Noctua Presents NH-D12L Low-height 120mm Dual Tower CPU Cooler

[Chrispy_]

I'm guessing you are not a big corsair guy? /s

They make some nice PSUs and cases that don't have lights, yet.

 

[jeffj7]

I dont get it, is 13mm really that big of a difference? Thats half an inch, I could see a full inch shorter being a big deal but not this.

yes it can be I had to change my idea's on case becuase the d15 i had wouldnt fit in a lain li o11d. had to get the XL

 

[freeagent]

The D15 and D15S have 6 heatpipes, the 7 heatpipes version with more surface area will come out later this year.

I have a 7 pipe cooler. On older, larger die CPU's it is fantastic. On smaller CPU's such as 7nm, it doesn't help as much as you would think. I say that because I have a Thermalright LGMRT on my 5600X, it is a 7x 6mm pipe cooler. Honestly, what would work better is 8mm pipes. They are fantastic, and provide better cooling than a 6mm pipe can. Hopefully they will copy Thermalright and use bigger pipes, or else they are just wasting their time imo.

 

[dyonoctis]

Nice! Now to wait 2 years until the chromax black version.

 

[konga]

I dont get it, is 13mm really that big of a difference? Thats half an inch, I could see a full inch shorter being a big deal but not this.

Yes. As the press release hints at, the U12A was incompatible with most/many 4U cases, while this cooler will be okay in those. That appears to be the primary intent behind this design as it allows you to use a noctua tower cooler with a whole new category of computer chassis.

edit: This also fits some SFF cases like the NR200 now. You could possibly fit a U12A in one if you were okay with a bulging side panel, but the D15 was completely out of the question. This is a dual-tower cooler that'll fit with room to spare now.

 

[Mistral]

At the price Noctua is asking... this is practically DOA against the Silver Soul 135 which, you guessed it, is 135mm tall and also packs a 120mm fan, not to mention a tag under $60.

At least it's a second valid option for anyone using the 280x case and wanting to go air.

 

[ir_cow]

That whole case setup with brown everything is so ugly. I use Noctua fans and towers, but geez...waaay to much brown.

 

[Chrispy_]

Noctua brown is probably hurting their sales.
I've avoided Noctua coolers on several occasions because someone didn't want "the ugly brown and beige thing".

 

[ARF]

How much heat can it dissipate? The Arctic Freezer 34 eSports Duo is rated up to 210 watts!

 

[Valantar]

How much heat can it dissipate? The Arctic Freezer 34 eSports Duo is rated up to 210 watts!

There's no agreed-upon standard for measuring that, so numbers like that are fundamentally incomparable even when they do exist.

 

[ARF]

There's no agreed-upon standard for measuring that, so numbers like that are fundamentally incomparable even when they do exist.

I am not quite sure that our statement is correct in any way.
You can always put the cooling system under stress at 100% CPU load and measure the wall power consumption, then measure the temperatures and define a safe threshold.

 

[tabascosauz]

At the price Noctua is asking... this is practically DOA against the Silver Soul 135 which, you guessed it, is 135mm tall and also packs a 120mm fan, not to mention a tag under $60.

At least it's a second valid option for anyone using the 280x case and wanting to go air.

DOA ......... the D12L is packing one less heatpipe but also has more height and more heatsink mass in both dimensions. And also as a D9L owner, heatsink mass definitely matters.

Only place to reliably find Silver Soul are places like Taobao and Aliexpress. And at that price the Noctua becomes the steal ($41 shipping anyone?). Noctua coolers I can always find locally on 1st-party Amazon or Newegg, not marked up unless it's a U12A Chromax.

And the SS135 is not some revolutionary messiah. Performance-wise it's just a scaled-up D9L, and didn't end up beating my C14S setup, which is itself barely better than a U12S. Unless you just *have* to have your windowed panel on NR200/NCASE/Cerberus

Regardless, both coolers have the same problems. You basically are stuck with one fan. They don't give you regular square clips (I bruteforced a pair of P12 redux on with stock SS135 clips). And even if you have clips, your second fan either sits on top of the RAM or the VRM heatsink, which prevents the side panel from going on. And if you have the clearance for it not to be a problem, there was no reason to use a D12L/SS135 in the first place......etc

Going back to C14S soon, so might pick up one of these if only for curiosity. But like the SS135, it won't cool my B-die, so who knows if I'll even try.

 

[Valantar]

I am not quite sure that our statement is correct in any way.
You can always put the cooling system under stress at 100% CPU load and measure the wall power consumption, then measure the temperatures and define a safe threshold.

Nope, there are far too many variables in play for that to be feasible. What cpu are you using? That introduces a lot of variance, both architectural and in terms of thermal density/hot spots/thermal dissipation even at the same wattage. What constitutes "100% load"? Is it some kind of AVX workload, or not? Different workloads can have vastly different heat outputs even if they all are "100%". And some are realistic, while others (Prime95, for example) produce a completely unreasonable amount of heat compared to nearly anything else. And is this in a case or on an open test bench? In what ambient temperatures? With what other surrounding fans or other disturbing factors? As I said: there is no standard for this, so numbers are inherently incomparable. The best solution is likely using a dummy heater like what Anandtech uses for their cooler testing, as that eliminates most variables, but it also means the numbers don't apply directly to any cpu.

Edit:spelling

 

[chrcoluk]

No offset model? On their standard height cooler I got the offset one, which without, my M.2 nightmares would have been even worse.

 

[tabascosauz]

No offset model? On their standard height cooler I got the offset one, which without, my M.2 nightmares would have been even worse.

This NF-A12x25r is not the NF-A15. That fan is a rounded 140mm, this one is a rounded 120mm. 120mm towers don't require an offset because they aren't big enough to intrude on the first PCIe slot.

 

[chrcoluk]

This NF-A12x25r is not the NF-A15. That fan is a rounded 140mm, this one is a rounded 120mm. 120mm towers don't require an offset because they aren't big enough to intrude on the first PCIe slot.

Fair enough, on the pic its right up to the top pcie card, but that is the slot above the GPU slot so as you said it has more room naturally.

 

[tabascosauz]

Fair enough, on the pic its right up to the top pcie card, but that is the slot above the GPU slot so as you said it has more room naturally.

idk if that's the perspective, but the higher res photos look fine. Looks about the same as my C14S (which is an offset cooler due to size) on some 1st slot boards. Fan clips might be close though.

On my Impact, the SO-DIMM.2 with both sides of the heatsink is right on the edge of slot 1, and there's still a good amount of wiggle room. Think that's the closest I've seen anything get to slot 1 - everyone stopped making those weird PCH-above-socket ITX boards. IIRC the back of PCIe cards usually don't extend to the edge of the PCIe bracket, even if they have a backplate.

 

[ARF]

Nope, there are far too many variables in play for that to be feasible. What cpu are you using? That introduces a lot of variance, both architectural and in terms of thermal density/hot spots/thermal dissipation even at the same wattage. What constitutes "100% load"? Is it some kind of AVX workload, or not? Different workloads can have vastly different heat outputs even if they all are "100%". And some are realistic, while others (Prime95, for example) produce a completely unreasonable amount of heat compared to nearly anything else. And is this in a case or on an open test bench? In what ambient temperatures? With what other surrounding fans or other disturbing factors? As I said: there is no standard for this, so numbers are inherently incomparable. The best solution is likely using a dummy heater like what Anandtech uses for their cooler testing, as that eliminates most variables, but it also means the numbers don't apply directly to any cpu.

Edit:spelling

How do you prove that any cooler will work, then? You have to state numbers in their specifications. Otherwise, no one will buy them on blind faith.

 

[aktpu]

Bit on a tall side, but this would just barely fit in Nouvolo Borg

 

[ARF]

What constitutes "100% load"?

The worst case scenario when there is highest power consumption measured at the wall..

 

[ThrashZone]

Noctua brown is probably hurting their sales.
I've avoided Noctua coolers on several occasions because someone didn't want "the ugly brown and beige thing".

Hi,
Secondary, most don't like noctua pricing more

I personally don't treat computers like fish tanks although they are water cooled

 

[Valantar]

The worst case scenario when there is highest power consumption measured at the wall..

That might be your definition (and it still leaves A LOT of variables not accounted for when you're talking about what a cooler can "handle"), but again, there is no broadly agreed-upon standard for this. That is simply a fact of the world as it exists today. Arguing for a specific understanding of this does not in any way affect the veracity of this fact.

 

[ARF]

That might be your definition (and it still leaves A LOT of variables not accounted for when you're talking about what a cooler can "handle"), but again, there is no broadly agreed-upon standard for this. That is simply a fact of the world as it exists today. Arguing for a specific understanding of this does not in any way affect the veracity of this fact.

When there are variables, you can use a set of numbers - for example from 100 watts when there is very high heat density (for example a 50 sq. mm die which is heavily overclocked) to 200 watts when the die, the IHS is large and the heat density is lower (for example when the dies are large but conservatively clocked)..

 

[Valantar]

When there are variables, you can use a set of numbers - for example from 100 watts when there is very high heat density (for example a 50 sq. mm die which is heavily overclocked) to 200 watts when the die, the IHS is large and the heat density is lower (for example when the dies are large but conservatively clocked)..

Yes, but... I don't understand why I have to repeat this so many times, but there is no widely accepted standard for this. I've never said a standard couldn't be made - there are several that exist already! - I've said that there isn't one that a sufficient number of companies agree on using. Which makes their numbers incomparable. That is all I've been saying. If cooler maker A says 210W and B says 160W, that tells us nothing at all about their relative performance, because the way these numbers were arrived at were different. You can (typically) compare these numbers within company A's coolers or company B's coolers, but never across companies with any reliability. It doesn't matter that we could make a way for them to be comparable, as in the real world today, they aren't. Period.

 

[claes]

Some resources on TDP:

AMD Ryzen TDP Explained: Deep-Dive on TDP Definitions & What Cooler Manufacturers Think | GamersNexus

stub Thermal Design Power, or TDP, is a term used by AMD and Intel to refer in an extremely broad sense to the rate at which a CPU cooler must dissipate heat from the chip to allow it to perform as advertised. Sort of. Depending on the specific formula and product, this number often ends up a...

www.gamersnexus.net

Noctua’s Standardised Performance Rating (NSPR) and compatibility classification for CPU coolers

Designed in Austria, Noctua's premium cooling components are renowned for their superb quietness, exceptional performance and thoroughgoing quality.

noctua.at

How Processor TDP Ratings Can Be Misleading

If you have ever been in the market for a CPU, there is a solid chance that you might have come across a little rating known as TDP. This is a rating that

appuals.com

Intel's Desktop TDPs No Longer Useful to Predict CPU Power Consumption

Intel's higher-end desktop CPU TDPs no longer communicate anything useful about the CPUs power consumption under load.

www.extremetech.com