Mineral Oil AND Water Cooling?

[VashTheStampede]

The oddest thoughts strike me at the oddest hours, this one hit at approximately 3:02 AM. What if you used BOTH mineral oil and water to cool your rig? Since oil and water do not mix, the water, or oil, would originally absorb the heat, while the other takes the heat off the of the lower liquid, and thus equal lower temps? I dunno, it seems weirder as I write it all out, but what do you guys think?

~RT~

[donny_paycheck]

They'd both be circulating really fast via a pump though, right? Cuz I dunno if that'd make much of a difference if the stuff was moving around so fast. Also, does mineral oil conduct heat very well? How is it in comparison to water?

[ButcherUK]

I don't imagine this would work to well - water is much less viscous than oil and better at heat transfer so there would be an imbalance there. Also the oil would tend to collect in the higher points of the system rather than flowing evenly.

[Wicked Klown]

Why not place the MoBo in a tub of oil and then watercool the CPU by a waterblock. Not really sure if this would work but it`s an idea.

[VashTheStampede]

How about this design then? Two-layered water/oil-block. Let me explain the picture below a bit. The gray rectangle is the CPU, the black filled rectangle is a rubber gromet to keep it water/air tight. The mineral oil block will have no bottom, the mineral oil will absorb directly from the die. Then transfer that heat through a thin copper plate to the water block on top, which will quickly take the heat off of the mineral oil block. Both liquids will go through two different pumps, so neither block is with stagnant water or oil.

~RT~

[DodgeViper]

Here is a photo of a display at Comdex using mineral oil. The board and power supply is in the oil.

[r0ckstarbob]

oils got great thermal capacity and really bad thermal conductivity. in tems of steady state heat transfer (like what we're doing) the capacity of a liquid isn't as important as it's conductivity more often then not.

in addition, by using this method you are increasing the heat transfer points from 2 points (die to WB to water) to 5 of them (die to oil block to oil to other side of the oil block to water block to water). this also means that due to the inefficiency of your oil to conduct heat, the oil block becomes the major choke point.

submerging the board in oil and then running a standard watercooling setup inside of it i think would be of more benefit but if you're not going to go sub-Z then i can't see the point. certainly interesting to think about though...

[The Overclocker]

i dont see the point of using water when you could use evaporative cooling and take adavntage of the oil, its freezing temp and the fact that their is no condensation

[Maximus Nickus]

Which sort of oil is non-conductive and has good heat transfer and is cheap?

I don't mind losing a degree or two for the knowledge that no matter how much of the fluid leaks (if it ever did) that there would be no damage and no BANG!!! of electrics.

So which could I use in watercooling?

[The Overclocker]

Originally posted by nick_cw
Which sort of oil is non-conductive and has good heat transfer and is cheap?

I don't mind losing a degree or two for the knowledge that no matter how much of the fluid leaks (if it ever did) that there would be no damage and no BANG!!! of electrics.

So which could I use in watercooling?

this minerel oil in question is very expensive and made by 3m, i dont know the dame, i have seen someone use cooling oil but, as it can be worked out. it was not nice. distilled water has less eletical conductivity

[Maximus Nickus]

What liquid which won't evaporate can I use which is totally NON-CONDUCTIVE...?

[hunter00]

im in the slow process of doing something exactly like this (soaked in mineral oil, and then standard double peltier setup in it)

apparently any oil should work, even like vegetable oil and stuff, but mineral oil is just a lot purer (=better?)

so if you guys want to see how it turns out you should answer all my questions :D

also if you have any ideas or warnings or anything like that tell me. thanks.

[CrystalMethod]

Originally posted by the overclocker


this minerel oil in question is very expensive and made by 3m, i dont know the dame, i have seen someone use cooling oil but, as it can be worked out. it was not nice. distilled water has less eletical conductivity
The liquid you're thinking of from 3M is called "FLORINERT™". It's not a mineral oil at all. It's a completley synthetic fluid that's got negligible electrical conductivity. Here in Canada, it sell for about $80 a pound (for some reason they measure quantity in pounds, and not volume). It weights in at a wee bit under twice the mass of water. I think me and my freind came to the conclusion it's take at least a 20 liter bucket of the stuff to have enought to cover all the necessary components and have enough to circulate through a radiator to cool it down.

[VashTheStampede]

Mineral Oil is about $30 a bucket, not sure of poundage or volume, but the guy using it said once bucket filled his 10 gallon aquarium about a 1/3 of the way.

~RT~

[ButcherUK]

Originally posted by hunter00
apparently any oil should work, even like vegetable oil and stuff, but mineral oil is just a lot purer (=better?)

Vegetable oil is a bad idea as it goes rotten after a while (as organic compounds tend to). You want a very thin precision instrument oil, or a thin oil for something like a compressor. Basically you want it as thin as possible.

[muddocktor]

You could try a poly alpha olefin or isomerized alpha olefin. They are synthetic oils that we use in oilfield drilling fluids down here in Louisiana. They are very pure and are almost as thin as water. They also shouldn't be too expensive, around $8 a gallon. I'm pretty sure that they are not conductive, but I'm not positive on that. The weight on them is about 7 pounds per gallon. That is a specific gravity of .84

[hunter00]

do you know where i could buy some of this stuff?

[wolfsid]

Originally posted by the overclocker
i dont see the point of using water when you could use evaporative cooling and take adavntage of the oil, its freezing temp and the fact that their is no condensation


I agree with him I would Use the oil in a tank put the board in it pretty much under complety then if U have a vapor chilling unit or a waterblock with 156 or 220 watt pelter then U should have Zero worries about ice forming and also you should acheive a lower temp... i was planing on building a costum case where the board and everthing else will fit perfect in not to big like koollance big enought to just fit the board and videocards and watercooling blocks cpu gpu and chipest then i close it all in there and make it moble to go brad at lan partys.....:D


hack the planet..

[Intraveinous]

Originally posted by wolfsid



I agree with him I would Use the oil in a tank put the board in it pretty much under complety then if U have a vapor chilling unit or a waterblock with 156 or 220 watt pelter then U should have Zero worries about ice forming and also you should acheive a lower temp... i was planing on building a costum case where the board and everthing else will fit perfect in not to big like koollance big enought to just fit the board and videocards and watercooling blocks cpu gpu and chipest then i close it all in there and make it moble to go brad at lan partys.....:D


hack the planet..

While you would have the coolest case at the lan party, plan on bringing a hand truck to move the thing... Water weighs in at about 8 lbs (~3.6 kilos) a gallon (~4 litres). Is Mineral Oil more dense (higher specific gravity) that water??? Regardless, even if it's got a pretty similar specific gravity, figure you'll need at least 5 gallons (20 litres as was said before) to cover everything. Thats 40 lbs (18 kilos) just in liquid weight, plus the container and all components... Something this extreme really isn't feasable for a lan party case, unless the lan party's at your house... :D
There was an earlier thread where someone, ButcherUK I think, mentioned that mineral oil can swell the o-rings on certain capacitors, which can lead to them bursting eventually after being submerged in oil for a while. Burst capacitor = dead whatever it was on... He sugested some kind of conformal coating over the capacitors before submerging...
Peace
John

[wolfsid]

Originally posted by Intraveinous


While you would have the coolest case at the LAN party, plan on bringing a hand truck to move the thing... Water weighs in at about 8 lbs (~3.6 kilos) a gallon (~4 litres). Is Mineral Oil more dense (higher specific gravity) that water??? Regardless, even if it's got a pretty similar specific gravity, figure you'll need at least 5 gallons (20 litres as was said before) to cover everything. Thats 40 lbs (18 kilos) just in liquid weight, plus the container and all components... Something this extreme really isn't feasable for a LAN party case, unless the LAN party's at your house... :D
There was an earlier thread where someone, ButcherUK I think, mentioned that mineral oil can swell the o-rings on certain capacitors, which can lead to them bursting eventually after being submerged in oil for a while. Burst capacitor = dead whatever it was on... He suggested some kind of conformal coating over the capacitors before submerging...
Peace
John


I read the same post I was going to put like 6 coats on both from and back.. as for the weight of the case I already thought about that and I was thinking of putting it on wheels with a little remote control so I don't even have to push it :D Why use muscle when you can use brain... not saying that I am weak just saying I use my brain more now...


hack the planet :cool:

[Monaco]

we discussed quite a bit of this already in this thread:
http://forums.overclockers.ws/vb/showthread.php?s=&threadid=47387

basically: oil is expensive, may destroy capacitors over time, and lighter than water. Also, you will need a special pump to move the oil- a water pump won't cut it.

still a really good fricking idea! I want to see someone try this.

[muddocktor]

Originally posted by Intraveinous



There was an earlier thread where someone, ButcherUK I think, mentioned that mineral oil can swell the o-rings on certain capacitors, which can lead to them bursting eventually after being submerged in oil for a while. Burst capacitor = dead whatever it was on... He sugested some kind of conformal coating over the capacitors before submerging...
Peace
John

Yes, I forgot about that aspect of PAO and IAO synthetic. They definitely will have some impact on rubber products and possibly some plastic products. You would have to seal the components on the board very well to eliminate the possibility of the oil deteriorating seals and such. As far as finding the oils you could check with the petroleum products distributors in your area. I think that most of the PAO and IAO synthetic used in the Texas/Louisiana region is manufactured by Chevron. One other bad trait about these products is that some people (including me:( ) are sensitive to them and will show a skin reaction to long-term contact with them (measured in hours, not days). Some of that stuff got under my watch wristband and burned the heck out of my skin, like a bad sunburn. If you wash it off with soap right away it won't affect you though.