Concept Sun Cool 150W TDP slim heatsink

XCreator

Trash Compacter
Original poster
Apr 7, 2018
42
33
Update 8-May-2018 Design change
Estimate performance will be same water cooler if equip with 92mm x 25mm fan or bigger.

Hi guy, I'm assistant from XCreator, established by Mr Rolexus, he now focus in design and other jobs, I'm in charge in media.
This Sun Cool is a full copper heat sink TPD 150W, Height 22mm, weight: 744g work with 92mm fan, total 36mm with Noctua 92mm slim fan.
This one will be used all our case in future.
Test result in simulation with CPU heat source 110W, air flow 57m3/h (33.5 CFM), ambient temperature 30C:

Max temperature : 62.4C
Test result in simulation with CPU heat source 150W, air flow 57m3/h (33.5 CFM), ambient temperature 30C:

Max temperature : 74.3C

Max temperature with 150W CPU, no fan is 115C.

Available time: TBD
 
Last edited:

loader963

King of Cable Management
Jan 21, 2017
660
568
I love the idea of high tdp coolers but I’m also curious about mobo compatibility. Specifically clearance specs. What motherboards have you guys tested with?
 
  • Like
Reactions: XCreator

XCreator

Trash Compacter
Original poster
Apr 7, 2018
42
33
I like the concept, but i have few questions:

1. Where are you going to manufcture this?
2. What do you expect it to cost?
3. What will be the supported socket?
Thank for interesting,
1.We are planing to manufacture this in Vietnam or Japan, but location can be change to China or Taiwan depend on actual condition.
2.There many things that effect to the cost, but at surely the cost will be less than 100$, the cost will be managed to be acceptable.
3.At this time, our cases model which mainly designed for air cooler only support ITX motherboard, so this design support socket 115X socket only, arm to Intel K-CPU series, we are considering to support other socket.
I love the idea of high tdp coolers but I’m also curious about mobo compatibility. Specifically clearance specs. What motherboards have you guys tested with?
The over hang side is in PCIE slot side so you can install it in any 115X motherboard with chipset heat sink height is lower than 12mm except Asus impact series. This one is designed to prevent hot air flow directly blow to mobo.
We surely will test the prototype.
I'm curious about the manufacturing process as well.
This one is fabricate by CNC milling.
Which fan are you basing the simulation off of? I think the NF A9x14 (the fan you mentioned) has an airflow rating of 50.5 m^3/h, and that's at max RPM. I'm also curious what impact being in a case and being surrounded by components will have on the heatsink's performance. I know it's early on in the process, but can you offer any info there?
This hear sink support all 92mm fans, and with 50.5 m^3/h, the simulation result is not much different. The simulation is fairly correct, but it is in idea case with idea air flow, in actual condition if we have good air flow, the result will not much different.

And I have some thing to share:
At this time, we can't provide a detail plan for our ideas, we have to finish preparing many thing (website, supply chain, workshop...) we want our product appear professionally, we are trying to finish preparing in 2019 Spring.
 

FullForceRainbow

Trash Compacter
May 12, 2016
45
47
Thank for interesting,
1.We are planing to manufacture this in Vietnam or Japan, but location can be change to China or Taiwan depend on actual condition.
2.There many things that effect to the cost, but at surely the cost will be less than 100$, the cost will be managed to be acceptable.
3.At this time, our cases model which mainly designed for air cooler only support ITX motherboard, so this design support socket 115X socket only, arm to Intel K-CPU series, we are considering to support other socket.

The over hang side is in PCIE slot side so you can install it in any 115X motherboard with chipset heat sink height is lower than 12mm except Asus impact series. This one is designed to prevent hot air flow directly blow to mobo.
We surely will test the prototype.

This one is fabricate by CNC milling.

This hear sink support all 92mm fans, and with 50.5 m^3/h, the simulation result is not much different. The simulation is fairly correct, but it is in idea case with idea air flow, in actual condition if we have good air flow, the result will not much different.

And I have some thing to share:
At this time, we can't provide a detail plan for our ideas, we have to finish preparing many thing (website, supply chain, workshop...) we want our product appear professionally, we are trying to finish preparing in 2019 Spring.

Thanks for the answer! I understand what you are saying about a case with good airflow, I was more concerned about the fact that the heat sink is going to be basically boxed in by the RAM, GPU, and VRM heatsinks.

Users on this forum have had issues cooling chips of only 95 watts in cases like the DAN-A4. A lot of their problems stemmed from how boxed in the heat sink was.

To my understanding, server style heatsinks like yours are normally intended to have very high RPM fans force air through them and vent out the back of a server rack, which solves the boxing issue at the expense of noise. I’m just worried that with a fan on top of the heat sink the warm air being forced out the GPU side and top side isn’t going to have anywhere to go.
 

caniplaymayo

Chassis Packer
Mar 28, 2018
15
9
I like the idea a lot, thin cases look amazing but the cooling has been keeping me away. My big question is about the simulations, are you applying the heat flux directly to the base and measuring the temps from there or do you also have a chip modelled with internal heat generation? Those are really good results, so I'm a bit skeptical.

CNC milling a heatsink like this might not be possible. I got quotes back from some CNC shop before and they all said that this won't be possible to achieve. You should check with some manufacturer that you want to work with to make sure they can do this.

With the base attached to all the fins like that it'd have to be extruded, unless there have been some big advancements recently. I guess you could mill the fans and the base separately, but there wouldn't be a great way to attach them. I wonder if replacing the cooler part of the heatsink with heat pipes to connect the fins would make manufacturing easier without sacrificing a lot of performance, although if they plan on mass producing this extruding it probably wouldn't be prohibitively expensive.

To my understanding, server style heatsinks like yours are normally intended to have very high RPM fans force air through them and vent out the back of a server rack, which solves the boxing issue at the expense of noise. I’m just worried that with a fan on top of the heat sink the warm air being forced out the GPU side and top side isn’t going to have anywhere to go.

This is a good point, I wonder if it would be worth it to try to simulate this with different motherboard models before committing to a prototype.
 

BernardoZ

Founder of Z-CASES
z-cases.com
Feb 7, 2018
264
490
www.z-cases.com
That's a nice concept, can't wait to see more!

Quick idea I just had, that might save you some time and money on prototyping: try to get an accurate 3d model of a known low-profile cpu cooler heatsink (Noctua L9i, for example), and throw it on that simulation software. Then, run the same tests you did with your unit, and see if the temp results compare to online benchmarks. This way you will know how precise your simulation is...

Good luck with the project ;)
 
Last edited:

Fruergaard

Trash Compacter
Feb 13, 2018
37
50
Interesting :-)
I believe you may have uploaded the 110W simulation two times?

Since it's CNC milled I assume that it is only a copper block?
So no heat pipes or vapor camber to transfer the heat to the open fin part (the part outside of the keep out zone, chipset area)?
Could you try running your simulations on the cooler that doesn't have that extra fin part (just the cooler that is within the 90x90mm keep out zone)

I'm just trying to compare this with Dynatrons offerings and what they are able to cool efficiently:
K129 is also just a copper block, a bit smaller but considerable more fins (35 vs 56) that also are a bit higher (490 grams with no fan).
And that is only rated for 95W TDP in a server solution with air blowing across the fins (as FullForceRainbow said).
https://www.dynatron.co/product-page/k129
More fins are better for higher airflow, whereas less fins can perform better in low airflow (and low pressure) situations.
As for the A9x14, remember that the 50 m3/t is in a free flow situation. When slapped on a heat sink that number will be lower due to the blockage from the fins. And the A9x14 doesn't have high static pressure. Not sure if you have incorporated that in the simulations. Also can it simulates the air stream?

EDIT: How big is the area of the heat source in your simulations?

And then there is the T318 that is a TDP 165W solution for socket 2011 (around same size but with more and higher fins). That is able to cool an undervolted 5820k CPU in the Dan A4-SFX reasonable, though with the A9x14 at max speed (Dondans test):

With no undervolt (~140W) the T318 is still able to cool the CPUs, check the forum for builds that uses these coolers.

The smaller R15 cooler (150W TDP according to Dynatron) is not able to efficiently cool a delidded I9 7980X (165 TDP) even with a Industrial 120mm noctua fan in a semi open bench setup:

The T318 was able to cool a 145W TDP Xeon E5-2699 v4 CPU with the A9x14 (with temps around 80C):
See temps a 7:38.

And these coolers utilize a vapor camber, which at least to my knowledge, is considerable better at "transporting" and distributing the heat from the base to the fins (and have more fins than your design (35 vs ~40).

All I'm trying to say is, that I have a hard time believing that a "small" pure copper block with a A9x14 @ 100% is able to cool 150W of heat in a 30C ambient room (according to your test parameters) to a CPU friendly level. And one thing is the temperature at the base of the cooler, another thing is the temperatures of the cores inside, as caniplaymayo also has concerns about.

Even more, I hardly believe it will be able to cool a 150W heat source to only 115C (in 30C ambient) with no fan (only airflow generated from the heat rising). I mean this big thing is only good for like 95W with high load temps after some time with no airflow (and in 21C ambient).

I would love to be proved wrong with a real life prototype, as I would love myself such a cooler.
But for now, I would advise you to consider removing the stated 150W TDP until you have a prototype that actual can cool 150W.

Best of luck :-)
 
Last edited:

XCreator

Trash Compacter
Original poster
Apr 7, 2018
42
33
CNC milling a heatsink like this might not be possible. I got quotes back from some CNC shop before and they all said that this won't be possible to achieve. You should check with some manufacturer that you want to work with to make sure they can do this.
Please wait to see what we should do with this "impossible" thing.
Thanks for the answer! I understand what you are saying about a case with good airflow, I was more concerned about the fact that the heat sink is going to be basically boxed in by the RAM, GPU, and VRM heatsinks.

Users on this forum have had issues cooling chips of only 95 watts in cases like the DAN-A4. A lot of their problems stemmed from how boxed in the heat sink was.

To my understanding, server style heatsinks like yours are normally intended to have very high RPM fans force air through them and vent out the back of a server rack, which solves the boxing issue at the expense of noise. I’m just worried that with a fan on top of the heat sink the warm air being forced out the GPU side and top side isn’t going to have anywhere to go.
See air flow design of this heat sink

So air flow to GPU and VRM HS is reduced some, over hang side can reduce temperature of hot air to GPU.
We have good air flow if hot air output not mix much with cool air intake, our cases almost design to support many ventilation fan in both side, so it should work well in my case. I think this is basic requirement for case with air cooler.
For other case that you mention, I think, to make it small, he cut one function is ventilation fan on top or bottom. But I think he still can add some 40/50 fans hole in top side.
You are right about sever fan, but this one is not really sever fan, it has bigger weight (744g), thicker fin, better for working with lower flow.
115C no fan is a example.
I like the idea a lot, thin cases look amazing but the cooling has been keeping me away. My big question is about the simulations, are you applying the heat flux directly to the base and measuring the temps from there or do you also have a chip modelled with internal heat generation? Those are really good results, so I'm a bit skeptical.
That's a nice concept, can't wait to see more!

Quick idea I just had, that might save you some time and money on prototyping: try to get an accurate 3d model of a known low-profile cpu cooler heatsink (Noctua L9i, for example), and throw it on that simulation software. Then, run the same tests you did with your unit, and see if the temp results compare to online benchmarks. This way you will know how precise your simulation is...

Good luck with the project ;)
Simulation is just a good referent, we will test with real one.
Interesting :-)
I believe you may have uploaded the 110W simulation two times?

Since it's CNC milled I assume that it is only a copper block?
So no heat pipes or vapor camber to transfer the heat to the open fin part (the part outside of the keep out zone, chipset area)?
Could you try running your simulations on the cooler that doesn't have that extra fin part (just the cooler that is within the 90x90mm keep out zone)

I'm just trying to compare this with Dynatrons offerings and what they are able to cool efficiently:
K129 is also just a copper block, a bit smaller but considerable more fins (35 vs 56) that also are a bit higher (490 grams with no fan).
And that is only rated for 95W TDP in a server solution with air blowing across the fins (as FullForceRainbow said).
https://www.dynatron.co/product-page/k129
More fins are better for higher airflow, whereas less fins can perform better in low airflow (and low pressure) situations.
As for the A9x14, remember that the 50 m3/t is in a free flow situation. When slapped on a heat sink that number will be lower due to the blockage from the fins. And the A9x14 doesn't have high static pressure. Not sure if you have incorporated that in the simulations. Also can it simulates the air stream?

EDIT: How big is the area of the heat source in your simulations?

And then there is the T318 that is a TDP 165W solution for socket 2011 (around same size but with more and higher fins). That is able to cool an undervolted 5820k CPU in the Dan A4-SFX reasonable, though with the A9x14 at max speed (Dondans test):

With no undervolt (~140W) the T318 is still able to cool the CPUs, check the forum for builds that uses these coolers.

The smaller R15 cooler (150W TDP according to Dynatron) is not able to efficiently cool a delidded I9 7980X (165 TDP) even with a Industrial 120mm noctua fan in a semi open bench setup:

The T318 was able to cool a 145W TDP Xeon E5-2699 v4 CPU with the A9x14 (with temps around 80C):
See temps a 7:38.

And these coolers utilize a vapor camber, which at least to my knowledge, is considerable better at "transporting" and distributing the heat from the base to the fins (and have more fins than your design (35 vs ~40).

All I'm trying to say is, that I have a hard time believing that a "small" pure copper block with a A9x14 @ 100% is able to cool 150W of heat in a 30C ambient room (according to your test parameters) to a CPU friendly level. And one thing is the temperature at the base of the cooler, another thing is the temperatures of the cores inside, as caniplaymayo also has concerns about.

Even more, I hardly believe it will be able to cool a 150W heat source to only 115C (in 30C ambient) with no fan (only airflow generated from the heat rising). I mean this big thing is only good for like 95W with high load temps after some time with no airflow (and in 21C ambient).

I would love to be proved wrong with a real life prototype, as I would love myself such a cooler.
But for now, I would advise you to consider removing the stated 150W TDP until you have a prototype that actual can cool 150W.

Best of luck :-)
We has already checked almost before design that one, we need a slim one <23mm, design for mount 92mm, work with low airflow.
Compare those models you provide, are for case fan not for mounting fan, their weight is much lower, fin is thinner, work better with extremely high air flow (sound like helicopter).
That why our heat sink (744g) simulation with no fan, temperature is 115C, but those one (433g) should be 146C.
This one design for intel K series 95W TDP (~110W electric power at 100%) work with more silencer fan, and design spare factor is good enough.
I will update simulation result when real one come.
Our design not only for A9x14, you can mount any 92mm fan on it, for example a delta 38mm 2.4A, and enjoy the supper power of a helicopter.
 

LjSpike

SFF Lingo Aficionado
Mar 20, 2017
140
72
CNC milling a heatsink like this might not be possible.

Pretty sure your typical CNC miller couldn't do this, however I wouldn't say it's impossible, or prohibitively costly, if the item is being mass produced. It'd really depend on the thickness of those fins and if they'd get teared off in milling. Extrusion would probably still be faster and/or cheaper though.
 

theGryphon

Airflow Optimizer
Jun 15, 2015
299
237
Confirm or elaborate:
1) You're applying heat on the heatsink through a small surface area where it makes contact with the CPU IHS. I assume the area you're using is the area of Intel 115X IHS.
2) You're measuring/reporting the highest temperature on the heatsink.
3) You assume the hot air will be exhausted out at 100% rate.

Assuming I'm correct with the assertions above (andI believe I am), you're not using the simulation to serve the purpose. Why? Because you're fatally ignoring...
1) The fact that the contact surface area of the heat source (the die itself) is much smaller than that of the IHS.
2) The fact that there is a TIM between the die and the IHS, and another TIM between the IHS and the heatsink.
3) The fact that what matters is the temperature on the surface of the die (aka Tdie, or Tjunction), not even the temperature on the IHS (aka Tcase), and forget about the temperature on the heatsink.
4) The fact that in this low-profile, top-down form factor, a lot -- and I stress, "a lot" -- of heat will be fed back to the heatsink through the air that bounces around the motherboard components, installed devices, and the PC chassis surfaces.

To correct your simulation and obtain more realistic results, I suggest you study the following picture and
1) use absolutely correct values for die area and IHS area
2) make accurate estimations on thermal resistance values for TIM and package (which includes the internal TIM and the IHS material), and incorporate them
3) model the operating environment: common motherboard components, installed devices, and either proposed or recommended PC chassis.
4) model the airflows and incorporate it's impact on heat transfer given the operating environment above
5) Use the silicon die as your heat source and report the temperature at the die (as pictured)



As a final note, without a vapor chamber, I do not believe you can attain a TDP support higher than 100W with this form factor, but I'd be very interested in seeing the results of the revised simulation.
 
Last edited:

caniplaymayo

Chassis Packer
Mar 28, 2018
15
9
Please wait to see what we should do with this "impossible" thing.

Awesome, does this mean you guys already have how it'll be manufactured all lined up/are close to a prototype for testing?

Simulation is just a good referent, we will test with real one.

I think the issue is the current simulations are giving unrealistic results. I like the design and want to see how it performs in real life, but I think you're setting people up for disappointment. That's not to say it will be bad, it just wont be at 64C w/ 110W. I think getting a more accurate picture could help you guys iterate the design before spending money on prototyping, as well as give people realistic expectations.
5) Use the silicon die as your heat source and report the temperature at the die (as pictured)

This would be the best way, but you can model server chips as a type of ceramic with internal heat generation to get pretty accurate simulation results with much less work. I'm not sure, but I believe it would also work for desktop cpus. If XCreator is interested, I can look up the exact material we used for these later today.
 

LjSpike

SFF Lingo Aficionado
Mar 20, 2017
140
72
but I believe it would also work for desktop cpus.

Be sure if it will or will not work, otherwise your making a product which you don't know will actually work, therefore the simulations were meaningless.

Or just simulate it as a silicon die heat source.
 

Fruergaard

Trash Compacter
Feb 13, 2018
37
50
We has already checked almost before design that one, we need a slim one <23mm, design for mount 92mm, work with low airflow.
Compare those models you provide, are for case fan not for mounting fan, their weight is much lower, fin is thinner, work better with extremely high air flow (sound like helicopter).
That why our heat sink (744g) simulation with no fan, temperature is 115C, but those one (433g) should be 146C.
This one design for intel K series 95W TDP (~110W electric power at 100%) work with more silencer fan, and design spare factor is good enough.
I will update simulation result when real one come.
Our design not only for A9x14, you can mount any 92mm fan on it, for example a delta 38mm 2.4A, and enjoy the supper power of a helicopter.

Yes, they are meant for server situation, and your design is pretty similar, no?
But if you see the links and videos I provided you can see those coolers used with a fan on top (like your design idea) and they have a hard time cooling 150W CPUs.

The weight of the cooler does very little in terms of heat dissipation to the surrounding air. Imagine two identical coolers, of same design and material, were only the weight differs. The higher weight cooler will be able to soak more heat before rising in temperature (less heat spikes), but after some time with continuous load it will reach the same temperature as the lower weight cooler.
Just because your cooler have a higher weight doesn't mean it is better at dissipating heat. If your simulation show that large of a steady state temperature difference (30C) between same design cooler with only weight difference, there is something very wrong!
It should be minimal otherwise then why not just make a large solid block of copper as your heat sink? I mean the weight could be 2 kg!

At steady state in a no airflow (no fan) situation the main thing of importance is surface area (considering same material coolers). Second thing is spreading the heat out over said surface area. Weight doesn't do shit here. Larger surface area = higher heat transfer to surrounding air = lower temperature. Again see that big chunk of metal (or this one) that have been designed for no fan situations. See that extreme amount of "free" surface area. And that is only for 95W CPUs! For 150W support it would need to be so much bigger.

Also thicker fins is not a selling point. It may help a tiny bit in internal heat transfer (meaning heat is transferred a bit further in the fins). But that advantage is minuscule compared to the potential loss of surface area compared to thinner but more fins.
Look at all the coolers out there. No one uses thick fins for a reason. It does not give you a good cooler.
Low airflow optimized heat sinks still uses thin fins, they just have a larger spacing between them (see the Silverstone HE02 linked above).

The main thing for a good cooler is surface area and getting heat spread out across that (and airflow but that is the fans job). Solid copper is a good thermal conductor, but it's not even close to heat pipes or vapor chambers. Again, look at what all other high end coolers use. Heat pipes to transfer the heat from the CPU out into the fins.
So I still don't see how your design can match that (or being better according to you?) of other similar designed coolers that have vapor chamber and larger surface area (T318 and R15)?

And for the fans and their airflow; 57 m3/h through the heat sink will not come silent with small fans. Remember that the stats supplied by the manufactures is in a "no obstruction" situation. Placing it on a heat sink will lower it a fair bit. So a good slim 92mm fan (which in general have low static pressure) will likely need above 4000 rpm to supply 57 m3/h through the heat sink. So if your simulations are with 57 m3/h going through the heat sink, you need a higher RPM fan than the A9x14.
Also, how is your airflow simulated? Remember that fans have a dead zone in the middle due to the hub.

Also theGryphon very nicely explained why I asked about the size of heat source in your simulations.
 
Last edited:

LjSpike

SFF Lingo Aficionado
Mar 20, 2017
140
72
Stumbled into a video by NFC showing a couple of mini-ITX coolers in action.


Thin fins. Heat pipes / vapour chamber. Noctua are the big boys when it comes to air cooling in my opinion so copy them bar their brown colour-scheme is a safe way to go.
 
  • Like
Reactions: XCreator

XCreator

Trash Compacter
Original poster
Apr 7, 2018
42
33
Confirm or elaborate:
1) You're applying heat on the heatsink through a small surface area where it makes contact with the CPU IHS. I assume the area you're using is the area of Intel 115X IHS.
2) You're measuring/reporting the highest temperature on the heatsink.
3) You assume the hot air will be exhausted out at 100% rate.

Assuming I'm correct with the assertions above (which I believe I am), you're not using the simulation to serve the purpose. Why? Because you're fatally ignoring...
1) The fact that the contact surface area of the heat source (the die itself) is much smaller than that of the IHS.
2) The fact that there is a TIM between the die and the IHS, and another TIM between the IHS and the heatsink.
3) The fact that what matters is the temperature on the surface of the die (aka Tdie, or Tjunction), not even the temperature on the IHS (aka Tcase), and forget about the temperature on the heatsink.
4) The fact that in this low-profile, top-down form factor, a lot -- and I stress, "a lot" -- of heat will be fed back to the heatsink through the air that bounces around the motherboard components, installed devices, and the PC chassis surfaces.

To correct your simulation and obtain more realistic results, I suggest you study the following picture and
1) use absolutely correct values for die area and IHS area
2) make accurate estimations on thermal resistance values for TIM and package (which includes the internal TIM and the IHS material), and incorporate them
3) model the operating environment: common motherboard components, installed devices, and either proposed or recommended PC chassis.
4) model the airflows and incorporate it's impact on heat transfer given the operating environment above
5) Use the silicon die as your heat source and report the temperature at the die (as pictured)



As a final note, without a vapor chamber, I do not believe you can attain a TDP support higher than 100W with this form factor, but I'd be very interested in seeing the results of the revised simulation.
I will send your information to my boss to refer, as I saw, each time do simulation his PC work like a PC form 1980.
Yes, they are meant for server situation, and your design is pretty similar, no?
But if you see the links and videos I provided you can see those coolers used with a fan on top (like your design idea) and they have a hard time cooling 150W CPUs.

The weight of the cooler does very little in terms of heat dissipation to the surrounding air. Imagine two identical coolers, of same design and material, were only the weight differs. The higher weight cooler will be able to soak more heat before rising in temperature (less heat spikes), but after some time with continuous load it will reach the same temperature as the lower weight cooler.
Just because your cooler have a higher weight doesn't mean it is better at dissipating heat. If your simulation show that large of a steady state temperature difference (30C) between same design cooler with only weight difference, there is something very wrong!
It should be minimal otherwise then why not just make a large solid block of copper as your heat sink? I mean the weight could be 2 kg!

At steady state in a no airflow (no fan) situation the main thing of importance is surface area (considering same material coolers). Second thing is spreading the heat out over said surface area. Weight doesn't do shit here. Larger surface area = higher heat transfer to surrounding air = lower temperature. Again see that big chunk of metal (or this one) that have been designed for no fan situations. See that extreme amount of "free" surface area. And that is only for 95W CPUs! For 150W support it would need to be so much bigger.

Also thicker fins is not a selling point. It may help a tiny bit in internal heat transfer (meaning heat is transferred a bit further in the fins). But that advantage is minuscule compared to the potential loss of surface area compared to thinner but more fins.
Look at all the coolers out there. No one uses thick fins for a reason. It does not give you a good cooler.
Low airflow optimized heat sinks still uses thin fins, they just have a larger spacing between them (see the Silverstone HE02 linked above).

The main thing for a good cooler is surface area and getting heat spread out across that (and airflow but that is the fans job). Solid copper is a good thermal conductor, but it's not even close to heat pipes or vapor chambers. Again, look at what all other high end coolers use. Heat pipes to transfer the heat from the CPU out into the fins.
So I still don't see how your design can match that (or being better according to you?) of other similar designed coolers that have vapor chamber and larger surface area (T318 and R15)?

And for the fans and their airflow; 57 m3/h through the heat sink will not come silent with small fans. Remember that the stats supplied by the manufactures is in a "no obstruction" situation. Placing it on a heat sink will lower it a fair bit. So a good slim 92mm fan (which in general have low static pressure) will likely need above 4000 rpm to supply 57 m3/h through the heat sink. So if your simulations are with 57 m3/h going through the heat sink, you need a higher RPM fan than the A9x14.
Also, how is your airflow simulated? Remember that fans have a dead zone in the middle due to the hub.

Also theGryphon very nicely explained why I asked about the size of heat source in your simulations.
You write too long.
You need to check Dynatron R25, surface and size... it's TDP is 160W, Air flow At Duty Cycle 100%: 22.28 CFM, also see some comparison video like ljspke provided to see more correct about each heat sink type.
Any way I also send your information to my boss to refer.

We got many unnecessary arguments when show concept without prototype, so some important designs we still keeping until make the real one.