DAN HSLP-48: A powerful sub 50mm heatsink

Sorry no, because I need the screw size of the backplate and the distance from pcb to topside cpu for a pefect contact pressure.
 
Dondan, how did you determine the amount of heatpipes or are you just going with four to be as close to the nexus as possible?
 
Feedback from Noctua. They said more than four will not increase thermal perfomance. For small heatsinks the surface is the problem. Not the count of heatpipes. The LP53 has only two heatpipes and perform nearly the same as the six heatpipe AXP 100.
 
I am still wondering, what version is better performer.
 
Thanks Vlad502, your question is the perfect introduction for my next update:

Some users in the German forum mentioned the same and they pointed me to an older forum comment of an heatsink engineer.

My initial question is: "Are there any high end premium heatsinks with direct touch?"
The answer is NO.

Mostly heatsinks in lower price regions use direct touch. The reason is that direct touch is cheaper in production than full copper block. The manufactures made a good job in marketing direct touch as the Holy Grail in cooling. The performance is not bad but it has some problems.

Here are the problems mentioned by the engineer in the German forum:

1.) It very hard to make smooth bottom plates with direct touch. So the contact between heatspreader and heatsinks will be not good as with copper block.
2.) Direct touch is only able to cool spots. For example on a heatsink with four direct touch heatpipes the middle ones will be directly above the DIE (core). The heatsink isn't able to share heat very good because of its thin layer. So the centered heatpipes will work at 100% load while the outer pipes will get less heat. With a thicker full copper block you share the heat to all pipes.
3.) Between the direct touch pipes are little gabs that are not good for transferring heat. Your CPU temp sensor will give you good temps, but in that zones where the gab is above the core, there will be very hot zones that aren't recognize by the sensor.

Here is picture to visualize the problems:

changes25xqm4.jpg


So I made the decision to use copper block instead of direct touch.
 
Hasn't it been proved that Low-7000 R2 is Better than the Falcon II? At least from the results people has given the Nexus one is the better CPU cooler.
 
What about 4x flattened 8mm heatpipes (lower base) instead 4x 6mm?
If solid copper base is more expensive version. But still contact with 2x8mm for small CPU Die:

To me the table explains that you can dissipate more heat with the 6mm circular pipes. In 20 mm you can dissipate 85.5W with the 3 circular pipes, while the flattened require 22.3 mm to dissipate 78W.
 
To me the table explains that you can dissipate more heat with the 6mm circular pipes. In 20 mm you can dissipate 85.5W with the 3 circular pipes, while the flattened require 22.3 mm to dissipate 78W.
Yeah but notice how its 3 pipes and the other one is 2, and Vlad is suggesting using 4 flattened but I rather just go with the full copper.
 
I have a little project in hands, I'm planning on getting liquid cooling inside a4-sfx.

The goal is to fit it in 47mm of total height.



I've found all the parts on internet!!:

· Rad: 106x81x25


1.jpg

It can be modified to be to be thinner by cutting the metal bezel, but it fits as it is.

https://es.aliexpress.com/store/pro...oling-system-devices/1358964_32263003396.html

There is a silverstone rad that's slimmer (22mm) and bigger that may fit over the ram using a smaller fan:
td03-slim-34right-top.jpg

http://www.silverstonetek.com/product.php?pid=598&area=en


· Fan: 80x80x10mm fan

1.jpg

It fit as it is, but Im planning on cutting all the plastic housing to insert it in the metal bezel of the rad to save some mm for better breathing. Ty100 may fit also modding it.

Cc.porsche.com/icc_euro/ccCall.do?rt=1490866599&screen=1920x1080&userID=SGS&lang=sp&PARAM=parameter_internet_sp&ORDERTYPE=991110&MODELYEAR=2018&hookURL=http%3a%2f%2fwww.porsche.com%2fspain%2fmodelstart%2f



· Cpu Block: 40x40x10mm

1.jpg

I've found this block, it's made out of copper, so it can be used with liquid metal thermal paste. There are a few better 10mm cpu blocks out there, atached to liquid pump kits, but this one is sold separately,

https://es.aliexpress.com/item/For-...ics-GPU-Endothermic-Head-Hot/32687537364.html

----------------------------------

Total height: 25+10+10=45mm

It fits over the cpu




· Pump:


1.jpg

The pump can fit below the cpu behing the power supply's cables.
I'm researching for the better pump, I think that anyone may fit.


-------------------------------------------

Notes:


· Extra milimeters are used for the cpu block mount, so no milimeters left,
· Using a naked cpu may save 1 or 2 mm more.
· Inserting the fan in the rad bezel may save another 1 or 2 mm or simply using a gpu fan.
· Cutting the rad's metal bezel may save another 1 or 2 mm (or fitting silversonte 22mm rad)




What do you think??

DonDan, why not to ask some vendor for a pre-built 20mm rad with a 10mm cpu block as this? It may be the following project after hslp.
 
Last edited:
Hey Dondan, I know I've been kinda riding your rear end on this one, but...

Could I get a close up of just the mounting system, (or even just the base of the cooler without the radiator or fan?)
 
Pedrofan I think you would get more/better input on the actual a4 thread not this one which is about an air cooler. That being said honestly the reason people use water cooling isn't because its inherently better/more efficient than air cooling (in endurance testing) its because you can take the heat somewhere else with another area for more heat dissipation space like having a 240. With a case this small putting a water block and a slim radiator inside it will most likely have minimal improvement over air cooling if any at all. It's likely a lot more trouble than its worth. The only practical application of water cooling the A4 (imo) is what user: M3ro did which was have quick disconnects that lead the tubing out of the case where a larger radiator can cool the line properly.
 
Between the direct touch pipes are little gabs that are not good for transferring heat. Your CPU temp sensor will give you good temps, but in that zones where the gab is above the core, there will be very hot zones that aren't recognize by the sensor.

Useful info about Continuous Direct Contact Heat pipes vs. Direct Contact Heat Pipes from Cooler Master patented CDC Technology
http://www.coolermaster.com/cooling/cpu-air-cooler/hyper-212-evo/
38c1c5a78d.jpg

71zFrtpnAXL._SX466_.jpg


The aluminum is holding the four pipes tightly together, yet the main feature is that all the pipes have been milled, without gaps between them. This raises the surface area which makes contact, and should also, in turn, increase the efficiency.
 
Last edited:
Yeah, but, those blocks are made of Aluminium, which doesn't transfer heat as Copper.
I think direct contact to the heat pipes would be better, but it has to be tested.
Definitely I would prefer an advanced AIO.
 
Yeah, but, those blocks are made of Aluminium, which doesn't transfer heat as Copper.
I think direct contact to the heat pipes would be better, but it has to be tested.
Definitely I would prefer an advanced AIO.
It is about direct contact with gaps vs. direct contact without gaps
 
Useful info about Continuous Direct Contact Heat pipes vs. Direct Contact Heat Pipes from Cooler Master patented CDC Technology
http://www.coolermaster.com/cooling/cpu-air-cooler/hyper-212-evo/

View attachment 21099
My understanding from Dondan's post was that the problem isn't with gaps between the direct contact heatpipes, but rather the concentration of heat on the CPU heatspreader. Regardless of whether or not there are gaps or no gaps in a direct contact design, the inner most heatpipes will absorb more heat than the outer ones, reducing the overall efficiency of the design. Using a copper block allows the concentrated heat from the heat spreader on the CPU to be dispersed more evenly throughout all the heatpipes, making the design more efficient.
 
Well, the heatpipes are copper and with CDC they would be touching each other, perfectly transferring the heat between each other, in turn perfectly achieving what a copper block is promising, potentially better since, well, there is no copper block to insert some inefficiencies...

When I had recommended 3x8mm (instead of 4x6) heatpipes with direct contact, I had assumed CDC to begin with. So, yeah, let me re-iterate, 3x8mm with CDC would be best with this type of heatsink IF CONVENTIONAL HEATPIPES ARE TO BE USED (more on this below***).

Why, because 3x8mm heatpipes, although at same total cross-sectional dimension as 4x6mm, would carry more amount of heat, OR same amount of heat faster (http://celsiainc.com/blog-design-considerations-when-using-heat-pipes-pt-1/).
Further, along with CDC, all 3 heatpipes would have shorter direct access to the heat source (which is the die) which increases heat transfer efficiency.
Furthermore, with 3x8mm heatpipes, the heatsink can be designed in a way that all the heatpipes are directly in the way of the fan's airflow where it's strongest (which is the region closer to the edges of the fan). Now the outer heatpipes are not getting direct airflow which will reduce the efficiency of heat dissipation from the heatpipes. This effect is very crucial in heatsink design: you will see that the best heatsinks are designed in a way to place the heatpipes in a way to directly receive airflow.


***
What I ultimately recommend here is the use of LHP (loop heat pipe) tech, which Calyos uses in their new fancy fanless solutions (http://www.calyos-tm.com/working-principles/):

http://opac.vimaru.edu.vn/edata/E-Journal/2005/Applied thermal engineering/v25su5-6.1.pdf
https://www.1-act.com/innovations/heat-pipes/loop-heat-pipes/
https://www.qats.com/cms/2014/08/04/understanding-loop-heat-pipes/

Coolermaster also has LHP-based solutions that would directly apply here: http://www.coolermastercorp.com/solution.php?page_id=22

In one of the references above, there is a mention of a CPU cooler that achieves 30W with only 50 grams of total cooler weight!
 
Last edited:
The problem with all that great ideas is, that there are patents on it. So if I like to use it, I have to cooperate with the inventors.
I think we should concentrate on a more simple solution. Working with basic technologies and developing something that is better than current solutions under 50mm
on a fair price level. I am not a thermal engineer and I am not a big company that has the man power to do a lot of testing and research. So my plan is to use simple existing stuff
and order it to a product that was not seen by bigger companies.


Maybe with newer and special technologies I could create a better heatsink, but this will end in a product that will be so expansive that nobody will buy it.
With this design we will have at least the performance level of the Nexus. But my version has more surface ~92.000mm² VS ~140.000mm² and will be made out of full copper.
So I expect a better performance.
 
The problem with all that great ideas is, that there are patents on it. So if I like to use it, I have to cooperate with the inventors.
I think we should concentrate on a more simple solution. Working with basic technologies and developing something that is better than current solutions under 50mm
on a fair price level. I am not a thermal engineer and I am not a big company that has the man power to do a lot of testing and research. So my plan is to use simple existing stuff
and order it to a product that was not seen by bigger companies.


Maybe with newer and special technologies I could create a better heatsink, but this will end in a product that will be so expansive that nobody will buy it.
With this design we will have at least the performance level of the Nexus. But my version has more surface ~92.000mm² VS ~140.000mm² and will be made out of full copper.
So I expect a better performance.
Has Cooljag given you a quote yet?
 
The problem with all that great ideas is, that there are patents on it. So if I like to use it, I have to cooperate with the inventors.
I think we should concentrate on a more simple solution. Working with basic technologies and developing something that is better than current solutions under 50mm
on a fair price level. I am not a thermal engineer and I am not a big company that has the man power to do a lot of testing and research. So my plan is to use simple existing stuff
and order it to a product that was not seen by bigger companies.


Maybe with newer and special technologies I could create a better heatsink, but this will end in a product that will be so expansive that nobody will buy it.
With this design we will have at least the performance level of the Nexus. But my version has more surface ~92.000mm² VS ~140.000mm² and will be made out of full copper.
So I expect a better performance.

You are right, it would be a greater undertaking with an LHP implementation...

Then with conventional heatpipes, still I hope you take the first half of my post under consideration ;)
 
I'm guessing that a fan can be included if the price is right? Maybe have a TY-100 and Noctua 92mm to be able to swap and or clear different items on the board.
 
Love this idea. Shaking up the market to make better options for this is needed, and it's fun to see just how much this forum is dedicated to making the market as good as possible.
 
I'm guessing that a fan can be included if the price is right? Maybe have a TY-100 and Noctua 92mm to be able to swap and or clear different items on the board.

The heatsink is not made by Noctua or Thermalright so it will be impossible to include that fans ^^. Currently I think of delivery without a fan, because 60% will swap it for Noctua or 120mm slim fan.
 
It would be very sensible to ship without fan for the majority of users such as ourselves. Even if this reaches the outer markets beyond the DanCase, its a niche heatsink that isn't planning to be available through retail channels, so buyers should be wall aware of what they are getting into.

I am all the more in supporting a fanless model to save costs that will most likely sit in the box somewhere forgotten and unused.
 
I would prefer it to be shipped with a fan, buying a fan on its own is quite expensive when you add shipping on top... I don't see why a manufacturers fan can't be sold with the cooler as a bundle as long as it is clear that the heatsink is not affiliated with them.
 
I would prefer it to be shipped with a fan, buying a fan on its own is quite expensive when you add shipping on top... I don't see why a manufacturers fan can't be sold with the cooler as a bundle as long as it is clear that the heatsink is not affiliated with them.

Branding.

A lot of marketing and business is kind of insane, but the one thing a company will fight for harder than anything else is it's brand.

Think about it; nobody buys MSI's motherboards, even though they make pretty goos ones. Why?

Because back when they first started making motherboards, they relied exclusively on Killer brand ethernet. If you've used Killer, especially back then, you know that their drivers suck.

Now, even though that wasn't MSI's fault, they got stuck with bad word of mouth because of it.

Same thing happens here. If they let their fans be used by some schmuck on a forum for the computer obsessed (Said with much love, Dan), there are going to be people who associate that heatsink with Noctua - even just someone watching a review of the cooler without paying much attention will see the brown and tan and go, "oh, I recognize those colors," without even thinking that it might not be associated with Noctua at all.

Now imagine if the cooler isn't good. What if it's a royal pain to mount, or it only cools 5 degrees more than the C7 for eight times the price? (Actually, I'd probably still buy it.)

I'm not saying that's going to happen, but in the eyes of the company in question, there's a chance that it could happen... And if it does, the people who associate this cooler with their company will talk about it.

Then all of a sudden, your reputation goes from, "expensive but OMG worth it," to, "overpriced junk." There's no way any company will take that risk. No matter how slim the chance, to them it's never going to be worth selling a few extra fans.

EDIT: what might be possible is a separate tier on the Kickstarter to add a boxed fan to your order - it would be more expensive than actually partnering and making it this heatsink's official fan, but it would cut down on shipping, I suppose?
 
Last edited:
I would prefer it to be shipped with a fan, buying a fan on its own is quite expensive when you add shipping on top... I don't see why a manufacturers fan can't be sold with the cooler as a bundle as long as it is clear that the heatsink is not affiliated with them.

1.) trademark law
2.) I am a competitor
 
I found a solution for pressure control and stain reliefs for transporting and shocks:

springsvuqgy.jpg

I assume this will cool better with the fan blowing on the heatsink, rather than away from it. Would a plastic duct (for venting outside the case) help performance? Is the heatsink tall enough that it doesn't need a duct, because it almost touches the case door?
 
I assume this will cool better with the fan blowing on the heatsink, rather than away from it. Would a plastic duct (for venting outside the case) help performance? Is the heatsink tall enough that it doesn't need a duct, because it almost touches the case door?

It should be practically touching the side panel, yes. The whole point of Dondan making a custom cooler is to use every scrap of space available for cooling. :)
 
And pulling air in from outside should give the CPU fresh cool air, while pushing air out will help exhaust hot air, but likely give the CPU hot air from the case internal temp, or even from the GPU.

We shall find out what works best when it is in our hands (if we do end up with one)
 
This is a phantek heatsink. How it does perform? I tested it years ago and was not happy with the results.
 
PH-TC12LS?
Can't be, can it? This has the fan underneath the heatsink, and that model has the heatsink stuck directly to the mounting plate.

Edit: whoops, got my models confused. Without the top fan, the PH-TC12LS is 1.9 inches tall, which comes out to 48.26mm. That could probably be squished enough that a sufficiently determined user could convince it to fit inside a case like this.

The only question is how it actually performs.
 
Can't be, can it? This has the fan underneath the heatsink, and that model has the heatsink stuck directly to the mounting plate.

Edit: whoops, got my models confused. Without the top fan, the PH-TC12LS is 1.9 inches tall, which comes out to 48.26mm. That could probably be squished enough that a sufficiently determined user could convince it to fit inside a case like this.

The only question is how it actually performs.

It is the PH-TC12LS, I have one too.
it fits in the case without any problems.

I am gonna test this one too but I had to wait for a replacement GPU since mine developed a horrible coil whine (I started testing all the heatsinks I bought). Currently I am sick though but I am gonna retest the old ones (I now have the reliding tool I waited for) and test the remaining ones as soon as I am better.
 
Back
Top