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

[theGryphon]

Hey dan, I just thought of a heatsink design, not sure if you considered... It would give better DIMM clearance (with standard height of 31.25mm) using the 120x15mm fan.

I haven't drawn anything but I'll try to put in words...

Think about an NH-L9i with 6 heatpipes... bend those heatpipe upwards to meet another fin array on top with just enough clearance for a 15mm fan in the middle (16mm clearance should be enough I presume).

Now, DDR4 DIMMs start at 2.4mm from the motherboard when installed. This means, the tip of the DIMM will be 31.25+2.4 = 33.65mm from the motherboard. So, let's shoot for 34mm clearance from the motherboard surface.

Assuming CPU surface starts at 8mm from the motherboard, this gives us 34-8 = 26mm distance from heatsink bottom plate to the fan. Shooting for 48mm height, and using a 15mm fan we have 48-26-15 = 7mm for the thickness of the upper fin array.
I'm going to assume that this upper fin array will have the same area (121x116.5) and density (59 fins) for comparison.

The bottom fin array (including the copper plate) will have 26mm to work with. Giving 1mm for fan clearance, let's say 25mm (which is still 2mm taller than NH-L9i). Considering heatpipe bend allowance, the bottom fin array should have an area of 95x85mm.
Assuming 3mm for copper plate, we have a 22mm height for the fins. Allowing some area loss for the cut-outs (to clear motherboard components, just like NH-L9i does) I'll assume 20mm (on average) fin height.
Further assuming the same kind of fin density as the top fin array, we would have 48 fins.

Total fin surface area would be: 7 x 121 x 59 + 20 x 95 x 48 = 141,000 (compare that to 114,000 with the current design)

TLDR; with a sandwich design (with a bottom fin array ala NH-L9i and a thin top fin array) it's possible to have more surface area AND much better DIMM clearance.

PS: The heatsink would install just like NH-L9i.

 

[Kmpkt]

Theoretically great idea, but I can pretty much guarantee you that the production cost of such a complicated heatsink would be nuts. The other thing to think about is that the upper 7mm thick array might actually hinder overall performance because it would significantly lower the delta between fin and air temperature in the lower array (kinda like sucking in recycled air).

 

[theGryphon]

Theoretically great idea, but I can pretty much guarantee you that the production cost of such a complicated heatsink would be nuts. The other thing to think about is that the upper 7mm thick array might actually hinder overall performance because it would significantly lower the delta between fin and air temperature in the lower array (kinda like sucking in recycled air).

Definitely fair point about the manufacturing costs but your concern about performance hindrance is not valid.

First of all, there would not be an actual air recycling, assuming the top fin array is at the case surface. Given that, laws of thermodynamics will tell you that it's not valid.

Let me explain. As we all know, the purpose of a heatsink is to do heat transfer, from the CPU surface to the air. Now, it's true that the air that goes in the bottom fin array will be warmer than that going in the top fin array. Simply because the top fin array did its job and transferred some of the heat to the air. The same air will still be cooler than the bottom fin array and its "total" heat carrying capacity will not have changed. The bottom fin array will also do its job and transfer some more heat to the air. By the time the air is leaving the heatsink, it will have off-loaded the fin arrays from as much heat as it can, given the total fin surface area. I'm thinking in terms of heat per cubic meters of air (of course the more air passes in a given time frame, heat transfer per time will increase).

Think of a heatsink (with its fins and pipes and fan) as a box (a closed system). It will ultimately depend on the heat carrying capacity (directly related to its temperature as well as some other factors) and the volume (per time) of the air coming in, rather than how the fins are arranged (assuming the same fin surface area).

On the other hand, I can argue that a thicker fin array on top of the fan creates a much stronger air resistance which decreases the airflow (per time) and thus heat transfer per time.

Edit: Another way to look at it is this: Consider the air flowing through Dan's current design of fins, which is 16mm. After the 7th mm, the air traveling is warmer than the air entering the fins, simply because the first 7mm of fins transferred some heat to the air. I hope you see that the same (that warmer air hits part of the heatsink) applies here too.

The air-resistance concern I mentioned above is real. That's why dual-tower heatsinks perform best (instead of a huge chunk of a heatsink).

 

[EdZ]

The improved performance from a dual-tower is because of the conductivity of bulk aluminium/copper. As you make a fin longer, the effectiveness of that extra length decreases:

To the point where adding extra length to the fin makes no real measurable change in the effectiveness of the heatsink. Bear in mind that here we are measuring fin length from the contacting heatpipe because adding heapipes allows for dramatically more efficient transport of heat.

With really big fins like in a 120mm tower heatsink, it's more efficient to have two stacks of 'short' fins than one single stack of extra-long fins. With a very small heatsink, you're well below that optimum length: adding a second stack rather than making the fins in the existing stack longer may sound like it will put more fin area in the 'highest efficiency' zone, but to add that extra stack you are taking away a fixed amount of fin area to fit the heatpipes in.

tl;dr for small heatsinks, it's more efficient to use as much volume on fin and as little as you can get away with on heatpipe.

 

[Kmpkt]

What I was getting at is that performance gain likely wouldn't be proportional to the gain in height due to the fact that air would be warmer that it would have otherwise been once it hit the lower 20mm of fin. I don't feel that this is quite the same as simply having taller fins (ie. 27mm) as in that situation you would have not only two pipes of heat transfer area, but also the cross sectional area of the fin at the 20mm point. Overall the split fin array solution would likely be less effective that a solid fin cooler of the same size. Furthermore I would have concerns about how much air the fan could intake with a fan array above and below it than with a fin array on only one side.

 

[bennyGrose]

Just preordered my Dan Case a4 SFX v2 today and couldn't be any more excited! Now I'm just thinking I need a CPU cooler to go along with it... any updates @dondan

 

[dondan]

Update:

Today I got an mail from Lian Li the new samples will be ready to ship on Oct. 26. There is a small delay because of Chinese Holidays. I think the samples will arrive early November.

 

[Fledder]

Fingers crossed. If I am correct my Dan Case a4 SFX v2 will be delivered December / Januari.
Let's hope for the best for the numbers of the 6 pipe version

 

[LeMoNH3aD]

Looking forward to an update, and hoping for an all copper version!

Take my money!

 

[gffermari]

@dondan

Have you thought the solution of a blower style cooler for A4-SFX??

I think blower fans have better CFM than traditional fans, are short enough letting you have a thick cooler plate and most importantly do not recycle the hot air by leading it out of the case.

 

[Kmpkt]

I think the problem with blower fans is the extremely loud noise. It's not uncommon for these to operate in the 40-50 dBa range.

 

[gffermari]

Yes, it would probably increase the sound of the pc, especially in case a cheap blower fan was used, but it could possibly solve the hot air recycing problem of A4-SFX and the hot cpu temperatures, in general.

I thought that, after watching my Asus GTX1080 Turbo (blower) which makes a bass sound, entirely not disturbing even at high rpm.

 

[dondan]

Here is a small update:

The heatsink manufacturer that works together with Lian Li for the HSLP-48 project is the same one that made the be quiet heatsinks.

They made two prototypes one with 6 heatpipes and the other with 5 heatpipes. Lian Li send me pictures and warned me that the quality isn't good and acurate because the manufacturer used electrical wire extrusion to made the samples.

I told them that I can't work with samples that are far away from the final product and that I will drop this project if I can't get better samples.
Heatsink samples are crazy expansive so I was very unhappy with the situation.

So Lian Li had a meeting with the manufacturer and I will get new CNC made samples in the next 10 days.
The quality will be very close to the final product that will be made by stamping.

So I think the new one will be much better. This is very important for the project because I don't have a big budget for it.

I am realistic with this project. I think it will be hard to get the MOQ of 1000 units for the campaign. Also if we can reach 1000 units this project is from a businesd point of view less interesting than a case project, because the marge is not that high and the work around it with developing, testing, manual making and logistics is nearly the same as for a case project. Increaseing the marge isn't possible because the HSLP-48 is already more expansive as competitors. I need to sell 4000 units to be on a level to say this project was worth all the work. Don't get me wrong also with 1000 units I will made profit but not that much. This is the reason why my budget for prototyping is low.

I will have more information in 10 days.

 

[Revenant]

Here is a small update:

The heatsink manufacturer that works together with Lian Li for the HSLP-48 project is the same one that made the be quiet heatsinks.

They made two prototypes one with 6 heatpipes and the other with 5 heatpipes. Lian Li send me pictures and warned me that the quality isn't good and acurate because the manufacturer used electrical wire extrusion to made the samples.

I told them that I can't work with samples that are far away from the final product and that I will drop this project if I can't get better samples.
Heatsink samples are crazy expansive so I was very unhappy with the situation.

So Lian Li had a meeting with the manufacturer and I will get new CNC made samples in the next 10 days.
The quality will be very close to the final product that will be made by stamping.

So I think the new one will be much better. This is very important for the project because I don't have a big budget for it.

I am realistic with this project. I think it will be hard to get the MOQ of 1000 units for the campaign. Also if we can reach 1000 units this project is from a businesd point of view less interesting than a case project, because the marge is not that high and the work around it with developing, testing, manual making and logistics is nearly the same as for a case project. Increaseing the marge isn't possible because the HSLP-48 is already more expansive as competitors. I need to sell 4000 units to be on a level to say this project was worth all the work. Don't get me wrong also with 1000 units I will made profit but not that much. This is the reason why my budget for prototyping is low.

I will have more information in 10 days.

Your cooler is becoming critical to anyone who wants to run a Coffee Lake CPU. The L9i isn't capable of properly cooling the 8700/8700k in the A4-SFX due to intels new method of rating TDP.

 

[loader963]

Thanks for your effort and time and effort in this Dan. Looking forward to seeing your results.

 

[Kmpkt]

Your cooler is becoming critical to anyone who wants to run a Coffee Lake CPU. The L9i isn't capable of properly cooling the 8700/8700k in the A4-SFX due to intels new method of rating TDP.

Not to steal thunder here, but it won't be the only option ; )

 

[DontPeek]

Not to steal thunder here, but it won't be the only option ; )

Awww snap. No details? That's harsh....

 

[Fernando93]

Not to steal thunder here, but it won't be the only option ; )

Post some info please!

 

[Colajelly]

When can I buy this amazing CPU cooler?, Has funding already been shut down?
My Ryzen Sentry is waiting

 

[dondan]

Small update:

Today I got the pictures from Lian Li. The two new prototypes looks very good. I should have it next week.

Here are some pictures:

Retention parts: