Pure hydrogen would be terrible for air cooling. Extremely low density means very little heat transfer. Methane is has about 2x the specific heat of air. Water in it's gaseous phase (steam) would also be more effective than air. There are issues with these solutions.
This is a real good sign for manufacturing in general, metal 3d printing must have taken a sharp drop in price for it to go from aerospace manufacturing to AIO water blocks.
Note this is in particular "water/solution based" ECAM, which to my memory are basically just copper and nickle. So not the often-desired stainless steel/aluminium/etc metals, though I do wonder if this method can do those more interesting metals? I am not finding any papers with my google-fu
@@EslamNawito I don't know how much I can say, but read up on Resin 3D printing, and most of those same complications seem to apply. So yes, there are challenges, and yes porosity is a concern. Though final-finish sintering is easy with pure copper to seal holes at whatever size you want, leaving any larger (intentional for flow) holes alone.
@@admalledd It should be able to do anything that you can electroplate with. Copper and nickle would be the most common but I can see other metals being used for some applications
I know it wasn't but even 2 years ago it fell l felt space age and niche. I am sure it had already started revolutionizing engineering and product sampling.
@@originalkhawk That depends on whether or not they can get a patent on an ML generated design. Which will depend on if they can convince a court that the AI was "just assisting" a human designer, or if the court (rightfully) recognizes that the design is fundamentally the work of the learning model.
@@originalkhawkwell they patented something as simple as having the pump on block, so they'll probably get something that prevents anyone else from utilizing 3D printing for blocks. They'll find a way to fuck every other company.
Forget 3D printing, if you can do this sort of fine-detail targeted electroplating you could probably manufacture low-volume custom PCBs without nearly as many complicated steps. Prototyping PCB designs can take a long time, because while they are pretty cheap per unit you typically need to order a larger quantity than you need and it's days or weeks between prototypes. But in theory this type of printer could dramatically speed that up.
Exactly this, for pcb prototyping id hazard this a borderline breakthrough if you can put it on any kind of proper substrate. Youd still need someway to do vias though, so its not all encompassing, but still. And with the resolution it offers, god only knows what kind of new shit you could do.
For some designs of PCBs there's already 3d printers that use conductive ink to create PCBs fairly quickly. Voltera is one brand. You're not going to make PCBs for super dense BGAs on a four layer board, but for some applications it's a viable alternative in the prototyping stage.
@@1vend7 Oh absolutely. I've been making my own PCBs at home for about 20 years using some pretty crude methods that work great once you're used to it. But sometimes it's nice to have a "proper" solution so to speak.
Huh I've never seen this version of metal 3d printing. Usually it's the powder bed, slm, or fdm based ones. Of course you also have the idiots who calles a cnc mill a 3d printer.
Those are really cool, on the close ups I recognised the gyroid infill, is it good for cooling though? I would imagine it would be given its surface area and that is is fully open. The other pattern they used looked even more complicated but it was difficult to see. How were these printed though? If it was a metal powder process then they would need to get all the powder out of the tiny fine details so how do they manage that? I should have just watched further, I am still surprised with how precise they can get, they need to have just the right concentration of copper and hope that there is enough dissolved copper where it is needed, then they need to tune the expose time too, if they leave it on too long it could just pick up too many copper atoms and basically electroplate it, but too short and they won’t build it up large enough.
I think the coolest part is the shape of the "fins" . The uprights look like a cool spire, with channels to flow not only around them, but through them too.
Lame, they had cases with an air conditioner inside LONG AGO, They used spray foam to keep condensation out. They also invented 10000 rpm 80mm fans LONG ago, and no I don't care about the noise, they're not as noisy as your mom or an M60A3/M1 tank. 10000 rpm fans are better than Noctua by far. I had them in my case waaaay back in 2008. The noise wasn't even bothering at all unlike your mom.
As cool as this is, I feel its only really viable for high end custom systems, overclocking benchmarks and maybe stuff like game consoles with multi year life spans. You're essentially trading the simplicity, affordability and universality of current designs for 2 degrees of thermal performance. Just imagine shopping for an AIO using this tech, you would need a different model for every CPU. Upgrading your CPU? Welp, time to buy another $200 AIO to go along with it.
Water blocks have already become absurdly expensive. This would just make it even worse, just for 2 degrees drop. 2 degrees drop from what? .... well, don't you worry about that. You will just have more 'degrees' of cooling, silly.
The moment you showed a copper SLA part, you debunked the video title. Working with oxygen free copper SLA machines is insanely expensive. I believe Asetek themselves said some of those prints cost well over 10k to produce. Meaning it wont have consumer applications anytime remotely soon.
When young 'Engineers' from the computery industry think they found something new that will bla Blaa bla, that has been utilized extensively in the energy, chemical, electric power, light industrial, petroleum, and other sectors. Benefits of titanium heat exchangers include their excellent heat transfer efficiency, smooth appearance without scaling layer, high strength, low specific gravity etc. As a result, it is frequently utilized in a variety of industries including aircraft, marine engineering, petroleum, chemical, medical and so on. 0OoH yeah & now the computer kids found them.
The only problem I foresee with this... if it moves to the DIY Water-cooling sector... is mixing metals (at all.) In cases with 3D-Printed cold-plates, the same, or compatible materials should also be used in the radiator and any other cold-plates in the loop. Corrosion within a cold plate like that would be devastating.
How do they get the crystalline structure right for maximum thermal conductivity when 3d printing? Also, isn't 3d printed metal inherently porous/less dense? Smells like buzz
Props to Alex, and the whole crew honestly, I can't imagine doing this. Huge loud crowds that are all saying the same thing, pressure from all sides to make the product look good, flashing lights everywhere, I would be very uncomfortable. Great work on the coverage so far.
So no tests as yet.....I feel you are not going to waver much more improvement out of watercooling even if the coldplate is made to the absolute best of innovation. You are still limited to the mere fact that it can only hold so much heat and only so much heat can be carried within the small confined space it has to move....Like if there is a 10degree difference from conventional over this,...then fine.
This sounds cool but I don't get how it will change water cooling forever. You have always been able to make custom cold plates with a CNC machine. We do not know if this process is even durable for CPU cooling long term. Water erodes, that is what it does a single drop of water over and over can erode a mountain and these complex 3d shapes seem like prime erosion targets.
No matter what your product is, you have to say that AI is involved, it doesn't have to make sense. It doesn't matter if deterministic models work better. In the current AI hype, humans cannot do anything good without AI.
Can someone clarify what I have seen in this video a bit further? I am a consumer, so I don't understand much of what is going on, does this mean that my cpu di will now have a better built in di built by intel or amd themselves, because I saw a lot of people achieve better cooling by removing these dis and attaching the coolers directly on the chip. With that said, assuming that these dis now have optimal contact on all the components of a said cpu, I do not see the point of a cpu di all together.... Hear me out, why not fabricate a die that will completly replace your current cpu die that has these built in contacts directly into the massive cooler? | think this will lead insane results! To clarify things further, what I mean with what I have said is, I want the water/air cpu cooler to replace my die with these cut outs.
Perfect niche market for 3d printing and ai. Not everyone wants to spend expensive early adopt untested gear so those COD bros can gloat their performance enhancing equipment.
wont this be irrelevant as arm takes over both mobile and desktop market and uses a lot less electricity, for super high end builds with an x86 then sure. but for everyone else aka 90% of the market it wont give any significant increases.
Although interesting tech, butAsus does not give me any good vibes considering their US support division is f-ing over their customers.. if you read it ASUS.. get your act together!
Cheap to make but will we the consumer see that for ourselves...Hell no! Asetek will price it ridiculously high for nothing other than greed. Screw them and their useless water cooling.
spends a lot of time in the video telling us how cheap the technology is. asks the company how much it will cost..."a premium." yeah of course. humanity's doomed
Nope, the future of gaming pc or high TDP hardware in general is the aquarium filled with non conducting water that cal cool the entire components of the pc and we will never need anything else for cooling....
so you hype the WCB up by saying they are using a relatively cheap design process to make them but then turn around and say that they are projected to sell for a "premium" price....
Not convinced by the variable density fins... Seems like you'd just get less flow through the dense fins at the same pressure as the rest of the system.
i hope this is cheaper or they at least find a way to make this cheaper then the standard process, yes the machines are probably crazy expensive but they probably dont have to waste any material and can use less in the final end product. Also there is probably less time on the machines since the old process for making these blocks takes forever. Prayers for a cheaper, cooler future.
I wonder if using AI for this is faster than a genetic algorithm. It probably is, since you don't need to run thousands of simulations to converge on a good result.
Hopefully someone decides to do this for open loop water blocks. The heat transfer between the CPU dies and the coolant is one of the limiting factors when it comes to open loop cooling and this could really help with that.
the surface area is also amazingly improved for water contact > that is huge and hopefully they start using that for the annode and cathodes in new l-ion batteries too, i have been waiting for the battery tech, and aio's beat them to it, lol
We use to do same kind of flow calculated CNC copper block work in 2002 for the XP1700 OC to 2400. Acrillic case over. I guess U r the new noob in town.
Gyroids have 2 surfaces. 2 liquids could seperately flow through that in a closed system, couldn't they? The print as a 3D membrane between those liquids could have super interesting electrochemical applications, where we currently use mostly membranes!
looks like it will be a pain to clean if you get contaminants in the coolant, bad enough with standard plates, but they look like they would be a scrap it and get another?
7:30 I am pretty sure the reason it is designed the way it is, is to be able to maintain a relatively constant water flow and pressure across the entire cold plate, in other words a larger volume of water is flowing across the entire cold plate.
