I only now found out you have made an english channel! I guess uploading EN and DE videos on single channel have been knocking you down in YT rankings... I greatly appreciate your dedication to making both versions of your videos though!
test the arctic Freezer 50 TR. It can handle ~450W-500W. The downside is you need standard size memory (for example, kingston hyperx without RGB or any thing).
I remember when Linus reviewed this cooler, he used a Threadripper system, not an Intel system. Because the Threadrippers have multiple dies under the heat spreader, and the cooler's Ice Giant's contact area is larger, the heat is transferred better from that CPU to the cooler's heat pipes.
yeah that's why he said linus's review was excellent. A lot of people sleep on LTT's reviews in favor of more standardized testing like say from Gamers Nexus (whom is amazing, no one is gonna say tech jesus is anything less.) But they're fairly accurate and thorough.
I like how when you do a teardown, you're not messing around. I don't see any other tech youtubers cutting through their brand new, extremely expensive (and even overpriced) cooler just to see exactly what's inside. Cheers man, keep it up. You're my new favourite tech guy.
What's more impressive is that he managed to put it back together so he could run a threadripper test at the end of the video, too. His skills are beyond compare.
Ehh, I wouldn’t say overpriced. It’s a small batch niche product that not everyone is going to want, need, or use. It’s kinda like small batch group buys for custom keyboards. Sure, you could call an $800 piece of milled and anodized aluminum overpriced, but when you look at the cost of production, design, advertising and UAE, that’s just what it’s gonna cost
Wendell from Level1 Techs did an interview with James prior of IceGiant few months ago about how this cooler was designed and details about the refrigerant used and more
hey man. discovered your channel last night watching some older Gamer's Nexus vids. Just want to say it's awesome to find another reviewer with similar attention to detail and objectivity. Liked and subbed, looking forward to more vids!
Hate to be that guy, but a thermospyhon/thermosiphon is in fact just a wickless heatpipe that relies on gravity for condensate return; think old-school single-pipe steam heating. Higher efficiency iff operating straight up and down, cheaper, and you can make them wider/longer more easily. Minding the iff, true thermosyphons must be [mostly] vertical. Horizontal, it's tossing nearly all of its transfer capacity out the window and relying substantially on conduction through the copper. I think the confusion here is that some HVAC people do call a simple density-based convective flow system a thermosiphon (I had to go look it up), but I wasn't even aware of that till you mentioned it, and I've been building the former version of thermosyphons for PCs for years. Loved seeing this thing chopped open nonetheless! Mucho kudos. [EDIT: BTW, I'd love to see a few other parts of the assembly not assembled... I'd buy it 'slightly used' ;-]
I don’t have access to any of the company’s data, but I’ve cooled a lot of different things with different materials in my career and the difference between aluminum and copper is usually close to the margin of error. It’s usually other factors that will amplify the cooling performance differential between the two materials. My theory, and it’s solely based on YT videos, is that the inside structure is ‘tuned’ to flow better with larger heat loads (bigger contact patch). If they really want to optimize performances on smaller contact patch, mainstream CPUs, they might need to release an optimized version. And I might be totally be wrong, as I can’t test any of the assomptions. Really like that video, kinda felt like you were going back to your roots.
In general yes, but contact surface size and contact material are critical, bigger contact sufrace gives bigger heat transfer, but if you use silicon thermal paste to connect two aluminium plates it will give weak heat conductor, If you use silver based thermal paste to connect two CU plates it will give better heat transfer than alu-paste-alu. Not as good as metal to metal connection - soldered or casted. I think in this case ALU is not sufficient for 250W, you can't cheat physics, CU is 70% better heat conductor. They should use CU head and pipes, and ALU fins.
You know, looking at the internal configuration, the reason why you were seeing rather average results may be from the orientation of the cooler. If the cooler was mounted on a vertical motherboard then those big holes would have been at the top end of the hotplate and the smaller holes would have been further down, you would see the gas going through the big holes at the top and the liquid coming back through the smaller holes leading to a circulation of sorts. Linus from LTT did run into a similar issue with some other similar cooler which was basically a AIO without a pump and having the orientation correct was the difference between instant overheating and somewhat decent cooling.
I'd say it was a premature pressure relief :D Would have liked to see it tested with ryzen packages too, where heat sources are much more spread out compared to intel's monolithic layout.
This is a very good result! Keep in mind this 'new tech' in the CPU cooling scene is compared here to heat pipes and AIOs, which are both extremely optimized to these applications
Did you lap the prosiphon cold plate? I have one of the early (pre-ordered) ice giants and mine has a brushed cold plate. Very deep grooves and it is nowhere near shiney, let alone the almost mirror look that yours has. It's a decent cooler nonetheless, but I was never getting the performance I thought it should. I'm wondering if I should lap the cold plate to an actually flat surface, or if it would make a difference at all. Thank you for the detailed video der8auer, you put in an immense amount of work to shoot these videos in both German and English and the quality never drops in the slightest.
It needs to have the mount point to be the lowest point, so the condensing liquid flows down to the CPU for evaporation. So both the horizontal mount on test bench and vertical mount with radiator above CPU in case will work.
It can be use only two orientations and horizontal in test bench is one allowed. If it is used in case, condenser must be above the evaporator, because it requires gravity.
Glad to know what its inside looks like! I own the ProSiphon Elite and it is not the best performing heatsink but still in the top range. It has one advantage which is why I got it. That it is the only high end air cooler (on market, when I purchased it) that blows air vertically and built for workstation platforms like Thread Ripper or Intel i9 X's (assuming common installation orientation). So with a good chassis air flow setup, you can just run two fans on it and makes a whisper quiet workstation. Which also makes maintenance a lesser issue especially when a few years down the road. But anyways, consider the price tag, I just cannot recommend it.
Sadly there is a scarcity of good Threadripper coolers. There are no proper TR-optimized AIOs on the market for example - I know Cooler Master has one with the right sized coldplate, but the microfins themselves are standard, and every version of the Enermax Liqtech TR4 has bacteria problems, there isn't even a NH-D15 for TR, only U14S-TR4-SP3, which isn't as massive. I'm still running a Kraken X62 on my 1950X, but the IceGiant Prosiphon is on my roadmap. That, or I'll finally bite the bullet and build a custom loop.
yep, still going to recommend Noctua for any friend building a PC. fantastic performance, support, and longevity. but always good to see new options providing market competition.
1:50 the second time he says evaporate should be "condense" (at the top of the pipe, when the liquid cools). I'm sure Roman knows this and was likely mis-spoken due to translation. That said this AIO looks super interesting and would love to try one if they go wide release, great video as always
I was surprised to see how much structure the base plate and tubing had inside. It's quite complex. Cast aluminum alloy usually is lower conductivity than pure aluminum, so a copper version might handle hot spots better. A wick is what made heatpipes different from thermo-siphons. They work without gravity, in space, that way. So this is a very complex thermo-siphon, but still no capillary action.
Was great seeing you open that big boy up. Fin stacks absolutely everywhere but like you say it would be interesting to see a version with some copper components.
I have the Ice Giant Prosiphon Elite on my 2700x and it works really well. Now I know this is a older chip, but I have a nice overclock on it . typically I idle around 40c, and when testing It usually works its way up to 60. I do agree that this cooler is definatly meant more for am4 and threadripper. I call it the mobo hider lol.
I preordered this cooler after the first LTT video. From what I understand it's best suited for HEDT chips. It kept my 9900k tame at 5.0 GHZ but it was nowhere near what I was hoping. Went custom water cooling and I'm not looking back. I have the Prosiphon in the closet as a spare cooler in case of a pump failure....etc.
I had one of these on pre-order. I am so glad I went with the NH-D15 after they failed to meet their deadline rollout. As far as it being liquid cooled I was told if there was any leaks to occur the liquid they are using would immediately turn into a gas once it comes in contact with air. They designed it that way to prevent motherboard/gpu related hardware damage. I do agree copper would be a much better choice.
@@obedulloa6219 I stuck with the preorder until the second target delay date then cancelled. I could not wait any longer that's when I decided to buy a NH-D15. I was not penalized for it either I was issued a full refund. I don't regret it.
where is your own cooler that linus did a video of when you were at that large computer show ? that thing looked really awesome , can we have an update about that?
A couple things I am curious about….. if you tested it with it sitting straight up rather than laying down would that have made it perform better?? 2. It may be underperforming due to the fact this is built for thread ripper and intel ihs’s concentrate heat differently so it may just work a lot better with a larger amd ihs.
I think calling it a “thermo syphon” is still valid, since it appears they intend to have circulation in DIFFERENT tubes, unlike heat pipe which is in the same tube. Furthermore, heat pipe have wick in side the tubes, which relies on capillary effect to transport condensed fluid back, the Icegaint didn’t have wick, it rely on density difference to transport the fluid back (which is exactly how thermo syphon works, circulation due to density differential caused by temperature)
I'm not eng but the fins on cold plate help heat distribution to heat up that liquid gas (if it smelled like strawberry its most probably cooling gas like 134a). then the gas would go through bigger hole for lower resistance, and descend in small holes. I was looking closely and it would only could work it the radiator its in a slight angle from the cold plate for the liquid circulate by gravity.
Even the Threadripper 3960X has a 280W TDP. To cool it with an air cooler and have it run quiet, there is no overclocking headroom, and you may even need to undervolt it, especially during the summer. I'm currently running a 3970X (same TDP), a custom loop and some undervolting that I did during the summer to keep the room a bit cooler.
A bit late at this point, but might the cooler orientation (and thus direction of gravity) have affected results? I would assume this is designed around use on a vertical motherboard. Of course this might not matter, but it would have been interesting to see.
Gravity returns do rely on orientation, Icegiant make this very clear in their manual, so with a vertical motherboard the fin array needs to be above the socket, but any horizontal orientation should work just as well because the liquid can still evaporate upwards and condense downwards.
This cooler does better on IHS with spread out chip set. not the center hot spot intel cpus.. one of the things they were going to address. imho a copper plate might help some.
Perhaps a crazy idea but how about installing a motor to act as an air pump with a pressure plate fan (blower or regular fan) inside the heat piping for faster heat delivery?
As a gravity fed unit, wouldn't laying it flat like that kind of ruin the capability. Don't the instructions say to have it orientated in a specific way so the gravity parts work properly? Wouldn't that position be in a standing side mount like how most computers are situated?
G'day Roman, that was really cool cutting it open to see how it works inside, I wasn't expecting that huge pressure release 😲, "it's just Fins, Fins, Fins, Fins everywhere" 🤔 it's like a holiday in Finland 😂
Seems it needs to be on a vertical motherboard instead of laying on a bench. heat going on upper tubes, cool down and returning on the lower tubes. With your setup, the flow can be stopped.
Did you only test in the horizontal mode? Site also shows a vertical mode. Would have liked that data really. Seems the fluid path has lower energy requirements to flow in vertical.
after seeing the contact plate internal structure i can see why they went with aluminum over copper otherwise that cooler would likely be in the 250-350 dollar range since it's a niche product.
What if Noctua took the basic NH-D15 design, but moved the two heatsink stacks NEXT to each other for a wider, thinner cooler? Some of the heatpipes would be LONGER and thus less efficient, but you also wouldn't be blowing the hot air from the first stack through the 2nd stack.
my watercooling loop failed, am being given one of these coolers(not sure which revision but its from their kickstarter, so early I assume), just need to pay for shipping. its just a 5700x, so I assume itll be plenty, just not sure its going to clear my TeamGroup XTREEM ARGB modules
its work like AC units so the motherboard should mount vertical like in the cases so the cooler works ( the cycle in it ) that's why its didn't worked here
Just asking out of pure memory here...I may be wrong, but aren't those coolers meant to be used horizontally?...I mean, it might be the reason for those poor results.
Yes i remember that. This cooler is designed for AMD in mind and not intel as far a s the dies go. I think this was aimed at thread ripper or a minimum 12 to 16 cores.
I think, IceGiant could be good for AlderLake OC also (or a little better than for LGA11xx/1200). There is a little higher size of flat. But propably I will waiitng for copper version of Elite cooler. Still, 360 Ryujin AIO or very solid Artic Cooling Liguid II 280 are fine on my benchtables.
@@sirmonkey1985 oh just rewatched their installation tutorial, there is a wrong way of installation when in a normal case (the condenser (longer part) has to show down, thats what it was - still strange the temps were so high
@@aleloAT because its an aluminum plate on a hot spot IHS like intels. this plate was originally designed for TR spread out chip set. a copper plate will help with the hot spot but the fins under the plate will limit heat transfer unlike the NHD15 with the heat pipes in the plate, closer to the IHS hot spot. that cooler will be better in this case..
Would it help to set load line calibration for gaming scenario with Siphon cooler? Also i remember someones video that the cooler needs to be in certain orientation for it to work.
I feel like if this had an F-gas inside it should come with environmental warnings on the box, just like an AC unit would. At least in the EU there are pretty strict rules regarding F-gasses.
Didn't you put it on backwards? From what I remember in the LTT videos this cooler has to be put in the right way to work at all. With the motherboard mounted like inside the case and the fans down towards where the GPU is.
I’m a little confused why you didn’t know what fluid is in there and why you’re surprised on lackluster performance on an Intel chip… That info is pretty well available online. I was very happy to see the inside of it tho, haven’t seen that one yet.
"I'm 99% sure that 99% of you know how heat pipes work"... (in his thoughts: And I'm 99% sure that 99% of you don't know the internal pressure of a heat pipe :p )
There's a large overlap between der8auer viewers and GamersNexus viewers, and GN Steve has explained it many times. And to add to it all of us are computer nerds.
smaller contact on the condenser side makes it perform worse. if you add a graphite thermal pad it may even perform better than with paste due to the high planar thermal conductivity, but even without a pad it should perform better on AMD CPUs.
WHY should that thing be better than an AIO? It has less Radiator Surface Space to get rid of the Heat. It has less volumen of Fluid inside than the AIO. And why would gravity based be better than pump based? Furthermore you can mount the Radiator of the AIO wherever you want. Put it to the front and you get cooler air from the outside. This one has the Radiator inside the Case like a normal tower cooler. So far, for me this looks like just a bigger Tower Air Cooler . . An AIO should be superior to this right?
Yes. This is about reliability, where you would be worried of a pump failure. So this works great where limited access/up time is more important than performance.
The radiator on it is almost twice as thick as an AIO radiator, so there's plenty of surface area. You don't need liters of water to do phase change cooling.
And if the aio's pump dies , it just dead... Its not like custom watercooling , pump faliure still a thing... but in the other hand depends on gravity...
I only now found out you have made an english channel! I guess uploading EN and DE videos on single channel have been knocking you down in YT rankings... I greatly appreciate your dedication to making both versions of your videos though!
Thanks!
I have the noctua threadripper cooler it can't handle 320W.
Good morning
@@Ayane13b you like boats?
Are u bragging on cooler size ? xD
@@ahoymateeez Would it be bragging if hes saying it CANT handle 320w??
test the arctic Freezer 50 TR. It can handle ~450W-500W. The downside is you need standard size memory (for example, kingston hyperx without RGB or any thing).
"This is going on eBay, slightly used" lmao x)
That was a great line lol.
I genuinely laughed out loud at that 🤣
Seller refurbished
more like brand new still
x demo
I remember when Linus reviewed this cooler, he used a Threadripper system, not an Intel system. Because the Threadrippers have multiple dies under the heat spreader, and the cooler's Ice Giant's contact area is larger, the heat is transferred better from that CPU to the cooler's heat pipes.
good point
yeah that's why he said linus's review was excellent. A lot of people sleep on LTT's reviews in favor of more standardized testing like say from Gamers Nexus (whom is amazing, no one is gonna say tech jesus is anything less.)
But they're fairly accurate and thorough.
@@ExarchGaming Gosh this did not age well lol
I like how when you do a teardown, you're not messing around. I don't see any other tech youtubers cutting through their brand new, extremely expensive (and even overpriced) cooler just to see exactly what's inside. Cheers man, keep it up. You're my new favourite tech guy.
What's more impressive is that he managed to put it back together so he could run a threadripper test at the end of the video, too. His skills are beyond compare.
Gamers Nexus, Steve doesn't mess around either.
Ehh, I wouldn’t say overpriced. It’s a small batch niche product that not everyone is going to want, need, or use.
It’s kinda like small batch group buys for custom keyboards. Sure, you could call an $800 piece of milled and anodized aluminum overpriced, but when you look at the cost of production, design, advertising and UAE, that’s just what it’s gonna cost
been using my Noctua D15 for three builds in a row. Longest lasting PC component I've ever purchased.
It's literally just a hunk of metal
I have one too and i love it. Centered my whole build around it. No hate
I put D15 in a Y60 so I could just look at it
Wendell from Level1 Techs did an interview with James prior of IceGiant few months ago about how this cooler was designed and details about the refrigerant used and more
hey man. discovered your channel last night watching some older Gamer's Nexus vids. Just want to say it's awesome to find another reviewer with similar attention to detail and objectivity. Liked and subbed, looking forward to more vids!
Thanks man 😬💪🏼
"This will go on Ebay. Slightly used." You got me man. Thanks for the laugh.
ya that was good lol
Hate to be that guy, but a thermospyhon/thermosiphon is in fact just a wickless heatpipe that relies on gravity for condensate return; think old-school single-pipe steam heating. Higher efficiency iff operating straight up and down, cheaper, and you can make them wider/longer more easily. Minding the iff, true thermosyphons must be [mostly] vertical. Horizontal, it's tossing nearly all of its transfer capacity out the window and relying substantially on conduction through the copper.
I think the confusion here is that some HVAC people do call a simple density-based convective flow system a thermosiphon (I had to go look it up), but I wasn't even aware of that till you mentioned it, and I've been building the former version of thermosyphons for PCs for years.
Loved seeing this thing chopped open nonetheless! Mucho kudos.
[EDIT: BTW, I'd love to see a few other parts of the assembly not assembled... I'd buy it 'slightly used' ;-]
Once the saw went in, I started to doubt that this particular one would be your next bench cooler.
I don’t have access to any of the company’s data, but I’ve cooled a lot of different things with different materials in my career and the difference between aluminum and copper is usually close to the margin of error. It’s usually other factors that will amplify the cooling performance differential between the two materials. My theory, and it’s solely based on YT videos, is that the inside structure is ‘tuned’ to flow better with larger heat loads (bigger contact patch). If they really want to optimize performances on smaller contact patch, mainstream CPUs, they might need to release an optimized version. And I might be totally be wrong, as I can’t test any of the assomptions. Really like that video, kinda felt like you were going back to your roots.
In general yes, but contact surface size and contact material are critical, bigger contact sufrace gives bigger heat transfer, but if you use silicon thermal paste to connect two aluminium plates it will give weak heat conductor,
If you use silver based thermal paste to connect two CU plates it will give better heat transfer than alu-paste-alu. Not as good as metal to metal connection - soldered or casted.
I think in this case ALU is not sufficient for 250W, you can't cheat physics, CU is 70% better heat conductor. They should use CU head and pipes, and ALU fins.
Ty for taking the time to record these in english man, your a legend :D
I woke up and just smoked one. I escaped into this explanation for a bit. Thank you. It was awesome!
So glad you've opened it :D
holy crap, that Datron CnC is a $200k+ machine!!! Nice work if you can get it.
You know, looking at the internal configuration, the reason why you were seeing rather average results may be from the orientation of the cooler. If the cooler was mounted on a vertical motherboard then those big holes would have been at the top end of the hotplate and the smaller holes would have been further down, you would see the gas going through the big holes at the top and the liquid coming back through the smaller holes leading to a circulation of sorts.
Linus from LTT did run into a similar issue with some other similar cooler which was basically a AIO without a pump and having the orientation correct was the difference between instant overheating and somewhat decent cooling.
I'd say it was a premature pressure relief :D
Would have liked to see it tested with ryzen packages too, where heat sources are much more spread out compared to intel's monolithic layout.
This is a very good result! Keep in mind this 'new tech' in the CPU cooling scene is compared here to heat pipes and AIOs, which are both extremely optimized to these applications
@Der8auer, wasnt it designed to be run sideways so the fans blow upward? If not the pumping action may not work.
It can be put horizontally and vertically so it should work that way too
Did you lap the prosiphon cold plate? I have one of the early (pre-ordered) ice giants and mine has a brushed cold plate. Very deep grooves and it is nowhere near shiney, let alone the almost mirror look that yours has.
It's a decent cooler nonetheless, but I was never getting the performance I thought it should. I'm wondering if I should lap the cold plate to an actually flat surface, or if it would make a difference at all.
Thank you for the detailed video der8auer, you put in an immense amount of work to shoot these videos in both German and English and the quality never drops in the slightest.
01:35 Temperature of evaporation must be set higher than ambient temperature (it is ambient temperature on cooler part what cause condensation).
I thought this cooler required to be in certain position to work, vertical (like in a case), not a test bench.
yes
yep
It needs to have the mount point to be the lowest point, so the condensing liquid flows down to the CPU for evaporation. So both the horizontal mount on test bench and vertical mount with radiator above CPU in case will work.
According to the manual and homepage it is designed to use in both orientations
It can be use only two orientations and horizontal in test bench is one allowed. If it is used in case, condenser must be above the evaporator, because it requires gravity.
Glad to know what its inside looks like!
I own the ProSiphon Elite and it is not the best performing heatsink but still in the top range.
It has one advantage which is why I got it. That it is the only high end air cooler (on market, when I purchased it) that blows air vertically and built for workstation platforms like Thread Ripper or Intel i9 X's (assuming common installation orientation). So with a good chassis air flow setup, you can just run two fans on it and makes a whisper quiet workstation. Which also makes maintenance a lesser issue especially when a few years down the road.
But anyways, consider the price tag, I just cannot recommend it.
Sadly there is a scarcity of good Threadripper coolers. There are no proper TR-optimized AIOs on the market for example - I know Cooler Master has one with the right sized coldplate, but the microfins themselves are standard, and every version of the Enermax Liqtech TR4 has bacteria problems, there isn't even a NH-D15 for TR, only U14S-TR4-SP3, which isn't as massive. I'm still running a Kraken X62 on my 1950X, but the IceGiant Prosiphon is on my roadmap. That, or I'll finally bite the bullet and build a custom loop.
yep, still going to recommend Noctua for any friend building a PC. fantastic performance, support, and longevity. but always good to see new options providing market competition.
1:50 the second time he says evaporate should be "condense" (at the top of the pipe, when the liquid cools). I'm sure Roman knows this and was likely mis-spoken due to translation. That said this AIO looks super interesting and would love to try one if they go wide release, great video as always
I was surprised to see how much structure the base plate and tubing had inside. It's quite complex. Cast aluminum alloy usually is lower conductivity than pure aluminum, so a copper version might handle hot spots better.
A wick is what made heatpipes different from thermo-siphons. They work without gravity, in space, that way. So this is a very complex thermo-siphon, but still no capillary action.
Тепловые трубки точно так же зависят от гравитации
Was great seeing you open that big boy up. Fin stacks absolutely everywhere but like you say it would be interesting to see a version with some copper components.
they are working on a copper version but the machining of the parts is more complicated
I have the Ice Giant Prosiphon Elite on my 2700x and it works really well. Now I know this is a older chip, but I have a nice overclock on it . typically I idle around 40c, and when testing It usually works its way up to 60. I do agree that this cooler is definatly meant more for am4 and threadripper. I call it the mobo hider lol.
I like that name. I'm using mine on a 3950x and it works great.
I preordered this cooler after the first LTT video. From what I understand it's best suited for HEDT chips. It kept my 9900k tame at 5.0 GHZ but it was nowhere near what I was hoping. Went custom water cooling and I'm not looking back. I have the Prosiphon in the closet as a spare cooler in case of a pump failure....etc.
@Bieber Ozinga Yea, I know.....expensive back up.
I had one of these on pre-order. I am so glad I went with the NH-D15 after they failed to meet their deadline rollout. As far as it being liquid cooled I was told if there was any leaks to occur the liquid they are using would immediately turn into a gas once it comes in contact with air. They designed it that way to prevent motherboard/gpu related hardware damage. I do agree copper would be a much better choice.
did you cancelled the preorder? I also preordered but used a "cheap" cooler until it arrived and I've been using it for about 10 months
@@obedulloa6219 I stuck with the preorder until the second target delay date then cancelled. I could not wait any longer that's when I decided to buy a NH-D15. I was not penalized for it either I was issued a full refund. I don't regret it.
Nice i really wanted to see the insides ever since Linus got one!
Same here, thanks Roman.
You should watch out with the fluid inside heatpipes. One refrigerant is HCl3, which boils at around 30 °C, and is carcinogenic.
Roman, what ever happened to the prototype hot water cooler that you were developing a few years ago ?
I sub this immediately, you are the only reviewer who tear open a cooler like this.
where is your own cooler that linus did a video of when you were at that large computer show ? that thing looked really awesome , can we have an update about that?
Der8auer having to explain stuff that people should have learnt on grade school
Good idea to split the channels, don't get the notifications twice because of the versions in beatiful and butterfly smooth german 🤣
A couple things I am curious about….. if you tested it with it sitting straight up rather than laying down would that have made it perform better?? 2. It may be underperforming due to the fact this is built for thread ripper and intel ihs’s concentrate heat differently so it may just work a lot better with a larger amd ihs.
This made me smile; getting the angle grinder(?) out is EXACTLY what my brother would have done!
Level1techs has some good videos on this as well but Wendell didn't open it up though. That was very interesting.
Precise as a surgeon indeed.
I think calling it a “thermo syphon” is still valid, since it appears they intend to have circulation in DIFFERENT tubes, unlike heat pipe which is in the same tube. Furthermore, heat pipe have wick in side the tubes, which relies on capillary effect to transport condensed fluid back, the Icegaint didn’t have wick, it rely on density difference to transport the fluid back (which is exactly how thermo syphon works, circulation due to density differential caused by temperature)
wow that's a really hot chip right there
Maybe the performance was lacking due to the orientation? It may require vertical case orientation?
lol Roman's face while opening that thing.
I'm not eng but the fins on cold plate help heat distribution to heat up that liquid gas (if it smelled like strawberry its most probably cooling gas like 134a). then the gas would go through bigger hole for lower resistance, and descend in small holes.
I was looking closely and it would only could work it the radiator its in a slight angle from the cold plate for the liquid circulate by gravity.
So the NH-D15 is on-par with a friggin 320mm AIO? I'm so buying it now!
yes, not much to think about there, one fan, one heatsink, less noise, nh d15 wins
Even the Threadripper 3960X has a 280W TDP. To cool it with an air cooler and have it run quiet, there is no overclocking headroom, and you may even need to undervolt it, especially during the summer. I'm currently running a 3970X (same TDP), a custom loop and some undervolting that I did during the summer to keep the room a bit cooler.
Always interesting stuff!! Have a nice day Roman.
Lol That cooler opening made the exact same sound an old mini fridge did when I accidentally puncture it years ago rofl
A bit late at this point, but might the cooler orientation (and thus direction of gravity) have affected results? I would assume this is designed around use on a vertical motherboard. Of course this might not matter, but it would have been interesting to see.
Gravity returns do rely on orientation, Icegiant make this very clear in their manual, so with a vertical motherboard the fin array needs to be above the socket, but any horizontal orientation should work just as well because the liquid can still evaporate upwards and condense downwards.
This cooler does better on IHS with spread out chip set. not the center hot spot intel cpus.. one of the things they were going to address. imho a copper plate might help some.
Perhaps a crazy idea but how about installing a motor to act as an air pump with a pressure plate fan (blower or regular fan) inside the heat piping for faster heat delivery?
Works like a charm on my 3970x
Sandybridge or threadripper?
@@tgdhsuk3589 Threadripper
As a gravity fed unit, wouldn't laying it flat like that kind of ruin the capability. Don't the instructions say to have it orientated in a specific way so the gravity parts work properly? Wouldn't that position be in a standing side mount like how most computers are situated?
Ok i missed something new channel?
der8auer EN! So the original will DE only?
G'day Roman,
that was really cool cutting it open to see how it works inside, I wasn't expecting that huge pressure release 😲,
"it's just Fins, Fins, Fins, Fins everywhere"
🤔 it's like a holiday in Finland 😂
What a video! Amazing stuff! Hefty price this ProSiphon has 169€.
Seems it needs to be on a vertical motherboard instead of laying on a bench. heat going on upper tubes, cool down and returning on the lower tubes. With your setup, the flow can be stopped.
Super interesting. That looks like it's expensive to manufacture with all of those fins everywhere.
At minute 2 it is condensation that occurs in the cooler part of the device, not evaporation.
Doesn’t the cooler only work in vertical position? I remember that from Linus’s video
Nice Datron CnC. Want one of those bad.
Alloying or working metals lowers their conductivity as it raises hardness. So it makes sense to use relatively pure, soft aluminum.
will the next model have a tiny compressor?
Finally receiving the English video correctly!
😁💪🏼
Did you only test in the horizontal mode? Site also shows a vertical mode. Would have liked that data really. Seems the fluid path has lower energy requirements to flow in vertical.
after seeing the contact plate internal structure i can see why they went with aluminum over copper otherwise that cooler would likely be in the 250-350 dollar range since it's a niche product.
Even at its current inflated price, copper is $5/lb. Either it didn't make a performance difference, or they were trying to save weight.
What if Noctua took the basic NH-D15 design, but moved the two heatsink stacks NEXT to each other for a wider, thinner cooler?
Some of the heatpipes would be LONGER and thus less efficient, but you also wouldn't be blowing the hot air from the first stack through the 2nd stack.
Will it make any circulation difference if you rotate yor motherboard vertical as it is in a pc case?
I might be off here, but I think it would work better upright in the case.
this was fun! and also disappointing cause I was really considering one of these.. but what I have works fine... so... probably not... Great vid! 😁
my watercooling loop failed, am being given one of these coolers(not sure which revision but its from their kickstarter, so early I assume), just need to pay for shipping. its just a 5700x, so I assume itll be plenty, just not sure its going to clear my TeamGroup XTREEM ARGB modules
Explore like a barbarian. I like it!
Aren't you supposed to have the cooler orientated in a specific way?
Cutting stuff = insta thumb up 👍
its work like AC units so the motherboard should mount vertical like in the cases so the cooler works ( the cycle in it ) that's why its didn't worked here
The new prototype they sent out fixed many of the problems of the prior version.
like the stuff. proces lasso fucked up with my 1200k. nothing destructive, but freezing.
Just asking out of pure memory here...I may be wrong, but aren't those coolers meant to be used horizontally?...I mean, it might be the reason for those poor results.
Yes i remember that. This cooler is designed for AMD in mind and not intel as far a s the dies go. I think this was aimed at thread ripper or a minimum 12 to 16 cores.
I think, IceGiant could be good for AlderLake OC also (or a little better than for LGA11xx/1200). There is a little higher size of flat.
But propably I will waiitng for copper version of Elite cooler. Still, 360 Ryujin AIO or very solid Artic Cooling Liguid II 280 are fine on my benchtables.
12:00 in did he realize why his results are bad? didnhebread the manual ? iirc only certain orientations are allowed (on desk isnt one of them)
i dunno did you read them? because if you did you'd know the orientation used was 1 of the 2 recommended..
@@sirmonkey1985 oh just rewatched their installation tutorial, there is a wrong way of installation when in a normal case (the condenser (longer part) has to show down, thats what it was - still strange the temps were so high
@@aleloAT because its an aluminum plate on a hot spot IHS like intels. this plate was originally designed for TR spread out chip set. a copper plate will help with the hot spot but the fins under the plate will limit heat transfer unlike the NHD15 with the heat pipes in the plate, closer to the IHS hot spot. that cooler will be better in this case..
Would it help to set load line calibration for gaming scenario with Siphon cooler? Also i remember someones video that the cooler needs to be in certain orientation for it to work.
This cooler is made for multi chiplet design like in AM4 or Threadripper.
I feel like if this had an F-gas inside it should come with environmental warnings on the box, just like an AC unit would. At least in the EU there are pretty strict rules regarding F-gasses.
I don't think that's the case here
3:45 Professor Roman, in todays class, thermal dynamics
hehehehehehhe
🤣
Didn't you put it on backwards? From what I remember in the LTT videos this cooler has to be put in the right way to work at all. With the motherboard mounted like inside the case and the fans down towards where the GPU is.
According to manual and website, both is fine
I’m a little confused why you didn’t know what fluid is in there and why you’re surprised on lackluster performance on an Intel chip… That info is pretty well available online. I was very happy to see the inside of it tho, haven’t seen that one yet.
Roman, I don't think you are going to be able to get the cooler replaced under warranty. :)
Could the orientation of the cooler affect the flow of the refrigerant throughout the cooler?
"I'm 99% sure that 99% of you know how heat pipes work"... (in his thoughts: And I'm 99% sure that 99% of you don't know the internal pressure of a heat pipe :p )
There's a large overlap between der8auer viewers and GamersNexus viewers, and GN Steve has explained it many times. And to add to it all of us are computer nerds.
smaller contact on the condenser side makes it perform worse. if you add a graphite thermal pad it may even perform better than with paste due to the high planar thermal conductivity, but even without a pad it should perform better on AMD CPUs.
Aren't you developing a true phase changing cooler? What's up with that?
Cooler will need to be upright to work correctly based off of the principle of where the hotter fluid has to be
WHY should that thing be better than an AIO?
It has less Radiator Surface Space to get rid of the Heat.
It has less volumen of Fluid inside than the AIO.
And why would gravity based be better than pump based?
Furthermore you can mount the Radiator of the AIO wherever you want.
Put it to the front and you get cooler air from the outside.
This one has the Radiator inside the Case like a normal tower cooler.
So far, for me this looks like just a bigger Tower Air Cooler . .
An AIO should be superior to this right?
Yes. This is about reliability, where you would be worried of a pump failure. So this works great where limited access/up time is more important than performance.
The radiator on it is almost twice as thick as an AIO radiator, so there's plenty of surface area.
You don't need liters of water to do phase change cooling.
The AIO coldplate is much smaller
And if the aio's pump dies , it just dead... Its not like custom watercooling , pump faliure still a thing...
but in the other hand depends on gravity...
isn't it supposed to be vertical motherboard or was that just on the first edition
Are you Dalton from the Conjuing? :D
Isn't your testing wrong. The cooling unit was not position properly as stated in installation for the "Gravity to work"