• Welcome to Overclockers Forums! Join us to reply in threads, receive reduced ads, and to customize your site experience!

Why not a matte finish

Overclockers is supported by our readers. When you click a link to make a purchase, we may earn a commission. Learn More.

Hoot

Inactive Moderator
Joined
Feb 13, 2001
Location
Twin Cities
We all know that the more surface area an HSF has, the more air contact and hence heat transfer. Most HSF fins you see have a shiny or at best (worse?) dull finish on them. Bead blasting metal leaves a rather granular finish. That roughness increases the surface area considerably. Wouldn't that improve heat transfer? Or, does the rougher finish reduce the velocity of the air and hence volume passing over a given spot? Or, does the rough finish cause surface turbulence and a resulting insulative "bubble" at the surface? Where are the Thermodynamic Engineers on this one?

Hoot
 
Just curious? Or are you thinking you might get a few more degrees cooling out of this?
 
the mirror finish is only for teh side contacing the cpu die or gpu or what ever. if the surrace isnt dead flat then theres less contact area between teh 2 surfaces... on the other sides of the heatsink im guessing that might work but yes something to do with airflow might play there.
 
I am not a total newbie to OC and I understand the difference between the core side of the HS baseplate and the top side along with the fins/pins. I was looking for input from knowledgeable people on the subject of aerodynamics. Will a rougher pin/fin surface yield lower temps? I don't know, but my "Engineering Technician" gut feeling tells me there must be a "sweet spot" for surface finish and heat transfer to air moving over that surface.

Postulation

As the surface finish is roughened, it's surface area increases. Increased surface area contributes to heat removal.

Question

At air velocities encountered in HSFs, at what point does a random pattern of roughness, on the surface of the HS fins/pins, adversely effect the flow of air over that surface with regards to removing heat through contact.

Process

I have an FOP-32 laying here that is not in service. It is already drilled for a thermal probe. I have a small sand blasting device that hooks onto my air compressor. I will measure CPU core temperature using my FOP-32 under a controlled setup and then sand blast the fins to a rough, granular finish. Trying to maintain the same controlled setup, I will then measure its performance again and report back to the group.

Purpose of this post

I was hoping to avoid a useless exercise in getting sand in my hair by polling any informed sources before I do this. But then, half the fun of OC'ing is modding and evaluating the result. I will wait awhile for any responses and then set off on this venture later today.

Hoot
 
Hoot - if you remove all the black finish won't that reduce it's effectiveness as it won't be able to "radiate" any heat ! Naw, just kidding. I don't think you'll see any difference, just like having a black heatsink doesn't help when you are force air cooling, the increased radiation contribution by a black surface is miniscule compared to what the fan is doing - I think they are black because that's how they used to be made when they didn't stick a fan on them.

Similarly, sand blasting the surface may well increase heat dissipation due to increased surface area, but it's such a small increase in the total surface area I doubt it will make a measurable difference. I may be totally wrong of course. :)
 
Well, this experiment won't happen. Not because it's a far-fetched idea, but rather a mechanical limitation. I grabbed a piece of aluminum bar stack from one of my junk boxes to test the sandblaster on. If you can direct the sand stream directly at the flat surface, you can obtain quite a nice "pitting" effect. The problem is, heatsink fins are very close to one another, with a small air gap between each one. The sand stream must be directed at an extremely obtuse angle to get down in between those fins and then the effect is not what I am seeking. There might be a way to do this with a strong caustic chemical etch, but it's a lot of trouble for a principle with no supporting background research. Oh well...

Hoot
 
Hoot-

Load the sandblaster up with corn flower. You will get a nice matte finish, not a pitted finish. We use this technique at the shop sometimes.
 
I've used corn meal to polish pistol and rifle brass in a tumbler, but the problem is not the medium. The problem is getting enough surface collision impact to rough it up, to the extent I wanted to, when you can't blast it straight on. I suppose "matte finish" was an understatement of what I was trying to accomplish. I wanted the surface to feel like sandpaper. IE a serious increase in surface area.

Hoot
 
Back