Anticipating the Noctua R-ANC successor to NH-D14

[BonzaiDuck]

Trying to remember who posted the Jakob Dellinger interview recently, but for my failing memory -- thanks for the heads up.

If we only look on the web a little, we find more information:

http://www.noctua.at/main.php?show=news_list&news_id=90

Apparently, the "dual U14S" as I called it -- replacing the NH-D14, is called the R-ANC. "ANC" is a noise-cancelling technology for new Noctua fans. The company seems to be obsessed about noise, but then maybe enthusiasts are obsessed with noise. It looks as though they may release either 120 or 140x38mm fan -- in the usual colors.

However, this new cooler is something to inspire impatience. They have developed a new heatsink base for the R-ANC and probably other coolers that is made of copper and diamond composite. This boasts an improvement in thermal conductivity of 25%:

I've been a proponent of the IC Diamond thermal paste for a long time, despite some skeptics here. This was for obvious and simple scientific reasons: the low thermal resistance or high thermal conductivity of diamond.

So if they boast a 25% improvement in heatsink-base thermal conductivity, it will be interesting to see how that translates into the usual "delta" measurements between high-load temperatures and either idle or room ambient.

 

[4960X]

It was me who posted the interview. Here's the link in case other people want to see it.

http://www.relaxedtech.com/interview/jakob-dellinger/1

Also, it won't be called "R-ANC". The "R" stands for RotoSub which is the company behind the A.N.C (Active Noise Cancellation) technology which Noctua and RotoSub have been working together for a few years to incorporate A.N.C into their future coolers. I remember reading somewhere that it was confirmed from Noctua itself that it will be called the NH-D15.

 

[BonzaiDuck]

It was me who posted the interview. Here's the link in case other people want to see it.

http://www.relaxedtech.com/interview/jakob-dellinger/1

Also, it won't be called "R-ANC". The "R" stands for RotoSub which is the company behind the A.N.C (Active Noise Cancellation) technology which Noctua and RotoSub have been working together for a few years to incorporate A.N.C into their future coolers. I remember reading somewhere that it was confirmed from Noctua itself that it will be called the NH-D15.

Age is taking its toll of my memory, 4960X!!

Anyway, I thought the text on the Noctua page referred to the "R-ANC cooler." It was their lack of clarity. So -- "NH-D15."

Originally, we were just thinking that it was a "double-U14S." But if the NH-U14S doesn't incorporate the Cu-C-composite heatsink base, then that's like an "ace-in-the-hole." I just don't know what to expect. It's going to be better than the D14, though. With or without the heatsink-base innovation.

 

[krnmastersgt]

I didn't think they were actually going to release the ANC stuff for some time. I remember in a little interview with Linus that they mentioned the biggest issue is that if you don't completely correctly place everything with their ANC set-up then it actually just amplifies the noise that the cooler creates.

 

[BonzaiDuck]

I didn't think they were actually going to release the ANC stuff for some time. I remember in a little interview with Linus that they mentioned the biggest issue is that if you don't completely correctly place everything with their ANC set-up then it actually just amplifies the noise that the cooler creates.

You mean -- "if you don't use their fan?" I don't see how it would be any worse than what I've done with my D14 and a different fan. The new one -- without their fan -- is just heatpipe fins and pipes. Or maybe they'll release this "noise-cancelling" fan with a little beefier airflow.

Somehow, I think the Jakob Dellinger interview (posted in another thread here) mentioned a release date for the "D15" cooler of April. Other news suggested "Qtr 2 or 3" of 2014. I'm guessing at worst, we'd have to wait until mid-summer to see reviews on the cooler itself.

All the reviews are "non-modded." If the cooler comes with a fan, they use it. That's why, with the D14, I believe the ranking of the cooler is lower than it could be.

 

[4960X]

Here are two videos from Noctua's YouTube channel that show the difference A.N.C makes.

http://www.youtube.com/watch?v=AikMzJc9gcY

http://www.youtube.com/watch?v=-atLQYr14zI

 

[krnmastersgt]

They use the same anc technology that headphone companies use by creating a low frequency counter wave to dampen any noise from a specific area or direction. But if you get the placement slightly off with the cooler or there's something slightly wrong with the fan or the fins then you won't correctly counter the wave and instead create 2 sound waves that don't overlap correctly and make more noise.

 

[4960X]

They use the same anc technology that headphone companies use by creating a low frequency counter wave to dampen any noise from a specific area or direction. But if you get the placement slightly off with the cooler or there's something slightly wrong with the fan or the fins then you won't correctly counter the wave and instead create 2 sound waves that don't overlap correctly and make more noise.

If Noctua pulls this off with their future coolers, it will be big!

 

[BonzaiDuck]

If Noctua pulls this off with their future coolers, it will be big!

It may seem a minor point to make, but it is important to distinguish the cooler from the fan.

I cannot say how other enthusiasts regard the inevitable noise of a computer system. I'm willing to compromise -- accepting air-turbulence under loaded CPU conditions above, say, 55C.

There are ways to deal with vibration, including rubber rivets.

If they ship the "D15" with a fan, the user can accept to use it, or replace it. But the cooler -- by itself -- doesn't generate more noise than any other cooler with a different fan.

It would be great to get optimum airflow with near-noiselessness. But I'm more interested in the copper-diamond composite heatsink base, and the effect of "double NH-U14S" towers. In this latter regard, it will be useful to examine more carefully the difference between the U14S single-tower, and the photo of the "D15."

In both cases, there are exactly six U-shaped heatpipes. So the anticipated D15 is not really a "double-U14S," but the heatpipe equivalent of a single U14S with twice the number of U14S fins or fin surface-areas.

If there's a 25% increase in heat-conductivity due to the materials used in the HSF-base, it would augment the cooling advantage for what would otherwise be less than "two NH-U14S's."

All this is still speculation, even with some crude estimation approach to what we would expect in terms of temperature reduction.

Saying all that, it will definitely be "big" for the cooling and probably noise reduction as well. But for me, the most important thing is the cooling.

And it will be big because the CLC coolers so far have only offered a slight advantage over the best heatpipe coolers. The difference is more like 5C in comparison reviews, even though some may boast 10C.

People forego custom water-cooling because of expense, the maintenance, the extra effort required for installation, risk-aversion and so forth. If they try CLC coolers, they may be looking for the best of both worlds. But this new Noctua entry may leave that question difficult to decide.

It always boils down to the 50-year-old NASA white-paper I found some time ago: MTBF for heatpipes can be a million years. A fan can die in your computer case, but cooling continues to some acceptable degree. A pump fails -- no cigar.

Maybe the CLC makers will come back with more marginal improvements in response. Where would that lead? It won't lead to inclusion of a reservoir, unless the reservoir is still "closed." Maybe their coolers will use more coolant. Maybe the radiators will be larger. But the larger the radiators, the more troublesome it becomes to meet the "ease of installation" goal of a CLC for a wider variety of computer-cases. Then, it becomes more like a "water-cooling kit."

 

[ICD7]

Very interesting post and I think marks a turning point on limits of performance. over the last 15 years I have watched HS performance improvements almost on a monthly basis Aluminum to copper to more airflow to heat pipes etc. The end product here 6/12 heat pipes, 2 120mm fans and is the size of a toaster.

This is what they call the end of the S curve where heat pipe tech is maxed out and the thermal conductivity of the copper base is at the material limit of it's ability to transfer heat so a transition to a new S curve is in order with the adoption of new tech and materials with ever increasing effort to maximize performance.

For my edification I did the chart below from data taken at Frostytech from the top 100 of his tests. they are probably the most consistent of any source with the most easily accessible numbers.

As you can see from the top performing sinks on the left side of the chart you are sitting at about .07 C/W so about 7 degrees C above ambient @ 100 W.

On a break out of a thermal cascade you might see resistances of .02 C/W on the thermal grease, .02 on the base plate, and .03 C/W on the radiator.

Say you double the radiator size and now have 4 120mm fans your performance can only increase to .015 C/W or a 1.5 degree improvement on 100 W. On the performance development side you are climbing the ambient infinite wall of diminishing returns.

I believe details in mounting as regards to contact and pressure will become increasingly important. Decision points might be spending time lapping sink and IHS flat for a 5 C improvement or a 140mm fan for a 0.5C improvement.

 

[BonzaiDuck]

Very interesting post and I think marks a turning point on limits of performance. over the last 15 years I have watched HS performance improvements almost on a monthly basis Aluminum to copper to more airflow to heat pipes etc. The end product here 6/12 heat pipes, 2 120mm fans and is the size of a toaster.

This is what they call the end of the S curve where heat pipe tech is maxed out and the thermal conductivity of the copper base is at the material limit of it's ability to transfer heat so a transition to a new S curve is in order with the adoption of new tech and materials with ever increasing effort to maximize performance.

For my edification I did the chart below from data taken at Frostytech from the top 100 of his tests. they are probably the most consistent of any source with the most easily accessible numbers.

As you can see from the top performing sinks on the left side of the chart you are sitting at about .07 C/W so about 7 degrees C above ambient @ 100 W.

On a break out of a thermal cascade you might see resistances of .02 C/W on the thermal grease, .02 on the base plate, and .03 C/W on the radiator.

Say you double the radiator size and now have 4 120mm fans your performance can only increase to .015 C/W or a 1.5 degree improvement on 100 W. On the performance development side you are climbing the ambient infinite wall of diminishing returns.

I believe details in mounting as regards to contact and pressure will become increasingly important. Decision points might be spending time lapping sink and IHS flat for a 5 C improvement or a 140mm fan for a 0.5C improvement.

So does that mean you think the 25% improvement in conductivity of the base will matter much, or would it be bottlenecked by the number of pipes and cumulative area of the fins?

 

[coffeejunkee]

"Made from pure diamond"

Damn, this thing is going to be expensive, even for Noctua standards.

(ok, ok, pure diamond and copper powder)

 

[ICD7]

So does that mean you think the 25% improvement in conductivity of the base will matter much, or would it be bottlenecked by the number of pipes and cumulative area of the fins?

I can not say as I have not really looked to hard at Heat Pipes nor done any in depth analysis on them. My thermal cascade comparison was kind of a mental averaging from some thermosyphon and water block tests a couple of years ago.

In any event you only have about 6-7C margin in temp rise on a 100 W and you can not get around the material resistance losses. Just take the grease alone as fixed at .02 C/W and make the broad assumption that it is a 25% improvement assuming no resistance losses on the base & radiator assy. of .05 C/W and you end up with a 1.25 C improvement in performance.

Increases in efficiency at this point would probably be better utilized to reduce size and cost after all how much is 1 C worth?

Like I said I think this is a turning point and time to switch over to a new development S curve of which this powdered base is one way to go.

one idea might be a TEC on the top side of the base may/might work (or perhaps it would screw up the phase change process I have not thought it through) or a TEC assist on the radiator by linking the heat pipes and adding a TEC there maybe worth 5 C or maybe not worth doing at all.

Point is I think you will see an introduction of new technologies or combinations thereof. New materials perhaps diamond like graphite replacing heat pipes and base,fins?

It was only 15 years ago and I remember heat pipes costing $100 USD, Now perhaps a dollar or two. I think it will be interesting to see what comes down the pike.

One thing that caught my attention on the new Copper/Diamond base was that I was doing a phase change project with a rather complicated machined structure for the boiling surface and in a effort to reduce costs i tested some of the powdered metal configurations but abandoned them as the heat transfer was poor vs cast or rolled pure copper.

The obvious idea here is if I was a water block manufacturer I would be all over this as it probably costs just a dollar or two make without compromising performance in fact I would not be surprised if they were already in development in some manufacturers lab.

 

[BonzaiDuck]

I can not say as I have not really looked to hard at Heat Pipes nor done any in depth analysis on them. My thermal cascade comparison was kind of a mental averaging from some thermosyphon and water block tests a couple of years ago.

In any event you only have about 6-7C margin in temp rise on a 100 W and you can not get around the material resistance losses. Just take the grease alone as fixed at .02 C/W and make the broad assumption that it is a 25% improvement assuming no resistance losses on the base & radiator assy. of .05 C/W and you end up with a 1.25 C improvement in performance.

Increases in efficiency at this point would probably be better utilized to reduce size and cost after all how much is 1 C worth?

Like I said I think this is a turning point and time to switch over to a new development S curve of which this powdered base is one way to go.

one idea might be a TEC on the top side of the base may/might work (or perhaps it would screw up the phase change process I have not thought it through) or a TEC assist on the radiator by linking the heat pipes and adding a TEC there maybe worth 5 C or maybe not worth doing at all.

Point is I think you will see an introduction of new technologies or combinations thereof. New materials perhaps diamond like graphite replacing heat pipes and base,fins?

It was only 15 years ago and I remember heat pipes costing $100 USD, Now perhaps a dollar or two. I think it will be interesting to see what comes down the pike.

One thing that caught my attention on the new Copper/Diamond base was that I was doing a phase change project with a rather complicated machined structure for the boiling surface and in a effort to reduce costs i tested some of the powdered metal configurations but abandoned them as the heat transfer was poor vs cast or rolled pure copper.

The obvious idea here is if I was a water block manufacturer I would be all over this as it probably costs just a dollar or two make without compromising performance in fact I would not be surprised if they were already in development in some manufacturers lab.

ICD7 [assume you are J.C. or associate]:

I'm guessing that graphite has promise for being carbon and likely already proven in making composites.

I might also wonder at how silver might be used for the base in a similar composite like copper. Perhaps the extra expense made it less feasible, or passing on the true cost in the market might deter potential consumers for the price. Can't say.

Parts would be more expensive, not sure by how much.

 

[4960X]

All I know is that my wallet is ready!

 

[ICD7]

ICD7 [assume you are J.C. or associate]:

I'm guessing that graphite has promise for being carbon and likely already proven in making composites.

I might also wonder at how silver might be used for the base in a similar composite like copper. Perhaps the extra expense made it less feasible, or passing on the true cost in the market might deter potential consumers for the price. Can't say.

Parts would be more expensive, not sure by how much.

Silver is only 5% (?) better than copper.

Heat pipes in of themselves are pretty damn good and transfer heat 100X - 200X that of diamond. The limitation of course ends up being the containment structure material like copper and the interface resistances of base and fins to the pipe.

So in a perfect situation with a zero resistance the best you could do is -5C with the fixed compound resistance. You can not go below the ambient without some kind of sub cooling involved.

Add in the interface resistances and perhaps with diamond heat pipes you might hit an overall of .03-.04 on a hundred W.

Overall increased capacity would be stunning perhaps 1000W but at the lower end you are still climbing an infinite wall as far as ambient temp is concerned.

 

[BonzaiDuck]

All I know is that my wallet is ready!

:thumbsup::thumbsup:!! Now that I'm dropping the plan to build an X79 IB-E, I'm looking at water-cooling kits of the more traditional variety as opposed to CLC. Those kits are "expandable." But in "CPUS & OC'ing," I explain some gains I've made in both temperature and voltage for "optimal" settings. At this point, water-cooling gives me a potential temperature improvement of maybe ~15C. I just can't say how many millivolts reduction it would provide, and it might not be much.

So the anticipated "D15" is going to say something. It's going to cost less than a decent WC kit, even if the price-difference is only $40-$50.

Someone else mentioned among recent threads a desire to "experiment with heatpipes," asking where he could buy some. I'd imagine the longer the heat-pipe, the better its performance -- but I really wouldn't know. I'd think it wouldn't matter whether they only have two bends in them with straight segments piercing the fins, or they were regularly bent this way and that to make the cooler shorter. In that latter case, the mfgr'g process would cost a heck-of-a-lot more just to install a set of fins with the same number and area. If you could buy straight heatpipes of some length for <$5 each as IDC7 suggests, you'd want some sort of plumbing tool to bend them -- assuming that it wouldn't affect the wicking material inside the pipe.

 

[ICD7]

:

Someone else mentioned among recent threads a desire to "experiment with heatpipes," asking where he could buy some. I'd imagine the longer the heat-pipe, the better its performance -- but I really wouldn't know. I'd think it wouldn't matter whether they only have two bends in them with straight segments piercing the fins, or they were regularly bent this way and that to make the cooler shorter. In that latter case, the mfgr'g process would cost a heck-of-a-lot more just to install a set of fins with the same number and area. If you could buy straight heatpipes of some length for <$5 each as IDC7 suggests, you'd want some sort of plumbing tool to bend them -- assuming that it wouldn't affect the wicking material inside the pipe.

I think to make a heat pipe is not to hard - take a tube and coat the inside with a light oil & then apply a copper powder with large enough particles so it is porous. or use a screen or some metal wool

heat it up to sinter the particles together. oil burns off

clean the inside of both ends and tin both inside dia. with solder.

Normally they use a special crimping tool that cuts/crimps and welds at the same time.

In this case you heat one end so the solder is liquid & then with an anvil and hammer bang it flat a couple of times, sealing one end.

add water, then heat the piece so the water is boiling - you can measure with thermocouple.

Then whack the other end flat with the hammer.

when it cools you have a vacuum. viola` heat pipe

I made a couple one day on a lark, not too hard. inside ends where you apply the solder have to be really clean otherwise they leak and only last a couple of weeks.

I think you would bend the pipe then add wicking material

 

[PliotronX]

Holy cow, Noctua taking the innovative pace from Thermalright! I hope this results in substantial gains, if so I would expect all bases to be of the same composite makeup (unless it was a patented idea or something).

 

[BonzaiDuck]

Holy cow, Noctua taking the innovative pace from Thermalright! I hope this results in substantial gains, if so I would expect all bases to be of the same composite makeup (unless it was a patented idea or something).

Well, like ICD7 said, the water-block manufacturers should be all over it like flies on a dead bunny. First, we'll see some significant heatpipe improvement; then, the water kits will advance by the same advantage.

I'll be interested to see how the "D15" stands up against the CM Nepton 280L CLC cooler.

AFTERTHOUGHT: No doubt the composite is patented. You'd think they'd seek patents everywhere they intended to sell it. The patent might just address the process for making the composite. You couldn't restrict the use of raw materials in making a similar heatsink base. Hard to say. But it's a new idea, and I wonder why nobody thought about it before. I'd seen silver waterblocks, and I often wondered if they couldn't enhance a HSF base with micronized diamond -- beyond the nano-diamond TIM you can buy.

 

[ICD7]

Evidently the idea has been around for a while.

http://en.wikipedia.org/wiki/Dymalloy

And much researched

Characterization of various metal matrix composite thermal management materials

http://udini.proquest.com/view/characterization-of-various-metal-pqid:1863925311/

My guess is like the $100 heat pipe, research has brought the price down enough to make it viable for the mainstream.

so with normal copper at 396 W/mK the new composite thermal conductivity ends up above 500W/mK. 1/4 what you would get out of pure diamond.

a pretty good jump over silver and copper alone.

 

[BonzaiDuck]

Evidently the idea has been around for a while.

http://en.wikipedia.org/wiki/Dymalloy

And much researched

Characterization of various metal matrix composite thermal management materials

http://udini.proquest.com/view/characterization-of-various-metal-pqid:1863925311/

My guess is like the $100 heat pipe, research has brought the price down enough to make it viable for the mainstream.

so with normal copper at 396 W/mK the new composite thermal conductivity ends up above 500W/mK. 1/4 what you would get out of pure diamond.

a pretty good jump over silver and copper alone.

I suppose you could call it a "Star Wars" spinoff. So I would imagine they refined the process or otherwise followed the limitations (per molten copper, for instance) cited in the wiki.

The thermal conductivity numbers you cite are order of magnitude comparable to copper and silver the last time I looked at a table of materials for thermal conductivity. If I remember correctly, diamond was higher than 1,000 -- possibly close to 2,000. So . . . that sounds about right, I think.

 

[ICD7]

One idea I like is the Stirling engine

http://www.tweaktown.com/news/9051/msi_employs_stirling_engine_theory/index.html

allot of people have been playing around with this for a hundred years and recently I think some breakthroughs on performance with solar applications.

I would like to think that somebody is working on a combo phase change/Stirling to cool a CPU

You can find about 100k Stirling videos on youtube, people are pretty imaginative- I could watch them all day.

http://www.youtube.com/watch?v=z7lUzKfdHd4

 

[BonzaiDuck]

One idea I like is the Stirling engine

http://www.tweaktown.com/news/9051/msi_employs_stirling_engine_theory/index.html

allot of people have been playing around with this for a hundred years and recently I think some breakthroughs on performance with solar applications.

I would like to think that somebody is working on a combo phase change/Stirling to cool a CPU

You can find about 100k Stirling videos on youtube, people are pretty imaginative- I could watch them all day.

http://www.youtube.com/watch?v=z7lUzKfdHd4

Interesting. And it works. The only question there: how does that resolve our forecast-future, with die-shrinks, smaller area to transfer heat and so forth. It already seems apparent that the die-shrinks lead to quantum-physics considerations, and there's less tolerance for OC'ing with the new chips. There's more leakage. To get them above 4.4Ghz, people have tried "direct-die" HSF application, "de-lidding," all sorts of things. There doesn't seem to be much consensus: some people can tweak their IB and Haswell cores to 4.6+ -- others can't. We may be able to get a phenomenal decrease in load temperatures by de-lidding and using Liquid Ultra or Pro, but the decrease doesn't help as much as the adventurers would think once they've done it.

But I will build a Haswell-E next year, if it doesn't break my bank. They should still use indium solder to install the IHS. For that, I'd be at a crossroads for water-cooling, and it all depends on the successful reviews of a Noctua "D15."

 

[ICD7]

Well as far as heat sinks are concerned we are pretty close to the ambient air version to Moore's law. water may be the interim solution for overclocking.

but billions are being spent to resolve this issue, some solutions on the exotic side

http://www.margaretmorrisbooks.com/diamond_microcircuitry.html

research will eventually pan out on some path that makes sense