Socket 775 and AM2 heatsink roundup

Mike D. 11 Dec, 2007 at 1:00am ET Article published in Tech

Reference Heatsinks

I decided to throw a couple of baseline measurements into the mix for good measure. I have both an AMD and Intel retail heatsink to use for the testing. Both are very different but are thought by their respective manufacturers to provide adequate cooling to their processors.

AMD Reference

AMD’s retail heatsinks have come a long way. I still recall the small aluminium model I received with my old Athlon XP 1800+. There are actually three different heatsink/fan combinations shipping with AM2 processors today. The one we’ll be focusing on for this roundup is what I like to refer to as the “Flagship” HSF. This is the four-heatpipe model that shipped with higher end dual-core processors like the FX-62. Since we’ll be using an X2 6000+ in this comparison, the lower-end aluminium based heatsinks would not be sufficient. This also gives the competition a higher reference mark for comparison, which is always good.

As you can see, the heatsink really does not look very OEM at all. With an 80mm fan and four copper heatpipes, there are after-market coolers that pale in comparison. The fan is interchangeable with just about any 80mm fan with a 25mm thickness. The heatsink mounts using the standard retention clips. The fin count is quite high for an OEM heatsink but it is very short in stature.

Intel Reference HSF

Intel’s latest reference cooler that bundled with my Q6600 is one of the new 4-pin models. It is an open fan and fin design that is intended for both low noise and cooling of the mainboard components surrounding the socket.

As you can see, there is an interesting placement of the thermal pads on the heatsink. It is spread horizontally in order to cover the area of the heatspreader in contact with the two dual cores dies underneath. Arctic Silver suggests applying their products in this similar way on Quad core processors.

Installation was rough to say the least. You think an OEM like Intel would strive to make mounting as painless as possible. The four-post push mount system requires a lot of force and literally bends the mainboard PCB more than I’m comfortable seeing. In my opinion the mount is too secure — the HSF does not budge and feels like it has become part of the mainboard. The bulge I observed at the rear of the socket proves that this statement is not far from the truth.

The open-fin design allows quite a bit of air to cool the mainboard, and the sink’s footprint would allow it to fit in even the shallowest cases.

« Previous Next page »

Pages: 1 2 3 4 5 6 7 8 9 10 11

Comments

16 Dec 2007 ~ 11:22pm shwaip hmm...the orders of the heatsinks on the graphs change...it's a little confusing. It would be easier to compare if the order was always the same, IMO.
16 Dec 2007 ~ 11:33pm Leonardo Thanks for well-done review. Information I can use. It was easy to read and digest.
17 Dec 2007 ~ 3:51am TheLostSwede Excellent review and ver well written Mike. I actually like the way you had the graphs written. The best one on top.
17 Dec 2007 ~ 6:04am Thelemech This article has perfect timing(in my world ), as it will help me to build my next machine. Great work Mike D and cheers for the editing Thrax!
17 Dec 2007 ~ 10:21am QCH Holy cow... that's one massive roundup. Very good job and I was amazed at the variety of the heatsink designs.

Also... where DID you get that fancy AMD heatsink.
17 Dec 2007 ~ 6:20pm lemonlime Thanks for the kind words all. It was a lot of work but I was pleased with the end result. I've already got two more heatsinks on the way that will be running through the same methodology.

QCH2002 wrote:

Also... where DID you get that fancy AMD heatsink.

Thanks again, Q!

Keebler wrote:

Also, big ups to brudda' Thrax for helping out with some serious editing business to get it marked up for the new CMS.

Thrax: Thanks very much for [strike]fixing my terrible grammar[/strike] tagging everything! really appreciate it.
17 Dec 2007 ~ 6:24pm lemonlime Also, didn't bother including this in the review but for anyone interested:

Ambient readings for all of the tests:
17 Dec 2007 ~ 7:18pm Leonardo Deviation from the mean ambient temperature did not exceed 0.5C. That's very good. I have nowhere in my house where I could accomplish that. I suppose I could in the garage on a sub-zero night if I were to open the garage doors and leave them open. But, I'd have to just leave my hands in my parka's pockets and just look at the computer parts, not actually doing anything with them.
18 Dec 2007 ~ 1:03am Garg That was one incredibly thorough and informative review, LL!

It was especially nice for me, since I stay comfortably behind the curve on hardware, and the SLK-900 was the last heatsink I used (before switching to equally antiquated water cooling systems, that is ).

Kudos! Great photography in tight places, too.
20 Dec 2007 ~ 1:56pm Snarkasm I've been meaning to ask: is the Mugen essentially just a rebrand of the Scythe Infinity, or is there some difference I'm not seeing? That's exactly what my Infinity looks like.
20 Dec 2007 ~ 4:17pm BuddyJ Same heatsink, different fans.
20 Dec 2007 ~ 7:10pm Snarkasm Thank you, sir. I replaced the fan anyway with a pair of higher-CFM ones. Great sink all the same.
20 Dec 2007 ~ 8:04pm Nightwolf Very helpful review!
23 Dec 2007 ~ 12:34am muddocktor Nice review, Mike. Too bad you didn't have an U-120 eXtreme and a regular U-120 to test too though as both of those are stellar performers, if you get them with a decent base. As for the Tuniq, you called it exactly right about the AM2 mount being an afterthought. Tuniq developed the TT120 before AM2 was on the market and the original versions (including mine) didn't have any way to mount to AM2 included at all. They cobbled together the AM2 mount some time after AM2 came to market. That's why the mounting for LGA775 and socket 754/939/940 is soooooo much better. BTW, try that TT120 out on a socket 939 Opty or X2 system and watch how well it works.
29 Dec 2007 ~ 8:39pm Ultra-Nexus Great article, specially the simple, but effective way of testing base flatness!
Thank you!
29 Dec 2007 ~ 9:53pm Leonardo Oh yes, I want to reiterate what Ultra stated about flatness testing. I used that test a week ago when I decided to lap CPUs and heatsinks. It saved me time, indicating which were not flat and which did not need servicing. As it turns out, the only truly flat heatsink base among my Q6600 rigs was the Zalman 9700 NT. BTW, the Zalman's base is nickel coated.