Heatsink base, polished grooves or smooth?

Cooling Processors quietly

Moderators: NeilBlanchard, Ralf Hutter, sthayashi, Lawrence Lee

Post Reply
EvoFire
Posts: 265
Joined: Tue Dec 21, 2004 10:06 pm
Location: Vancouver, BC
Contact:

Heatsink base, polished grooves or smooth?

Post by EvoFire » Sun Jun 12, 2005 8:52 pm

I dunno if this hase been posted or not, and I feel that I should ask before I do anything to the sink.

The heatsink I have right now has these really fine grooves on the base of the whole heatsink which seem to be very nicely polished. Now everyone had said lap the sink for better temperature, but what about these grooves that seem to be very smooth? Would the grooves not increase surface area and increase heat transfer? Or would a absolutely flat and smooth base be better? Some one please enlighten me.

ATWindsor
Posts: 285
Joined: Mon Apr 18, 2005 11:53 pm

Re: Heatsink base, polished grooves or smooth?

Post by ATWindsor » Sun Jun 12, 2005 9:00 pm

EvoFire wrote:I dunno if this hase been posted or not, and I feel that I should ask before I do anything to the sink.

The heatsink I have right now has these really fine grooves on the base of the whole heatsink which seem to be very nicely polished. Now everyone had said lap the sink for better temperature, but what about these grooves that seem to be very smooth? Would the grooves not increase surface area and increase heat transfer? Or would a absolutely flat and smooth base be better? Some one please enlighten me.
Polishing heatsinks has close to no effect, thats my guss, there will always be micocopic holes all over thats not filled, with polishing you risk not making the surface level, which should be a much bigger problem.

AtW

The Greek
Posts: 18
Joined: Fri Jun 03, 2005 6:51 am
Location: South Florida, USA

Post by The Greek » Sun Jun 12, 2005 9:13 pm

If it is dirty then polishing will make a difference but out of the box and clean, not even slightly noticable.

And I wouldn't modify the base of the heat sink, they design it that way for a reason, it allows more surface area to contact the thermal conpound between the heatsink and CPU.

lenny
Patron of SPCR
Posts: 1642
Joined: Wed May 28, 2003 10:50 am
Location: Somewhere out there

Post by lenny » Sun Jun 12, 2005 9:44 pm

I thought there has been tests conducted with lapping (polishing to make flat) the heatsink bases? I recall reading that it made a difference, but I couldn't help feeling that the author started off wanting to prove that it makes a difference...

I think lapping does make a difference, if done properly. Doing it properly is non trivial. And the difference with a god heat sink is minor. I wouldn't go through all that trouble for 1-2 degrees C (if that) of difference. But if you feel so inclined, google "lapping" and I'm sure you'd come up with a whole bunch of instructions. Just be prepared to trash the heatsink if lapping doesn't go well.

frankgehry
Posts: 1424
Joined: Sun Dec 12, 2004 5:00 am
Location: New York, NY

Post by frankgehry » Sun Jun 12, 2005 10:33 pm

Of course its already been posted. It should be flat but it doesn't have to have a mirror finish. Look at the www.frostytech.com testing procedures.
http://www.frostytech.com/testmethod.cfm

EvoFire
Posts: 265
Joined: Tue Dec 21, 2004 10:06 pm
Location: Vancouver, BC
Contact:

Post by EvoFire » Sun Jun 12, 2005 10:48 pm

Sorry if I am at all offensive... but I'm not getting the answers.

I know about not needing to be a mirror finish... I'm not looking for whether if there is a difference between a mirror finish or not, but rather an absolutely flat base compared to a base with fine grooves that are of a same size.

lenny
Patron of SPCR
Posts: 1642
Joined: Wed May 28, 2003 10:50 am
Location: Somewhere out there

Post by lenny » Sun Jun 12, 2005 11:25 pm

Are you referring to the Arctic Cooling heatsinks? I notice that their base is kinda rough - you can feel it with your nails.

My thoughts on this, not necessarily correct, is a flat surface is better. Reasons:

1. A flat surface will provide maximum contact between the two metallic contacts.

2. Thermal compound is used to bridge the gaps where contact is not perfect.

3. Having grooves that need to be filled with thermal compound degrades heat transfer. You don't get more surface area unless you have a perfectly matching set of grooves on the CPU.

EvoFire
Posts: 265
Joined: Tue Dec 21, 2004 10:06 pm
Location: Vancouver, BC
Contact:

Post by EvoFire » Sun Jun 12, 2005 11:55 pm

Hmm... I don't have an Arctic Cooling. I have a Spire WhisperRockIV.

1 Isn't the CPU ceramic or something?

2 I know

3 Really? I didn't know compound degrades transfer, I though it was just used to fill gaps. I did know as least as possible is the best.

frankgehry
Posts: 1424
Joined: Sun Dec 12, 2004 5:00 am
Location: New York, NY

Post by frankgehry » Sun Jun 12, 2005 11:59 pm

This is the link I meant to give you. http://www.frostytech.com/articleview.c ... 572&page=4

absolutely flat base vs base with uniform grooves?

It should be flat as determined by a machinist square or drafting triangle. Grooves do not help, but are an artifact of the machining process. The TIM can fill small groves and imperfections assuring there are no air gaps if applied correctly. Center contact with the cpu is most important as that's where most of the heat is transferred. You describe well finished fine grooves.
That sounds ok but does that mean like the fine grooves on a thermalright or a swiftech. Nickel plated or copper. The link has a surface comparator where you should be able to find something similar to yours.

EDIT: All beginners want to lap their heatsinks including myself. If its an old heatsink and has lots of scratches, maybe, but generally don't worry about it as long as its flat. If its new and seems warped or damaged have it replaced. Chances are if you lap it you'll only make it worse.

lenny
Patron of SPCR
Posts: 1642
Joined: Wed May 28, 2003 10:50 am
Location: Somewhere out there

Post by lenny » Mon Jun 13, 2005 9:54 am

EvoFire wrote:1 Isn't the CPU ceramic or something?

3 Really? I didn't know compound degrades transfer, I though it was just used to fill gaps. I did know as least as possible is the best.
Ah, I was referring to the metal heat spreader on P4 / AMD64. Bad assumption on my part. CPU is of course silicon ;-) I have no idea what's that material on top of the CPU. You're probably right about it being ceramic.

Still, my reasoning is that direct contact between two surfaces will result in best transfer of heat. Having thermal compound in between is preferable to having air, but not as good as having direct contact. Since it is impossible to have perfect contact between two surfaces, we use thermal compound. But we use as little as necessary. Having grooves that require (relatively) a lot of thermal compound to fill is not optimal.

Rusty075
SPCR Reviewer
Posts: 4000
Joined: Sun Aug 11, 2002 3:26 pm
Location: Phoenix, AZ
Contact:

Post by Rusty075 » Mon Jun 13, 2005 11:07 am

In the grand scheme of things:

Pefectly flat, perfectly smooth HS base meeting perfectly flat, perfectly smooth CPU may be the "ideal" situation. But flat is much more important than smooth. If the base is flat, I wouldn't worry about its smoothness unless it was so bad that you can easily feel the individual grooves with your fingernail.

Unless you're very careful, you're more likely to make things worse instead of better when trying to lap a heatsink, by making the base smooth, but concave or convex.

cpemma
Posts: 351
Joined: Sun May 11, 2003 2:31 pm
Location: S Yorks, OK
Contact:

Post by cpemma » Thu Jun 16, 2005 2:49 pm

My old Taisol sink came with quite a marked grooving pattern from its machining; the sinks had a very good reputation but it did worry me a bit. I lapped it, but as a lab metallurgist I was polishing steel samples every day for microscopic examination, scratch-free at 400x.

If you do decide to lap, get a piece of plate glass to fasten the wet & dry to, keep it wet, start on 240, then 400, and don't bother going finer than 600.

And take it slow & steady, bend the elbow not the wrist. Fast = rocking and curvature.

Likif
Friend of SPCR
Posts: 189
Joined: Tue Nov 26, 2002 11:40 am
Location: Copenhagen, Denmark

Post by Likif » Thu Jun 16, 2005 6:18 pm

for the life of me, I can't understand why no at least semi-respected site has made a thorough test of home vs factory lapping on different heatsinks, and mirror vs rough finish, with scientific methodology and that. One such test would put a lot of ghosts to rest.

Nudge, nudge.

Rusty075
SPCR Reviewer
Posts: 4000
Joined: Sun Aug 11, 2002 3:26 pm
Location: Phoenix, AZ
Contact:

Post by Rusty075 » Thu Jun 16, 2005 6:23 pm

Mostly because the margin for error in such a test is greater than the variablility in the testing itself. How you apply the TIM has a bigger affect than lapping the base.

EvoFire
Posts: 265
Joined: Tue Dec 21, 2004 10:06 pm
Location: Vancouver, BC
Contact:

Post by EvoFire » Thu Jun 16, 2005 11:49 pm

I dunno now, I will have to get a tube of heat transfer material first.... I'm out after redoing the videocard and CPU a month ago

cpemma: that is the simplest lapping method I have seen, will try soon as I have 220, 400, and 600 sheets at home.

Rusty075: that would be too bad... I would have wished for some good comparison... Oh wellz....

Post Reply