Noctua NH-D12L

[AusWolf]

Wow! A fairly decent tower cooler that fits into my Corsair 280X case!

Only if the fan could be changed / wasn't brown.

 

[..0]

i think where it shine is the performance at 25db, that should be the gold standard testing.

 

[Vario]

This unit doesn't seem to perform well relative to the price. I like the idea of a low height tower cooler for the narrower cases, but it should be performing better to justify its $90 price tag. The product does look good in terms of aesthetics and build quality.

 

[JustBenching]

Definitely would love to see Frost Commander 140 reviewed, I don't think there's a single professional English language review. The price is about 1/3 cheaper than D15 and Dark Rock P4, and if it can match them, shouldn't people know that?

I think the FC140 is the best cooler right now for high wattages. Probably alongside the jonsbo hx7280 and the assassin 3

 

[Mussels]

i think where it shine is the performance at 25db, that should be the gold standard testing.

I like the way you think


This is too pricey for common use, but SFF builds dont have much choice - lower sales means higher prices, with niche coolers

 

[MIRTAZAPINE]

Fuma 2 does not fit the NR200P without pushing on the glass panel. The vented panel isn't the problem, fits even a U12A fine.

edit: seems to be some tolerance differences there. Runs the risk of contacting the glass.

I kind of wish Scythe kept the old Fuma 1 height of 149mm when they updated to Fuma 2. I got an extra Fuma 1 knowing that it was discontinued. Scythe used to make 150mm coolers but they have bum them all to 155mm for current coolers.

 

[AsRock]

Its not a value proposition cooler. It's a niche/special purpose cooler for SFFs. No one is going to buy it because its a good value.

So funny how Noctua could easily have made this Chromax but said nah you're going to buy this anyways.

As you say some are just going to buy it, and you know if a new socket is released they will release a new system to be able to fit it.

And yes they could of but they are a business after all, this does well enough they will release a Chomax version.

 

[crazyeyesreaper]

I like the way you think


This is too pricey for common use, but SFF builds dont have much choice - lower sales means higher prices, with niche coolers

i think where it shine is the performance at 25db, that should be the gold standard testing.

When you say db iam guessing you mean 25dBA

You both realize I test with no side panel at 15 cm / 1/2 ft away right? if a cooler registers say 50 dBA at 15 cm at 30 cm it will register roughly 48 at 60 cm or 2 ft it will drop to 44-45. Add a Side Panel ie tempered glass panel and it drops to 43ish dBA. thats a roughly 7 dBA drop by going from 15 cm to 60 cm open bench to cheapo case. Keep in mind i test at a noise normalized 45 dBA at 15 cm. that 7 dBA drop means at 2ft with a cheapo case a cooler that is 45 dBA in my chart will register roughly 37-39 dBA dBA in reality. That is 2-4 dBA above the ambient noise of an empty home with nothing else on or running.

Thus why there is a linked article with testing methodology. which also has this:

Keep in mind to actually measure down to 25 dBA would likely require a specialized room and a multi thousand dollar dBA meter. Even then it doesn't really make sense.
For example W1zz for GPUs used to do testing with a Bruel & Kjaer 2236 sound-level meter that at the time was around $4000 and to get GPU noise measurements at idle of 26 to 30 dBA he was testing from 100 CM or 3ft. This means the 25 dBA is arbitrary since you set no standard for distance etc.

If your not talking about db level then I guess I wasted my time lol.

 

[Mussels]

I do in fact, get confused with db and dba

 

[AusWolf]

When you say db iam guessing you mean 25dBA

You both realize I test with no side panel at 15 cm / 1/2 ft away right? if a cooler registers say 50 dBA at 15 cm at 30 cm it will register roughly 48 at 60 cm or 2 ft it will drop to 44-45. Add a Side Panel ie tempered glass panel and it drops to 43ish dBA. thats a roughly 7 dBA drop by going from 15 cm to 60 cm open bench to cheapo case. Keep in mind i test at a noise normalized 45 dBA at 15 cm. that 7 dBA drop means at 2ft with a cheapo case a cooler that is 45 dBA in my chart will register roughly 37-39 dBA dBA in reality. That is 2-4 dBA above the ambient noise of an empty home with nothing else on or running.

Thus why there is a linked article with testing methodology. which also has this:

Keep in mind to actually measure down to 25 dBA would likely require a specialized room and a multi thousand dollar dBA meter. Even then it doesn't really make sense.
For example W1zz for GPUs used to do testing with a Bruel & Kjaer 2236 sound-level meter that at the time was around $4000 and to get GPU noise measurements at idle of 26 to 30 dBA he was testing from 100 CM or 3ft. This means the 25 dBA is arbitrary since you set no standard for distance etc.

If your not talking about db level then I guess I wasted my time lol.

I only use noise data as a comparison to other coolers that I know. Pure "db" or "dba" values are useless to me, just like reviewers' opinions are.

For example, reviews that I read / watched about the Evga 2070 Black praised it for being quiet with a relatively basic cooler (without comparison), so I bought one only to discover that it's too noisy for my taste. Now I have an Asus TUF 6500 XT in my PC that achieved one of the lowest values in TPU's noise testing, and it's really that quiet, so I'm happy.

 

[efikkan]

I only use noise data as a comparison to other coolers that I know. Pure "db" or "dba" values are useless to me, just like reviewers' opinions are.

For example, reviews that I read / watched about the Evga 2070 Black praised it for being quiet with a relatively basic cooler (without comparison), so I bought one only to discover that it's too noisy for my taste. Now I have an Asus TUF 6500 XT in my PC that achieved one of the lowest values in TPU's noise testing, and it's really that quiet, so I'm happy.

Yeah, those scales only tells us the relative noise level compared to other coolers.
But still, most reviews are conducted on open benches, which is far different from a closed case, especially if that case is underneath a desk. Coolers also have different noise profiles, which can be either dampened or even amplified when put into a case.

Underneath my desk I have one desktop on either side, both sitting in Fractal Design 7 XL cases. One is an old i7-3930K with a BeQuiet Pure Rock and the other a fairly new Ryzen 9 5900X with a Noctua NH-U14S. Guess which one is the most noticeable? It's clearly the old one with the Be Quiet Pure Rock, both at idle, medium and max load. That Noctua one has a much more pleasant "hum", and isn't audible under low or medium loads, while the other is always audible. My old computer used to sit in a different case (Corsair Carbide 500R), and the Be Quiet cooler was anything but quiet back then. So this shows that both the cooler and the case matters a lot in terms of perceived noise.

 

[F-man4]

Ideal option for the NR200(p)?

No.
U12A with side HDD frame removed is better.

Heatsinks taller than 140mm are CONS but not Pros.
Too many mid-sized ITX cases (10-15L) have 140mm limit which can’t fit 145mm heatsinks.

 

[shaolin95]

and if you are space limited for sure there is better cooler for same or less money with better performance. This cooler should be less than $80

Can you name some that are better and the same height or lower?

 

[AsRock]

Why the reviews on fan percentage and not a fixed RPM ?, i ask as some fan spin faster\slower than the others so in this case it's a 2kRPM and then you have the Fuma which is a 1.5kRPM fan.

Would it not be easier to read if they were set to a actually RPM for noise ?, as there is a greater chance of a higher fan speed generally makes more.

 

[claes]

Think it’s worthwhile to include both data points but percentages make sense for noise normalized. Up to the reader to figure out what value this data has to them. @VSG uses fixed rpm’s though, so maybe it’s on the editors?

 

[crazyeyesreaper]

Think it’s worthwhile to include both data points but percentages make sense for noise normalized. Up to the reader to figure out what value this data has to them. @VSG uses fixed rpm’s though, so maybe it’s on the editors?

Its fairly simple. Every cooler has a different fan every fan has a different RPM curve. By using PWM % at a set 25% 50% 75% and 100% you get essentially a plotted curve of the fans RPM over its entire range along with its noise level. for example 500 / 1000 / 1500 / 2000 RPM at at 25/50/75/100 but you also get noise levels of for instance. 35 / 42 / 49 / 56 so using a few brain cells we can predict that if the fan is 35 dBA at 500 and 42 at 1000 that means at 750 its noise level would also likely fall somewhere in the middle at 38-40 same goes for 1750 RPM if 1500 is 49 and 2000 is 56 a reasonable assumption would be that 1750 would likely split the difference. so 53 +/- 1 dBA.

As for noise normalized testing I go by the overall noise output not RPM directly I have to find the happy medium that results in a 45 dBA reading. In my testing anything below 45 dBA is essentially unobtrusive or nearly silent in a proper chassis. It is what I would consider an acceptable level of noise you can hear it but if your gaming etc it will easily drown out the system no problem. There is also the fact that motherboard to motherboard there is some level of variation. So by going by pure noise output over RPM it eliminate that excess variable. Therefore i am not using %PWM to check 45 dBA noise normalized performance but instead I am checking pump noise / fan noise / etc to find that perfect mix. Typically that is roughly the same RPM but it requires a bit of fine turning between boards. So rather than go by those I go by noise output alone.

That said with current testing parameters in place testing a single cooler takes 19 hours that doesn't include photos / writing and editing. Just an FYI.

 

[claes]

Sorry if my comment felt like a slight, I think your process is good and that percentages are useful, as well as noise normalized. I was only pointing out that VSG tests differently. I appreciate both methods

 

[crazyeyesreaper]

Sorry if my comment felt like a slight, I think your process is good and that percentages are useful, as well as noise normalized. I was only pointing out that VSG tests differently. I appreciate both methods

Nah just pointing out the why I do things the way I do. And I just happen to be an idiot because I meant to quote @AsRock and epically failed at it on my phone.

 

[AsRock]

Nah just pointing out the why I do things the way I do. And I just happen to be an idiot because I meant to quote @AsRock and epically failed at it on my phone.

Hehe shit happens, i was just curious on your reasoning . I can get the data kinda still from what you post just takes a little extra effort is all.

 

[shaolin95]

Can you name some that are better and the same height or lower?

I figured the guy was just spilling nonsense and has nothing to offer in spite of his previous claim. Typical