Any Watercooling components made from non-metals?
Moderators: NeilBlanchard, Ralf Hutter, sthayashi, Lawrence Lee
Any Watercooling components made from non-metals?
just wondered if gold-plating was the be all and end all of corrosion or if there is a market for say ceramic waterblocks (ie no metal in contact with fluid)
what do you experts think
what do you experts think
Those would come from corrosionating metal. There are very good anti-corrosion fluids on the market (Innovatek Protect for example) which, if used correctly, make corrosion a non-issue. They are cheap, too, so I doubt that there is a market for your suggested materials.where does the water pick up its impurities from? metals, tubes or what?
-
- *Lifetime Patron*
- Posts: 1288
- Joined: Sat Oct 25, 2003 3:21 pm
- Location: 15143, USA
- Contact:
Tap water is full of impurities, distilled water is filtered to leave a negligible amount. With most watercooling components, it seems they leave the factory dirty from the manufacturing process, so if you don't clean them before use, you will intorduce those substances into your loop as well. The radiator is most likely to be the source of substantial contamination to the coolant stream.Myth! wrote:dont know what acetal is....
where does the water pick up its impurities from? metals, tubes or what?
Corrosion will occur when you have a loop with parts that are made from two or more reactive metals (eg aluminum and copper). As soon as water touches any metal, it will obtain some ions and become somewhat conductive iirc.
Like I said, all the good blocks, rads, etc are made from copper, brass, or some nonmetal like acetal (aka Delrin) or plexiglass. With those components in a properly-cleaned loop, a negligible amount of corrosion will occur.
Gold plating like anodizing aluminum requires an even, unscratched surface to remain noncorrosive. That's not likely to happen in wear and tear systems like wc loops.
Anyhow, don't take my word for it, I failed high school chem
Anodizing has nothing to do with gold (maybe that's not what you meant) and it'll still work with scratched parts.ZeroR3D wrote:Gold plating like anodizing aluminum requires an even, unscratched surface to remain noncorrosive.Myth! wrote:dont know what acetal is....
where does the water pick up its impurities from? metals, tubes or what?
When you scratch aluminum, the scratched surface still oxidizes. Anodizing just makes the oxide layer thicker.
i see, so the acetal is the rigid plastic stuff.
The thing i'm still seeing though is...... alot of preparation and adding stuff to keep the metal parts happy, I was just posing the question of what materials would make tap water feasable and no added extras.
are there non-metals that insulate heat as badly as copper? can these be worked easily or made thin enough etc, just wondered if anyone had seen anything to this effect.
Was reading something the other day about CO2 starting to be popular in the brewing industry as the coolant, not sure the loudness of the compressor (if needed) but i think it said it cools at a higher temp than the dew point which means about 15C I believe.
The compressor noise is probably prohibitive for SPCR but its an interesting development
The thing i'm still seeing though is...... alot of preparation and adding stuff to keep the metal parts happy, I was just posing the question of what materials would make tap water feasable and no added extras.
are there non-metals that insulate heat as badly as copper? can these be worked easily or made thin enough etc, just wondered if anyone had seen anything to this effect.
Was reading something the other day about CO2 starting to be popular in the brewing industry as the coolant, not sure the loudness of the compressor (if needed) but i think it said it cools at a higher temp than the dew point which means about 15C I believe.
The compressor noise is probably prohibitive for SPCR but its an interesting development
Unfortunately it's hard to beat the metals.. silver does a bit better than copper, but additional 20% of conductivity does not exactly justify the 250-fold price except for die-hard enthusiasts.Myth! wrote:are there non-metals that insulate heat as badly as copper?
Diamond is a lot better heat conductor than silver, thou, but industrial diamond production is not yet capable of producing it cheaply enough. However, this is likely to change in the future, as diamond is just 100% carbon, ie. when somebody comes up with a cheap way to put the carbon atoms to the correct grid, you are not going to run out of the ingredients. Diamond also is almost impossible to get to react with anything, so there's not much fear of corrosion or anything like that..
Naturally the material researchers can come up with something still better or easier, but it will take quite a while to get them into production.. So for now, gold-plated copper is you best bet.
Distilled water is distilled not filtered (hence the name). Distilling is the process of evaporating the water off then collecting the steam and condensing it back to water. The impurities tend not to evaporate so are left behind.ZeroR3D wrote:Tap water is full of impurities, distilled water is filtered to leave a negligible amount.
It will regardless of the type of water you use. The only type of water that is non-conductive to sufficient levels to not fry a computer is ultra-pure water. And anything that's been near a typical water cooling loop (or hit the months of built up crud on a typical mobo) is nothing like ultra-pure. If nothing else, the copper (or Al) of your water blocks will ionize the water making it conductive in fairly short order.ZeroR3D wrote:One more reason to never use tap water: if you get a leak, your electronics can short due to its conductive nature.
Something that would be an issue with a diamond waterblock is that diamond is relaitvely brittle - you'd probably shatter the waterblock if you dropped it.
I think they are aiming to have diamond replace silicon as a base material for microprocessors. That way you could scale the temps up quite a lot..bobo5195 wrote:flourinert direct die cooling would also work and be cheaper than diamond
Anyway, with the thermal load the current microprocessors put to the cooling system, proper waterblock has it's place. Die surface area is small, and few millimeters of copper will not really affect heat conductance enough to compensate for diminished thermal conductance area. Sure, in certain conditions direct die cooling can work, but it's not a solution for masses..
Admittedly direct die fluorinert is pretty poor but it scales a lot better than normal water cooling as there is less in the way. I believe that cathar said that the storm g7 was doing 137 KW/(cm^2) while I worked out ages back that a direct die fluorinert could do about 50 KW/(cm^2) under similar conditions which is not all that impressive. The reason for this is water is much better in the region under the jet where the heat transfer from the jet in water can be huge; a lot higher than the heat transfer from a block of copper for example.
The reasons why you might want to do direct die in this case are
1) focused cooling of areas of the die. Work our the hot spots and you can tune the block to it. If you were clever and accurate enough you could get much higher overclockerability.
2) Removes thermal interface as much as possible (I don’t think thermal interface was included in the above calculation). Coppers not the problem but thermal paste most definitely is.
3) Scales very well at very high flow (8LPM +). At high flows it will always beet any block out there.
4) Can achieve a lot more than the 50KW per cm^2 by jet pulsing (doubling of performance) or cross flow (jet impingement onto water travelling parallel to the block (heard of tripling of performance).
5) Sub zero cooling would work quite well with cheapish water chiller.
6) Could be cheap as you don’t need much fluid, may also use it as part of a two phase loop with a heat exchanger from fluorinert to water to cut down on the expensive FC77.
Diamond is always going to be expensive and hard to produce.
The reasons why you might want to do direct die in this case are
1) focused cooling of areas of the die. Work our the hot spots and you can tune the block to it. If you were clever and accurate enough you could get much higher overclockerability.
2) Removes thermal interface as much as possible (I don’t think thermal interface was included in the above calculation). Coppers not the problem but thermal paste most definitely is.
3) Scales very well at very high flow (8LPM +). At high flows it will always beet any block out there.
4) Can achieve a lot more than the 50KW per cm^2 by jet pulsing (doubling of performance) or cross flow (jet impingement onto water travelling parallel to the block (heard of tripling of performance).
5) Sub zero cooling would work quite well with cheapish water chiller.
6) Could be cheap as you don’t need much fluid, may also use it as part of a two phase loop with a heat exchanger from fluorinert to water to cut down on the expensive FC77.
Diamond is always going to be expensive and hard to produce.
Not *so* brittle. It's not exceptionally good, but still plenty good. If I googled correctly, fracture toughness values for certain materials are (MPa*m^1/2):Butcher wrote:Something that would be an issue with a diamond waterblock is that diamond is relaitvely brittle - you'd probably shatter the waterblock if you dropped it.
Steel alloy: 50
Aluminum alloy: 36
Natural diamond: 4-5
Delrin: 2.4-3.4
Polymethyl methacrylate (PMMA/acrylic/plexiglass): 1
Glass: 0.83
Concrete: 0.2-1.4
These are the forces needed to *shatter* those materials, usually by some sharp impact like dropping or hitting with hard object. Naturally substances like steel will give in to a lot lower forces. So I'd say that fragility of the diamond is not exactly a problem. While it does not break any records, being more than three times as hard to shatter than concrete and twice as hard as delrin, I do not think the toughness of the diamond will be an issue in cooling.
Price, yes, is high, now. But diamond is 100% carbon, and carbon is one of the most abundant elements on Earth, so it's just up to the scientists to determine how expensive it will be in the end. Some guys are already just few years away from mass-producing cheap artificial diamonds: http://www.wired.com/wired/archive/11.09/diamond.html. Technology is there, they just need to get big enough "seed" grown.
References:
http://en.wikipedia.org/wiki/Fracture_toughness
http://www.lbl.gov/Ritchie/Library/PDF/ ... nd_APL.pdf
http://www.tech.plym.ac.uk/sme/interact ... oblem2.htm
http://plastics.dupont.com/plastics/pdf ... DELDGe.pdf