Prototype This is the best solution for AIO. Thac Nuoc L1

[LegoFab]

This is one of accessories of LegoFab T1, but can be used in other normal case.
With this one you can easy move your motherboard to other cases without remove CPU cooler.
Size is same with medium size air cooler (92mm fan) 100-120mm height. Performance is nearly same with 240mm AIO. Forget the big size bulky AIO and huge air cooler!

Test result with intel 2678 V3 12 core, 120W TDP CPU, ambient 25C degree. 3 min stress test 100% usage:

All fan running less than 50% very silent.
I can provide this solution soon with low MOQ.
Please make this solution become perfect by leave your feedback.

 

[Allhopeforhumanity]

From what it looks like, is this is a 92mm rad mounted directly to a pump/cold-plate combo, like an apogee 2 or aio block?

Based on other results I've seen, I have a hard time believing that a 92mm system would come anywhere close to a 240mm solution at moderate to high thermal loads (120 -> 150 W).

Is this test showing a continuous 120 W load? And 3min seems awfully short to thermally soak the fluid. Can you retest for 15-20 minutes where you're more likely to reach thermal equilibrium between the cooler and the fluid?

 

[LegoFab]

From what it looks like, is this is a 92mm rad mounted directly to a pump/cold-plate combo, like an apogee 2 or aio block?

Based on other results I've seen, I have a hard time believing that a 92mm system would come anywhere close to a 240mm solution at moderate to high thermal loads (120 -> 150 W).

Is this test showing a continuous 120 W load? And 3min seems awfully short to thermally soak the fluid. Can you retest for 15-20 minutes where you're more likely to reach thermal equilibrium between the cooler and the fluid?

It is a big size 92mm copper radiator 16 FPI.
Xeon 2678 V3 in the test is a powerful 120W TDP CPU.
I just remove 240mm copper radiator 22FPI, performance is close, In some case, it will be more powerful than normal 240AIO with aluminium radiator.

 

[BonfireOfDreams]

This seems genuinely dishonest. I'd like to be wrong. Can you do a 30 minute stress test for both this 92mm AIO & the 240mm AIO, providing relevant data?

 

[LegoFab]

This seems genuinely dishonest. I'd like to be wrong. Can you do a 30 minute stress test for both this 92mm AIO & the 240mm AIO, providing relevant data?

Some one should say unbelievable, but the 92 mm radiator thickness in this test is 54 mm with high pressure fan and good ventilation case.
This system is used for my work so I will do stress test in long time with other system, even 3 min is over enough for system to reach stable temperature.

 

[LegoFab]

I just modified the bracket, now looks more beautiful:

 

[Allhopeforhumanity]

The specific heat of water is ~4 kJ/(kg*K) or 4,000,000 J/(L*K). Say the loop has 0.001 L (1mL) of fluid, that would mean that at 120 W, in 3 minutes (180 sec) the water would only heat up by 21600 J / (4000 J/K) = 5.4 K in the complete absence of heat dissipation.

From testing I've seen, most 120mm AIOs saturate after about 15 minutes, which is why I suggested running the test longer, as all other things being equal the 92mm should be marginally worse than the same thickness and FPI 120mm.

 

[LegoFab]

The specific heat of water is ~4 kJ/(kg*K) or 4,000,000 J/(L*K). Say the loop has 0.001 L (1mL) of fluid, that would mean that at 120 W, in 3 minutes (180 sec) the water would only heat up by 21600 J / (4000 J/K) = 5.4 K in the complete absence of heat dissipation.

From testing I've seen, most 120mm AIOs saturate after about 15 minutes, which is why I suggested running the test longer, as all other things being equal the 92mm should be marginally worse than the same thickness and FPI 120mm.

I have some ideas about cooling solution that shown before, use professional simulation software.
For sure, I will test in hardcore condition.
I think about 120mm radiator before but I decided to select 92mm one because of compatibility, easy to install and powerful enough to handle flagship CPU in good temperature.

I also can provide bracket separately for the one who can assembly himself.

 

[Fruergaard]

This have been up a few times in your threads; 3 min is not near enough time to get to a steady state situation as Allhopeforhumanity shows and therefore it only "seems as powerful as a 240 aio" since you only see the effect of heat uptake by the water and not how well it dissipate heat.

This seems like the sun cooler situation all over again with wild postulates and no valid data to back them up?

Concept - Sun Cool 150W TDP slim heatsink

Update 8-May-2018 Design change Estimate performance will be same water cooler if equip with 92mm x 25mm fan or bigger. Hi guy, I'm assistant from XCreator, established by Mr Rolexus, he now focus in design and other jobs, I'm in charge in media. This Sun Cool is a full copper heat sink TPD...

smallformfactor.net

 

[Allhopeforhumanity]

I have some ideas about cooling solution that shown before, use professional simulation software.
For sure, I will test in hardcore condition.
I think about 120mm radiator before but I decided to select 92mm one because of compatibility, easy to install and powerful enough to handle flagship CPU in good temperature.

I also can provide bracket separately for the one who can assembly himself.

I'd be curious to see the simulation boundary/initial conditions and results. I do quite a bit of thermal modeling myself, I just wanted to run the back of the envelope numbers to show why I think that 3 minute load testing is likely insufficient to bring the system to equilibrium.

I agree, I think that 92mm is likely better for compatibility in most cases, but I think its yet to be seen that it is sufficient to handle modern flagship CPUs without any throttling. My 4790K for example, can pull ~160W under AVX type loads and around ~130W running numeric simulations; so uncapped I'd have to imagine that 6,,8,12 core options are only going to drive that thermal load higher.

 

[LegoFab]

This have been up a few times in your threads; 3 min is not near enough time to get to a steady state situation as Allhopeforhumanity shows and therefore it only "seems as powerful as a 240 aio" since you only see the effect of heat uptake by the water and not how well it dissipate heat.

This seems like the sun cooler situation all over again with wild postulates and no valid data to back them up?

Concept - Sun Cool 150W TDP slim heatsink

Update 8-May-2018 Design change Estimate performance will be same water cooler if equip with 92mm x 25mm fan or bigger. Hi guy, I'm assistant from XCreator, established by Mr Rolexus, he now focus in design and other jobs, I'm in charge in media. This Sun Cool is a full copper heat sink TPD...

smallformfactor.net

I told that this system is used for my work, every one know that if you do the stress test in long time it is not good for the system, that why I will do that kind of test in other system.
I have final design for that (sun cool), extremely powerful design, but it costs a lot for development and not easy to make the prototype as a mini case so I launch T1 first.

 

[ignsvn]

30 minutes of testing should not harm your system. Worst case it'll just throttle down the speed.

Also, please test with case panel on & off.

 

[LegoFab]

30 minutes of testing should not harm your system. Worst case it'll just throttle down the speed.

Also, please test with case panel on & off.

Don't worry, I got a Z370 motherboard, looking for a 8700K or 9900K ES/QS version, it is more popular and easy to compare, will be ready to burn system with any test.

 

[LegoFab]

In the waiting time. I made a video test in 13 min with current system. Ambient temperature 25C. Close case
In first 10 min test with silent mode, CPU temperature reach about 49C.
After that, I switch all fan / pump to 90-95%. temperature decrease to 44-45C.
Decrease after 10 min, that means it will not keep increasing more.

 

[Allhopeforhumanity]

Decrease after 10 min, that means it will not keep increasing more.

I mean you increased the pump/fan speeds, and that lead to the temperature drop, but that doesn't mean that thermal equilibrium was necessarily reached in that duration. Do you have a plot of temperature over time showing an asymptotic regression to the claimed max?

Also, it would be beneficial to utilize a software suite that also shows CPU load % and power draw; even better if you can plot them over time. Power draw vs cooling power (pump/fan speed) is especially useful in comparing the coolers efficiency with other options in apples-to-apples comparisons.