New (?) waterblock block idea...

[Fraggle]

Crikey, I`m a busy boy on the forums today aren`t I :o)

Anyway, I was just looking at my generic tower-type cooler when it occurred to me just what a great waterblock it would make if you ditched the fans and engineered a flow of water over the fins instead - tricky but do-able I`m sure. The surface area would clearly dwarf even the most finely milled traditional block and boost flow to boot.

Thoughts?

[Conumdrum]

Still have to rely on the efficiency of the heatpipes which are not as good as a waterblock.

Still, if you could do it it would be better temps than the basic fans. Water will be able to remove more heat due to the higher density of water.

[Fraggle]

Still have to rely on the efficiency of the heatpipes which are not as good as a waterblock.

This is where it gets me pondering though - obviously we know that the thermal conductivity of a well-designed heatpipe is loads greater than that of copper, so maybe with one of those heat-pipe-directly-against-the-heatspreader-type coolers WOULD push the heatspreader/water delta below that of an ordinary block, as the interface would be a heatpipe rather than a lump of copper (except the negligible thickness of the wall of the heatpipe obviously).

Worth a try at least...

[Fraggle]

Can`t upload a pic - it just says it`s not a valid image file no matter what I do. Tried PNG, JPG, everything. Must be MSPaint or something.

[Toolius]

Hello .. try and use a image hosting site like imgur or something..
Once the image uploads , Copy the link under the Linked BBCode (message boards) option and paste it here.
That should do it

[Fraggle]

There it is.

Thoughts?

[defect9]

The reason heatpiping is more efficient than a chunk of copper with fins is in the efficiency in which it moves the heat away from the heat source and spreads it out over a larger area, to be cooled by air.

The thing is, that's the same thing you're doing with water. moving the heat away from the heat source and then cooling it with air. Water cooling is basically a big-ass heatpipe. The reason it's better at cooling than a heatpipe system is because the radiator allows you to get a much larger surface area than a heatpipe-based cpu cooler. That's why the tiny single-fan radiator systems barely out perform a heatpipe based air system (and sometimes they don't).

Then there's the fact you can adjust the flow of the watercooling system to flow more water to allow for larger heat sources. You can't adjust a heatpipe system so well.

[Fraggle]

Agreed but I`m thinking purely of the IHS to water interface, nothing else.

[Church]

Extra layers/members in heat transfer will only make things worse. Better follow KISS principle and get normal waterblock. Will be simpler, more compact, better looker, will cool better and will be cheaper.

[Martinm210]

It's been tried, doesn't work well at all. Water specific heat value and thermal conductivity properties are drastically different and why waterblock designs are so micro. Heat pipes act like the water flow but not quite as good. You end up with a tower of fins that are all water temperature and the only conductivity occurring near the base. No need for all that surface area under water. The extra material end up being more of a bottleneck and insulator.

HSF use heatpipes to move heat to the air exchange. In WC you use the water. Huge difference between water and air you have to think on two different scales.

[Fraggle]

Fair enough then x

[CrazyNutz]

I second what martinm say's



Converting Air heatsinks to waterblocks have been tried several times back in the day, all with disappointing results.

[Fraggle]

...still plugging away at this one...

1). We know direct-die water doesn`t work because water`s thermal conductivity is too low to grab the heat in time.

2). Therefore, we use a heatspreader made of copper with as much surface area as we can mill. OK.

3). We don`t use use silver or diamond for obvious reasons. OK.

4). Heat-piping has thermal conductivity way beyond even diamond.

=> use heat-piping as the heat-spreader.

So the problem is that that heat-piping is not homogeneous in the way a lump of material is: as stated above, all the action would be at the bottom and the vapour would never reach the top of the pipes.


I will continue thinking about this some more...

[MpG]

Have you taken a look at vapor chambers, often found in GPU heatsinks these days?

And don't forget the spreading effect of micropin arrays in waterblocks. I believe it's already been established that even making the pins taller, hits the point of minimal returns extremely quickly. There may be less improvement possible in that department than we'd like.

[Conumdrum]

Vapor chambers are heat pipes. Just a new marketing name. Works just the same.

[Fraggle]

OK, had an idea - watch this space...