So the large radiator has a peltier element or something similar to facilitate sub-ambient cooling? Or am I missing something? Anyhow, won't this run the risk of getting condensation inside the PC? Still, a neat concept.

ValantarSo the large radiator has a peltier element or something similar to facilitate sub-ambient cooling? Or am I missing something? Anyhow, won't this run the risk of getting condensation inside the PC? Still, a neat concept.

I was just going to say, I prefer my PC without condensation so I'll take a pass on this one. But I agree that it is a neat idea.

Franzen4RealI was just going to say, I prefer my PC without condensation so I'll take a pass on this one. But I agree that it is a neat idea.

I suppose with sufficient flow and a fine-tuned temperature sensor or two you could keep it exactly at ambient or just above, though the peltier element would still get mighty cold under heavy loads as it would have to run continuously at some point. Suppose submerging it in warm liquid with good flow might alleviate that? Still, quite skeptical so far. Suppose we'll see if it ever comes to market.

Franzen4RealI was just going to say, I prefer my PC without condensation so I'll take a pass on this one. But I agree that it is a neat idea.

condensation goes a bit far... the element seems to reside in the big radiator so it just cools the water but the water stays water, it just makes it cold.

Here's a drawing I made that explain why it doesn't condensate

ZoneDymocondensation goes a bit far... the element seems to reside in the big radiator so it just cools the water but the water stays water, it just makes it cold.

You misunderstand completely. Anything below ambient air temperature, even hoses containing chilled water, will attract condensation on its outside-facing surface from moisture in the surrounding air. This has nothing to do with the substance within the tubes, other than the temperature in said substance. In other words, if any part of this assembly dips below ambient temperatures, water will condense there. If that place is placed so that this water drips or runs onto powered components - uh oh.

altcapwnHere's a drawing I made that explain why it doesn't condensate

If that setup cools the radiator or coolant to anything below ambient, it will still condensate along the radiator-to-CPU tube and on the cool side of the CPU block. Also, that setup is purposely making the peltier system inefficient by drawing hot air from the hot side over to the cool side. In other words: purposefully done bad design. Yeesh. It would be far better to insulate any part of the cold side not contacting water and regulate the peltier through a high-frequency thermal probe and regulator setup. At least then it wouldn't need its own waste heat to avoid condensation. Then again, the sketched setup could provide a pleasant, cool warm mist for the user if placed in a suitable location. I guess that's something :p

Seriously, though: the heat output of a peltier is always higher than its cooling effect (laws of thermodynamics and all that). As such, a setup like the one you've sketched up will, except for at points where the cool side is directly contacting the water pipes, effectively warm the water, not cool it. Unless the warm side of the peltier is cooled entirely separately from the rest of the system. If the waste heat from the warm side is drawn into the radiator with any efficiency, it will cancel out any cooling effect of the cool side.

So aluminum pump.... aluminum rad and copper block ... if so violates the 1st law of water cooling.

Use preheated interior case air to cool rads ? There goes the 2nd law.

Pump and tubinng will accumulate condensation.

Hot air actually holds more moisture.

ValantarYou misunderstand completely. Anything below ambient air temperature, even hoses containing chilled water, will attract condensation on its outside-facing surface from moisture in the surrounding air. This has nothing to do with the substance within the tubes, other than the temperature in said substance. In other words, if any part of this assembly dips below ambient temperatures, water will condense there. If that place is placed so that this water drips or runs onto powered components - uh oh.


If that setup cools the radiator or coolant to anything below ambient, it will still condensate along the radiator-to-CPU tube and on the cool side of the CPU block. Also, that setup is purposely making the peltier system inefficient by drawing hot air from the hot side over to the cool side. In other words: purposefully done bad design. Yeesh. It would be far better to insulate any part of the cold side not contacting water and regulate the peltier through a high-frequency thermal probe and regulator setup. At least then it wouldn't need its own waste heat to avoid condensation. Then again, the sketched setup could provide a pleasant, cool warm mist for the user if placed in a suitable location. I guess that's something :p

Seriously, though: the heat output of a peltier is always higher than its cooling effect (laws of thermodynamics and all that). As such, a setup like the one you've sketched up will, except for at points where the cool side is directly contacting the water pipes, effectively warm the water, not cool it. Unless the warm side of the peltier is cooled entirely separately from the rest of the system. If the waste heat from the warm side is drawn into the radiator with any efficiency, it will cancel out any cooling effect of the cool side.

You're right there's something wrong here or there's something we don't understand. I hope somebody gonna ask them somewhere on youtube or will get more info for this cooler. The idea is good but, as you said, the heat output is always higher than its cooling effect. Hmmmmmmmmmmm.


And with a good shielding, the pipe and pump won't condensate.

I havent read anything about this yet, but if CM doesn't go full retard, I'd guess the engineers would put a temp sensor so the hot plate wouldn't go on if temps are like less than 5 degrees higher than ambient or cpu temp... otherwise yes, that would be an expensive moisturizer

Judging from the pictures of the radiator, the setup here is like this:

-The big "radiator" has the peltier unit on the water intake side. The majority of the "radiator" is cooling for the peltier unit, with the large heatpipes clearly seen in the picture. If I were to guess, the only water circulating in there is over the cold side of the peltier and back out.
-The small radiator looks like a regular radiator. The purpose of this? Unknown, but depending if it's on the cool or warm side of the CPU, there are different possibilities: on the cool side, it'd raise water temps to ambient or close to it, reducing condensation inside of the case; on the warm side, it'd cool off the warm water somewhat, making the peltier's job easier. My money is on the former. The K-type thermocouple in the radiator in the picture (and the fact that it's at 21C, which I'd guess is 5-6-7 degrees below ambient in Taipei) also leads me to think it's the former.

Nice to see something new and innovative instead of last year's product with twice as many lights.

It's a shame they missed the VRM fan.

altcapwnHere's a drawing I made that explain why it doesn't condensate

So they are cooling the radiator with TECS and not the actual CPU at the bottom?

It kind of sounds like effiency-loss to me if you do not cool the CPU indirect with a TEC and head for the radiator to chill the water.

TECS where fun in the days; i used to have a huge 300W TEC laying around eating up to 30Amps to keep it running. Those things really got cold. But the heat output of both CPU & tec combined was enormous. And the extra power requirement of something chewing up 30amps a second made it not really last.

I've played with vapochills, compressors and dry ice. All fun.

altcapwnHere's a drawing I made that explain why it doesn't condensate

Image #3. Red and black wire pair inside the white ribbon is the power cable for TEC element. It's inside the CPU block.

ValantarSo the large radiator has a peltier element or something similar to facilitate sub-ambient cooling? Or am I missing something? Anyhow, won't this run the risk of getting condensation inside the PC? Still, a neat concept.

It's much simpler than that. All you need is a pair of sensors (one for cold side and one for ambient), and keep the ΔT below the dew point for, let's say 75-80% relative humidity(to be on the safe side).

Still don't like the concept, cause it's not going to "maximize the performance" by reducing system efficiency by almost tenfold. TEC is good for space-constrained applications, or for competitive overclocking. Definitely bad for daily use.
They've already tried sticking Peltier modules inside the air cooling before, and we all know good of a commercial success that was:
www.coolermaster.com/cooling/cpu-air-cooler/v10/

silentbogoImage #3. Red and black wire pair inside the white ribbon is the power cable for TEC element. It's inside the CPU block.

That wire pair doesn't actually go to the CPU block, it's running behind the acrylic plate its mounted to. My guess is it's running to a PSU somewhere. Also, wouldn't a peltier element worth its salt need thicker power cabling than that? Also, if that's the case, how do you explain the large radiator being made up of heatpipes?

silentbogoIt's much simpler than that. All you need is a pair of sensors (one for cold side and one for ambient), and keep the ΔT below the dew point for, let's say 75-80% relative humidity(to be on the safe side).

That doesn't sound especially simple. Where do you find a reliable ambient temperature reading point without requiring the user to install a probe somewhere? Where do you put the controlling electronics? Are they user accessible?

ValantarAlso, wouldn't a peltier element worth its salt need thicker power cabling than that? Also, if that's the case, how do you explain the large radiator being made up of heatpipes?

Sorry, forgot to pay attention to the large rad behind the smaller one(with thermal probe sticking out). If that's the case, it's even more retarded than sticking TEC into the CPU block.

ValantarThat doesn't sound especially simple. Where do you find a reliable ambient temperature reading point without requiring the user to install a probe somewhere? Where do you put the controlling electronics? Are they user accessible?

It's not like a pair of thin-film thermistors and a tiny microcontroller takes up a lot of space. You can stick it near the pump if TEC is inside the CPU block, or inside thje radiator housing if it's the CM way of doing things.

If we abstract away from air/liquid thing and just call it "cooling the hot side", here's what you'll most likely find inside of it (once again, using V10 as a reference):

silentbogoSorry, forgot to pay attention to the large rad behind the smaller one(with thermal probe sticking out). If that's the case, it's even more retarded than sticking TEC into the CPU block.


It's not like a pair of thin-film thermistors and a tiny microcontroller takes up a lot of space. You can stick it near the pump if TEC is inside the CPU block, or inside thje radiator housing if it's the CM way of doing things.

If we abstract away from air/liquid thing and just call it "cooling the hot side", here's what you'll most likely find inside of it (once again, using V10 as a reference):

That's a very convoluted setup. Wow.

My guess is they're using the water as a large thermal mass to avoid dramatic temperature fluctuations and thus avoid going too far below ambient other than directly on the cool side of the peltier unit - which I assume is pretty much directly covered by water. With the small water-only radiator ensuring a basic level of cooling even with the peltier off, and the large heatpipe+fin stack dissipating heat from the hot side of the peltier, this could be quite effective all in all. But as we've both said, it's entirely reliant on precise, well placed and frequently polled thermal sensors controlling it to avoid over-cooling and condensation. Still, if it can maintain ambient or slightly below on a CPU with significant wattage, that's better than water.

Still too risky and complicated for me, though.

ValantarBut as we've both said, it's entirely reliant on precise, well placed and frequently polled thermal sensors controlling it to avoid over-cooling and condensation. Still, if it can maintain ambient or slightly below on a CPU with significant wattage, that's better than water.

Yep, but at power consumption over 10 times higher than a typical H2O setup of this size.

The problem with Thermoelectric cooling is that to avoid large amounts of energy usage, you have to use two materials with excellent thermoelectric cooling properties. However, my guess is that they're going to go with Bismuth chalcogenides, most likely Bismuth Telluride, because they've been used the longest, are most likely the cheapest for this application, and all the other materials generally used would probably raise a lot of issues to be used in consumer devices. Also, it allows more control/options for the designers/electric engineers, and they excel at room temperature cooling. In the first picture provided by techpowerup, you can see the silver plate at the top of the radiator, this is most likely a sandwich of ceramic substrate and in the central layer, there's probably an array of P and N type pellets that have the flow of electricity passed through them [FIGURE 1]. Honestly, I'm not sure if this unit actually cools the fluid itself, it may just cool the air entering the radiator, or with those heatpipes, it may just conductive thermodynamics to cool the radiator fins itself, but the TE unit isn't touching the fluid directly in any way. and I really doubt,[FIGURE 2] with such a small TE unit, and wanting to NOT use a ton of power which TE units are know to do, it's probably going to be like I said....I know, that doesn't sound too efficient, but, in my opinion the unit is NOT designed to actually get the fluid sub-ambient, it is just designed to get the fluid a lot cooler than any other 240mm radiator. 70 degrees Fahrenheit is approximately 21 degrees Celsius, and most 240mm AIOs only get the CPU down to low 70's perhaps 60's celsius, correct? (60C = 140F) So, if this unit can get the CPU down to 50 degrees C or (and this would be amazing, perhaps too amazing) 40 degrees C so, that's at the lowest 104F. If this unit can get the fluid itself to be at 35 degrees C/86F (But, probably not, I'm thinking 40C+), that might be enough to beat all the other AIO's by 10 degrees on the CPU package, and that's all Cooler Master has to accomplish here. Basically, trying to get the fluid sub-ambient presents a whole lot of trouble, for results they don't have to achieve, remember, they don't have massacre the competition to sell units, they just have to beat them by about 5 to 10 degrees C to obviously be the best, and that would sell. And this is capitalism that pretty much decides what gets made and what doesn't, meaning, this Coolermaster product has to be cheap enough to sell a bunch of units with a good margin of profit for coolermaster to want to do it, I really don't think they're doing this to make a uber-halo product that is only going to sell 1000 units at $500 each, do you?

Nice idea, I hope it flies, but change those hoses first, they are ugly.

Hardware Canucks' video covering this cooler confirms my hypothesis that the second (120mm) radiator is there to "balance out" temperatures and avoid condensation close to hardware. They also say it's fully capable of sub-ambient temperatures.

Anyhow, it's confirmed that the flow is like this:

CPU block/pump ->TEC (240mm rad) -> 120mm rad -> CPU

Apparently the TEC is only going to switch on after liquid temps reach a certain threshold, which sounds like there's one or more liquid temperature sensors in there. Sounds good.

Oh, and apparently it's launching before the end of the year for ~$300.

ValantarOh, and apparently it's launching before the end of the year for ~$300.

Well, that's not too bad at all.
I thought it was going to be no less than $500 (cause TEC used to be overpriced as hell all the time).