The case of computer heatsink/cooler missing important specification

[mirage_pt]

I have been wondering for quite a long time as to why PC heatsink manufacturers do not specify their product's thermal performance (thermal resistance), every heatsink built for any purpose that specification is probably one of the most important however it is missing from every manufacturer i have seem except raijintek(however i would like more 1 digit of resolution to better compare their products, but it is still nice)

Another thing than somehow still is not popular in computers are anodized CPU/GPU air heatsinks (H²O radiators all are), for electronics use they are usually black for better performance but even when it is not worth to use it(forced convention) i do still think than it is missing for the people than enjoy the aesthetic part of PC building

Bellow are 2 datasheets of electronics heatsinks and 2 air coolers (asetek didn't even bother to post any specifications, raijintek 240mm is 0.1°C/W)
http://www.cui.com/product/resource/hse-b18x-035h.pdf
https://media.digikey.com/pdf/Data Sheets/Ohmite PDFs/S_Series.pdf
http://www.raijintek.com/images/Themis-DataSheet.pdf (page 4)
http://noctua.at/media/blfa_files/infosheet/noctua_nh_d15_datasheet_en_1.pdf

From the LED heatsink datasheet i can see than the 25mm anodized option was more than enough to cool a cree 3.6W LED even without thermal paste (hardly gets warm to touch after 1 hour use), for most CPU coolers available i am not capable of knowing before doing my own testing/checking reviews how the cooler actually performs and if it is adequate for my use case including passive and active airflow (at various airflow numbers).

To conclude with this wall of text. Would thermal resistance and/or anodized heatsinks be something people would like to have/know?

 

[FreedomEclipse]

Would thermal resistance and/or anodized heatsinks be something people would like to have/know?

Not really. I think it would be more important for thermalpaste though

 

[micropage7]

i just think that coz its not their point, people buy coz of how it can handle the heat but never ask about it. so why you make it on your description

 

[rtwjunkie]

Are we talking an air cooler's TDP? if so, I've been able to find this for most coolers I have used or considered using.

If not, just forget you read this.

 

[lilhasselhoffer]

So...

The short answer is that the average user has no idea about how to do thermodynamic calculations, and even less about what the variables mean. Coolers are sold on TDP, and people somehow get the false conclusion that as long as TDP cooler>TDP processor everything is fine.

If you're about to go onto a tangent, and explain why this is valuable, best of luck. Whenever I do the math people say that I'm...being obtuse is probably the most positive way I've had it put to me.

In the world of electronics you generally have a power envelope and minimum required heat dissipation for a component. In the world of computing, the concept has been muddied by the built-in ability for processors to thermally throttle and overclock themselves. As such, giving thermal conductivity or resistances are generally not deemed useful, and therefore not provided. While it would be excellent to see them, the associated work and testing is not something that companies deem a good return on investment, when they can just slap a TDP in Watts onto a cooler and the consumer just has to make sure one does not exceed the other.


Mathematically, conductivity is the inverse of resistivity. It's necessary to understanding how much heat can be transferred, given a thermal differential. It's largely how TDP is calculated, but without the fudge factors companies build it. If you know conductivity, temperature differential, and output power you can calculate temperature (assuming constant power output). Nobody seems to like this math.

 

[FreedomEclipse]

Are we talking an air cooler's TDP?

Nope.

We're talking about the thermal resistance of the heatsink itself.

@mirage_pt -- if youre still concerned just google the thermal resistance of Copper or Aluminium (or nickel even as some manufacturers do use those to make heatsinks) to give you a general idea of their thermal resistance

 

[EarthDog]

I could care less about the details. All i want to know is how it performs. Thermal resistance clearly plays a role in this but end results can still vary depending on many different factors.

I file this under minutia.

 

[Frick]

In the world of electronics you generally have a power envelope and minimum required heat dissipation for a component. In the world of computing, the concept has been muddied by the built-in ability for processors to thermally throttle and overclock themselves. As such, giving thermal conductivity or resistances are generally not deemed useful, and therefore not provided.

This is the crucial thing. CPU coolers (for instance) are pretty far from the electronic circuits themselves, and as such you don't really need those numbers anyway.

 

[Vya Domus]

It wouldn't hurt.

 

[Sasqui]

Are we talking an air cooler's TDP? if so, I've been able to find this for most coolers I have used or considered using.

If not, just forget you read this.

Good question, since the thermal resistance is MOOT if the total package can't dissipate whatever amount of heat you're trying to move into the surrounding air.

The thermal resistance varies within components of a heatsink:

1. The transfer of heat from the IHS to the heatsink base (thermal goo),
2. The transfer of heat from the base to the cooling tubes
3. Transfer of the heat from the cooling tubes to the fins (this varies a lot with different types of passive heatwick tubes as well plus temperature deltas and ambient temps)
4. Transfer of the heat from the fins to the surrounding air (which varies wildly with fan type and speed)

I wish people knew more about the laws and principles of thermodynamics, there's seems to be so much misinformation.

 

[mirage_pt]

The concept of heatsink TDP is very convoluted because it would vary with the accepted temperature delta so for 1 manufacturer it can mean 35°C and for another 65°C. As a example i will take the CUI 64mm one in natural convention and ideal junction to heatsink, the TDP for 60°C and 90°C(from 25°C ambient) would be 7.8 and 14.5W respectively, which are way too different to actually mean anything if the manufacturers do not show the thermal resistance and the spec TDP temperature raise.

Yesterday i saw another example of the insanity with arctic's i32 and 33 series in the marketing material they put a hyper 212x at 0.131°C/W and their options at 0.1 and 0.106°C/W never to be seen again in another place like the spec sheet and "technical data"(the main reason i started again to question it again, someone mentioned than they actually do give thermal performance), i do believe all manufacturers do proper testing of the heatsinks but for some reason (marketing probably) most do not mention it anywhere, but somehow thermal paste seems to usually have their actual specs listed(only saw thermal grizzly, noctua, arctic and CM)

I am only mentioning the thermal resistance on the heatsink and not the rest (because it would't be only it and the thread is about them and not junction to heatsink thermal resistance), with the huge power consumption of the new high-end CPUs from both manufacturers thermal selection should't be of the type: "it is a cooler, the fan spins, it cools something", and ideally should be "this cooler would't cool the CPU enough for maximum performance so i will see another one(probably more expensive)"

 

[EarthDog]

Minutia... just read reviews on how they perform. A label like this on the box/website can only be misleading... it could hurt.........See sasqui's post.

And for proof, see Mr Genius' post about thermal throttling... LOLOLOLOL

 

[Bill_Bright]

CPU heatsink fan assemblies are just that, assemblies. And it is likely Company A buys the fan from Company B and mounts it on the heatsink from Company C. We are probably lucky just to see the CPUs it supports.

I say that's what review sites are for.

 

[eidairaman1]

Why is this a thing?

Fyi Most coolers within a class (140mm, 120 mm, etc) typically are within 1-5 degrees of each other so it really doesn't matter.

Thermal performance all relies on how cool a room is anyway, only peltier and refrig style coolers it doesn't matter much.

Remember 1 thing these companies are in it to make money, no matter what they do, they can't make everyone happy.

So please write them letters of your concerns.