Lapping

[Guest Blooz1]

Hey,that's just different ways of saying the same thing, t_ski!

[Redrum]

I couldn't really tell you a temp difference because as soon as I take the cooler from its packing I check the base with a straight edge before I do anything else. I would say truthfully that I have never found one that suited my taste right out of the box. They all needed lapped. And I have tried quite a few before finding the correct cooler for each different CPU and board configuration. The one on this machine right now on a naked 146 Opteron with a low speed panaflo in a 70F room idles at 27C and full load at 38C using sound98 MBM data files. And its the quietest computer I think I have ever assembled.

[Jodym4]

It definitly Helped but I'll get bact to you when my artic Silver 5 Cures. I used 800 Grit then 1000 Grit. I polished it with Braso temps still up at like 73C on Othos was 75C.

 

 

[General Septem]

I didn't think the TIM has to adhere to anything. If there aren't any valleys for it to find its way into (which there always will be anyway), so much the better. The important part is that you want complete surface contact. If that contact is 90% metal-to-metal, that will transfer heat much better than if it were 50% metal-to-metal and the rest was filled in by TIM.

 

The stock Opteron cooler definitely needs a lapping. The machine marks are so deep in that thing, I can see them at three feet with my glasses off.

[Jarman]

Havent read the entire thread but here's my 2 cents:

 

A surface wants to be flat as possible to conduct heat, ok, agreed.

 

So from that sentence we can assume a greater surface area contact increases heat transfer yes?

 

Ok, so, a "scratched" base has more surface area than a "mirror finish" one, so i wouldnt worry about getting a mirror finish, just a flat sink - let AS5 do the rest.

 

A new material for heatpipes and waterblocks based on this sort of principle is going to be released soon.

[radodrill]

So from that sentence we can assume a greater surface area contact increases heat transfer yes?

 

Ok, so, a "scratched" base has more surface area than a "mirror finish" one, so i wouldnt worry about getting a mirror finish, just a flat sink - let AS5 do the rest.

 

Mirror finishes will actually make a difference in heat transfer as the critical factor is Contact{/B] surface area. Having a scratched surface will reduce the direct contact surface area.

Additionally, the more thermal interfaces (i.e. IHS-AS5-HS == 2 thermal interfaces) there are and the thickness of TIM/AS5 will reduce the efficiency of the heat transfer.

 

Moral is: the ideal is a perfectly flat mirror finish on both the IHS and the HS; such that there is only a micron of TIM between the two.

 

The purpose of AS5 is to fill the ridges, resulting in better heat transfer than just having air in those gaps.

[TouGe]

I lapped both my x6800 & Tuniq Tower using 240 grit 3M wet/dry sand paper and some water of course. I'm not a believer of the shinny surface but try to get it as flat as possible. I got the IHS to the point were I had to remove the mounting bracket on the motherboard in order for the HS to make contact with the processor, so I can imagine there is very little copper between the actual CPU and the HS.

 

I just finished lapping this afternoon so the AS5 has not had time to cure yet but my temps are already lower than they were prior to lapping with cured AS5.

 

non-lapped/cured AS5

Idle: 32-33

Load: 48-50

 

lapped/non cured AS5

Idle: 29-31

Load:45-46

 

I've always experienced about a 2 degree difference between non cured and cured AS5 but I would be really suprised to see idle temps drop below 28-29 on air. I'm just looking forward to seeing how my load temps turn out.

[Retrospekt]

Average temp drop is 4-10, as reported by TheInformationator on overclock.net