This has LTT video written all over it: In socket, no lid cooling

[Misanthrope]

The Ncore V1-naked die cooling waterblock has hit Kickstarter

 

https://twitter.com/OC3D/status/981560424259780608

So there's a kickstarter for this water block that took this TR4 style system to mount not the CPU but the cooler itself:

 

Quote

This is where the NUDEcnc Ncore V1 and V1D naked die liquid coolers come in, offering users the opportunity to cool their CPU dies directly to offer up to a 10% improvement in thermal performance over already delidded CPUs and up to a 38% improvement over non-delidded hardware.  

The Ncore V1 will release with two models, the V1 which is designed for delidded CPUs only and can be mounted directly to your CPU's IHS using a unique "in-socket" tool/screw free installation process. The V1D can be installed on delidded and non-delidded CPUs and comes with a more traditional CPU mounting bracket, which can also be used to delid LGA1151 CPUs. 

 

Now I will admit: I am mostly an (un)educated viewer here when it comes to this kinds of overclocks. But I actually do know enough to realize that delidding is already kind of a bad prospect that carries a lot of risks. Now if you go beyond that into naked die contact well you're increasing efficiently but multiplying risks. If on top of that you trust the tolerances of the flimsy 115x mounting system and this kick-starter seems to be the kind of thing Linus would love to dedicate a video to. In fact I'm not too sure he isn't secretly just trollingly making this kickstarter himself. Already on it I'm being told.

 

Source: https://overclock3d.net/news/cases_cooling/the_ncore_v1-naked_die_cooling_waterblock_has_hit_kickstarter/1

[tarfeef101]

About that....

(those who are fans of aquatic aviation will get it)

[Jurrunio]

probably worth a tiny bit of notice... Laptops and modern desktop graphics cards use this method for quite a long time already, I guess it's just new to the desktops?

[DocSwag]

It seems it has some supports that sit on the PCB as well. As long as you don't scratch the die it looks fine... one problem with direct die is you're putting tons of stress on the die itself, but with those supports on the side it might be ok.

 

Also it's copper so LM is an option . I just hope it doesn't cost too much...

[RedAV8R]

Hmm... I wonder... I bet something will come down the pipe soon enough.

[Misanthrope]

2 minutes ago, DocSwag said:

It seems it has some supports that sit on the PCB as well. As long as you don't scratch the die it looks fine... one problem with direct die is you're putting tons of stress on the die itself, but with those supports on the side it might be ok.

 

Also it's copper so LM is an option . I just hope it doesn't cost too much...

Ok that's fair, I guess this guys thought of, and I quote, "non-believers" like me:

 

 

https://www.kickstarter.com/projects/341758082/ncore-v1-naked-die-cooling-waterblock-designed-by/description

 

[NoxiousOdor]

I believe @Damascus already posted this

[Taf the Ghost]

Let's just hope Linus doesn't drop it, or splash all over the place in haste.

[Damascus]

9 hours ago, Misanthrope said:

Ok that's fair, I guess this guys thought of, and I quote, "non-believers" like me:

 

 

https://www.kickstarter.com/projects/341758082/ncore-v1-naked-die-cooling-waterblock-designed-by/description

 

Linus has a video up on floatplane talking about this, it beats a supremacy evo on a delidded 7700k by 8-10c (10% ish) making it one of, if not the best cpu blocks on the market.

 

9 hours ago, Taf the Ghost said:

splash all over the place in haste.

He dumps water all over the motherboard in the video on this thing

[Taf the Ghost]

2 minutes ago, Damascus said:

He dumps water all over the motherboard in the video on this thing

I've seen the video, that's the joke. Floatplane through the website itself doesn't show here.

[Blademaster91]

19 minutes ago, Misanthrope said:

Now I will admit: I am mostly an (un)educated viewer here when it comes to this kinds of overclocks. But I actually do know enough to realize that delidding is already kind of a bad prospect that carries a lot of risks. Now if you go beyond that into naked die contact well you're increasing efficiently but multiplying risks. If on top of that you trust the tolerances of the flimsy 115x mounting system and this kick-starter seems to be the kind of thing Linus would love to dedicate a video to. In fact I'm not too sure he isn't secretly just trollingly making this kickstarted himself.

The 115x mounting system is fine,as nice as Ryzen is i still prefer it to AMD's system which if the TIM acts as an adhesive you can rip the CPU out of the socket, though i'm assuming 115x isn't meant to be clamping down a copper block.  Securing that water block down to the bare die would still have me really nervous. Interesting to see how this works for extreme overclocking otherwise i don't see the point.

[Damascus]

2 minutes ago, Blademaster91 said:

The 115x mounting system is fine,as nice as Ryzen is i still prefer it to AMD's system which if the TIM acts as an adhesive you can rip the CPU out of the socket, though i'm assuming 115x isn't meant to be clamping down a copper block.  Securing that water block down to the bare die would still have me really nervous.

The stock IHS is a copper block  this just replaces it for top tier thermals

[Blademaster91]

8 minutes ago, Damascus said:

The stock IHS is a copper block  this just replaces it for top tier thermals

True, nickel plated copper although seeing that water block move a bit as its being shut into the socket seems sketchy to me.

[Damascus]

1 minute ago, Blademaster91 said:

True, nickel plated copper although seeing that water block move a bit as its being shut into the socket seems sketchy to me.

Fair enough, though a deluded gpu will have the same wiggle if you don't glue it back together (best performance)

[Notional]

Ah yes, the good old days, where installing the CPU cooler a little too hard on one corner, would crack the cpu core, rendering it useless (this is also one of the reasons I got a Pentium 4 Northwood rather than an Athlon back in the days):

 

 

And all because Intel is greedy using mud found in the backyard as tim.

[Misanthrope]

13 minutes ago, Notional said:

Ah yes, the good old days, where installing the CPU cooler a little too hard on one corner, would crack the cpu core, rendering it useless (this is also one of the reasons I got a Pentium 4 Northwood rather than an Athlon back in the days):

 

-Athlon CPU pic was here-

 

And all because Intel is greedy using mud found in the backyard as tim.

What intrigues me is that they obviously tests them all to bin them. So while they're binning, why not designate the non-k parts for mud and the parts they plan to sell as K chips to be fitted with an actually decent compound?

[Taf the Ghost]

8 hours ago, Notional said:

Ah yes, the good old days, where installing the CPU cooler a little too hard on one corner, would crack the cpu core, rendering it useless (this is also one of the reasons I got a Pentium 4 Northwood rather than an Athlon back in the days):

 

And all because Intel is greedy using mud found in the backyard as tim.

I remember when people used to convex mirror polish their CPU coolers to get even better thermal performance.

[Blademaster91]

9 minutes ago, Misanthrope said:

What intrigues me is that they obviously tests them all to bin them. So while they're binning, why not designate the non-k parts for mud and the parts they plan to sell as K chips to be fitted with an actually decent compound?

The locked i7 8700 has a 4.6 turbo so i'd be surprised if they got binned that well. Intel likely figures the mud works fine for 95% of people but must know by now that people delid, soldering the chips can't cost much more but i'd pay another $10-20 for a soldered IHS.

[Coaxialgamer]

Ooh , i like this

 

But i do see a few problems though :

-changing CPU requires a loop drain ( although that's not necessarily much of an issue because you'd never get a bad mount or need to change TP )

-leaks ?

-What about the capacitors and other stuff around the die ? The'd need to be properly insulated , lest they be shorted by the coolant.

 

8 hours ago, Misanthrope said:

What intrigues me is that they obviously tests them all to bin them. So while they're binning, why not designate the non-k parts for mud and the parts they plan to sell as K chips to be fitted with an actually decent compound?

It seems that while the paste used is certainly sub-par , a non-insignificant part of the problem us caused by the z-axis height , meaning contact isn't ideal .

Deliding has the benefit of removing the glue between the IHS and package.

 

8 hours ago, Blademaster91 said:

The locked i7 8700 has a 4.6 turbo so i'd be surprised if they got binned that well. Intel likely figures the mud works fine for 95% of people but must know by now that people delid, soldering the chips can't cost much more but i'd pay another $10-20 for a soldered IHS.

The problem isn't so much the cost of solder . It's not really feasible to not solder the mainstream chips  . The S dies are some of the smallest intel has had in decades . Even with 6 cores , the 8700k is barely pushing 150mm² ( smaller than broadwell or ivy bridge ). You simply can't use solder for chips that small , 200mm² is the tipping point. The area isn't large enough for the chip to properly dissipate the heat associated with the process .

But there is absolutely no excuse for skylake X though...

[Notional]

27 minutes ago, Misanthrope said:

What intrigues me is that they obviously tests them all to bin them. So while they're binning, why not designate the non-k parts for mud and the parts they plan to sell as K chips to be fitted with an actually decent compound?

I believe they are lidded before they get binned. Or maybe that was just intels bs excuse for using cheap mud as tim.

[AHaskin14]

So am I really the first floatplane member to say that this video is already out and on its way to youtube?

[Ginger_]

48 minutes ago, AHaskin14 said:

So am I really the first floatplane member to say that this video is already out and on its way to youtube?

Nah, it's been mentioned a couple times already 

[RotoCoreOne]

1 hour ago, AHaskin14 said:

So am I really the first floatplane member to say that this video is already out and on its way to youtube?

Spoilers yo...

[Bananasplit_00]

3 hours ago, Jurrunio said:

probably worth a tiny bit of notice... Laptops and modern desktop graphics cards use this method for quite a long time already, I guess it's just new to the desktops?

"new"(also extremely old, before the IHS i think there were some on die coolers) to those that arent extreme overclockers, yes

3 hours ago, Damascus said:

Linus has a video up on floatplane talking about this, it beats a supremacy evo on a delidded 7700k by 8-10c (10% ish) making it one of, if not the best cpu blocks on the market.

ok well there we go, im getting one. screw this EK shit, never liked this block anyway, build is crap and so are the temps.

2 hours ago, Coaxialgamer said:

Ooh , i like this

 

But i do see a few problems though :

-changing CPU requires a loop drain ( although that's not necessarily much of an issue because you'd never get a bad mount or need to change TP )

-leaks ?

-What about the capacitors and other stuff around the die ? The'd need to be properly insulated , lest they be shorted by the coolant.

1: not with soft tube

2: yah, like with all watercooling this is an issue

3: i never bothered when i delided(yah pretty bad idea tbh) so this is probably even more fine

1 hour ago, AHaskin14 said:

So am I really the first floatplane member to say that this video is already out and on its way to youtube?

no, Damascus beat you

[Coaxialgamer]

7 minutes ago, Bananasplit_00 said:

1: not with soft tube

2: yah, like with all watercooling this is an issue

3: i never bothered when i delided(yah pretty bad idea tbh) so this is probably even more fine

 

-The cpu die will litterally be surrounded by water , and the block is held on by the socket mounting . Soft tubing doesn't change much ; if you try to remove the cpu the water will just spill out

- Agreed , but this is a different concern .I'm concerned because of the retention system . a normal loop is supposed to be completely sealed , it's a closed system . The only seal here is pretty much the pressure between the package and cooler .There's no roper seal.

-The coolant will likely be water-based , so conductive . Isolation when deliding is only needed when using a conductive paste ( metal ) , normal tim isn't an issue . If water touches those resistors/caps , they WILL be shorted , as they are within the cooling chamber