Sign in with
Sign up | Sign in
Your question

Submerged + Water loop?

Last response: in Overclocking
Share
December 18, 2009 11:28:08 AM

I've made up my mind that I'm going to give a mineral oil submerged build a shot. I've looked at the reactor system, and while their prices have dropped to reasonable levels I hate getting locked into proprietary hardware, especially considering how inactive their forum and CS is.

My second look was at the aquarium kits. If I were running a 2 or 3 year old system that was just being used to browse the web then perhaps that would be ok. Otherwise their 6 gallon solution (even with rad/pump) seems inadequate.

I'm setting up a i7 920 that I'm going to push to the 4ghz range. I'll then be slapping a 5870 (possibly 5970) onboard. If I do go the 5870 route CF will be on my mind down the road. Eitherway we're talking about some pretty massive heat signatures.

Now I'm thinking of running from a 12 gallon tank and adding a 120.3 loop for the gpu/cpu. Mind you, I'd be pumping water, not oil. The oil would remain passively cooled. Unlike the Puget setup I'll have my psu mounted externally to reduce tank heat. All that's left is nb, memory, and whatever heat the loop doesn't carry away.

I haven't seen anyone else give this a shot. Thoughts on it's efficiency? If the oil does start to creep up just from the residual heat not radiating fast enough I could add a 2nd rad/pump strictly for cooling the oil.

I'm sure the first question people will ask is why even bother with the oil if I'm going to have the water loop. Well, hotspot management mostly. But dust is also another major deciding factor, as it's unmanageable here. That, and tbh... I really just love the concept. I'd really appreciate any thoughts on if a water loop would be as effective in oil as out, and if there are any angles I'm missing here.

I had briefly considered a chilled solution by converting a small 5600 btu window unit. It's still rattling around in my head, but honestly I'm not sure if I want the headache of the window units noise and the external nature of it all. There are certain benefits (no condensation in oil) but... I dunno, maybe down the road.

More about : submerged water loop

a b K Overclocking
December 18, 2009 5:35:39 PM

According to technical reviews I've read the past few days both the Intel Core i7 920 and the 860 can be pushed to 4.2Ghz on air cooling alone without much of a problem. They can be pushed to 5.2Ghz or more on liquid nitrogen.
m
0
l
a b K Overclocking
December 18, 2009 5:50:10 PM

Not a very good idea IMO i think it would be just fine if you only did a water cooling.

A good i7 920 can hit approx 4.3-4.5 on water.
m
0
l
Related resources
December 18, 2009 10:32:09 PM

overshocked said:
Not a very good idea IMO i think it would be just fine if you only did a water cooling.

A good i7 920 can hit approx 4.3-4.5 on water.


Could I ask you to elaborate as to why you feel this isn't a good idea?
m
0
l
a b K Overclocking
December 18, 2009 10:50:55 PM

^ I'll spare overshocked some typing and say this:

1. It's a PITA to upgrade a oil cooled PC

2. You'd have to buy a pump + radiator + tubing to cool the oil anyways (granted, it will be a cheap heater core + cheap aquarium pump). May as well stick with WCing.

3. If you plan to get a chiller, why not go below ambient with phase change/cascade or chilled water?

4. I doubt you'll get any cooler or better (may be a little) OC by doing this method vs just plain WCing.

m
0
l
a c 155 K Overclocking
December 18, 2009 10:51:20 PM

Why bother when you can hit 4.3 Ghz on air cooling ?
m
0
l
a b K Overclocking
December 18, 2009 10:56:53 PM

^ Just so you could say you'v done it :D . Not many can say that. I did this oil cooling on a Atom build, and it worked fine, but swapping out parts is a MAJOR PITA.
m
0
l
December 19, 2009 9:54:33 AM

Shadow703793 said:
^ I'll spare overshocked some typing and say this:

1. It's a PITA to upgrade a oil cooled PC


I dunno, from my perspective it seems easier to upgrade (admittedly not by much) You don't have to touch a single case screw. Just lift out of the oil, plug and go. If you're talking about something along the lines of a cpu I can see your point, but I've yet to replace a cpu w/o also upgrading the mobo in the process.


2. You'd have to buy a pump + radiator + tubing to cool the oil anyways (granted, it will be a cheap heater core + cheap aquarium pump). May as well stick with WCing.
said:

2. You'd have to buy a pump + radiator + tubing to cool the oil anyways (granted, it will be a cheap heater core + cheap aquarium pump). May as well stick with WCing.


Initially I only plan to pump and radiate the water. *IF* the oil requires additional cooling then yes, but it's not as if I'm going to be surprised by this. I'm not adverse to spending more to get this setup if the benefit is worthwhile. At this point, avoiding dust, reducing noise, avoiding condensation (if I go with a chiller), and most important stabilizing the system by reducing mobo heat spots seem like a worthwhile investment.


3. If you plan to get a chiller, why not go below ambient with phase change/cascade or chilled water?
said:

3. If you plan to get a chiller, why not go below ambient with phase change/cascade or chilled water?


Have you looked at peltier's and how ineffecient they are? Running a cascade isn't a weekend project. Anything involving a TEC is going to be a MAJOR undertaking. A simple chiller will run sub-ambient (sub-freezing if you set it up correctly) This however is a project for down the road, the oil cooled system has the benefit of not requiring massive amounts of foam insulation running into your case by eliminating condensation. The issue here of couse is oil density and how it gels (thus becoming VERY difficult to pump) at lower temps. There are some cost effective ways around this, Voltesso 35 is a relatively cheap oil (4 bucks a liter) that has a pour point of -40c, low viscocity increase on chilling, and a boiling point of 260c. If I were to decide to run a chiller from my old a/c unit, I'd upgrade the oil accordingly.



4. I doubt you'll get any cooler or better (may be a little) OC by doing this method vs just plain WCing. said:

4. I doubt you'll get any cooler or better (may be a little) OC by doing this method vs just plain WCing.


Now this I can agree with 100pct. As a matter of fact I probably won't see any greater OC'ing at all considering I'll be cooling the gpu/cpu with a waterloop. What I do hope to accomplish is greater system stability by keeping the other (mosfets, memory, nb) components cool.




Why bother when you can hit 4.3 Ghz on air cooling ? said:

Why bother when you can hit 4.3 Ghz on air cooling ?


Let me just say that I'd have to see it to believe it. Even running premier water solutions it's tough to reach that. The i7's seem to have a natural limit right around the 4-4.2 range unless you just get spectacularly lucky with your binning.



I'm submerging. The question I pose to everyone here is simple...
Am I missing some unforseen issue by combining a submerged system with a water loop?
m
0
l
December 19, 2009 12:13:55 PM

Since you are set to move forward anyway:

1) Make all possible effort to test your water loop thoroughly before you submerge.

2) Design your loops with contingencies in mind. Be ready in case you do have a major leak. You are lucky in the sense that water and oil don't mix, but be prepared in some way to separate the water from the oil. Also be aware that most tubing, whether PVC, rubber or plastic, will allow for some osmotic loss of fluid from your loop. It's not much, but enough that you will probably end up with a thin layer on top of your oil. Just be aware of this or you will spend a lot of unnecessary time looking for a leak. Only be concerned if there is a lot of water on top of the oil.

3) No matter what you do, it will be messy if you do anything. You will have oil everywhere. I hope you will have this machine in a room that doesn't have carpet, because a spill or a leak will probably ruin that carpet. Otherwise your wife/significant other/roommate/parents will probably be very upset as well.

4) Unless you seal the system, you will end up with "junk" in your oil. Dust, hair, bugs and other small animals. Seal your case unless you want to be fishing said junk out of the oil.

5) From talking to folks who do submerge, don't submerge your hard drives. The oil will break down the rubber gaskets that seal the drive, and you will end up with oil in the drive itself, which will ultimately end with the drive failing.

Good luck, and let us know how it turns out. Conundrum, shadow and I can give you some pointers with regards to the water cooling (and some more extreme as well), but the only advice I can offer for the submersion is to read read read, and then read some more.
m
0
l
December 19, 2009 1:40:42 PM

Houndsteeth said:
Since you are set to move forward anyway:

1) Make all possible effort to test your water loop thoroughly before you submerge.

2) Design your loops with contingencies in mind. Be ready in case you do have a major leak. You are lucky in the sense that water and oil don't mix, but be prepared in some way to separate the water from the oil. Also be aware that most tubing, whether PVC, rubber or plastic, will allow for some osmotic loss of fluid from your loop. It's not much, but enough that you will probably end up with a thin layer on top of your oil. Just be aware of this or you will spend a lot of unnecessary time looking for a leak. Only be concerned if there is a lot of water on top of the oil.

3) No matter what you do, it will be messy if you do anything. You will have oil everywhere. I hope you will have this machine in a room that doesn't have carpet, because a spill or a leak will probably ruin that carpet. Otherwise your wife/significant other/roommate/parents will probably be very upset as well.

4) Unless you seal the system, you will end up with "junk" in your oil. Dust, hair, bugs and other small animals. Seal your case unless you want to be fishing said junk out of the oil.

5) From talking to folks who do submerge, don't submerge your hard drives. The oil will break down the rubber gaskets that seal the drive, and you will end up with oil in the drive itself, which will ultimately end with the drive failing.

Good luck, and let us know how it turns out. Conundrum, shadow and I can give you some pointers with regards to the water cooling (and some more extreme as well), but the only advice I can offer for the submersion is to read read read, and then read some more.


Tyvm for your reply. I'll make sure and run the loop for awhile on dry newspaper to make sure it's tight. As far as the submersion itself I'll only be doing the mobo and all that it entails. PSU and HDD's will be mounted outside. The HDD mounting has to do with air equilibrium but can still be done (VERY cool setup btw, guy did a spectacular job). That's beyond my level of craftsmanship or desire however.

I'm really not a tinkering kinda guy. I build my systems, then they sit until they get ready to be replaced. This is a functional build so once I get everything setup I don't plan on messing around on the inside anymore. I'll definitely have a lid covering it all, not sure about sealing however.

Anyways, thanks for the advice once again, and once I actually start putting it together I'll run a build log.
m
0
l
!