An Enthusiast Review of CoolClouds' Revolutionary CPU Cooler Prototype

TheHobbyist

Hugs Hard Johnnies [H]ard
Joined
Apr 8, 2008
Messages
456
If you read the rest of the thread, the new testing I was proposing was requested.

+ There was concern that by testing with the side panel off, I somehow made the Corsair H80i look worse in the comparison.
+ There was interest in the cooler's ability to cool passively, and it is easy to test.
+ Some people are concerned about pump failure. I argued that modern CPU's throttle or shut down, but the concern persisted. Also, some wanted to know if they could still use their computer in the event of a pump failure while waiting for RMA.
+ I agree the cooling effect on system components is about the case performance. However, the argument is coming up about whether the AIO form factor is superior to the motherboard mounted form factor.
+ Based on the form factor of the cooler and radiator placement, one form factor may be a better choice over the other for overall system health.

There is always the argument of synthetic testing vs trying to emulate "real world" testing. I'm trying to cater to the crowd here. Anandtech uses a very scientific method in their latest AIO cooler roundup. However, the coolers may not stack up that way in a "real world" scenario, which is what consumers will be experiencing. Hardocp tries to look at "real world."

"The cooler performance is about how well cooler can cool CPU based on air temperature going into cooler / radiator" Obviously the cooler form factor can influence the "air temperature going into the cooler / radiator." This is a question people seem to want answered as CoolClouds is making an AIO cooler with the motherboard mounting form factor.

What do you think?
 

doyll

[H]ard|Gawd
Joined
Feb 4, 2012
Messages
1,222
People will always have concerns; some founded, some not. Side panel on or off is not a founded concern.
People will always have interests; some founded, some not. Water coolers are not designed to be passive.
Concern about pump failure is a valid concern. Yes, CPUs will throttle and/or shutdown. But a water cooled system simply will not work without a pump. Test it if you want but it's beating a dead horse.
It's not AIO versus motherboard mount. It's water cooling versus air cooling. The case and how well it's airflow is optimized / tuned is a far more important issue then AIO / CLC versus air.
Same as above.

"Very scientific" is highly debatable .. as is "real world" The only "real world" is the one each of us has .. and each "real world" is different from every other "real world". We all have different cases, different fans, different PSUs, different motherboards, CPUs, RAM,.GPUs, coolers in different temperatures, different barometric pressure, different humidity .. and they all affect the airflow and cooling.

I believe the best way to test most components is on an open bench test station. At least it removes many of the variables that case mounted testing exacerbates.

Delta temp is a joke. 15c air cools very differently than 25c air .. and the difference is not 10c. We can test in a 21-25c environment but even that requires +/-1 -2c tolerance.

Yes! Cooler performance has to be based on the air temperature going into the cooler or radiator, not the room! But cooler form factor is not a big factor if case is designed and airflow is setup properly. I have an old Define R2 with ENGTX580DCII and 980 under a TC14PE with only 3x 140mm intakes. Cooler intake air temperature is never more than 2-3c above room temperature .. and at times it runs for hours under 100% or near 100% load. Maximum CPU is high 40s to low 50s dependent of room temp. Summer with 28c cores peaked between 47c and 53c with maximum fan speed of 1200rpm. When system goes to work maximum temperature is reached in less than 4 minutes .. the same 4 minutes it takes fans to find their rpm to heat balance.
Here's an example. No idea why temps dropped a couple degrees after 12-14 minutes.
 

M76

[H]F Junkie
Joined
Jun 12, 2012
Messages
11,187
I think the biggest advantage of water cooling is that you can take the heat exchange unit away from the source of the heat. Plain: You can take the radiator near the vents of the case or outside even, to remove the heat from the case.

With this we loose that advantage, so basically it functions as an air cooler where the heatpipes were replaced with water circulation. It might be better at absorbing the heat, but that heat is then transferred to the air inside the case, so you're back at base 1. Since it uses the air inside the case to cool. Whereas with a traditional AIO setup, the heat is either vented outside the case, or the radiator is cooled with the air from outside the case.

It's an interesting concept, but it has no advantages. And the lack of tubes means the amount of coolant in the system is also much less, which is also a disadvantage cooling wise.

Of course by testing with an open case, you've removed the biggest disadvantage of the cooler from the equation.
 

doyll

[H]ard|Gawd
Joined
Feb 4, 2012
Messages
1,222
While I agree with much of what you are saying M76, cases setup with proper airflow and air coolers cool just fine for all but the Banzai overclockers and multi GPU gamers.

The small amount of coolant is not a disadvantage. Cooling is the same regardless of amount of coolant. More coolant only means it takes longer for the coolant to reach equilibrium.

Using an open case does not remove any disadvantage, it only helps us know how the coolers performance compares instead of how the system cooler in in performs. As my system is different from any tester's system I prefer knowing the coolers' performance rather than the systems' performance. :D
 

W.Feather

Hard DC'er for Aug 2012 / Feb 2019 / Mar 2020
Joined
Nov 10, 2009
Messages
11,431
In that case, the better benchmark is outside of a case, in a controlled environment, rather than just with a case side on / off .... as there is still a different airflow pattern than your case.....
 

doyll

[H]ard|Gawd
Joined
Feb 4, 2012
Messages
1,222
Indeed.
And some coolers require extra work even on an open bench test system to get cool air.
Downflow / pancake coolers are notorious for eating their own heated exhaust air. :D Fan blows air down hitting mobo, turns out hitting RAM GPU, I/O housings, etc., turns up past cooler, and gets sucked back into fan. Turning fan so they become upflow coolers often give 4-8c better CPU temps and also lowers mobo component temps.
 

TheHobbyist

Hugs Hard Johnnies [H]ard
Joined
Apr 8, 2008
Messages
456
People will always have concerns; some founded, some not. Side panel on or off is not a founded concern.
People will always have interests; some founded, some not. Water coolers are not designed to be passive.
Concern about pump failure is a valid concern. Yes, CPUs will throttle and/or shutdown. But a water cooled system simply will not work without a pump. Test it if you want but it's beating a dead horse.
It's not AIO versus motherboard mount. It's water cooling versus air cooling. The case and how well it's airflow is optimized / tuned is a far more important issue then AIO / CLC versus air.
Same as above.

"Very scientific" is highly debatable .. as is "real world" The only "real world" is the one each of us has .. and each "real world" is different from every other "real world". We all have different cases, different fans, different PSUs, different motherboards, CPUs, RAM,.GPUs, coolers in different temperatures, different barometric pressure, different humidity .. and they all affect the airflow and cooling.

I believe the best way to test most components is on an open bench test station. At least it removes many of the variables that case mounted testing exacerbates.

Delta temp is a joke. 15c air cools very differently than 25c air .. and the difference is not 10c. We can test in a 21-25c environment but even that requires +/-1 -2c tolerance.

Yes! Cooler performance has to be based on the air temperature going into the cooler or radiator, not the room! But cooler form factor is not a big factor if case is designed and airflow is setup properly. I have an old Define R2 with ENGTX580DCII and 980 under a TC14PE with only 3x 140mm intakes. Cooler intake air temperature is never more than 2-3c above room temperature .. and at times it runs for hours under 100% or near 100% load. Maximum CPU is high 40s to low 50s dependent of room temp. Summer with 28c cores peaked between 47c and 53c with maximum fan speed of 1200rpm. When system goes to work maximum temperature is reached in less than 4 minutes .. the same 4 minutes it takes fans to find their rpm to heat balance.
Here's an example. No idea why temps dropped a couple degrees after 12-14 minutes.
<snip></snip>
doyll, thank you for the post. I agree with most all of what you say. You have a good understanding of how things ACTUALLY work =)

I'm taking your feedback toward my future testing decisions. For some of the questions posed though, I think the only way to solidify the truth is through demonstration. I think it is a case of "the world is round" "no, the world is flat." If I performed the testing, it would at least give non-believers a place to learn.

Thanks!
 

TheHobbyist

Hugs Hard Johnnies [H]ard
Joined
Apr 8, 2008
Messages
456
I think the biggest advantage of water cooling is that you can take the heat exchange unit away from the source of the heat. Plain: You can take the radiator near the vents of the case or outside even, to remove the heat from the case.

With this we loose that advantage, so basically it functions as an air cooler where the heatpipes were replaced with water circulation. It might be better at absorbing the heat, but that heat is then transferred to the air inside the case, so you're back at base 1. Since it uses the air inside the case to cool. Whereas with a traditional AIO setup, the heat is either vented outside the case, or the radiator is cooled with the air from outside the case.

It's an interesting concept, but it has no advantages. And the lack of tubes means the amount of coolant in the system is also much less, which is also a disadvantage cooling wise.

Of course by testing with an open case, you've removed the biggest disadvantage of the cooler from the equation.
Thank you for your thoughts! I think doyll explained my viewpoint nicely in the post that followed yours.

Cheers!
 

TheHobbyist

Hugs Hard Johnnies [H]ard
Joined
Apr 8, 2008
Messages
456
In that case, the better benchmark is outside of a case, in a controlled environment, rather than just with a case side on / off .... as there is still a different airflow pattern than your case.....
Indeed.
And some coolers require extra work even on an open bench test system to get cool air.
Downflow / pancake coolers are notorious for eating their own heated exhaust air. :D Fan blows air down hitting mobo, turns out hitting RAM GPU, I/O housings, etc., turns up past cooler, and gets sucked back into fan. Turning fan so they become upflow coolers often give 4-8c better CPU temps and also lowers mobo component temps.
Thanks gentlemen!

I'm still deciding what the next move is, and working a lot so I haven't had much time to tinker =)

Appreciate the feedback!
 

GonzoP

2[H]4U
Joined
Jul 3, 2006
Messages
3,660
Interesting stuff Hobbyist. I'm actually considering upgrading my cpu cooler. Debating on AIO versus something like the D14. I followed the IndieGoGo link and it appears your cooler is available starting in November? $99 puts you right where the popular AIO's are. I was considering the D14 since I have a older 1366 mobo and the D15 isn't compatible. NE had the D14 on sale last week for $69. Not sure I wanna wait 3 months and spend $30 on your cooler which might only net me a couple ºC.
I'd love to see a comparison between your product and the D15 as they are both priced the same. Given the margin the D15 has over the D14 it ought to be in the same ballpark as your cooler.
So while its unlikely I will personally patron your business I wish you well. The biggest winners in a competition like this is the consumer
 

doyll

[H]ard|Gawd
Joined
Feb 4, 2012
Messages
1,222
Really, a CoolCloud's production cooler is needed to get true comparison testing against other top coolers.

Honestly, I would use PH-TC14PE and/or NH-D14 as the "reference" cooler for comparisons .. they are the most common hi-performance cooler is use right now.
 

GonzoP

2[H]4U
Joined
Jul 3, 2006
Messages
3,660
Really, a CoolCloud's production cooler is needed to get true comparison testing against other top coolers.

Honestly, I would use PH-TC14PE and/or NH-D14 as the "reference" cooler for comparisons .. they are the most common hi-performance cooler is use right now.
^^^
Agreed
 

Unknown-One

[H]F Junkie
Joined
Mar 5, 2005
Messages
8,903
Water coolers are not designed to be passive.
This was already brought up. There are plenty of liquid cooling systems designed to operate passively (no pump).

They actually use passive liquid cooling on nuclear reactors. The temperature differential drives convection currents that move the coolant through the system.
 

doyll

[H]ard|Gawd
Joined
Feb 4, 2012
Messages
1,222
This was already brought up. There are plenty of liquid cooling systems designed to operate passively (no pump).

They actually use passive liquid cooling on nuclear reactors. The temperature differential drives convection currents that move the coolant through the system.
You need to get a grip on reality mate. :rolleyes: You are on a computer forum where we are dicussing CPU cooling. Cooling nuclear reactors is something I doubt anyone but you is thinking about. :D

That said, Captherm is developing a hi-tech multiphase liquid cooler. the MP1120
 
Last edited:

Unknown-One

[H]F Junkie
Joined
Mar 5, 2005
Messages
8,903
You need to get a grip on reality mate. :rolleyes: You are on a computer forum where we are dicussing CPU cooling. Cooling nuclear reactors is something I doubt anyone but you is thinking about. :D
What, exactly, is making you question my grip on reality?

This is simple thermodynamics. If you have a temperature differential, you can generate a thermosiphon, which can then be channeled into moving coolant through a loop without the need for a pump.

This works just as well with the heat generated by a CPU as it does with the heat generated by a nuclear reactor. There are already a few examples of people who have built perfectly functional PC water cooling loops without any kind of pump.

I brought up nuclear reactors as an example to prove that the effect works even when scaled-up massively, and is also highly reliable.
 
Last edited:

doyll

[H]ard|Gawd
Joined
Feb 4, 2012
Messages
1,222
I was joking with you.

But to put it simply, I don't think we can cool a CPU chip using thermodynamics and thermosiphon principles. The area of CPU generating heat is too small. It cannot move enough water across the heat exchanger on CPU fast enough to remove the heat without the use of a pump or phase change like Captherm MP1120 does. .
 

Unknown-One

[H]F Junkie
Joined
Mar 5, 2005
Messages
8,903
I don't think we can cool a CPU chip using thermodynamics and thermosiphon principles. The area of CPU generating heat is too small. It cannot move enough water across the heat exchanger on CPU fast enough to remove the heat without the use of a pump or phase change like Captherm MP1120 does. .
Dude, there's a link in my previous post of a PC, being water cooled, with no pump. The only thing moving coolant through the loop is the temperature differential created by the CPU / GPU / other components.

It works.
 

doyll

[H]ard|Gawd
Joined
Feb 4, 2012
Messages
1,222
Okay, we can do almost anything if we try hard enough. But I don't see it as practical.
That is on a 60w TDP CPU. I would be surprised if it would cool a overclocked 130w TDP CPU using 200+W of power .. and he's still using fans to cool other components.
 

Unknown-One

[H]F Junkie
Joined
Mar 5, 2005
Messages
8,903
Okay, we can do almost anything if we try hard enough. But I don't see it as practical.
That is on a 60w TDP CPU. I would be surprised if it would cool a overclocked 130w TDP CPU using 200+W of power .. and he's still using fans to cool other components.
You've got it totally backwards. The thermosiphon effect becomes stronger with larger temperature differentials.

It's actually more surprising that the system was so functional with such a limited heat source. A hotter chip would actually drive coolant through the loop faster, improving cooling efficiency.

This effectively makes the system self-regulating. Add more heat, the coolant automatically moves faster. Cool the radiator better, the coolant automatically moves faster.
 

doyll

[H]ard|Gawd
Joined
Feb 4, 2012
Messages
1,222
You've got it totally backwards. The thermosiphon effect becomes stronger with larger temperature differentials.

It's actually more surprising that the system was so functional with such a limited heat source. A hotter chip would actually drive coolant through the loop faster, improving cooling efficiency.

This effectively makes the system self-regulating. Add more heat, the coolant automatically moves faster. Cool the radiator better, the coolant automatically moves faster.
The assumption that putting 130w or 200w of heat into CPU water block and the water flow rate increasing proportionally to keep it as cool or cooler is as assumption that may or may not prove true.

We could just as easily assume the flow rate will remain similar to what it is now because the waterblock will flow the same amount regardless of added heat. Or assume the added radiator needed will have as much resistance to flow as the added heat give in increased flow .. giving use the same flow rate we now have.

But my educated guess is the CPU waterblock's fitting size and flow channels would / will be the limiting factor.
 

Unknown-One

[H]F Junkie
Joined
Mar 5, 2005
Messages
8,903
The assumption that putting 130w or 200w of heat into CPU water block and the water flow rate increasing proportionally to keep it as cool or cooler is as assumption that may or may not prove true.
If cooling efficiency couldn't increase proportionally to the size of the heat source, this type of cooling wouldn't be viable for nuclear reactors...

Boom, conversation comes full circle! :p

Seriously though, the effect is more than powerful enough for computing use. Its been used in computers, combustion engines, whole-house heating systems, etc...

But my educated guess is the CPU waterblock's fitting size and flow channels would / will be the limiting factor.
That's easily solvable. Heck, air coolers designed for fan-less operation solved this same problem ages ago.

More space between the fins allows convection to work more effectively. I'm sure there are some fairly non-restrictive blocks out there that would work well for this.
 
Last edited:

doyll

[H]ard|Gawd
Joined
Feb 4, 2012
Messages
1,222
If cooling efficiency couldn't increase proportionally to the size of the heat source, this type of cooling wouldn't be viable for nuclear reactors...

Boom, conversation comes full circle! :p

Seriously though, the effect is more than powerful enough for computing use. Its been used in computers, combustion engines, whole-house heating systems, etc...


That's easily solvable. Heck, air coolers designed for fan-less operation solved this same problem ages ago.

More space between the fins allows convection to work more effectively. I'm sure there are some fairly non-restrictive blocks out there that would work well for this.
The 3/8" fitting ID is very limiting .. as is the small area of a CPU heatsink waterblock. Not much room to "spread the fins" in to transfer the heat from CPU to heatsink to water.

It may be possible. But it is definitely not a slam-dunk that it will. ;)
 

Unknown-One

[H]F Junkie
Joined
Mar 5, 2005
Messages
8,903
The 3/8" fitting ID is very limiting .. as is the small area of a CPU heatsink waterblock. Not much room to "spread the fins" in to transfer the heat from CPU to heatsink to water.

It may be possible. But it is definitely not a slam-dunk that it will. ;)
What's with all this "may be possible" stuff? There are plenty of examples of this type of cooling system in active day-to-day use.

And don't forget, it's the temperature differential that drives the effect, not the component temperature itself. The bigger the difference between the hot and cold coolant in the loop, the stronger the siphon.
You could always attach a thermoelectric cooler to the top of the loop instead of a radiator and generate substantial flow (even with a very minimal heat source in the system) by dropping the coolant to sub-ambient temperatures.

You might consider this configuration a "solid state" pump, of sorts.

Anyway... It's totally POSSIBLE, and on no uncertain terms. What you seem to be attempting to make an example of is weather or not it's PRACTICAL.
 
Last edited:

M76

[H]F Junkie
Joined
Jun 12, 2012
Messages
11,187
While I agree with much of what you are saying M76, cases setup with proper airflow and air coolers cool just fine for all but the Banzai overclockers and multi GPU gamers.

The small amount of coolant is not a disadvantage. Cooling is the same regardless of amount of coolant. More coolant only means it takes longer for the coolant to reach equilibrium.
Well I went trough a lot of enthusiast cases recently and airflow still is a problem for most, if you don't want loud high cfm case fans. Only a few very good cases offer the type of airflow that would allow this cooler to shine.

As for coolant iI assume normal working conditions where the cpu is rarely if ever at 100%, and is constantly flipping between idle and load. So the more coolant the better, as it won't reach equilibrium at all. But I agree this is only a very slight advantage maybe not even measurable.
Using an open case does not remove any disadvantage, it only helps us know how the coolers performance compares instead of how the system cooler in in performs. As my system is different from any tester's system I prefer knowing the coolers' performance rather than the systems' performance. :D
If you want to test the coolers plain performance sure, but that is as irrelevant as overclocking with liquid nitrogen. As you'll always use the cooler in a system. Even with the best of the best, you'll always have the temp higher inside the case than outside. And a cooler that uses air from inside the case will perform worse.
 

doyll

[H]ard|Gawd
Joined
Feb 4, 2012
Messages
1,222
Well I went trough a lot of enthusiast cases recently and airflow still is a problem for most, if you don't want loud high cfm case fans. Only a few very good cases offer the type of airflow that would allow this cooler to shine.
I said cases optimized for airflow, not off-the-shelf. Very few off-the-shelf have even decent cooling. I can optimize a Define R4 to run less than 3-4c warmer than room ambient at cooler intakes at extended 100% load, and Define R4 is not known as a "good cooling" case.

As for coolant iI assume normal working conditions where the cpu is rarely if ever at 100%, and is constantly flipping between idle and load. So the more coolant the better, as it won't reach equilibrium at all. But I agree this is only a very slight advantage maybe not even measurable.
I don't agree. My system runs at 80-100% for hours at times. Having a system dependent on coolant volume is building for failure .. because at times most systems will run at full load for extended periods.

If you want to test the coolers plain performance sure, but that is as irrelevant as overclocking with liquid nitrogen. As you'll always use the cooler in a system. Even with the best of the best, you'll always have the temp higher inside the case than outside. And a cooler that uses air from inside the case will perform worse.
No, testing air coolers on an open bench to see how their performance compares is nothing like overclocking wiht liquid nitrogen.

Testing coolers in your own case and system for your own use is all find and good, but don't expect others to have the same or even similar results is in their case and system. That is an unrealistic expectation. This is shown by the wide range of temps between coolers different reviewers get. When reviewing coolers for the public the testing should be done outside a case, or at the very least the cooler intake air temperature should be be monitored and used to determine delta temp. As I've said before, there is no "real world" environment to test in that is the same as my "real world", your "real world" or anybody elses' "real world".
 
Top