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

[Kmpkt]

Also @dondan, are any of these in your possession yet? I'm dying to see benchmarks, especially comparing the copper and aluminum versions.

 

[CubanLegend]

Here are some shots of the Lian Li sample from the HSLP-48

Holy crap that full-copper looks amazing. This needs to be tested on the ASUS z270i!

Honestly though @dondan I'm rooting for your HSLP-48 to beat the NHL9i and the LP-53, then I'll be upgrading my LP53 to that HSLP-48.

One question: the Lian-Li version seems to have these air-gaps on top, between the fins, right on the top of the heatpipes... where the CoolJag version doesnt have this airgap, is this an error in production or a fabrication shortcut?

 

[dondan]

Some fans for the HSLP-48 test

 

[dondan]

Sorry guys I must leave this here very quick.

Today I got the Lian Li samples and currently my system is running Prime95 with a crazy room temp of 25°C.

The copper edition with the thermalright 100mm fan perform nearly on the same level as the Asetek AIO with a room temp of 21°C. So the HSLP-48 is 3°C better as the AIO and 13°C better as the Dynatron T318.

Now It will try 4Ghz

 

[jeshikat]

Heatsink discussion split to a new thread since it's a separate project.

Edit: But if you would prefer they be a single thread @dondan I can merge them back.

 

[dondan]

This is ok thank you

 

[McTeags]

Wow, a 13°C difference between the HSLP-48 and your previously recommended air cooler. That's substantial. What CPU are you using? i7-7700k? This will really help people who are using high end CPUs in combination with high end open air GPUs like the EVGA GTX 1080 TI FTW 3.

You've already sold thousands of A4-SFX cases, between people who want this specifically to get the best CPU cooling out of the A4 as well as individuals who plan to use it for other projects you should already be looking at a pretty successful Kickstarter.

 

[jmsantos1983]

Copper!!! Love copper.
Will they be 25mm or 15mm?
Regular fan or slim fans?

 

[dondan]

Here is a small update what I plan to do with the HSLP-48 in the next weeks.


I will test the following setups all CPUs @ default clock:

• AM4 (1700x)
• 1151 (7700k)
• 2011-3 (5820K)

I will test the HSLP-48 against these heatsinks:

• Asetek 92mm AIO (AM4, 2011-3, 1151)
• Cryorig C7 (AM4, 1151)
• Noctua L9i (AM4, 1151)
• Dynatron T318 (2011-3)
• Thermalright AXP 100 (AM4, 2011-3, 1151)

I will test the HSLP-48 Copper and Alu version following fans:

• A9x14 (92mm)
• TY-100 (100mm)
• SY1212SL12H (120mm)
• A12x15 (120mm)

I will test the following fan modes:

• under the heatsink sucking through it
• under the heatsink blowing through it
• on the heatsink blowing through it
• sandwich mode
• different fan speeds

What I can say about the Lian Li samples, the quality is great for a prototype, but they forgot to include the 2011-3/AM3 retention bracket. They forgot to add a thread on the 1151 brackets so I can't use them. Therefore I use my Thermalright retention kit that I have for all sockets. What luck the Thermalright retention kits are full compatible to the HSLP-48

I have to improve the fan clips because the two ends that go inside the heatsink are very short. I have to make them longer to better secure the fan. I also plan to add more holes for the clips to better support different thickness of fans.

I think I will get the CoolJag samples tomorrow, I hope they did a better job on the retention brackets.

I am thinking of outsourcing the retention kits to Thermalright if they can offer a good price. The brackets are the most complicated part on this project and it could be a good idea also for future compatibility. I will ask Lian Li if they can get in touch with them and I don’t have to take the route over the German contact.



Preview

Here is small preview of tests on the copper version I did today. All values are not comparable to older result, because I run the CPU (i7 5820K 140W TDP) on default clock (3.4GHz) and voltage. Older tests was made with 3GHz and lower voltage. The Dynatron T318 will reach ~80°C on default clock and voltage.


Prime95 v26.6 8K Test – 15min each

Heatsink------------Fan---------------Position-----Mode----------RPM--------Room Temp--------Core Temp
HSLP-48 Copper.....A9x14...............Under......Pushing..........2600..............25........................74 °C
HSLP-48 Copper.....A9x14...............Under......Sucking..........2600..............25........................65,8°C
HSLP-48 Copper.....TY-100..............Under......Pushing..........2400..............25....................... 65°C
HSLP-48 Copper.....TY-100..............Under......Sucking..........2400..............25........................63,8°C
HSLP-48 Copper.....SY1212SL12H..Under......Sucking..........2000..............25....................... 65,2°C
HSLP-48 Copper.....A12x15..............Top..........Pushing.........1800..............25........................63,5°C
HSLP-48 Copper.....A12x15..............Top..........Pushing...........900..............25........................73,5°C


Maybe you wonder why the A9x14 performs not good in pushing mode. I think the fan has a bad pressure and isn't able to push enough air through the fins in this mode. What I think is realy crazy that the heatsink is able to cool a six core i7 in Prime95 with the new A12x15 on top with only 900 RPM at 73°C. I am nure sure if any other heatsink with a height of 63mm is able to do that.



Some pictures

 

[Kmpkt]

Wow, a 13°C difference between the HSLP-48 and your previously recommended air cooler.

Pretty sure the T318 was the heatsink of choice because it is the only Narrow ILM heatsink on the market. Talking to people that have used it and then compared to other units, it's actually not a terribly great cooler. I'll be curious to see how the HSLP-48 does against some of the other top downdraft SFF heatsinks like the L9i and LP53.

 

[jeshikat]

Pretty sure the T318 was the heatsink of choice because it is the only Narrow ILM heatsink on the market.

*the only one that fits in the A4-SFX.

 

[Boil]

@dondan

Any plans to test with the NF-A12x15 UNDER the heatsink, or do you not have VLP RAM to make said test feasible...?

 

[Kmpkt]

That's what I meant

 

[CubanLegend]

@dondan can you test your HSLP-48 with the ASUS Z270i mobo? and see if it fits with the A12x15, with some VLP-DDR4? Maybe you have to remove the mosfett heatsinks, but I'd like to know if your cooler will fit my motherboard with VLP DDR4!

 

[dondan]

I will get a low priced MSI board with H110 chipset.

 

[CubanLegend]

Pretty sure the T318 was the heatsink of choice because it is the only Narrow ILM heatsink on the market. Talking to people that have used it and then compared to other units, it's actually not a terribly great cooler. I'll be curious to see how the HSLP-48 does against some of the other top downdraft SFF heatsinks like the L9i and LP53.

Yes I agree, I'd like to see the temperature comparison of the HSLP-48 against the NHL9i and the LP53 as well... here were my results with the LP53 "The LP53 beat the NHL9i in the same configuration without the VRM heatsinks, it dropped my idle temps by 3c, and my load temps dropped between 2-5c."

 

[dondan]

Here is a small preview of my tests today with the HSLP-48. (Copper vs. Alu)

The benefits of copper vs. alu will be smaller on lower RPMs.

i7 5820K default clock (3,4Ghz) and volatage, Room temp 25

HSLP-48 Copper...|...TY-100...|...Pull...|...2400RPM...|...63,8°C
HSLP-48 Alu.........|...TY-100...|...Pull...|...2400RPM...|...67,2°C
HSLP-48 Copper...|...TY-100...|...Pull...|...2000RPM...|...66,5°C
HSLP-48 Alu.........|...TY-100...|...Pull...|...2000RPM...|...69,3°C
HSLP-48 Copper...|...TY-100...|...Pull...|...1600RPM...|...70,2°C
HSLP-48 Alu.........|...TY-100...|...Pull...|...1600RPM...|...72,8°C
HSLP-48 Copper...|...TY-100...|...Pull...|...1400RPM...|...75°C
HSLP-48 Alu.........|...TY-100...|...Pull...|...1400RPM...|...77°C
HSLP-48 Copper...|...TY-100...|...Pull...|...1200RPM...|...80,5°C
HSLP-48 Alu.........|...TY-100...|...Pull...|...1200RPM...|...82,3°C

In the middle of all values the difference is 2,52°C (max 3,4°C, low 1,8°C)

 

[dondan]

Bad news – from hero to zero

Today I made some tests of the HSLP-48 against the Thermalright AXP 100H and Dynatron T318 all with TY-100 fan. I did all tests on an open bench table and not inside the A4-SFX. In my last post I wrote the HSLP-48 is ~15°C better as the T318, but I made this statement based on the values I had in my memory of the T318 mount inside the case. It looks like these values aren’t comparable with the values on an open bench test. On an open bench table the difference is only 2°C (63,8°C vs. 66°C). Wow this is a real setback. I think it could be different inside the case, because the T318 recycle a lot of air.

Also it looks like the AXP-100 and AXP-100H both with the TY-100 fan on top performing 4-5°C better than the HSLP-48. Maybe it was a wrong decision to select 4 heatpipes instead of 6. Keep in mind the AXP-100 has a height of 58mm and the AXP-100H of 65mm and will not fit inside the A4-SFX or Sentry.

On weekend I will get a rent setup (i7 7700k and MSI H110I PRO) from Caseking. Before I order more parts I will start with a test HSLP-48 vs. C7 to see if it is worth to continue. I will also do some tests mounting the parts inside the case to see if the heatsink will perform better, because it recycle less air.

But it looks like I have to go back to the drawing board and create a version with 6 heatpipes. I think the problem is not the surface because it is big enough but maybe the 4 heatpipes are too slow in transfer heat.

Maybe the CoolJag samples are better, but I don’t believe in this.

 

[AleksandarK]

If you add 6 heatpipes it will move heat 50% faster compared to 4 heatpipe version. I dont know how well does it scale but it will reduce temperature by at least 20%.
Btw how big is the gap between fins.

 

[Kmpkt]

Honestly I think the Achilles heel of your design right now is the bend radius at the top if quite sub-optimal for a 6mm heat pipe. If you could widen it you might be able to stick with four pipes.