Using a oil in your PC cooling system

[bork]

... . It's impossible to have pure water in a loop. As soon as the water runs through blocks, rads, fittings, metal ions will be present.
[/quote]

But what is missing is important, the minerals and salts that are in normal tap water. 

[bloody000]

... . It's impossible to have pure water in a loop. As soon as the water runs through blocks, rads, fittings, metal ions will be present.

But what is missing is important, the minerals and salts that are in normal tap water. 
[/quote]
The point is that distilled water can still fry electronics and cause galvanic corrosion.

[zat0x91]

Uhh, corrosion?  Don't fucking mix metals like copper and aluminum and you shouldn't have that problem.

Otherwise, yes, distilled water isn't 100% non-conductive, but it's certainly much safer than other stuff, as other people have recovered video cards, and other components, probably not their motherboard though.

[bloody000]

Jet plates in most CPU blocks are made of steel. The pump also have some steel components.

see table for potential difference of steel-copper vs. aluminium-copper:
http://en.wikipedia.org/wiki/Galvanic_corrosion#Galvanic_compatibility

[zat0x91]

Although it could corrode over time, but it's they're usually made with 316 stainless steel (at least with the EK Supreme) so it won't happen for a long time.  If that really does bother you then put in an inhibitor or just use a different waterblock.

[bloody000]

Although it could corrode over time, but it's they're usually made with 316 stainless steel (at least with the EK Supreme) so it won't happen for a long time.  If that really does bother you then put in an inhibitor or just use a different waterblock.

it does not bother me. I simply do not think distilled water is that great.

[SupraGuy]

Like I said, there's a reason why water (which actually does cause a lot of difficulties with engines) is used as a coolant. It's efficient, and simple.

I suppose that if you wanted to deal with a compressor and expansion valve, you could try using refrigerants, which would vaporise in the event of any leak and not conduct, but the power input for cooling would probably start being counter-productive quickly.

Pulling a few numbers on the use of a moderate density oil, I figured that in terms of cooling potential over the size of a CPU or graphics chipset die, using oil becomes rapidly less efficient than using air, which is still safer, easier and cheaper.

[kitamesume]

theres lighter oil and when oil is heated up, it becomes less dense or more fluid, try it on a frying pan, pour a light oil, something like light olive oil, it`ll be abit dense but when you heat up the pan, it becomes so fluid that it coats the pan quickly. and oil is quite nice as a substitute if you're paranoid about leaks, alcohol in a fact, is non-conductive and is still a good substitute, heck, why would you be scared of a burn when you know that you dont have any leaks in the first place? its not like alcohol flashes below 80c, who would run a comp higher than 80c to begin with, and theres other substitute tubing than acrylic imho theres plastic and rubber tubings <,<

[AnimeJanai]

You cannot use the common freon refrigerants that consumers can buy since they typically have some petroleum-based lubricant in it for the compressor's moving parts.  It also creates high pressure when hot, so your connectors would spring leaks.  As for alcohol, you would have a potential fire hazard right there.  If a leak occurred, a spark from a switch or relay would set the alcohol on fire and that small fire would follow up and melt the tubing and the rest of the alcohol would be pumped out making a nice huge fire.  You know what, when using a PC coolant system indoors, do a mental check first and assume that it will leak.  If that leak doesn't bother you, then it is an acceptable cooling approach.   As for ammonia, the type used as refrigerant would not be the usual ammonia you used to be able to buy in the stores.  This is anhydrous ammonia which is very dangerous.  But as danger goes, never fear, it is next to impossible for consumers to get strong ammonia now because it is an ingredient used in making many types of drugs (such as methamphetamine).  If a consumer went to buy some from an industrial supplier, their first thought would be that you were going to use it for making drugs.

[kitamesume]

well true but, theres nothing that could spark off from a computer aside from you plugging/unplugging usb all the time, i havent seen relays or any contact switches on a modern computer since when? i forgot, year 2000 maybe? all i see on them are oscillators and transistors which are solid state and doesnt make any sparks. oh and, have you tried igniting rubbing alcohol with just a minor spark toy? i mean those clicky tease sparkers even beer doesnt get ignited with those clicky sparkers exept the drinks that has 70%++

on another thing, what are paint tinners made of? they feel like alcohol when i soak my hands on them =)

[AnimeJanai]

Some power supplies can have relays inside.  At least I have seen a relay in one when looking thru the holes of one PS at Fry's Electronics.  A power surge or temporary black out might cause something to happen.  Anyways, having volatile vapours around any electrical device doesn't seem to be a good idea.  Your 8-inch goliath tarantula pet might wander by and create a static spark as it gets near the metal of the case.

on another thing, what are paint tinners made of? they feel like alcohol when i soak my hands on them =)

It depends what kind of paints you use but turpentine doesn't feel like alcohol to me and it is the most powerful paint cleaning solvent I would liberally use on my hands.  Acetone can be sparingly used on hands.  hexanes and xylene are noxious and you should not breath or touch it for extended periods.
house paint:  water for either stains, porous stain coats, or latex.  You used turpentine then.
cheap model paint:  triple-distilled water  (don't use alcohol in the solvent)
primer coating: acetone or xylene
better model paint:  acetone (not very noxious but still toxic), nhexane (toxic, carcinogenic)
best model paint:  xylene (toxic, carcinogenic)

[kitamesume]

so meaning, paint tinners arent just one substance but a whole bunch of them, i usually soak my hands on tinner when i get paint on them, soap just aint enough to wash paint off >,>

[bork]

Oils in general do not have the ability to absorb heat like water does.  The term is called "specific heat capacity" and is defined as the amount of heat required to raise a unit of mass by one degree.  Water requires about twice the amount of heat as compared to oil to have the same change temperature.  What this means in simple terms, the amount of oil being move through the system will be a little bit more than twice that of water to have the same cooling effect.  Also oil will have a thicker surface boundary causing heat to transfer less efficiently and that will require larger surface areas in the radiator and the CPU block (they get bigger).

In short, you use anything besides water in your system you can assume that you just toasted your CPU.  Just get a decent air cooled heat sink if you have concerns about a water leak.

[AnimeJanai]

Popular Mechanics current issue showed how they installed a PC into one of those rolling Sears Craftsman tool boxes that garage mechanics use.  It used water cooling in order to physically isolate the PC in another chamber to prevent it from getting shorted out by fine metal dust or conductive grit that exists in a home workshop and automotive area.

[SupraGuy]

yeah, I have a PC in my workshop, and it's in a sealed chamber, water cooled.

The motherboard sits inverted, so if any water leaks from the cooling system it drips away from the board. Not the most efficient method since the coolant nearest the CPU is highest, and so gets the hottest coolant next to it. The inlet and outlet for the graphics card is at the "top" of the card, so the bottom in the orientation that the machine is in for the same reason. It's certainly not impossible for this to cause problems in the event of a leak, but it's less likely at least.

I'm using automotive antifreeze/water mix as a coolant, since I have lots of it, and the workshop can get cold in winter (It's in the garage.) The radiator is a heater core from one of my old cars before I sent it to the crusher, with the car's heater fan blowing through it. (It's robust, tolerant of a hostile environment, moves a lot of air, and does so fairly quietly.)

In any event, I would not use oil.

Oh, for a refrigerant, you ADD oil to it for the compressor, but it should only be a very small component of the refrigerant. The leaks are most likely to occur on the "hot side" in any event, which would be well away from the circuit board components (So, between the compressor and the expansion valve) I have the components from an old refrigerator/freezer which I've been thinking of making use of in that regard, but need to ensure that I have compatable blocks and lines (Aluminium most likely) before I can make use of it. The worst things about this are that they're heavy, need inflexible hosing, so the PC won't be easy to move, and the possibility of condensation on the cold side putting moisture where I don't want it. Probably all this adds up to an unuseable system, but I'll probably try it someday anyhow.

[AnimeJanai]

How do you make sure condensation doesn't form if it gets too cold?

[kitamesume]

wrap an absorbent around the tubing and the parts that gets condensations. easiest but not really that safe either, when the absorbent gets too damp it`ll start dripping =), maybe wrap a small heater around the tubing? loool nicrome wires

[AnimeJanai]

The solution then is to set an alarm in the utility you use to monitor CPU or fan temperature to a temp that is close to the point where condensation forms.   With all the parts they sell for water cooling users, maybe there is a temperature measurement device you can hook into the coolant loop.  The hotter it gets, the faster the radiator fan runs.

[zat0x91]

The solution then is to set an alarm in the utility you use to monitor CPU or fan temperature to a temp that is close to the point where condensation forms.   With all the parts they sell for water cooling users, maybe there is a temperature measurement device you can hook into the coolant loop.  The hotter it gets, the faster the radiator fan runs.

There's a few since you can go with one from Koolance or Aquacomputer.  I don't know if the software will actually be able to control the speed of the fans depending on the water temperature though.

[SupraGuy]

How do you make sure condensation doesn't form if it gets too cold?

The worst things about this are that they're heavy, need inflexible hosing, so the PC won't be easy to move, and the possibility of condensation on the cold side putting moisture where I don't want it.

So really, I haven't got that far yet, however if I can seal the refrigerated area, I can probably use a dessicant (To be changed regularly) in order to ensure that the air in the chamber is dry enough to avoid it.

The other thing that can be done is leave the expansion valve right before the cooling block, which makes the line hot right up until it enters the cooling block. It could still condense on the ling after the block, but if the line is insulated and run next to the hot line, then it would probably avoid issues.

Another option is to assume that it WILL drip, and set up catches to route water into a pan (Similar to the way that refrigerators work) which would be over the hot side so the water could re-evaporate.

Anyway, it's probably going to be literally years before I actually do anything with this, so it's mostly academic.