Bigger, heavier heatsink not always = lower temps

[Tech Tweaker]

Well, in an effort to lower the noise coming from my backup rig when I have to run it for one reason or another I set about installing an old Zalman CNPS7000 series heatsink I had here that I wasn't using for anything else.

I was hoping that this one would also run cooler at load, because it is larger, heavier, has a bigger fan, and is made entirely from copper, whereas the HSF it is replacing is made from mostly aluminum and has a smaller fan spinning at 5-7k rpm that is annoyingly loud even at it's lowest speed.

Results were not encouraging.

Stock heatsink load temps: 48-52°C, Zalman CNPS7000 load temps: 54-61°C max. Ouch...



Seems I'm still looking for a new HSF, and will have to find something else to do with this one. At least I do still have two or three other candidates here.

What remains are an Arctic Cooling Freezer Pro 64, a Zalman CNPS9500 LED, and either a Thermalright Ultima-90 or an Ultra 120 eXtreme (depending upon which one I feel like and whether or not I can find a way to fit the TRUE into my 775 rig (ran into fitment issues the last time I tried)).

 

[Deltafan909]

Are you sure you seated it correctly...

 

[Janus67]

what thermal paste did you use? and I second Deltafan's question of it it was seated 100% correctly.

 

[Tech Tweaker]

Are you sure you seated it correctly...

Absolutely, with 100% certainty.

I wonder why everyone always asks me if I've seated a heatsink correctly when I report high temps? I have mounted heatsinks dozens, or perhaps hundreds of times, and I do know what I'm doing in that area.

Anyway, I found that the mounting system on this thing is so easy it would be almost impossible to screw up.

what thermal paste did you use? and I second Deltafan's question of it it was seated 100% correctly.

Arctic Cooling MX-2

 

[Deltafan909]

Without knowing your background, I'm sorry. It's a common mistake seen.

 

[muddocktor]

TT, I fully believe those results with that cooler. That cooler was all right back in the day, but is woefully inadequate for today's processors. I used one of them to cool a Merom mobile C2D installed on an Asus P4C800 board using the Asus socket adapter that let you run a Merom on several models of Asus P4 (socket 478) boards. It worked great for cooling an overclocked Merom, but since it doesn't have heatpipes, it's limited in the amount of heat dissipation it can do.

Of the ones you named, I would try either the TRUE or the Ultima 90.

 

[mjw21a]

Personally I can't believe anyone is using a CNPS7000 heatsink. These things are ancient, I was using one about 7-8 years ago on my old Barton core Athlon XP 2500+.

They were a great heatsink in their day, though seriously, any half decent tower style heatsink will beat the living cr@p out of it performance wise, and probably beat it on price too.

You'd have been much better off getting a Cooler Master 212+ or 212 EVO. Their weight isn't bad and their performance is excellent considering their price point. It'll eat that poor old Zalman cooler for breakfast.

AMD hasn't changed their retention mechanism for heatsinks in one seriously long time. Almost any current socket AM2/AM3 cooler will fit the older socket 754/939 mainboards. The upside is that you can use current coolers on older chips. The downside is there are still plenty of ancient coolers floating around for sale which should have been binned years ago.

 

[Robert17]

The specific heat of aluminum is over double that of copper which is why it is more frequently used these days. Aluminum has become cheaper while copper has become more expensive as well. But aluminum is less ductile than copper which is why copper is used for fluid handling. In fact, aluminum is regarded as brittle, not malleable. In thin sheets, such as are what the fins on heat sinks are formed from as well as aluminum foil (alloy with more malleable elements), aluminum behaves well.

Expect heat sinks to become less exotic for PC's as the TDP of various chips goes down. All the "players" that compete for your (and Wall Street's) computing dollar are driving the TCO (total cost of ownership) down with each succesive generation of advance. The first HP mainframe I played with fit in a 3'x3'x6' cabinet in a heavily air conditioned room and was still only "comfortable" to be near. That same computing ability today can be had in a Gameboy or less at a fraction of the heat generation.

 

[mjw21a]

Perhaps, though I'm yet to see the performance of top end heatsinks actually go down, although I think that higher performance could be had with higher density fins + higher airflow fans. The thing is, thats not required. Fins are now packed less densely allowing the use of lower performance fans and we're STILL getting higher performing coolers than those of yesteryear.

 

[Mjolnir]

Arctic cooling freezer pro should do you fine for that CPU. Assuming you just want something that works and is cooler and quieter than stock. I would go the Thermalright if you want to start OC'ing however.

 

[mjw21a]

Arctic cooling freezer pro should do you fine for that CPU. Assuming you just want something that works and is cooler and quieter than stock. I would go the Thermalright if you want to start OC'ing however.

Hmmmm, I'd still go with a Cooler Master Hyper 212+ or 212 EVO myself....

 

[Mjolnir]

Hmmmm, I'd still go with a Cooler Master Hyper 212+ or 212 EVO myself....

In his first post he listed a bunch of coolers he already had laying around.. my choice was from them

 

[mjw21a]

Ah, got ya..... Yes, I've run the Freezer Pro and it is (slightly) better than the CNPS7000, the the difference isn't really that great.

Problem is that the CNPS7000 used to corrode over a year or so. Once it had turned brown or even green, performance just wasn't what it was at the beginning of its lifespan. Unfortunately Zalman used almost pure copper for this heatsink, and that wasn't good in the long run.

 

[Ivy]

Its not only material, its aswell the surface of the cooler. But generally they should own the new pipes, they truly have improved a lot nowadays and dissipate much more effective to the sink aswell.

What kind of stressing program did you use? Load isnt always load, there is different kind of...

 

[Tech Tweaker]

Its not only material, its aswell the surface of the cooler. But generally they should own the new pipes, they truly have improved a lot nowadays and dissipate much more effective to the sink aswell.

What kind of stressing program did you use? Load isnt always load, there is different kind of...

Prime95 Small FFT's and Blend tests.

 

[mjw21a]

There can be a few other factors as well, such as efficiency of the heat pipes, the process used to attach the fins to the heatpipe, how densely the fins are packed, fan performance, flatness of the base etc etc. In the end, reviews comparing sinks is the only true test of performance.