Polishing Blocks/Heatsinks After Lapping, Good or Bad?

[Fuzzy_3D]

Ever since I started lapping my heat-sinks, one of the main steps was the final polishing, which brings the base to a mirror like finish, thus making it extremely smooth, and by extension providing excellent contact to the core.

After hearing on one of the forums that polishing a heat-sink actually decreases its thermal conductivity, I found this article (http://www.overclockers.com/tips458/), supporting that same theory.

Has anyone here got any ideas on the matter?

[CedricFP]

Polishing may decrease the potential surface area for the block.

However, this effect may be minimal as I polished my Maze 4 and did not yield a decrease in temps after re-mounting.

[[XC] DragonOrta]

Cathar has stated previously that the best finish on a Heatsink with hand-lapping is with 600 grit sandpaper. And I don't think that a polish would have the same thermal conductivity as copper or aluminum, which just creates another barrier for heat to travel through. I think that putting Arctic Ceramique or AS5 on a heatsink and rubbing it in like a polish using a plastic bag might help better.

[javascripterror]

Polishing also significantly reduces the flatness obtained by lapping so it defeats all the lapping work done

[Smalltimer]

Ever since I started lapping my heat-sinks, one of the main steps was the final polishing, which brings the base to a mirror like finish, thus making it extremely smooth, and by extension providing excellent contact to the core.

After hearing on one of the forums that polishing a heat-sink actually decreases its thermal conductivity, I found this article (http://www.overclockers.com/tips458/), supporting that same theory.

Has anyone here got any ideas on the matter?
Here is an interesting phenomenon...

If you lap using precision instruments you can make two surfaces so true that the atoms between them will begin to bond / shared with each other (like a magnet) If this were the case, there would be no room for thermal paste or otherwise.

I do not know if this condition takes place between a smaller and a larger surface though. I have witnessed this with two machining blocks ground and lapped, I don't remember the exact no's on the surfacing but if memory serves me right it was in the 10 000th of an inch range. Under conditions the two blocks were inseperable by pulling appart alone, they had to be sheered appart.

[Torin]

Just a question, why is a mirror-like finish a concern on the base-plate of a heatsink that no one will ever see?

From what I've read from reputable sources like Cathar who have the time, resources and expertise to do proper testing, that more often than not, hand-lapping a machine-lapped heatsink does more harm than good. A mirror finish is by no means an indicator of a perfectly flat surface, which is ultimately what you want for superior contact and heat transfer.

But on the topic at hand, yes, a polish can only decrease the flatness and thus the heat transfer capabilities of said surface, as it not only disrupts the flatness of the surface, but it also provides an extra layer of material that heat has to transfer through, at obviously a lesser rate of transfer than you'd get from pure metal.

[XS Janus]

I remember reading somewhere that a 1200 grit finsh is perfect for AS5.
Mirror finish is nothing matter than 1000 IMO because you must use thermal paste anyway, but it does add that feeling of accomplishment :D
________
Lesbian milf (http://www.:banana::banana::banana::banana:tube.com/categories/739/milf/videos/1)

[Fuzzy_3D]

But on the topic at hand, yes, a polish can only decrease the flatness and thus the heat transfer capabilities of said surface, as it not only disrupts the flatness of the surface, but it also provides an extra layer of material that heat has to transfer through, at obviously a lesser rate of transfer than you'd get from pure metal.

Are you suggesting that polish actually creates/leaves a layer of something other than the material you are polishing? My understanding of polish was that it removes the extreme peaks, leaving the actual material smooth.

[[XC] DragonOrta]

Here's (http://www.xtremesystems.org/forums/showthread.php?p=1361080#post1361080) what Cathar has to say about lapping.

[adamwinn]

It all has to do with the molecule size of the thermal compound you are using. There are several formulas and I don't recall them at this time, the only stat that I remember off-hand is that Arctic Ceramique's molecule size responds best to a 1000grit finish.

[Thrilla]

Since I use liquid metal (sometimes ceramique as well) mirrror is what I do :P However I use polish compound with 2000 grit lapping to it's flat, cuz polished doesn't mean flat (my ring is hella shiny btw :P)

Ya I do recall reading it somewhere that 2 really flat and smooth pieces of metal sticks together, that was like 2 years ago in grade 8 science, lol

[Stigma]

Nearly all polishes leave some sort of compound on the surface. Most of them use some kind of wax. Naturally this wax layer acts as an insulator. Unless you polish with abrasives, then you aren't actually making the surface any smoother, compound just fills in the "pore" holes and makes the surface look shiny and reflective instead of rough.

Thus, it is undoubtably best to ONLY SAND the surface and not polish it. If you wanna make it look pretty, feel free to polish every surface except the bottom.

I would assume that the finer grit you used, the better it would be for thermal contact, but I don't have any actual research data on this, and also it would of course depend on the granularity of the thermal interface compound you use. If you polished with a finer grit that the compound, then the particles would fit easily in a meaningful way between the surfaces. that could potentially cause a loss of contact rather than an increase. I'd probably sto lapping in the 1200 to 2000 area, just based on my own intuition that is.

-Stigma