EK Water Blocks May Be Preparing a Thermoelectric CPU Cooler

[Uskompuf]

In a recent YouTube video from Linus Tech Tips titled "The Fastest Gaming PC in the World!...For Now!" a new cooler from EK Water Blocks is shown. The cooler is paired with a binned Intel Core-i9 10900K running 5.4 GHz and 360 mm front radiator. The cooler was largely censored but we can see a number of cables coming off the block including a PCI-E power cable which helps the suggestion that the cooler is a thermoelectric cooling device (TEC) utilizing the Peltier effect to transfer heat from the CPU. Thermoelectric coolers require significant power to run with the EK cooler in question being used in a 1600w power supply system. Thermoelectric coolers aren't a new invention but haven't taken off in the PC realm due to power and cost concerns so it will be interesting to see if EK is able to buck the trend.

View at TechPowerUp Main Site

 

[Just Some Noise]

Lets hope they do not......

If you need more power to cool your CPU, than the CPU needs for itself, then there is something really wrong.

TEC's are only useful for some rare cases, but for the most users, even for Enthusiasts, it is dumb.

 

[ebivan]

Havent seen that ist 15 years... Wonder what solution they have for condensation. I remember constructions with socket heaters to counter water built ups from condensation

 

[Rob94hawk]

15 years is a long time to R&D smaller & more efficient technology.

 

[ZoneDymo]

thought this did not work at all?

 

[ebivan]

15 years is a long time to R&D smaller & more efficient technology.

Well, thermoelectric efficiency has not gone up in the last 15 years, its still bad.

 

[xtreemchaos]

very interesting for sure but is it needed ? i can see why he needed that huge PSU those thermoelectric plates eat power like its going outa fashion.

 

[Kohl Baas]

TEC is dead tech. It's only viable when you need it to be small, otherwise standart gas/compression cooling is just beating rounds on it in both efficiency and performance. Who wants to pump 300Ws to cool down 150W and then cooling tha summarized 450 for CPU only? That is insane! Not to mention the problem of going under dew point. If you want to be able to go subzero, just get yourself a chiller, insulate the system and start watercooling with glykol. Much cheaper on the long run.

 

[DemonicRyzen666]

Tec's aren't bad.
using the TEC at full power trying to get it's full temperature differential is BAD.
They're much better at 10-15C differential and use much less energy. with less voltage and wattage

 

[ExcuseMeWtf]

15 years is a long time to R&D smaller & more efficient technology.

Unfortunately, laws of thermodynamics just don't seem to budge.

 

[Caring1]

TEC and Liquid cooling combined should in theory be more energy efficient.

 

[silentbogo]

They're much better at 10-15C differential and use much less energy. with less voltage and wattage

with 10-15C delta you can do the same thing with a massive passive heatsink that uses no energy.
Unfortunately, we only have these conditions when CPU is idling and produces less than 10W of heat...
Having high ΔT like on overclocked 10900K, this hybrid cooling setup definitely has a potential of becoming a new Cooler Master V10 with dead peltier elements, shorted PSUs and RMA nightmare.

15 years is a long time to R&D smaller & more efficient technology.

TEC and TEG are inherently inefficient. You can't change physics.
Maybe a century later, when we get some new super-duper-magnets and cheap room-temperature superconductors, we'll be able to achieve something more adequate, like being half as efficient as phase-change... For now, let's leave it where it belongs.

TEC and Liquid cooling combined should in theory be more energy efficient.

Nope. It's still as inefficient as its components separately. The only thing that'll maybe improve is heat transfer, because the hot side of TEC is gonna be hotter than CPU, hence warmer water, and faster dissipation on the rad side (assuming it's not gonna get saturated)

 

[Tsukiyomi91]

"The Fastest Gaming PC in the World!...For Now!" just experienced a BSOD. hah.

 

[Haile Selassie]

Well, thermoelectric efficiency has not gone up in the last 15 years, its still bad.

Ditto. Last time this was really really useful was in the Slot 1 Pentium III days.
This is probably the dumbest 'invention' of this year and it appears someone skipped the whole Peltier-Seebeck theory. Reinventing hot water so to say.

A ~ 180W TEC (50x50mm unit, which is probably what they are using, given the size of the whole contraption) has a cooling/pumping capacity of ~ 100W in best case scenario. It's not a magical device - you can only pump as much as the TEC allows you to. Once you go past this the thermals go to sh%& and the temperatures will creep up until thermal throttling kicks in as two basic metals used in TEC are not good thermal conductors.

According to the photographs and video the TEC is in direct contact/sandwich between coldplate and hotplate. Unlike Cooler Master TEC cooler this cooler has TEC as a sole thermal joint so once it gets overwhelmed it will just cook. This is a recipe for disaster. Even if you turn of TEC the bad thermal construction of TEC makes for a huge dT.

Now, imagine pairing this with a 10900K in a modern multi-threaded game like Flight Simulator 2020 (or god forbid actual production software like Blender or Premiere). I am surprised this actually went past basic sanity check, let alone internal QC.

 

[Valantar]

Hasn't this been tried, and then tried again, and then tried again and again and again and again and never worked? I guess directly water cooling the TEC is a decent enough idea given the massive cooling needs of a TEC with the capability of keeping a contemporary high end OC'd CPU cool, but as plenty of people above have said: this is a bad idea.

Edit: looking at the video and that secondary box that they aren't showing makes me wonder if that's some kind of management system for the TEC to make it run better - keeping everything above freezing, avoiding condensation, scaling the cooling in line with the thermal load, etc. Still doubt that can overcome the limitations of the TEC itself though.

 

[Blueberries]

Not sure why everyone seems to think Peltier's don't work. They do, they just require a lot of wattage to be effective and isn't really necessary for day-to-day computing, for overclocking it makes complete sense.

Der8auer even setup a peltier chiller that got to subzero temps.

 

[X71200]

If you're using peltiers to get to subzero, you're doing it wrong. As was pointed, since you do need insulation for subzero, a compressor is inherited superior. Just because he's an extreme overclocker, doesn't mean he's doing everything sensibly for his intend.

I still see peltiers on second hand sites by random people. Mostly cheap ones with stupid things written in the ad about how it would improve your cooling. Such a hot wash.

You're dumping A LOT of heat into your ambient and while doing so, you're doing it over a plate. As such, things can go wrong. EK and Linus are the geniuses that would do this as I have seen a lot of QC issues from EK watercooling products. My thick 280 EK rad smelled like galvanize for days, and the AIOs were much worse with bubbles stuck in the rads.

 

[TechLurker]

It's kind of mind-boggling to see that despite all the years of R&D, they haven't yet moved the efficiency needle by even 2% since the last major wave of peltier products back in the early 2000s (CPU coolers, GPU coolers, fancy car foodkeepers that could switch between cooler or warmer modes, specialized PC cases using a "Peltier Air Conditioner", etc).

That said, "intelligent" Peltiers that simply regulate to sub-ambient (just before reaching the condensation point), would be a workable (if not-exactly-efficient), but niche cooling option. Still going to dump a lot of heat into the room regardless, but the same can be said of water chillers and those phase-change cooling systems that were also popular around the same time.

 

[CheapMeat]

I'm pretty sure with the right management and algorithms, they can manage the cooler much better. I think some on here are thinking about the basic "100%" on TECs or other heatsinks that either work or not, on or off. It's 2020 though. Software wise, they can make sure each element is doing what it needs to within certain zones or circumstances. I don't think the TEC will be 100% on. Why are others thinking so? EK has been doing cooling for a long time. They're not complete morons. I'm NOT saying this is 100% feasible, necessary or useful for the majority or that I even want it. I know people will spin what I'm saying. And another thing is why is this community constantly AGAINST options. I see it time after time, post after post. The community is toxic against lots of hardware. PC community is amazing because of so many parts and build options. Yet, the community talks smack when you want to do it differently from the status quo or do niche things.

 

[X71200]

You are dumping heat with phase-change as well, but compressor tech has probably moved a lot more in all these years. Look at the figures from an old fridge to a new one.

It's fine to go for doing niche things, but when you aren't really improving the products most people go for (EK-AIO with loud fans and parts going back and forth between Slovenia and China), and spending your R&D into pointless investments, you're bound to get negative comments.

 

[TheUn4seen]

It's kind of mind-boggling to see that despite all the years of R&D, they haven't yet moved the efficiency needle by even 2% since the last major wave of peltier products back in the early 2000s

Well, damn physics didn't change in the last few years, just as it didn't since the universe came to existence (at least as far as we know).

But then again, using a Peltier device just to, as they say, "take the edge off" might be feasible. TECs by themselves are kind of a thermal insulator, but if you build the heat exchanger in a way that uses the TEC just to aid in the heat transfer and not be the main transfer path, it might be useful. I can only guess that EKWB doesn't hire ignorants and they found a way to make it at least useful.

 

[rethcirE]

Back when I was building a Socket-462 system this seemed to be all the rage on the message boards. I guess that was 15 years ago...

 

[Beertintedgoggles]

I think they'd be better off using the TEC as a reservoir chiller. Stick the cold side on the reservoir and slap a good air cooler on the hot side. And yeah, I know this has already been done before as well. Hell, they should go all out and include a bong cooler for that complete retro feel!

 

[SamuelL]

Peltier coolers work. They are horrifically inefficient, yes, but they do work. The problems with bringing these to market were always setting expectations vs cost (CoolIT Systems...) and products/ideas destined for failure due to cheap implementation and low tolerances (cooler master v10...).

I think there is a small market for a peltier today, even given the drawbacks. I don’t see the issue (assuming a well-engineered product is made) so long as buyers are aware of the added cost up front and to the electric bill AND everyone understands the intent is likely a 5-10+* drop (not subzero temps or other ridiculous claims).

Back when I was building a Socket-462 system this seemed to be all the rage on the message boards. I guess that was 15 years ago...

Yep, fond memories of socket A and the “Wild West” overclocking days

 

[bug]

Tec's aren't bad.
using the TEC at full power trying to get it's full temperature differential is BAD.
They're much better at 10-15C differential and use much less energy. with less voltage and wattage

Yeah, just stack 3 of those and enjoy silent, efficient cooling

Though yes, keeping the differential down help with the other TEC problem: condensation.