The Perfect Block..

[badbaker`]

Just curious, if money was no option, what would be the perfect block?
I think I read somewhere diamonds have better thermals then copper or other metals used, or that may of been inregards to something else with computers.

[gejmer]

Aquacomputer Cuplex Kryos
The Perfect block and perfect price.

[NKrader]

Aquacomputer Cuplex Kryos
The Perfect block and perfect price.

+1 and i have full nickel hk3.0

[badbaker`]

I'm sure with unlimited resources you would make a better one, Not having to worry about making a profit off it.

I'm just curious as to what is out there, that would be better, but is held back by price and isnt avaliable for the general public.

[Kayin]

There's a good bit more surface area available even in copper, but the production techniques are out of this world.

[frank anderson]

I think the "Perfect Block" would have to be a solid 24K gold plate custom made with a copper top for threading purpose since solid gold is pretty soft.

or if you have so much money you don't know what to do with, then a custom Diamond or Graphite block would work also... and yes, diamond has a higher thermal conduct then copper or aluminium, if I recall correctly, it's even higher then gold..

[badbaker`]

Thats the kind of answer I'm looking for.

So, like a base of diamonds would work well, With a graphite top to thread the fittings into would be the best?

[zanzabar]

I think the "Perfect Block" would have to be a solid 24K gold plate custom made with a copper top for threading purpose since solid gold is pretty soft.

or if you have so much money you don't know what to do with, then a custom Diamond or Graphite block would work also... and yes, diamond has a higher thermal conduct then copper or aluminium, if I recall correctly, it's even higher then gold..

why would u make it out of gold (its worse than copper), silver is much better at conducting heat and even when it corrodes and gets the sulfur layer on it, it is still almost as good and better than copper. if u did make the plate out of gold the pressure from mounting would be enough to deform it and the heat load would also make it really soft..

[defect9]

I think some people assume that electrical conductivity = thermal conductivity

[TJ Tom]

They're not equal but there is a relationship between them.

http://www.physlink.com/education/askexperts/ae432.cfm

[zanzabar]

I think some people assume that electrical conductivity = thermal conductivity

gold isnt the best for electricity ether, its just cheap (when used as 16-18k plating) and corrosion resistant, silver and copper are both better at electricity as well. gold is jsut a little better than aluminum at heat and electricity and is a good chunk worse than copper and copper is just a smidgen worse than silver

[pokazene_maslo]

Talking about silver, I know about one CPU waterblock that's made of 99.9% pure silver.
http://www.cawedog.cz/cs/products/12-ZOUY-Ag-INTEL-AMD

review: http://www.cttl.cz/index.php?gid=10438&kap=0

[Church]

material isn't everything. i'm guessing silver can at most give 2-3C better (but is it worth quadrupling price?), so if block internals themselves are outdated and inefficient silver as material certainly won't fix that.
Top efficiency wise silver kryos might be good block .. but will it still be best, when EK/Koolance/Swiftech recently had released updated block versions with enhanced performance and diminished difference (and at mere 1/4th of price?). And kryos' mounting sys .. "best" should have everything best or at least good. While AC's block shines at some areas it still is far from perfect in others.

[bds71]

diamond is the best thermally conductive material (of those in discussion - graphene/SWCNT are better, but not **really** an option...at all) - but is it possible to make a block out of industrial composite diamond? perhaps???? (i'm not talking about naturally occuring 1000$/gram diamond - i'm referring to the synthetic stuff used by industry for coating metal cutting bits). now would it be possible to coat a copper block/base with this material and have a superior block? (or silver since it is slightly better - assuming you *can* even bond diamond to those materials)..... who knows - it's all theoretical.

FYI:
thermal conductivity (W/mK): (higher is better)
silver - 429
copper - 401
gold - 310
aluminum - 250
diamond - 2200 (pure, natural)

electrical resistivity (resistance, not conductivity) p(Ohm/M): (lower is better)
silver - 1.59 x10-8
copper - 1.68 x10-8
gold - 2.44 x10-8

benefits/detriments:
copper oxidizes in air (effects outward appearance, does not effect bulk (inside) not exposed to air, but can cause loss of conductivity in small guage wires)
silver tarnishes (again mostly only an issue in apprearance; does not effect bulk material, but can cause loss of conductivity in small guage wires)
gold is soft, but does not tarnish/corrode (which is why most electronic components are coated with 24k gold - to prevent loss of conduction is small/fine wires that can oxidize/corrode and lose electrical conductivity)

[3lfk1ng]

I'm curious where Carbon Nanotubes would stand.

[Kayin]

http://www.frostytech.com/articlevie...id=2424&page=1

I've been playing a bit with one of these technologies, but I won't say which. Hopefully I'll have results soonish.

[3lfk1ng]

^oooOOOOoo

[gmat]

The figure given for diamond here is for the monocrystalline form, for the size of a waterblock you cannot get a monocrystal big enough i think, and much less shape it (you would have to chisel it, i guess ?) or there are electrolytic ways of growing diamond crystals of the proper shape maybe... And a polycrystalline block wouldnt have as much heat properties, be horribly expensive for little to no interest over a block of copper. Diamond coating for the waterblock base maybe ? And then the base has to bow, which wont play nice with a layer of hard, non flexible diamond.

[Church]

IIRC industrial artificial diamond making needs insane temperatures alongside insane pressure. No simple & cheap electrolytic ways. And that's just a half of story. After that you need to shape into base of waterblock by using other diamond coated tools. Heck, buying 1K of cpus and selecting best overclockable, or making 24*7 LN cooling probably is cheaper.
Maybe we should keep discussion a bit closer to reality?

[Panther_Seraphi]

I like the idea of a vapourchamber comined with water cooling as IIRC evapouration cools better than water running over something.

[NKrader]

best block = no baseplate. have the top seal straight to the ihs, BAM direct water to ihs no more thinking about best conductor.

[badbaker`]

Doesn't need to be closer to reality. I'm just curious about what the possiblities are for blocks if say, Bill Gates wanted a custom block. Current ones are confined to cost or manufacturing, so what would be posible if we didn't have thoes barricaides.

[badbaker`]

Think outside the box people.

[gmat]

Direct die cooling has been tried, and turned out (predictably) to be not very good due to the very small surface of CPU cores.
If Gates was to get a waterblock he would have one engraved to his initials or gold plated or something in that spirit. Peltiers are usually an expensive but not too bad solution (with dew point temperature control) - but that's already beyond the scope of just a waterblock. Current top waterblocks are pretty much "the perfect waterblock" as far as single phase watercooling is concerned.

[Kayin]

I'd say we're close, but not there yet. Let me see what I can cook up. I know there are some gains still to be had over such as the Supreme HF, but not all are in the block. Transfer fluids, fans, and radiators all have some room too. But I promise there are more companies besides me that are playing with these technologies, and there's still some room for improvement.