Air Cpu Cooler Recommendations

[damian]

This is indeed an excellent HSF and comes highly recommended, as does the very quiet Noctua fan. If it wasn't up against the Thermalright, then it's a fantastic upgrade over a stock HSF. But it's no better than the Therm - this has been shown in many head to heads, one here too http://www.overclockersclub.com/reviews/noctua_nh_u12p/5.htm

 

So although the Noctua is a great product and a worthy recommendation, it won't be any better. Lapping, case cooling improvement, cable management or water cooling is the only thing that will get your temps below what they are now.

Trying another $60-70 HSF would (IMHO) leave you even more frustrated and a few dollars short!

 

GG

 

are you sure about that? look at the reviews on anandtech and xbitlabs

Edited May 31, 2008 by damian

[IVIYTH0S]

it seems that the "true" seems to be the best heat sink that money can buy but i must know a few things...

how come it's aluminum, wouldn't it be better off copper for cooling?

what size fans can mount up to it?

is the heatsink so tall that you can not run the system unless it is positioned on the case's side?

i have a nice all copper heatsink now, i'm not sure the make since my friend picked it out for me but i'm very interested in that dual fan push/pull system in that picture.

 

In a nutshell is that the best heatsink out pretty much? i may sell this heatsink to a friend along with my old conroe but was wondering whether to just give him the stock intel heatsink or give him my copper and upgrade.

 

Thanks

[FatalMelody]

lol. i have the exact same cooling as you do, but different mobo and cpu (abit p35 pro xe and q6700), and i also got the exact same temp, with no ac my cpu is at 30-35 and on load its about 60.

 

im also afraid to lap my cpu and hs, cuz im afraid ill screw it up, is there any chance i might lower the temp without lapping? i want to overclock my machine but afraid my temp will be more than 60c

[Great_Gig]

are you sure about that? look at the reviews on anandtech and xbitlabs

 

I'm not quite sure what you mean, as you have not clarified your comment or provided a link as an example?

 

I think you may mean this review http://www.anandtech.com/showdoc.aspx?i=3268&p=5 if it is, then there is no marked difference in their test between the two, so spending any more money won't see a significant gain, as I remarked. However, the Noctua was tested in a 'push-pull' configuration as suggested by Rehit, which does see an improvement of 3 degrees C. BUT, what they didn't do was a 'push-pull' test with the TRUE, so it's not a fair comparison as they have moved the goal posts in favour of the Noctua. What they should have provided was a like for like test. At stock Idle the TRUE is just beaten by the ZEROtherm NV120, so their test would to me suggest the ZEROtherm was in fact the best HSF in stock form at Idle, under their test conditions?

 

 

The Xbitlabs review is less conclusive, as there is no direct comparison between the two HSF's http://www.xbitlabs.com/articles/coolers/d...rs_2.html#sect0. So as I said, there is little difference between the two, unless you 'push-pull', but they didn't compare in that configuration, so that review is less informative than Anandtech's.

 

So as we can see from Review article research there is little to choose from between the ZEROtherm, Noctua and TRUE? All I suspect would provide similar results in a 'push-pull' setup, logic dictates it as they are all similar designs. And I have no doubt at the moment, these are going to be the best you can get, as research and development teams labour long and hard to find ways of squeezing another few points of performance from their design. You just make your choice based on compatibility, price and preference.

 

So I stand by my original statement, that little or nothing would be gained by purchasing a replacement HSF. Spending another $50.00+ dollars on a not guaranteed degree or two drop in temperature is false economy to me. There are other cheaper steps you can take, as suggested by others here?

 

GG

Edited May 31, 2008 by Great Gig

[PeterStoba]

under their test conditions?

 

That is the key, it may not be the same for other machines

[Puck]

Idle temps are useless unless you plan on keeping it stock volts and clocks.

 

Any pos heatsink can keep a processor cool at idle - it is once they begin to get saturated with heat where the high end heatsinks shine. Two heatsinks may have the same exact idle temp, and one may be a degree or two cooler at say 3.6ghz since it can hold a higher saturation of heat(like a big solid copper brick for example) - but at higher temps, say 4.5ghz, the other heatsink may then have better temps since it can bleed off the heat quicker(like a copper/alu mix).

 

Just an basic example since it is a whole lot more complicated then that when you start getting into the "why"s, but the moral of the story is that with high heat output chips with high overclocks the TRUE 120 is as good as it gets for air cooling.

 

You may benefit from a lapping and putting a penny between the base and mounting clip...that helped mine seat better and drop a few c's.

[Rehit]

that is exactly what i was trying to point out by posting the image in my last post.

show me any heat sink that will perform like that under 100% load, on all 4 cores, with an overvolted quad chip.

i will put my TRUE setup against all comers.

we could have a battle of the heat sinks..

i get even better temps with my Xeon E3110 at 1.7 volts...

but it does run cooler in general.

[Great_Gig]

That is the key, it may not be the same for other machines

 

Absolutely right, as is Puck, I should have posted the Anandtech Load test as well, so hear it is -

 

 

Again they have got the bias toward the Noctua, as it's the only figure quoted under load, with the HSF in a 'push-pull' configuration. Seems an odd way to review a product by heavily stacking the odds in favour of one manufacturer? Maybe they were afraid to 'push-pull' the TRUE 120 . . .

 

I came across the below vid on YouTube and it seems a bit extreme (and unlucky!), but it's a good example of why you should consider lapping some HS? And again, all HSF are not the same, all conditions are not the same, so temps will vary depending on how you load/OC your system and how effectively you dissipate the extra heat that OC generates. And strike me down for suggesting such a thing, but not all reviews are 100% impartial - you need to take a balanced view as I am sure most do.

 

YouTube - Why bother lapping your heatsink?

 

But bottom line Rokkaholik, is I don't think you need to change your HSF for another one. Which is something we all seem to be in agreement with (I think?)

 

GG

[PeterStoba]

They did test the noctua with one fan however, and it is 1 degree coole than the true.

 

Basiclly, get either of the three and get the same temps basically (1 degree either way), so get whatever is cheapest / fits?

[Hienrich Jager]

Well I have that same processor and under four instances of SETI I get temps of 35C (full load).

 

I am using this cooler:

Zalman

 

Jager

 

..:EDIT:..

Ambient temp is 23C

[Rehit]

you have a QX6800 running at those temps under 100% load?

that is amazing even if it is running at stock speed and voltage...

Intel did not recommend using these chips with air cooling when they were released due to the high heat output.

[Warby35]

are you sure about that? look at the reviews on anandtech and xbitlabs

 

Make your mind up Damian! your the guy who first recomended the Noctua to me, remember?

 

Warby