I feel like I'm overlooking something obvious here. Looking at the Recommended Heatsinks list, I don't see why it doesn't fare nearly as well for the AMD as it does for the P4.
I really like the Zalman, at least by design and style, but it doesn't seem to do as well in lower airflow for the AMD compared to others.
Zyzzyx
- now pondering the SLK-800 w/ 5/7 volt Panaflo
Zalman CPU cooler: AMD vs Intel models
Moderators: NeilBlanchard, Ralf Hutter, sthayashi, Lawrence Lee
-
ChiefWeasel
- Posts: 262
- Joined: Sun Aug 11, 2002 3:26 pm
- Location: Worcester, UK
-
ez2remember
- Friend of SPCR
- Posts: 809
- Joined: Thu Nov 28, 2002 5:07 pm
- Location: London, UK
AMD Processors do generate more heat than Intel rivals. But the tiny contact area of the AMD CPU is it's biggest downfall. Intel CPU have very large contact area with the HS.
Intel also have a better locking mechanism than AMD 3 lugs on either side method. It holds it far tighter than AMD counterparts, which explains the bowling you get on the motherboard with some Intel HS.
Another point is that the CPNS6000 (AMD model) heatsink does not have a copper base, most of it is made out of aluminium with only a small strip of copper on it. Copper conducts more heat than aluminium, silver is even better but not practical because of the cost. I would say another downfall of the Zalman HS (AMD) is that the fins are divided into two sets with the middle gap designed for the retentiion clip to fit between. This gap is probably the hottest but yet the fins don't have any contact with this area.
That's my opnion, I could be proven wrong.
Intel also have a better locking mechanism than AMD 3 lugs on either side method. It holds it far tighter than AMD counterparts, which explains the bowling you get on the motherboard with some Intel HS.
Another point is that the CPNS6000 (AMD model) heatsink does not have a copper base, most of it is made out of aluminium with only a small strip of copper on it. Copper conducts more heat than aluminium, silver is even better but not practical because of the cost. I would say another downfall of the Zalman HS (AMD) is that the fins are divided into two sets with the middle gap designed for the retentiion clip to fit between. This gap is probably the hottest but yet the fins don't have any contact with this area.
That's my opnion, I could be proven wrong.
-
MikeC
- Site Admin
- Posts: 12285
- Joined: Sun Aug 11, 2002 3:26 pm
- Location: Vancouver, BC, Canada
- Contact:
agreed with most of ez2remember's points...
...except that the P4 doesn't REALLY have a larger contact area -- it has a heat spreader that protects the CPU from mech damage and actually adds another 2 layers (aluminum and thermal interface material) between HS and core. This actually decreases the cooling power of the HS!
What Intel also has in the P4 is all kinds of embedded tricks to cool itself down when no being pushed, a good thing. And a thermal senor located in a far corner of the core that reads as much as 15C below the highest point. Maybe not so good. So you and I see the monitored diode reading and shrug -- oh it's only 55C -- when in fact the hottest part of the core could be 70C. Deliberate? Smoke and mirrors? Hard to say, but I wouldn't put it past them. See page 2 of SPCR's Unique Heatsink Testing Methodology
There are 2 version of the 6000 BTW - cu and alcu. The copper version is better, for sure.
Aside from the 6000's big gap down the middle and the questionable clip, the big difference vs the 6500 models (for P4) is size: the latter are twice the mass, at least 25% larger in area.
...except that the P4 doesn't REALLY have a larger contact area -- it has a heat spreader that protects the CPU from mech damage and actually adds another 2 layers (aluminum and thermal interface material) between HS and core. This actually decreases the cooling power of the HS!
What Intel also has in the P4 is all kinds of embedded tricks to cool itself down when no being pushed, a good thing. And a thermal senor located in a far corner of the core that reads as much as 15C below the highest point. Maybe not so good. So you and I see the monitored diode reading and shrug -- oh it's only 55C -- when in fact the hottest part of the core could be 70C. Deliberate? Smoke and mirrors? Hard to say, but I wouldn't put it past them. See page 2 of SPCR's Unique Heatsink Testing Methodology
There are 2 version of the 6000 BTW - cu and alcu. The copper version is better, for sure.
Aside from the 6000's big gap down the middle and the questionable clip, the big difference vs the 6500 models (for P4) is size: the latter are twice the mass, at least 25% larger in area.
-
ez2remember
- Friend of SPCR
- Posts: 809
- Joined: Thu Nov 28, 2002 5:07 pm
- Location: London, UK
Copper, Silver... Diamonds!
Ahh, good ol' heat transfer in action!ez2remember wrote:Copper conducts more heat than aluminium, silver is even better but not practical because of the cost.
While Silver may be marginally better (around 10% or so) than Copper, CVD Diamond is about twice as good as the Silver, and up to 10 times as good as Copper (depending on operating conditions).
Check out this article for a detailed description of the heat transfer abilities of CVD Diamonds.
http://www.ddk.com/PDFs/the%20Fraunhofe ... burg-1.pdf
Also, here's a good article that lists materials and their thermal conductivities in a handy-dandy table.
http://www.p1diamond.com/prop_ref.html
Cheers,
Bryan
-
ez2remember
- Friend of SPCR
- Posts: 809
- Joined: Thu Nov 28, 2002 5:07 pm
- Location: London, UK
Interesting.. But the real question is the cost of CVD diamond. Is it really practical even in small quantities? Such as using it as a inlay or base for a heatsink. I have read numerous reports about aluminium, copper, and silver heat conductuctivity which all gives differenent percentages and figures. But they all end up agreeing silver is the best out of the 3 followed by copper and aluminium respectively.
I mentioned silver is not practical but I should of said not practical in large quantities, but there are HS already starting to use silver instead of copper such as the Silvardo Mountain, although not the best of designs.
Transformation is happening from Aluminium ---> Copper ---> Part Silver (most likely silver inlay in the base of the HS or the base completely made of silver). I would love to see thermalright SLK-800 made of a silver base / silver inlay and the same with pure copper flower cooler for the P4. I just wonder how much difference this will make in real life?
I mentioned silver is not practical but I should of said not practical in large quantities, but there are HS already starting to use silver instead of copper such as the Silvardo Mountain, although not the best of designs.
Transformation is happening from Aluminium ---> Copper ---> Part Silver (most likely silver inlay in the base of the HS or the base completely made of silver). I would love to see thermalright SLK-800 made of a silver base / silver inlay and the same with pure copper flower cooler for the P4. I just wonder how much difference this will make in real life?