DAN HSLP-48: A powerful sub 50mm heatsink

OmenemO said:
I disagree, Dan is not doing us all a favor, it's a business venture since he's in this to make money. I was following this thread hoping to see the best cooler designed and tested for usage in his case, not an open air environment.
Click to expand...

It would be very hard for me to believe this project is only profit motivated. Dan is a small form factor enthusiast who probably doesn't want to starve working on a labor of love.
 
OmenemO said:
I disagree, Dan is not doing us all a favor, it's a business venture since he's in this to make money. I was following this thread hoping to see the best cooler designed and tested for usage in his case, not an open air environment.
Click to expand...

He tryed Open air for its easier to mount and switch, also to se the best performance for all cases.

Then he have not get VLP memories yet to try it in Case i dont even now if he order them because the bad test results on Lian li version that i think is because he used Thermalrights mounting kit to it for they didnt send out 2011 mounting kit but Nexus Cooler version did that.
 
dondan said:
Here are some ideas for prototype 2:

Heatsink: Downsize width from 130mm to 120mm
Heatsink: Maybe increasing the fin spacing +0,2mm and reducing the fin count from 60 to 54

Heatpipe: Increasing bend radius from 10mm to 15-18mm
Heatpipe: One sample with 4 heatpipes and one with 5-6 heatpipes

CPU-Plate: For both samples moving the heatpipes closer together.

Clip: Optimise clip mounting I will focus on fans with a thickness of 12mm and 15mm.
Click to expand...

On the 4 Heatpipes Version that Cooljag send out is any of this changes implemented ??

And what is the plan after the new better test results on the Cooljag version 1
 
OmenemO said:
I disagree, Dan is not doing us all a favor, it's a business venture since he's in this to make money. I was following this thread hoping to see the best cooler designed and tested for usage in his case, not an open air environment.
Click to expand...
Well, why dont you do us all a favor and design a better, cheaper, faster production Cooler to save us all from "Dan's greediness"???
Really, ill be the first to buy the cooler if you can achieve it...

PD: Also, if you are able to make a better case than the A4 at a lower MSRP achieving no profit, i'll buy one too
 
Last edited:
Update:

Today I tested the heatsink inside the case again and also did some special tests:

Normal tests:

5820K 6x3,4Ghz - Bench Table
Prime95 v26.6 8K Test – 20min each (AIO 60min)

Heatsink-----------------------Fan-----------Position---Mode-------RPM--------Room Temp-----Core Temp
HSLP-48 Alu (CoolJag)...........TY-100..............under..........suck-in......2400..............25........................68,5°C
HSLP-48 Alu (CoolJag)...........TY-100..............under..........push-out....2400..............25........................68,5°C
HSLP-48 Copper (CoolJag)... .TY-100..............under.........push-out....2400..............25........................69,5°C
HSLP-48 Copper (Lian Li)........TY-100.............under.........push-out.....2400..............25........................71°C
HSLP-48 Alu (Lian Li)...............TY-100.............under.........push-out.....2400..............25........................73°C
Asetek 92mm AIO ...................A9x14...............Top............suck-in.......2200..............25........................74°C


The interesting part of this test is, that the perforce between CoolJag and Lian Li is now very closely. I think the reason for this is
the heat level inside the case. Furthermore this must be also the reason why the the CoolJag alu version is better as the copper version, because
copper absorb heat better as alu. (It could also be that this is a measuring inaccuracy affected by room temp.)
Keep in mind that all tests was not made with the Noctua 120mm fan, because I need to order low profile ram and low profile 8pin ESP connector to fit on my motherboard.


So from now it looks like the HSLP-48 will perform 7-10°C worse inside the A4-SFX. Maybe the 120mm Noctua fan can change this. For the noise level I can say that with the
TY-100 fan on full speed the setup it is far away from being silent. This can be also changed with the Noctua A12-15 fan.



Special tests:

I also did some special tests with a paper covering the bottom and side part of heatsink excepting a hole for the fan. A forum user some pages ago
came with this idea. This solution will reduce the temp for 0.5°C so it looks like the fan does not recycle hot air so a special duct will not help to get better temps.


I also tested the influence of removing the I/O shield on the cooling performance.

5820K 6x3,4Ghz - Bench Table
Prime95 v26.6 8K Test – 20min each

Heatsink-----------------------Fan-----------Position---Mode-------RPM--------Room Temp-----Core Temp
HSLP-48 Alu (CoolJag)...........TY-100..............under..........suck-in......2400..............25........................64,8,0°C (no I/O shield)
HSLP-48 Alu (CoolJag)...........TY-100..............under..........push-out....2400..............25........................67°C (no I/O shield)

As you can see the suck-in mode can benefit from it and temps are 3,7°C lower.


 
Last edited:
Just curious - is the fan also that loud at 2400 RPM in open case / open test bench? I wonder if there's any similarity with C7 being loud when the side panel is on.
 
These are incredibly good results. We can save money with aluminum and still beat a water cooler.

The only questions now are: does a different fin density and heatpipe count make a difference, and does a better fan make a difference. Super work dan!

Edit: If just opening the io cover saves 4 degrees when sucking hot air in, that means an output fan would help too. You could retest with fans in the various output positions. You could also reverse the PSU fan, because PSU components can tolerate high heat, and it might be better to have a hotter PSU in trade for cooler CPU. Or, put another way, it might be possible to have the PSU and CPU fans run at closer speeds, rather than one much noisier....

dondan said:
Update:

Today I tested the heatsink inside the case again and also did some special tests:

Normal tests:

5820K 6x3,4Ghz - Bench Table
Prime95 v26.6 8K Test – 20min each (AIO 60min)

Heatsink-----------------------Fan-----------Position---Mode-------RPM--------Room Temp-----Core Temp
HSLP-48 Alu (CoolJag)...........TY-100..............under..........suck-in......2400..............25........................68,5°C
HSLP-48 Alu (CoolJag)...........TY-100..............under..........push-out....2400..............25........................68,5°C
HSLP-48 Copper (CoolJag)... .TY-100..............under.........push-out....2400..............25........................69,5°C
HSLP-48 Copper (Lian Li)........TY-100.............under.........push-out.....2400..............25........................71°C
HSLP-48 Alu (Lian Li)...............TY-100.............under.........push-out.....2400..............25........................73°C
Asetek 92mm AIO ...................A9x14...............Top............suck-in.......2200..............25........................74°C


The interesting part of this test is, that the perforce between CoolJag and Lian Li is now very closely. I think the reason for this is
the heat level inside the case. Furthermore this must be also the reason why the the CoolJag alu version is better as the copper version, because
copper absorb heat better as alu. (It could also be that this is a measuring inaccuracy affected by room temp.)
Keep in mind that all tests was not made with the Noctua 120mm fan, because I need to order low profile ram and low profile 8pin ESP connector to fit on my motherboard.


So from now it looks like the HSLP-48 will perform 7-10°C worse inside the A4-SFX. Maybe the 120mm Noctua fan can change this. For the noise level I can say that with the
TY-100 fan on full speed the setup it is far away from being silent. This can be also changed with the Noctua A12-15 fan.



Special tests:

I also did some special tests with a paper covering the bottom and side part of heatsink excepting a hole for the fan. A forum user some pages ago
came with this idea. This solution will reduce the temp for 0.5°C so it looks like the fan does not recycle hot air so a special duct will not help to get better temps.


I also tested the influence of removing the I/O shield on the cooling performance.

5820K 6x3,4Ghz - Bench Table
Prime95 v26.6 8K Test – 20min each

Heatsink-----------------------Fan-----------Position---Mode-------RPM--------Room Temp-----Core Temp
HSLP-48 Alu (CoolJag)...........TY-100..............under..........suck-in......2400..............25........................64,8,0°C (no I/O shield)
HSLP-48 Alu (CoolJag)...........TY-100..............under..........push-out....2400..............25........................67°C (no I/O shield)

As you can see the suck-in mode can benefit from it and temps are 3,7°C lower.
Click to expand...
 
Last edited:
bAMtan2 said:
These are incredibly good results. We can save money with aluminum and still beat a water cooler.

The only questions now are: does a different fin density and heatpipe count make a difference, and does a better fan make a difference. Super work dan!

Edit: If just opening the io cover saves 4 degrees when sucking hot air in, that means an output fan would help too. You could retest with fans in the various output positions. You could also reverse the PSU fan, because PSU components can tolerate high heat, and it might be better to have a hotter PSU in trade for cooler CPU. Or, put another way, it might be possible to have the PSU and CPU fans run at closer speeds, rather than one much noisier....
Click to expand...

Turning the PSU in the A4 means it is now squished between a flexible plastic shield and the backside of the GPU. I don't think it will draw in any air whatsoever which would be very dangerous.
 
mathijs159 said:
Turning the PSU in the A4 means it is now squished between a flexible plastic shield and the backside of the GPU. I don't think it will draw in any air whatsoever which would be very dangerous.
Click to expand...
I think bamtan2 meant just the fan inside the PSU, not the entire unit.
 
dondan said:
The interesting part of this test is, that the perforce between CoolJag and Lian Li is now very closely. I think the reason for this is
the heat level inside the case. Furthermore this must be also the reason why the the CoolJag alu version is better as the copper version, because
copper absorb heat better as alu. (It could also be that this is a measuring inaccuracy affected by room temp.)
Click to expand...

This is not how heat flow works, if you put a hot heatsink in a hot (but cooler) environment, the heat will still flow from the hotter heatsink to the less hot case. Ergo, the difference in thermal conductivity would lead to an advantage (however small) for the Cu even in a hot case. I think the delta you see is just measurement uncertainty or due to changes in the ambient temperature as you mentioned.

Steve Burke from GN mentioned that he asked Cryorg to send him the new Cu coolers such that he can compare them to the current Al versions. His initial thoughts are that copper will lead to a more stable fan profile due to the higher thermal capacity. The same effect will also have benefits in short burst loads as the cooler can absorb more energy for the same temperature change (as compared to Al). However, he assumed that the steady-state performance would be equal. Nevertheless, he is not certain, ergo the request to Cryorg.
 
Last edited:
This talk from Gamers Nexus, shows that testing is very complicated.
Lets remember that Aluminium has a lower Heat Transfer Coefficent than Copper.
That is a number that show us how the material behaves, in transferring heat through it, how it absorbs and to expels heat. At some point it will get "saturated", like AIOs or liquid cooling. (Thats why thermal cameras are useful)
So the test (I think) should last for a while to get all things at working temps. (20 min will not be enough, even at full load)
Another thing is the DAN A4-SFX Case, which will get hotter because of the back of the video card and the back of the MoBo are very close.
 
Last edited:
Interesting that CoolJags aluminum sink is better than their copper one.

Keep up the good work dondan. Do you have a 7700K to do any tests on? (Can we all buy you one?)
 
Not to be lazy, but dondan would it be possible to have a table of the most up-to date testing results in the OP? Would be really nice to see where things are at performance wise in one place rather than having to jump back and forth to make performance comparisons. Thanks!
 
KMPKT said:
Not to be lazy, but dondan would it be possible to have a table of the most up-to date testing results in the OP? Would be really nice to see where things are at performance wise in one place rather than having to jump back and forth to make performance comparisons. Thanks!
Click to expand...

  1. Cooljag Aluminum
This is the most important part of the table so far, based on Dan's post above ;)
 
dondan said:
Update:

Special tests:

I also did some special tests with a paper covering the bottom and side part of heatsink excepting a hole for the fan. A forum user some pages ago
came with this idea. This solution will reduce the temp for 0.5°C so it looks like the fan does not recycle hot air so a special duct will not help to get better temps.
Click to expand...


Thanks for the additional test but I was trying to also say that the bottom of the cooler needs to be profiled in a way so that the smaller fan also gets air to the full width of the cooler with the shroud.

With the paper shroud you describe anything wider than the fan, in one direction, is essentially passive.

Maybe as a test you could try the nexus with and without the paper shroud. The bottom shape of the nexus would allow full heatsink usage with a ~90mm fan and a shroud.

Obviously this would be a significant effort but I think it's a good test. Thanks for all your work to date, it's all looking very positive.
 
Got my VLP DDR4 RAM all ready for this cooler DonDan! This may have been asked previously so please forgive me but will the HSLP-48 be compatible with the ASUS Z170I Pro Gaming?

FullSizeRender.jpg
 
Looking good noctua. Now let's make the HSLP-48 perform better than this bad boy.

aHR0cDovL21lZGlhLmJlc3RvZm1pY3JvLmNvbS81L0kvNjg5NjcwL29yaWdpbmFsL0wtVHlwZS5KUEc=
 
Ram clearance. My guess is that the rational behind is if it is high performance that they want they won't be limiting themselves to a super compact form factor.

I was merely showing off the awesome design that dondan and all of us backers are looking for. A clean fanless top hiding a tornado of cooling underneath.
 
I guess it's really early to give a date but when is this thing expected to be on kickstarter? Less than a year?
 
QuantumBraced said:
If HSLP-48 outperforms this new Noctua cooler and the T318 by more than 2C, then serious kudos.
Click to expand...
It already outperforms the t318. Dondan was running his originally at 3GHz with the t318. He later tried the asetek 545lc and was able to clock it higher. Now the alu fin version performs better than the 545lc so we are doing real good in the performance department.
 
So about a month ago I contacted Innodisk about getting a batch of RAM done with Samsung B dies at 2400 MHz in black and they finally got back to me. Looks like about 102.50 per 8GB DIMM. While I did get a few emails and PMs from interested people, it definitely wasn't enough to meet the MOQ of 100 units. As such I'm not going to bother following up on this.
 
Does it really have to be VLP and will 31mm Corsair Vengeance LPX work with this heatsink?
 
The LPX should work with the heatsink, but if you want to use a 120mm fan instead of a 92mm fan, you'll need VLP.
 
Has there already been a test of this cooler inside the dan a4?
 
Shaav said:
Has there already been a test of this cooler inside the dan a4?
Click to expand...

Yes. Scroll up.

dondan said:
Update:

Today I tested the heatsink inside the case again and also did some special tests:

Normal tests:

5820K 6x3,4Ghz - Bench Table
Prime95 v26.6 8K Test – 20min each (AIO 60min)

Heatsink-----------------------Fan-----------Position---Mode-------RPM--------Room Temp-----Core Temp
HSLP-48 Alu (CoolJag)...........TY-100..............under..........suck-in......2400..............25........................68,5°C
HSLP-48 Alu (CoolJag)...........TY-100..............under..........push-out....2400..............25........................68,5°C
HSLP-48 Copper (CoolJag)... .TY-100..............under.........push-out....2400..............25........................69,5°C
HSLP-48 Copper (Lian Li)........TY-100.............under.........push-out.....2400..............25........................71°C
HSLP-48 Alu (Lian Li)...............TY-100.............under.........push-out.....2400..............25........................73°C
Asetek 92mm AIO ...................A9x14...............Top............suck-in.......2200..............25........................74°C


The interesting part of this test is, that the perforce between CoolJag and Lian Li is now very closely. I think the reason for this is
the heat level inside the case. Furthermore this must be also the reason why the the CoolJag alu version is better as the copper version, because
copper absorb heat better as alu. (It could also be that this is a measuring inaccuracy affected by room temp.)
Keep in mind that all tests was not made with the Noctua 120mm fan, because I need to order low profile ram and low profile 8pin ESP connector to fit on my motherboard.


So from now it looks like the HSLP-48 will perform 7-10°C worse inside the A4-SFX. Maybe the 120mm Noctua fan can change this. For the noise level I can say that with the
TY-100 fan on full speed the setup it is far away from being silent. This can be also changed with the Noctua A12-15 fan.



Special tests:

I also did some special tests with a paper covering the bottom and side part of heatsink excepting a hole for the fan. A forum user some pages ago
came with this idea. This solution will reduce the temp for 0.5°C so it looks like the fan does not recycle hot air so a special duct will not help to get better temps.


I also tested the influence of removing the I/O shield on the cooling performance.

5820K 6x3,4Ghz - Bench Table
Prime95 v26.6 8K Test – 20min each

Heatsink-----------------------Fan-----------Position---Mode-------RPM--------Room Temp-----Core Temp
HSLP-48 Alu (CoolJag)...........TY-100..............under..........suck-in......2400..............25........................64,8,0°C (no I/O shield)
HSLP-48 Alu (CoolJag)...........TY-100..............under..........push-out....2400..............25........................67°C (no I/O shield)

As you can see the suck-in mode can benefit from it and temps are 3,7°C lower.
Click to expand...
 
Here is a small update:

I don't ordered the 6 heatpipe design because it need some more detail work. I changed the fin design a bit so it doesn't look like a Noctua copy. Furthermore I am working on the retention kit to get a save pressure and a more confortable way to mount the heatsink:

v1.6hwosr.jpg
 
Last edited:
dondan said:
Here is a small update:

I don't ordered the 6 heatpipe design because it need some more detail work. I changed the fin design a bit so it doesn't look like a Noctua copy. Furthermore I am working on the retention kit to get a save pressure and a more confortable way to mount the heatsink:

v1.6hwosr.jpg
Click to expand...
Still in love with the full copper version...
 
You must log in or register to reply here.
Back
Top