GPUs.

EDIT: I forgot to mention that "Near silent" is the first and most important goal.
 

·
Registered
Joined
·
2,838 Posts
You will get much more performance utilizing a AC unit, and should have a lesser power bill.
You can always insulate and run direct die with a ac unit

If you know tig welding and all of that making a phase change should be easy?
Phase change usles less watts and does more cooling, EG 300W input for 700W of cooling
For Peltiers, you'll put in 1000W for 300-600W of cooling, depending on what voltages you run them at and what delta of hot and cold side. Once you get icy cold delta of hot cold side should be quite great and you'll probally end up with 0w of cold cooling output. While phase change stays linerar at all times

Unless you have a great cpu chip and 1000W or more of peltier cooling. I do not think going from custom loop to a TEC chiller can give you 5.5GHZ stable on prime 95
 

·
Registered
Joined
·
168 Posts
Discussion Starter #3
So I've been perusing the threads here and have found some interesting stuff.
I found one thread where both the condenser and evaporator are multi-plate liquid to liquid heat exchangers. I have some of those. If I can use them then I can use the geothermal loop to cool the condenser. They are aluminum/stainless; will that present a problem with the computer loop?
Thinking of putting the UT60 radiator in the computer loop to act as the case chiller. Does that sound like a good idea?
Planning on using r-134a refrigerant as that's what I have. Should I consider another refrigerant or will this do? Can get r-22 if necessary.
 

·
Registered
Joined
·
2,838 Posts
So I actually do not know much on how a phase change works, so possibly ask someone else about it.

I'm also confused or don't know what the heat exchager stuff is. But if your thinking of using peltier to cool down your condeser for your phase change its not a good idea.

If you want cold then I recomend doing a direct die of phase change is good and even colder you can go cascade phase change
I don't think a sub zero case chiller with a vacum case would yeild much performance

If you were to vacuumm your case and replace the air with ln2 then yeah, you'll get some extreme cooling performance I reckon. But possibly the ln2 may heat up and you'll be slowly going back to room temp

If your worried about how that pcv/arcrilc is going to perform in subzero then just go stainless steel tubes

Performance wize I'll say phase change.
Peltiers is less work and with more peltiers and running them on 25% of maximum wattage will let you have more cooling per watt, but you will suffer as the delta of hot and cold side increaces your cooling out put will decreacse. So you wouldn't be able to achieve anything extreme.
(Peltiers is just hooking up to a waterblock cooling loop, power supply and done)
 

·
TEC Veteran
Joined
·
2,007 Posts
"Near Silent" screams TEC. I think the TEC with the geothermal cooling has a lot of potential, but the phase chillbox will for sure be a lot more powerful.
Quote:
Originally Posted by Iwamotto Tetsuz View Post

S
Peltiers is less work and with more peltiers and running them on 25% of maximum wattage will let you have more cooling per watt, but you will suffer as the delta of hot and cold side increaces your cooling out put will decreacse. So you wouldn't be able to achieve anything extreme.
(Peltiers is just hooking up to a waterblock cooling loop, power supply and done)
TEC will never have more cooling per watt then phase, and the delta being big is a good thing. Cooling output fails when the TEC delta decreases, causing the cold side to warm up. You've got it flipped around.
 

·
Registered
Joined
·
2,838 Posts
When you get too cold like -20c delta of hot cold side would surely be very high
 

·
Registered
Joined
·
1,045 Posts
I like the idea of liquid cooling the condenser, uses your prior investment in the geothermal loop so its not a waste, will help to keep the noise down as well, but as you can mount the compressor outside your noise is low anyway, I'd say with that A/C unit you have you are best going in that direction. I don't think you will be satisfied with what you can achieve from the TEC's, at least not with your 8350 as they can be run at high volts and frequencies provided you can cool it, but as the power/heat is so high phase is the better option.

the aluminium/stainless heat exchanger is not ideal, its fine in the condenser/hot loop, but in the evap/PC chiller loop it could create problems, should be able to avoid that with the right mix of fluid. I'd say methanol will stop most problems but you should probably ask orthello and Giveitup about that stuff for confirmation. Your 240 rad is perfect for using inside the chill box, its all that is required. Keep the inside of the chill box as small as possible while still being big enough to work in easily, that will ensure faster pull down of air temp so less likely to form condensation. make sure the rad is the first component in the loop order inside the chill box.

as for the nitrogen vacuum, I'd say it's unnecessary and not practical, you don't want to have to charge the chill box with nitrogen every time you change a component, that's for you to decide though, I think it is not necessary though, once again, check with orthello. The chill box will be in a vacuum/lower than atmospheric pressure state due to the internal temp drop, so there are no issues with dew/condensation. why do you want to use nitrogen, just to remove all risk of condensation?

I will say that TEC's are extremely fun and easy to mess around with, they are very versatile and can do many many things well, but for pure brunt cooling power, your A/C will get you further, and that particiular A/C unit you have removes the noise/size issue as you can mount the bulk of it outside, those 2 things are my biggest beef with phase units. My TEC direct die chiller is portable though, which is nice as I don't own the home I live in, but for you, permanent is fine (I take it you own the house because of your investment in the geothermal loop, that thing is seriously sweet by the way, that alone is more of an investment/extreme cooling set up than I've ever seen on here haha)
 

·
Registered
Joined
·
168 Posts
Discussion Starter #8
Yes I own my home. The flooding we had back in October has made me question that decision during the recovery phase. They just finished insulating the crawlspace today; which was the last thing to be done.

I really like the chiller box idea. You and I talked about a mineral oil bath but this sounds like this will accomplish the same task without the mess.

You hit the nail on the head with the nitrogen. I wouldn't leave it under a vacuum but instead; would vacuum out the air and replace it with nitrogen. Eliminating condensation is the primary reason but from what I've read, an alcohol based fluid will be needed in the loop. I was thinking that a low oxygen environment would help in the case of a fluid leak. It looks like a 70-80% alcohol mix will be needed to eliminate the possibility of freezing. Would hate to burn down the house right after I got it put back together
wink.gif


I'm not sure of the volume I will need for the condenser heat exchanger. I could daisy chain a couple together if needed but will have to research that. It gets HOT here in SC and I don't want my condenser temperature influenced by ambient temps. I guess I could build a box for the systems existing condenser and seal it as best as I can. I would have to check it regularly for evaporation loss but that wouldn't be a big deal. BTW, the geothermal system cost less than $200. Some PEX, copper tube, fittings, and insulation. And all told it only took about 6 hours of labor.

It would seem that it is time to move my build from the TEC section to the experimental one. Darn, I really like it here...
 

·
Registered
Joined
·
1,045 Posts
by all means keep us updated in here haha. but you might be better served in the phase section, not the experimental section. I would just start posting in this thread http://www.overclock.net/t/1533164/the-24-7-sub-zero-liquid-chillbox-club/600_100 Phase section is full of knowledge and they are all happy to help, NOL, Orthello, GiveItUp and many more will all guide you in the right direction about how best to build your phase system and the sizing of components etc. Also I beleive R22 is the prefered refrigerant, not R134a, as the boiling point of R22 is much lower than R134a, so you would be better with a charge of R22. (not sure if you can get it without a licence, most guys just buy existing old A/C units that have R22 in them.

Orthello is a wealth of knowledge on the chill box and is more than happy to pass on any tips. The addition of the nitrogen might even be something all the guys with chill boxes may try as they all have vacuum pumps etc already and could easily get some nitrogen. Or they may just say don't bother too haha.

That's a great value system you built with that geothermal loop, far more economical than the number of radiators you would need to achieve the same coolant temp in your loop. Good to know
biggrin.gif


I live in Brisbane in Australia, I know all about how bad floods can be, terrible time for a lot of people when it flooded here about 5 years ago, the clean up took months and the repairs are only really being completed now.
 
1 - 9 of 9 Posts
Top