Prototype DAN HSLP-48: A powerful sub 50mm heatsink

[readeh]

Btw here are some results of the Black Ridge made on an erlay sample I god month ago from EKL:

L9i @ 2400rpm = 63°C
Black Ridge (default 92mm fan) @ 2400rpm =61°C
Black Ridge + A12x15 (120mm fan under fin package) @1800rpm = 52°C
Black Ridge + A12x15 (120mm fan above fin package) @1800rpm = 50°C

Hmm.. I preordered, but can't really figure out of it's worth the "upgrade" from
LP53 or my Cooljag Falcon II .. Not sure it even is an upgrade.
VLP ram is needed for 120mm fan?

 

[dondan]

Yes you need VLP ram for the 120mm fan under the fin package.

 

[fabio]

The reason for this is quality variation. If you buy 10 times C7 there will be some samples that are better then other. The reason for this is soldering quality and heatpipe performance variation. The variation for heatpipes is ~ 8% for each. So maybe you have a bad copper sample but a good aluminum sample.

I found out about this problem while testing heatsinks. This is also the reason why in some reviews the L9i is better and in other the C7 is better. You will have this on every heatsink. But to be fair the difference on copper vs. alu on heatsink is tiny. Depending on heatsink size also only ~8%. So in most cases you will only compensate quality variation on the heatsinks. So a copper version can work like a perfect manufactured aluminum version.

Conclusion: A review will be only good if they preselct 10 samples of every heatsink a test with the avarage value.




Btw here are some results of the Black Ridge made on an erlay sample I god month ago from EKL:

L9i @ 2400rpm = 63°C
Black Ridge (default 92mm fan) @ 2400rpm =61°C
Black Ridge + A12x15 (120mm fan under fin package) @1800rpm = 52°C
Black Ridge + A12x15 (120mm fan above fin package) @1800rpm = 50°C

Thanks for the info DAN. Could you tell us which CPU/s have you used for these temp test? Are temp under load (I know, dream mode on) or in generic tasks?

Thanks again!

 

[dondan]

The tests was made on an Ryzen 1700 in Prime95 8k.

 

[fabio]

The tests was made on an Ryzen 1700 in Prime95 8k.

AMAZING! Thanks a lot!

 

[Nanook]

The reason for this is quality variation. If you buy 10 times C7 there will be some samples that are better then other. The reason for this is soldering quality and heatpipe performance variation. The variation for heatpipes is ~ 8% for each. So maybe you have a bad copper sample but a good aluminum sample.

I found out about this problem while testing heatsinks. This is also the reason why in some reviews the L9i is better and in other the C7 is better. You will have this on every heatsink. But to be fair the difference on copper vs. alu on heatsink is tiny. Depending on heatsink size also only ~8%. So in most cases you will only compensate quality variation on the heatsinks. So a copper version can work like a perfect manufactured aluminum version.

Conclusion: A review will be only good if they preselct 10 samples of every heatsink a test with the avarage value.




Btw here are some results of the Black Ridge made on an erlay sample I god month ago from EKL:

L9i @ 2400rpm = 63°C
Black Ridge (default 92mm fan) @ 2400rpm =61°C
Black Ridge + A12x15 (120mm fan under fin package) @1800rpm = 52°C
Black Ridge + A12x15 (120mm fan above fin package) @1800rpm = 50°C

It’s amazing that the poor manufacturing tolerances can have this much impact.

 

[Choidebu]

This is not how thermal properties of materials work. It is a common misconception (and an urban myth) though.

When you say , you are saying two contradicting things.

Rapidity of "letting go of heat" is thermal conductivity. In addition to that, association of any thermal property with density is erroneous, as it has nothing to do with it:

The density of diamond (for example) is little more than a third of the density of copper (3.51 vs 8.96 g/cm3), yet the thermal conductivity of natural diamond is about 2200 W/(m·K), which is approximately five times greater than copper. Isotopicallly pure diamond has the highest thermal conductivity of any known solid at room temperature at 3320 W/(m·K). Specific heat of natural diamond is 0.520 joule/gram K, which is closer to copper at 0.385 joule/gram K than aluminium at 0.870 joule/gram K, even though it's density is much closer to aluminium than to copper.

The speed of heat transfer is defined by thermal conductivity. This is, for example, why, when you walk over ceramic, glass, stone or metal, it feels colder than walking over cloth, wood, polymers, etc. Materials with higher thermal conductivity conduct heat from your body faster, which makes them feel colder at the exact same temperature than materials with a lower value.

With the basics of thermal conductivity out of the way, let us address and reintroduce the, often misunderstood, "elephant in the room", which is, of course, specific heat, otherwise known as thermal (or heat) capacity.

Specific heat is the amount of heat per unit mass required to raise the temperature by one degree Celsius. The temperature change is required to take effect without phase change, as heat released or gained during a phase change does not change the temperature.

The misconception is that because aluminium has a higher value than copper (0.870 vs 0.385 joule/gram K) it is better at releasing heat. This, of course is incorrect, as specific heat has no relation to the speed of heat transfer either way (gain or loss). Pure distilled, deionised, demineralised water has the highest specific heat value of the common substances at 4.186 joule/gram K . This is why water is the best liquid for any heat transfer applications where the operating conditions permit.
What this means is that a gram of water can hold the most heat before being saturated and thus elevates its temperature.

All of the above means that substances with higher specific heat value change temperature slower (with mass being equal), however, this is only because they can hold more heat per unit mass.

Fun fact: inhomogeneous systems that do not meet the strict definition of thermodynamic equilibrium, such as stars and galaxies have negative specific heat values. This property can result in negative temperature.

So speaking of specific heat, if I understood that correctly, there should be little difference of Cu and Al in how much heat it can keep before being saturated in same amount of volume/size? Since Al is lighter but have lower specific heat.

And speaking of thermal conductivity, I looked up pure Cu and pure Al, and Cu is almost twice Al. The unit is complicated though - W/(m K) - suggesting that some other factors play into the equation, e.g surface area (for contact) and material thickness (for fins).

To sum up, in sff world when size (volume) is everything (we don't need to worry about tower coolers sagging for example) cooler volume (hence specific heat) is our limiting factor, not material (Al vs Cu). The competition hence relies on engineering the most efficient way of releasing the heat into surrounding air. Or is this just same goal regardless sff or not?

 

[Geran Brown]

Yes you need VLP ram for the 120mm fan under the fin package.

Any recommendations?

 

[Supercluster]

Had to snip the GIF. Proofreading this post many times with it gave me hallucinations.

Very dense post. I felt that my answer lacked something, yet couldn't understand what that was. Thank you for pointing this out and making me whip out the pen and paper.

So speaking of specific heat, if I understood that correctly, there should be little difference of Cu and Al in how much heat it can keep before being saturated in same amount of volume/size? Since Al is lighter but have lower specific heat. /snip

Firstly, aluminium has roughly 2.25 times higher specific heat per unit of mass than copper: 0.870 joule/gram K for Al vs 0.385 joule/gram K for Cu.

Regarding the volumetric specific heat:
If we take as our volume 1 cm3 (1 ml). In 1 ml we have 8.96 grams of copper, which gives us a value of 3.4496 joule Kelvins (8.96 g x 0.385 j/g K). In another 1 ml we have 2.7 grams of aluminium, which gives us 2.349 jK (2.7g x 0.870 j/g K).
This allows us to conclude that copper has approximately 1.47 times higher specific heat than aluminium per unit of volume.

Summary:
In the same volume copper stores 47% more heat than aluminium, however, in the same mass aluminium stores 125% more heat than copper.

/snip And speaking of thermal conductivity, I looked up pure Cu and pure Al, and Cu is almost twice Al. The unit is complicated though - W/(m K) - suggesting that some other factors play into the equation, e.g surface area (for contact) and material thickness (for fins). /snip

Considering real world values for alloys used, yes, the thermal conductivity of copper is twice that of aluminium.

As for the unit of thermal conductivity:
The W/mK (Watt per meter Kelvin) represents the rate of heat transfer in a cubed volume of homogeneous material. This means that a material with a value of 1 W/mK will transfer heat at a rate of 1 Watt per degree of temperature difference between the opposite faces in a 1 meter cube (1m x 1m x 1m) of that material.

/snip To sum up, in sff world when size (volume) is everything (we don't need to worry about tower coolers sagging for example) cooler volume (hence specific heat) is our limiting factor, not material (Al vs Cu). The competition hence relies on engineering the most efficient way of releasing the heat into surrounding air. Or is this just same goal regardless sff or not?

I believe you are mixing specific heat and thermal conductivity here:
Higher value of specific heat is good. It enables more heat to be stored in the heat sinking body before it can be dissipated. The higher the value, the smoother the heat load spikes.

With that said, in the end it's the highest thermal conductivity that, as you said it, is "the most efficient way of releasing the heat into surrounding air".

P.S. Thanks to chyll2 for prompting me to add a "summary" section.

 

[VegetableStu]

man supercluster I should pick up your formatting style O_O

 

[Supercluster]

man supercluster I should pick up your formatting style O_O

Thanks man. I am always researching ways to ease reading on the brain and the eyes.

I remember you, particularly your previous profile picture, back when Bit-Tech forums were the place to be. I was not very active, as in commenting, but man, have I spent thousands of hours combing through and picking up features from the best builds. Some amazing people and works happened there in the past ~15 years, some of them sadly not with us anymore.

 

[eme]

So not even option to try 2066? @dondan any 2066 plates left?

 

[VegetableStu]

I remember you, particularly your previous profile picture, back when Bit-Tech forums were the place to be.

eh? o_o was I on bit-tech? (or did you mean you've read my comments in the cerb thread during the time you frequented bit-tech more?)

sorry if I misunderstood, had to skip coffee today ,_, gonna have it as soon as i hit home

 

[Choidebu]

With that said, in the end it's the highest thermal conductivity that, as you said it, is "the most efficient way of releasing the heat into surrounding air".

That's what I meant, sorry if it wasn't that clear. In short there's only a certain amount of volume we can work with within cpu/gpu keepout zone and case height in sff therefore restricting specific heat factor so making optimal use of thermal conductivity is the priority here.

 

[Supercluster]

eh? o_o was I on bit-tech? (or did you mean you've read my comments in the cerb thread during the time you frequented bit-tech more?)

sorry if I misunderstood, had to skip coffee today ,_, gonna have it as soon as i hit home

I've made a mistake then. I guess it was after all only the picture you shared.

I apologize for derailing the thread.

 

[Hifihedgehog]

Do we have a science thread or subforum? Man, I'd love reading through one here for things like this just as I do the sound science subforum at Head-Fi. This was a fun and informative read for me. Thanks for sharing, @Supercluster!

 

[TudorAdrian]

So, y'all got any of them benchmarks yet? Dying to see some numbers. I own a C7 CU in a Sentry and am eagerly looking forward to see how the Black Ridge performs.

 

[fabio]

First preview!

 

[D90]

First preview!

Can anyone translate into English? I saw the thermal chart in the video but don’t know wjat was used for testing or was stock fan/upgraded fan

 

[fabio]

Can anyone translate into English? I saw the thermal chart in the video but don’t know wjat was used for testing or was stock fan/upgraded fan

They replied to you on the youtbe comments!
http://translate.google.com/transla.../tests/kuehler/alpenfoehn_black_ridge/s05.php

BTW, the test system is:

• AMD R5 2400G
• ASUS Crosshair VI Extreme X370
• 8G DDR4
• Seasonic Fanless 460W
• Windows 10 64 bit
  

Now if the GPU and CPU TDP are dissipated together, the results are good! But I was hoping to see a more extreme setup with a 95W TDP Cpu and a GPU inside a case. Anyway is preordered! Let's see!