"Spongy" heat sinks

[grenert]

Got this from Elektor today:
https://www.elektormagazine.com/news/chip-gone-spongy?mc_cid=5ba92a4080&mc_eid=96f2993431

A supposedly "significantly better" heatsink with an open, sponge-like structure.  Sounds fishy (read: like marketing) to me.  I can't see how air would really circulate inside the sponge.

What do you think?

[Pjotr]

Think Elektor's PC clock is a 4 months off....... 

[MagicSmoker]

The general premise behind a metal foam heatsink is that the reduction in thermal conductivity due to lower density is more than offset by the increase in the rate of convection due to higher surface area.

Of course, the improvement in convection declines as the much rougher surface of the metal foam gets clogged with dust over time...

[Zero999]

Hmm it sounds like BS. Generally foams have better insulating properties than solid materials or open air because the air is a very good insulator and cell structure of the foam inhibits convection.

[codeboy2k]

Hmm it sounds like BS. Generally foams have better insulating properties than solid materials or open air because the air is a very good insulator and cell structure of the foam inhibits convection.

Yeah, I'm with hero.  Any closed cells inside the material will act like an insulator, with LESS copper available for conduction to the surface, where the open-air cells need to convect the heat away.  And again, with trapped air in the open pockets, depending on orientation of the heatsink, I would expect this to be less efficient than a simple pin design which will have much better convective air flow.

Someone show me this is better than a pinned / finned heatsink of the same mass.

[TheMG]

These are very small heatsinks passively dissipating small amounts of power. Just look at the thermal resistance values in the article...

Testing of the LPH00xx range with an applied 5W load indicates a thermal resistance value of 17.4°C/W for the LPH0010 (40x40x5mm) heatsink, and with a 2W load the thermal resistance of the LPH0004 (20x20x5mm) product is 35.8°C/W.

For such small heatsinks and low power dissipation, conductivity isn't as important as being able to radiate or convect the heat away.

I think the whole point of it is to dissipate as much heat as possible through emissivity, while minimizing the amount of material (metal) required to do so.

When you take that into consideration, it definitely looks plausible.

Of course, this technology would not work for forced-air heatsinks or large convection-cooled heatsinks dissipating large amounts of power. That's where other technologies such as heatpipes come in.

[amyk]

How well do they work when they're covered in dust, and more importantly, how easy to clean?

That looks like a nightmare to clean.

[rs20]

The copper foam is then coated with a thin, hard layer of high temperature copper oxide to improve emissivity.

Translation: "as an unintended side-effect of the manufacture process, it's corroded on the surface."

[Pjotr]

For radiation an electrical non-conductive surface is a lot more efficient than an electrical conductive. So the oxide is a bonus But radiation is only effective when it is into free space, NOT into other parts of the same heat sink. The pores will only help at the outer surface to raise surface area. Such small heat sinks are almost always used with some forced airflow. There will be no air flowing through that thing I guess, even new and completely clean due to the very large flow resistance.  The board itself is a better heat sink.

It looks more a metallurgical development looking for an application.

[rs20]

The pores will only help at the outer surface to raise surface area.

As far as radiation is concerned, this isn't even true. A concave pore may have more surface area, but a corresponding proportion of rays emanating from that surface will hit another part of the same pore. The end result is that you simply can't cheat the system at all when it comes to radiative surface area. Think about it: imagine if this thing were glowing red hot -- would you expect a more porous thing to look brighter? No, black body radiation is diffuse, the intensity of light you see is independent of surface normal.

[Pjotr]

True, it will be no more effective than the surface of the chip's plastic package itself, apart from being a bit larger.

[Simon]

What a load of bollocks, it's as good as saying that the honeycomb material will conduct more electricity than solid wire. The only place such am increased surface will help is on the outer layer of the heatshink to help dissipate, but otherwise the same rules as electricity apply, more cross section of lowest resistance material means more conduction.

[rs20]

What would be good is if it were slightly more open-pored, and you paired it with forced air (and an air filter). The annoyance of an air filter aside (although I kind of wish all computer heatsinks had a filter since a filter is so much easier to clean...), I'd not be surprised if that had better performance than a standard finned heatsink of the same dimensions, once you actually force air past all those surfaces.

[MagicSmoker]

What a load of bollocks, it's as good as saying that the honeycomb material will conduct more electricity than solid wire. The only place such am increased surface will help is on the outer layer of the heatshink to help dissipate, but otherwise the same rules as electricity apply, more cross section of lowest resistance material means more conduction.

You are correct that a foamed copper heatsink will have poorer thermal conductivity than a solid one, but the relative thermal conductivity of those two is irrelevant compared to the ~1000x lower thermal conductivity of the typical plastic IC encapsulation.

Foamed copper heatsinks appear to be intended for passively cooling (ie - by natural convection/radiation) relatively large ICs producing moderate amounts of waste heat - in other words, devices with a relatively low thermal flux (W/m²). For example, the Northbridge IC or maybe even the DRAM modules in a mobo might be good candidates for cooling with a foamed copper heatsink, but not the CPU or GPU.

[Zero999]

The copper foam is then coated with a thin, hard layer of high temperature copper oxide to improve emissivity.

Translation: "as an unintended side-effect of the manufacture process, it's corroded on the surface."

Whether it actually improves emissivity or not depends on whether copper oxide is transparent or opaque to far IR.

[Nerull]

http://www.mouser.com/ProductDetail/Versarien/LPH0001/?qs=%2fha2pyFaduhODbnQtkZaN4Zdm9CdDdH54%2fTYEVeaGNYMlWvFgyTUoQ%3d%3d

For a whopping $2 someone could do some actual tests.

[dom0]

Cooper foam heatsinks aren't a new idea (at least 10 years old) and can work well in certain circumstances (e.g. only filtered air with little to no dust).

[Pjotr]

http://www.mouser.com/ProductDetail/Versarien/LPH0001/?qs=%2fha2pyFaduhODbnQtkZaN4Zdm9CdDdH54%2fTYEVeaGNYMlWvFgyTUoQ%3d%3d

For a whopping $2 someone could do some actual tests.

I wouldn't be surprised if the results are comparable with a piece of black anodised aluminium of the same size. Or even a piece of natural anodised aluminium. In the "datasheet" they compare it to blank aluminium, that makes no sense.