• Welcome to TechPowerUp Forums, Guest! Please check out our forum guidelines for info related to our community.
  • The forums have been upgraded with support for dark mode. By default it will follow the setting on your system/browser. You may override it by scrolling to the end of the page and clicking the gears icon.

Pumpless AIO that actually works, tested by Der8auer

It's not a new concept. Silverstone demonstrated this many years back I believe, but still cool nonetheless.
 
dgianstefani said:
Hmm, I should amend my first post

The mixed metals and general cheap quality with a limited intended service life still remain inherent to AIOs.
Click to expand...
My Cooler Master Nepton is still going strong. What specific AIOs are you talking about?
 
kapone32 said:
My Cooler Master Nepton is still going strong. What specific AIOs are you talking about?
Click to expand...
Yet again you cannot understand true generalisations, and instantly get defensive over your own purchase choices.

AIO pumps eventually fail. It's a matter of when, not if.

When the wear accumulates, they steadily get noisier and lower performing. It's not unusual for an old AIO to struggle cooling a load less than half of which it could when new, due to this issue and others, including non refillable coolant.

When they do fail, you throw the whole unit away, becoming just more ewaste.

Air cooling is absolutely fine for most users. They don't need AIOs for a couple degrees of performance. For those that want the performance but are too technically inept or can't be bothered to build/pay for custom loop, this pumpless solution is a step in the right direction for better cooling than air, without one of the drawbacks current AIOs inherently have.
 
dgianstefani said:
Fixes one of the main issues I have with AIOs - weak pumps that wear out and are noisy.

Cools 200 W no problem, 36000 points instead of 36300 with a corsair 240 AIO, prototype, still room for improvement.

Custom loop will still be much better, but this fixes the reliability and noise issues with AIOs, while eliminating cost from manufacturing.

Click to expand...

I assume it works by convection; if so then the radiator should be length wise vertical for best operation.
 
dgianstefani said:
Yet again you cannot understand true generalisations, and instantly get defensive over your own purchase choices.

AIO pumps eventually fail. It's a matter of when, not if.

When the wear accumulates, they steadily get noisier and lower performing. It's not unusual for an old AIO to struggle cooling a load less than half of which it could when new, due to this issue and others, including non refillable coolant.

When they do fail, you throw the whole unit away, becoming just more ewaste.

Air cooling is absolutely fine for most users. They don't need AIOs for a couple degrees of performance. For those that want the performance but are too technically inept or can't be bothered to build/pay for custom loop, this pumpless solution is a step in the right direction for better cooling than air, without one of the drawbacks current AIOs inherently have.
Click to expand...
One thing you're missing here is permeation, which adds to the noise issue. It's not really the pump failing, but coolant evaporating, causing more air to enter the system and make the whole thing noisier. I'm betting that is a bigger issue than pump failure. Permeation is the primary reason why AIOs perform worse over time, assuming we're sticking to issues unique to water cooling systems.
 
280 with slightly thicker radiator and better fan pressure seems like the right direction for this prototype. I'd say a 360 wouldn't be bad either, but it probably add to build costs a bit more and won't fit in as many cases.
 
BoggledBeagle said:
Should not the rad be sloping towards the liquid return hose?
Click to expand...

Remember those old cars with the tall grill? they had no pump and so needed the cooling water to sink.
 

Attachments

  • tall grill.jpg
    tall grill.jpg
    52.8 KB · Views: 59
Shrek said:
Remember those old cars with the tall grill? they had no pump and so needed the cooling water to sink.
Click to expand...
Not sure why you talk about cars with water cooling loop, when the cooler we talk about has some phase change coolant in it, the coolant enters radiator as vapour, then it condenses on the radiator walls and the droplets form a stream of liquid coolant leaving the rad on the other side. That is why I asked if the rad should not be sloped to facilitate the flow of the coolant in liquid form.
 
kapone32 said:
My Thermalright 240mm keeps my 5950X pretty cool. Could be the 3300 RPM pump speed though. AIOs are still fine to use. For most AIOs the source of noise is not the pump but the fans. The fans on my Cooler Master Nepton 280 were just as loud as a hair dryer. That led me to Phanteks fans.
Click to expand...
Experimenting with a 5950x using NH-L9a recently got me thinking why have we become so obsessed with keeping the CPU's so cold when they are now designed to throttle so gracefully.

dgianstefani said:
Fixes one of the main issues I have with AIOs - weak pumps that wear out and are noisy.

Cools 200 W no problem, 36000 points instead of 36300 with a corsair 240 AIO, prototype, still room for improvement.

Custom loop will still be much better, but this fixes the reliability and noise issues with AIOs, while eliminating cost from manufacturing.

Click to expand...
I wonder if the fluid is non-toxic. If so I'd consider getting something like this. Only loses by 300 points, I'll take that tradeoff for zero pump noise.
 
Last edited:
This has been posted as front page news now:

www.techpowerup.com

Wieland Pumpless AIO CPU Cooler Tested

Wieland AIO CPU cooler is a prototype closed-loop liquid CPU cooler that lacks a pump, or any form of active coolant flow between the heat-source (the CPU block), and the sink (the radiator). The cooler works on the principle of thermosiphon, where the temperature differential between the source...
www.techpowerup.com www.techpowerup.com

I'll not move the posts over, as last time I did, it put the news OP in date order and it wasn't the header anymore. I'll leave this open for now.

Edit: I've asked W1zzard if he can merge them.
 
That's essentially tower air cooler performance, if not worse, nothing groundbreaking here. Single-phase thermosiphons simply do not work for the power flux emitted by modern day CPUs, regardless of the elaborate heat exchanger design.

As a secondary HX heat pump it does work well (as proven by solar where you normally get 1kW/m2), it just doesn't work with CPUs which have about 1500x higher heat flux (up to 150W/cm2).
 
Haile Selassie said:
That's essentially tower air cooler performance, if not worse, nothing groundbreaking here.
Click to expand...
An innovation should bring some improvement. The trouble is, that a small, cheap and light 4 heatpipe 1 fan air cooler can easilly cool 180W - that is exactly the same performance of the "innovation" cooler after 30 minutes.

Somebody mentioned, that it is just a prototype: even a prototype can have the raw performance, just lacking some user friendly and design elements.
 
Last edited:
It sounds good but they still can do planned obsolescence with the liquid quality.
No one wants to make anything that works forever.
 
dgianstefani said:
Fixes one of the main issues I have with AIOs - weak pumps that wear out and are noisy.

Cools 200 W no problem, 36000 points instead of 36300 with a corsair 240 AIO, prototype, still room for improvement.

Custom loop will still be much better, but this fixes the reliability and noise issues with AIOs, while eliminating cost from manufacturing.

Click to expand...
This looks very interesting! I am fascinated!

I would like to see a 280/420 version of this.
 
Last edited:
BoggledBeagle said:
An innovation should bring some improvement. The trouble is, that a small, cheap and light 4 heatpipe 1 fan air cooler can easilly cool 180W - that is exactly the same performance of the "innovation" cooler after 30 minutes.

Somebody mentioned, that it is just a prototype: even a prototype can have the raw performance, just lacking some user friendly and design elements.
Click to expand...
Air cooling directs the heat over other components, liquid cooling directs it outside of the case.
 
dgianstefani said:
Air cooling directs the heat over other components, liquid cooling directs it outside of the case.
Click to expand...

True, but I still worry about the VRMs not getting air flow, albeit warm.
 
Shrek said:
True, but I still worry about the VRMs not getting air flow, albeit warm.
Click to expand...
Normal case airflow is more than enough for that.
 
dgianstefani said:
Air cooling directs the heat over other components, liquid cooling directs it outside of the case.
Click to expand...
Yet you can see installation examples even from AIO manufacturers to install the radiator as an air intake in the case.

And also most CPU air coolers just send the warm air to the back of the case, where it is expelled by the case fan/s.
 
That looks cool and all (excusing the pun), but can it handle two points of heat source?... I'd want my currently watercooled GPU in the same loop (although I suppose I could accept dual loop; one radiator at the top & one in the front of case.)

12700k & 3080 10G in an NR200 original bog standard mesh-sided case, vertical watercooled GPU, dual radiator custom loop, with 280mm rad in the bottom and 240 slim rad in the top. (I had to dremel the back of the case slightly, to vertically re-align the GPU between the two radiator&fans setup).

The loop goes CPU pump/block->bottom rad->GPU->top rad->back to CPU pump block... That way I use the top radiator as the reservoir, so it's sorta like a custom-AIO...

I have M-F quick-connects on the tube from top radiator to CPU and a separate reservoir with M&F quick connects, which I can temporarily insert to fill and bleed the loop...

40% pump, 40% fans, undervolted and slightly overclocked CPU and GPU, 67C max in gaming across CPU, GPU die & VRAM (after a few hours, they all pretty much equalise to this temperature), virtually silent operation (I really have to concentrate on listening for the noise)...

I even went as far as doing a Nocpooa fan-replacement-mod the power supply because that was the only noticable noise, and somewhat frustrating after everything I'd done to make the system silent XD
 
This does not seem to be configurable in any way, it is a closed system with a special working fluid with probably elevated pressure inside.
 
Neat prototype, but I don't expect to ever see high volume production of these.

With how Icegiant's ProSiphon turned out, and the apparent lack of interest in by the bigger players in the market, I suspect there are some inherent hurdles to this sort of cooler that don't make it suitable for mass market.
 
masterdeejay said:
It sounds good but they still can do planned obsolescence with the liquid quality.
No one wants to make anything that works forever.
Click to expand...
Personally, I think this is why we've not seen these coolers come to market.
(Other than the fans) There isn't anything *to* wear out. (Toxicity and Regulations on choice refrigerants, is the other)

It's more/less a
"Heat Pipe Cooler, with extra steps"

The pressures and refrigerant involved necessitate tubing and fittings that have little-to-no wear (other than ground-level Ozone, UV exposure, etc).
I could see the condensing heat exhanger eventually succumb to thermal cycling (cracking) but, is that even an issue in fixed (low vibration) installations?

Haile Selassie said:
That's essentially tower air cooler performance, if not worse, nothing groundbreaking here. Single-phase thermosiphons simply do not work for the power flux emitted by modern day CPUs, regardless of the elaborate heat exchanger design.

As a secondary HX heat pump it does work well (as proven by solar where you normally get 1kW/m2), it just doesn't work with CPUs which have about 1500x higher heat flux (up to 150W/cm2).
Click to expand...
You are clearly more familiar with the maths involved than I am.

However, from my 'special interest' observations, such a broad statement could not be true; there's simply too much variation in refrigerants.
Examples:
R-718 is seemingly physics-breaking in its thermal capacity.
R-717, 'most efficient', but is corrosive and toxic.
Both, are components in urine :laugh:
 
Last edited:
@ShiBDiB
just because "we" dont want a pink tricicle, doesnt mean no one will.. ;)

i care about silence, but i also like that i can dump cpu/gpu heat outside my case, dropping case/chipset/vrm/drive temps by +25C,
and my gpu doesnt reach 43C under gaming load, so i dont have any boost "throttling", not happening with an aircooler,
unless in a completely separate chamber/external enclosure, and that's way more expensive.
and improper installation (as intake) has to do with the user, not the hw, besides that all AIO i used/installed,
offered more than 1 way of installing it (not just intake).

ignoring the difference in weight, vs a same perf air cooler.

@Space Lynx
maybe those, unless something changed, both have copper blocks:
for "higher" fan pressure
Arctic
better this, as rad is copper
Alphacool

@dgianstefani
technically speaking, all pumps will fail at some point.
glad you didnt say "all" (aio):

the alphacool has G1/4 connections, so replacement/upgrade is possible,
and it doesnt cost anywhere close to the EKs, ignoring i can start with the AIO,
and make a loop later, without having to shell out the money (right away),
or parts ready, to make it work..


@BoggledBeagle
only if the intake port would be on the other end.

@Shrek
more of a case temp/vrm cooler issue.
on a decent board, i havent seen mine hitting 70C under full load, even with case fan doing 600rpm and PBO.
 
You must log in or register to reply here.
Back
Top