Noctua NH-D15 G2

[ARF]

I do not know if many of you commenters noticed that the cooler body vibrates and a fin buzzes. I have never seen any cooler with a buzzing fin before.

I compared the vibrations with each one and both fans running and one fan seems to vibrate more. So the fans you are complimenting so much can be faulty.

These can be harmonics.

BTW, they can improve the fins stack by making them from copper, thicker, and with more spaces in-between.

 

[Ruru]

@Ruru
Like two thirds of the price ARE the fans. The heatsink itself is much of muchness with the old NH-D15/S. “Core” version would be grabbing one of those used and slapping the new fans on when they come out this Fall, I would think. Or doing the same with a TR cooler, I guess.

Here the fans are like half the price

 

[fevgatos]

@Ruru
Like two thirds of the price ARE the fans. The heatsink itself is much of muchness with the old NH-D15/S. “Core” version would be grabbing one of those used and slapping the new fans on when they come out this Fall, I would think. Or doing the same with a TR cooler, I guess.

Oh I agree, I wasn't implying that noctua coolers outlast the completion (although they kinda do whne it comes to fans and socket updates) but that essentially a 100 euros extra for a product that is going to last a lifetime isn't such a big consideration.

 

[kapone32]

Noctua have never been about value. This is the best air cooler you can buy and the graphs confirm it. You can't use more than 1 cooler on a CPU, so the amount of coolers you can get for this is somewhat irrelevant. This cooler was never going to be about value, it was about Noctua engineering the very best air cooler, and they achieved that.

You're paying for the engineering that has gone into making this the best air cooler you can buy(R&D isn't cheap). You're also paying for top notch customer support, and if the past is anything to go by, future socket support at no-cost. That being said, this costs a bit too much with how close to the competition it is. I think a more reasonable price would be $100-130, and I think it'll get down to that ~$130 within the year.

But the difference is less than 2 degrees vs Thermalright in every chart in the review. With the rise of core components in price, splurging on a CPU cooler is not as popular as before and cases have multiple Rad support because we have gone to AIOs anyway. Otherwise I would still have my Raven 02.

 

[maxfly]

It's not 2014. The CPU cooling market has changed drastically from the days of only two or three really solid heatsinks. There are more excellent cooler choices than ever now. Hence, more and more people treat cooler purchases like disposables, they no longer reuse them build after build. With so many great cooler choices available now, people regularly buy a matching heatsink or AIO to fit their next new build/aesthetic. More so now than ever with the growing popularity of rgb, glass panels and AIOs (despite what we old fogies try to profess). The days of plain aluminum heatsinks have been coming to a close for many years now.

 

[HTC]

So you presume that Noctua will still exist in the next decades and supply several mounting uprages for their decades old coolers? And do you really believe that those plastic fans will last for decades?

What would they sell if they were only supporting their old coolers?

My current cooler is from 2011 (or 2010??), so "it wasn't made with AM4 in mind".

It was discontinued FOR YEARS when i applied for the AM4 mounting kit AND i had misplaced the invoice, but Noctua STILL sent an AM4 mounting kit FREE OF CHARGE, after doing the necessary steps.


I'm not saying they don't exist, but i've yet to HEAR about Noctua and "bad customer support" in the same sentence, even once.


That said, would i buy NH-D15 G2 @ THIS price? Like i sad earlier in the topic, ABSOLUTELY NOT.

 

[BoggledBeagle]

The days of plain aluminum heatsinks have been coming to a close for many years now.

I believe this open design does not help with performance either. For example modern looking cooler as Deepcool Assassin IV looks 1000% better and probably performs better as well, thanks to its closed design an no air escaping where it should not be.

 

[ARF]

I believe this open design does not help with performance either. For example modern looking cooler as Deepcool Assassin IV looks 1000% better and probably performs better as well, thanks to its closed design an no air escaping where it should not be.

Focused, directed airflow is extremely important. I guess noctua is extremely overhyped, and its coolers' aerodynamic characteristics are poor, with disturbances and turbulences which worsen the temperatures.
Not to mention that extremely aggressive brown colour, which doesn't fit to anything anywhere seen.

 

[Onasi]

I believe this open design does not help with performance either. For example modern looking cooler as Deepcool Assassin IV looks 1000% better and probably performs better as well, thanks to its closed design an no air escaping where it should not be.

Doesn’t seem so.

 

[ARF]

Doesn’t seem so.

The graph doesn't measure the true efficiency, yes, the cooler is big, bulky, and with its extreme raw power it manages to somehow beat the others by a degree...

 

[Onasi]

@ARF
What the hell is “true efficiency” when talking about coolers? What, heat load capacity per gram of weight? Per square cm of surface area? Because yes, in this regard air coolers are inherently limited and suffer from diminishing returns in size. In this regard, single tower 120-140mm coolers would technically probably be “most efficient”.
Assassin IV has the same weight as the D15 G2 though and is barely smaller, so I am not sure what your point is.

 

[fevgatos]

@ARF
What the hell is “true efficiency” when talking about coolers? What, heat load capacity per gram of weight? Per square cm of surface area? Because yes, in this regard air coolers are inherently limited and suffer from diminishing returns in size. In this regard, single tower 120-140mm coolers would technically probably be “most efficient”.
Assassin IV has the same weight as the D15 G2 though and is barely smaller, so I am not sure what your point is.

But assassin looks better therfore it's more efficient. Or something

 

[HTC]

Doesn’t seem so.
View attachment 355683

The problem with this chart is the "noise normalized" bit: some coolers don't show their TRUE performance because they produce A LOT MORE NOISE to do it.

Instead, the review should show the MAX performance of the coolers, the % they had relative to the reviewed cooler (like it currently has) AND the noise they made to reach that performance right next to it (in % of the reviewed cooler).


ALSO, there should be some way for the reader to know IMMEDIATELY if the cooler is an air cooler or not WITHOUT knowing specific cooler names: different shade of color, perhaps?

For example, i look @ ... say ... "Endorfy Fortis 5 Dual Fan" and i don't know if it's an air cooler or not FROM THE NAME ALONE.

My 2 cents.

 

[Han44]

Yes, spend 150 euros and fix it, good . Bubble and that's it, so they didn't check and test it? 150 euros plus silicone and other crap to make it work hehehe .

 

[Onasi]

Instead, the review should show the MAX performance of the coolers, the % they had relative to the reviewed cooler (like it currently has) AND the noise they made to reach that performance right next to it (in % of the reviewed cooler).

There is a reason why reviews went the way of db normalized results - because this actually does exactly what you suggest. It shows the true performance of the whole package (HS and fans) on a specified thermal load while also having an averaging measurement via said normalized db. It actually shows what the cooler can do in reasonable use. Just posting maximum possible performance at max RPM is functionally useless. It will just show what everyone knows already - a fan running at a bajillion RPM can salvage even a mediocre HS/radiator design. The best then becomes “the most surface area with the highest airflow/pressure fan at most RPM” which is of little practical consideration. Nobody sane (or at least not many) are running their coolers full tilt. The question for the end-user is “how cool can this get my CPU while being reasonably quiet”. Which is exactly what noise normalized charts reflect.

 

[HTC]

It actually shows what the cooler can do in reasonable use. Just posting maximum possible performance at max RPM is functionally useless. It will just show what everyone knows already - a fan running at a bajillion RPM can salvage even a mediocre HS/radiator design.

It's supposed to show the highest OUT OF THE BOX performance of the coolers: NOT what the coolers can do AFTER being modified in any way (changed fans, for example).

 

[ARF]

Just posting maximum possible performance at max RPM is functionally useless.

You probably forget overclocking, and some extreme CPU utilisation cases (high ambient temperature, mining sessions, computing, Furmark, stress tests, you name it) during which that is the thing that matters.

@ARF
What the hell is “true efficiency” when talking about coolers? What, heat load capacity per gram of weight? Per square cm of surface area? Because yes, in this regard air coolers are inherently limited and suffer from diminishing returns in size. In this regard, single tower 120-140mm coolers would technically probably be “most efficient”.
Assassin IV has the same weight as the D15 G2 though and is barely smaller, so I am not sure what your point is.

Too much aluminium, too high volume, too few heat pipes. When it is the heat pipes which carry the heat away.
You need a heatsink with less fins, but with more copper heat pipes..

 

[Onasi]

It's supposed to show the highest OUT OF THE BOX performance of the coolers: NOT what the coolers can do AFTER being modified in any way (changed fans, for example).

No fans are being changed. Nothing is modified. I am not sure what you think the test is. The reviews, TPUs among them, lays out the methodology quite clearly and provides all results - noise normalized, peak RPM, noise produced and all that at several set TDP targets. The overall performance summary is made using noise normalized results summed up, which makes sense for accuracy. I am not sure why you are confused.

You probably forget overclocking, and some extreme CPU utilisation cases (high ambient temperature, mining sessions, computing, Furmark, stress tests, you name it) during which that is the thing that matters.

Irrelevant in practice. Again, @crazyeyesreaper has his methodology layed out in this very review and all of the reasonings for it. Just read it.

 

[ARF]

ALSO, there should be some way for the reader to know IMMEDIATELY if the cooler is an air cooler or not WITHOUT knowing specific cooler names: different shade of color, perhaps?

I get what you want to say.
The thing is that it is heat pipes vs. direct liquid cooling, and the performance of the two methods is relatively close.

 

[BoggledBeagle]

Gamers Nexus states in 200W AMD full speed test that DC Ass. IV is better than all these new Noctuas. For almost half the price. Just saying.

 

[ARF]

Maybe they next need to innovate in direction of using freon gas, and direct refrigerator means, Peltier element, and AirJet.

 

[HTC]

No fans are being changed. Nothing is modified. I am not sure what you think the test is. The reviews, TPUs among them, lays out the methodology quite clearly and provides all results - noise normalized, peak RPM, noise produced and all that at several set TDP targets. The overall performance summary is made using noise normalized results summed up, which makes sense for accuracy. I am not sure why you are confused.


Irrelevant in practice. Again, @crazyeyesreaper has his methodology layed out in this very review and all of the reasonings for it. Just read it.

In the performance summary section, i was thinking more along the lines of THREE charts, using these three "variables", and ordering the charts by them:

Instead, the review should show the MAX performance of the coolers, the % they had relative to the reviewed cooler (like it currently has) AND the noise they made to reach that performance right next to it (in % of the reviewed cooler).

The closer a cooler is from being the best IN ALL THREE charts, the better it is. I SERIOUSLY DOUBT any cooler leads IN ALL THREE, but they can be CLOSE to the top in all of them, no?

I get what you want to say.
The thing is that it is heat pipes vs. direct liquid cooling, and the performance of the two methods is relatively close.

How close the performance of air or AIO coolers is, is irrelevant.

Many people simply AVOID any AIOs, while others prefer them.

Having a way to tell if it's an air cooler or an AIO without being familiar with the product (like the example i gave) is a plus, no?

 

[Onasi]

@HTC
Ah, I see. You could probably petition the reviewer for something like that. I don’t think it would be too useful overall, but I suppose no harm. But there is a correlation - the G2 is the best air cooler in noise normalized results AND the best at peak RPM at highest set TDP. So there at least everything lines up nicely.

 

[ARF]

How close the performance of air or AIO coolers is, is irrelevant.

There is no simply "air" cooling. Its name is wrong. It's heat pipes cooling based on boiling and condensation vs. heat transfer from hot to cold object/thermal equilibrium.

Many people simply AVOID any AIOs, while others prefer them.

Because of the risk fear.

Having a way to tell if it's an air cooler or an AIO without being familiar with the product (like the example i gave) is a plus, no?

It is a plus, but physically not possible at this level of heat pipes coolers development.

@HTC
Ah, I see. You could probably petition the reviewer for something like that. I don’t think it would be too useful overall, but I suppose no harm. But there is a correlation - the G2 is the best air cooler in noise normalized results AND the best at peak RPM at highest set TDP. So there at least everything lines up nicely.

Yeah, and the perfromance per dollar graph it also wins, so that the top one is 475% better.

 

[HTC]

@HTC
Ah, I see. You could probably petition the reviewer for something like that. I don’t think it would be too useful overall, but I suppose no harm. But there is a correlation - the G2 is the best air cooler in noise normalized results AND the best at peak RPM at highest set TDP. So there at least everything lines up nicely.

Imagine a cooler leads in two but it's REALLY CRAPPY in the last: would you prefer it over another that ... say ... is 3rd in one, 5th in another, and 4th in the last?

It's an hypothetical, but i think you can see what i'm getting @.


Forgot one IMPORTANT "variable": price. So it's FOUR charts, instead of three.

On this last "variable", it's safe to say NH-D15 G2 would be "far from the top".