• Welcome to Overclockers Forums! Join us to reply in threads, receive reduced ads, and to customize your site experience!

? the best s478 hsf

Overclockers is supported by our readers. When you click a link to make a purchase, we may earn a commission. Learn More.

deeman

Member
Joined
Apr 27, 2003
what is the best hsf for a socket 478 ive looked around and the Zalman CNPS7000A-Cu looks the best any others i should look at?? Also some links would be great Thanks.
 
Last edited:
the zalman is prety nice, and very quiet, but the slk is the "best" so far although some people said the zalman comes pretty close.
 
I own both an SLK-900U and a zalman CNPS7000CU. For me, my SLK-900U had a horribly finished base, and I lapped it. At the same noise levels, the Zalman seems to do about 10C (stability) better than the Slk-900U. With a torando on the Slk-900U for me, It preformed slightly beter than the CNPS 7000CU at a rediculous noise level. I also like the fact all zalmans have a mirror like base :)
 
looking like Zalman is the way to go. I heard the slk 900u was very heavy so really didnt want to go that way unless it was thought as the "best"but i see the Zalman is 773g to the 587g of the slk. but would sooner not have to lap if i dont have to.
 
Last edited:
Yeah, the Zalman is the way to go, but really gotta be careful about the weight. You might have to end up removing it when moving your computer around.

You also have the AlCu option, as its quite a bit lighter, but that sacrifices some performance.
 
According to the tests on the main page, the Zalman 7000Cu is outperformed by several other heatsinks including two ThermalRights (SLK800 and SP94) and a Coolermaster IHC Heatpipe design.
The benefit of the Zalman may be low noise, but any heatsink can be used with more silent fans, whereas the Zalman is pretty much permanently bonded to its proprietary fan that cannot achieve top performance even on the highest setting...

I'd get a ThermalRight sink and use it with a fan that I find tolerable, just my opinion.
 
With a silent fan on the 900 the 7000 would outperform it.
So if you don't care about noise go with the 900 and a tornado.
If you don't like noise and wan't to have good cooling go for the 7000-Cu.
 
the SLK had a 92mm (66cfm) fan.

http://forum.oc-forums.com/vb/showthread.php?s=&threadid=204813&highlight=7000

http://forum.oc-forums.com/vb/showthread.php?s=&threadid=225322&highlight=Zalman+cnps+7000

And in this tread higher up.

AMD_Me said:
I own both an SLK-900U and a zalman CNPS7000CU. For me, my SLK-900U had a horribly finished base, and I lapped it. At the same noise levels, the Zalman seems to do about 10C (stability) better than the Slk-900U. With a torando on the Slk-900U for me, It preformed slightly beter than the CNPS 7000CU at a rediculous noise level. I also like the fact all zalmans have a mirror like base :)

Still searchin for some rewievs that shows the same thing.
 
Last edited:
I will post my temps here (all temps in celsius), recorded using Asus probe on a P4C800 Deluxe In a 23c room with 25c case temp on a P4 2.4C @3585MHz 1.58v. 20Minutes of Prime 95 is load temp. BIOS temp before windows load is Idle.

CNPS7000Cu on High (25dBA) 36Idle
41 Load

CNPS7000Cu on Low: (19dBA) 38 Idle
49-52 Load

Lapped (mirror finish)SLK-900U with Panaflo L1A (22dBA) 42 Idle
Lockup (over 60*c) Load

SLK-900U with Sunon 49 CFM Fan (33dBA) 36Idle
45-46 Load

SLK-900U with Tornado 34 Idle
38-39 load

UPDATE
After Ceramique setting in, the CNPS on high Idles @ 33c
Load @ 37c
UPDATE
 
Last edited:
AMD_Me said:
I will post my temps here (all temps in celsius), recorded using Asus probe on a P4C800 Deluxe In a 23c room with 25c case temp on a P4 2.4C @3585MHz 1.58v. 20Minutes of Prime 95 is load temp. BIOS temp before windows load is Idle.

CNPS7000Cu on High (25dBA) 36Idle
41 Load

CNPS7000Cu on Low: (19dBA) 38 Idle
49-52 Load

Lapped (mirror finish)SLK-900U with Panaflo L1A (22dBA) 42 Idle
Lockup (over 60*c) Load

SLK-900U with Sunon 49 CFM Fan (33dBA) 36Idle
45-46 Load

SLK-900U with Tornado 34 Idle
38-39 load

is this on the same computer? if not many things could be variables like every cpu is slightly different, as is the application of as or whatever thermal compound you use.
 
svenrune said:
the SLK had a 92mm (66cfm) fan.

http://forum.oc-forums.com/vb/showthread.php?s=&threadid=204813&highlight=7000

http://forum.oc-forums.com/vb/showthread.php?s=&threadid=225322&highlight=Zalman+cnps+7000

And in this tread higher up.
Still searchin for some rewievs that shows the same thing.

Ozziegn's pics are down, apparently. In any case I have to say I am not convinced. Here's why:

- there are two user experiences available comparing the two heatsinks. A small sample in itself, but putting that aside, they use rather uncontrolled testing methods. There is a lot of variation hiding in things like ambient temp, load level, heatsink mounting, TIM application etc. All of these are pretty much not addressed (yet?). Results obtained under such circumstances may not represent the true values at all.

- temperature monitoring on mainboards is notoriously inaccurate. Differences that do exist may not get picked up at all, and erratic variation can get introduced as well. Using software like Asus PC Probe is also not advisable; this program has been known to report pure fantasy temps.

- against these user reviews, one of which shows the Zalman *outperforming* the SLK900 on high, there is the Heatsink ranking results on the main page. This roundup of tests does include some measures to control variance. The SLK/ThermalRight heatsinks are shown to perform quite a bit better than the Zalman.

As a sidenote; the copper vs aluminium debate has been discussed many times, and the notion that "alu dissipates heat faster and is thus better than copper" has been shown to be false. Copper *IS* better.

I am not biased towards either product nor manufacturer- I have no products of nor shares in either company. I'm not discrediting the available user experiences, but trying to provide objective, and constructive, comments on them. Good information benefits us all.
 
The SLK-900 also required extensive lapping (200 grit to start).

The Heatsink performance rankings are the highest performance for each HS. Look at comparable noise levels, even from BillA, and you will find, at approximately 25dBA, the Zalman is better by around .07C/w
 
AMD_Me said:
I agree mainboard sensors of the p4 diode are innacurate, but are consistently innacurate on the same mainboard.

That is possible, but the opposite is too, unfortunately. Furthermore, there is no saying what will be detected at all and what not. It remains problematic to measure in a regular user's PC environment.

Regarding the noise vs performance issue: the heatsink ranking does not provide a clear answer. Looking at the DIE simulator results, all that is shown is that the SLK900 scores .29 C/W at 56dBA 3244RPM, the quietest setting tested. The Zalman 7000 Cu scores .32 C/W at 2632RPM. If you compare the 600RPM vs the .03 C/W, it's not clear that either of them is better in this (relatively) low-noise scenario.
 
Back