IBM developing more efficient way to transfer heat away from the processor

Sunday, October 29 2006

Most of the time, when people see cuts or grooves in their heatsinks or Integrated Heat Spreader (IHS), they think that it will really reduce thermal transfer. However, IBM is working on a technology that actually uses those grooves, and organizes them into a system of canals. By carving these canals into the IHS of a processor, or a heatsink, IBM actually increases the efficiency of thermal paste. The technology is still in prototype stage.

Source: The Register

posted by zekrahminator - 5:38 PM | Related News |

User comments

by v-zero (October 29th - 5:40 PM) - Reply

Oh.

by xylomn (October 29th - 5:42 PM) - Reply

i'll believe it when i see actual results

by Jimmy 2004 (October 29th - 5:57 PM) - Reply

Probably nothing huge I'd expect, but we'll see. IBM have done good stuff in the past (obviously :D).

by spectre440 (October 29th - 6:04 PM) - Reply

by: xylomn
i'll believe it when i see actual results

agreed.

by tigger (October 29th - 6:13 PM) - Reply

thats it,i'm slicing up my ihs.:D

by xylomn (October 29th - 6:14 PM) - Reply

by: tigger69
thats it,i'm slicing up my ihs.:D

shall i start dialing 999 now?

by tigger (October 29th - 6:15 PM) - Reply

nah,it aint got no rights man.

by DanTheBanjoman (October 29th - 6:17 PM) - Reply

by: Jimmy 2004
Probably nothing huge I'd expect, but we'll see. IBM have done good stuff in the past (obviously :D).

They still do, they request and get the most patents per year.

by xylomn (October 29th - 6:17 PM) - Reply

by: tigger69
nah,it aint got no rights man.

i meant for yourself when you put in your pc and blow yourself up :p

by cdawall (October 29th - 7:20 PM) - Reply

thats nice so ibm is making a way to cool down there chips since they happen to be some of the hottest runnning since prescott problem meet solved

by EastCoasthandle (October 29th - 8:29 PM) - Reply

Hmm, i found that a smooth IHS yielded the lowest cpu temps vs. ihs with grooves or unfinished ihs.

by Agility (October 29th - 8:30 PM) - Reply

Wow i'm going to cut my IHS into bits and pieces like those 1MM sizes. Weeeeee.

by pt (October 29th - 8:55 PM) - Reply

time to start cut my IHS :D

*gone look for a chainsaw*

by xylomn (October 29th - 9:01 PM) - Reply

by: pt
time to start cut my IHS :D

*gone look for a chainsaw*

chainsaw :eek:

by pt (October 29th - 9:33 PM) - Reply

by: xylomn
chainsaw :eek:

too powerfull? :rolleyes:

by RaDaR (October 29th - 10:22 PM) - Reply

Make sense to me

by infrared (October 29th - 11:34 PM) - Reply

Copper conducts heat far better than thermal grease, i don't see how cutting groves in the copper is going to help.

by KennyT772 (October 30th - 3:21 AM) - Reply

i think it is to reduce the amount of thermal paste between the highpoints on the ihs and the cooler. that way if too much tim is used it will be squished into the channels and wont interfear with the heat transfer.

by kakazza (October 30th - 7:28 AM) - Reply

More area = More heat can be transfered. The same way heaters work, just that his is for cooling ;p

So if there is more area on the IHS it can transport more heat to the copper cooler. And yes I think this will work. Not like -20°C but it will give some advantage.
Why? a) IBM scientists better know what they are doing, b) the 'more area = more heat dispense' has been used for decades anywhere where heat dispension is important.

by infrared (October 30th - 8:26 AM) - Reply

I think kenny's got it figured. Nice thinking mate ;)

by Slater (October 30th - 10:40 AM) - Reply | Edit

by: xylomn
i'll believe it when i see actual results

Actually when you look at it on a molecular level yes, because they are removing a little bit they increase the surface area, and since it doesnt require a thicker layer of thermal paste, it will increase thermal transfer.

__________________ vs. ~~~~~~~~~~~~~~

Thats how it prob is

My way is right. :)

by olstyle (October 30th - 12:16 PM) - Reply

In some test of an Arctic Cooling CPU-cooler it was said that in its manual Arctic says they use the same method on the bottom of this Cooler.

by Firedomain (October 30th - 1:22 PM) - Reply

well i guess it'd make more surface area between the thermal paste & IHS.....
but wouldn't it also make more air pockets?

We'll see what they manage to do with it!

by Jimmy 2004 (October 30th - 3:05 PM) - Reply

Probably no air pockets if it's applied correctly.

by Alec§taar (October 30th - 7:44 PM) - Reply

by: kakazza
More area = More heat can be transfered. The same way heaters work, just that his is for cooling ;p

So if there is more area on the IHS it can transport more heat to the copper cooler. And yes I think this will work. Not like -20°C but it will give some advantage.
Why? a) IBM scientists better know what they are doing, b) the 'more area = more heat dispense' has been used for decades anywhere where heat dispension is important.

You're "dead on right" & it's the SAME principal heatsinks use really... increase of surface area, via folding the surface & allowing for more of a "heat transferral event horizon".

APK

P.S.=> Personally, I wouldn't expect 'gigantic/knockout' blow-your-mind levels of better cooling, but it should work... IBM imo @ least, STILL has some of the finest most experienced minds out there in this field... apk

Post your own comment

Username:
Password:
Register Here!