Cooling Petition for short tube sff specific 120mm AIO

[Thehack]

In that case, something like this would only have one strike against, vs the 645lt which has both nonoptimal install and elbows. This will also be primarily for the sff community. I'd like to reach out to other hardware companies but asetek is so far the only one show interest.

Does asetek have an in radiator pump design?

 

[alexs]

IMO "shorter" tubing is not the way to go, especially in SFF.

The users should be able to cut tubes to length, to fit a particular build perfectly. Using G1/4 threads throughout should be enough to address maybe half the points from the first post. Fittings are easily available, and more importantly threaded fittings can be combined to get around layout limitations. It wouldn't be a "true" AIO, more like a water cooling kit, but who cares.

I did something similar with a BeQuiet AIO, but with larger clearances, and with a 240mm radiator.
BQs have ports on the flat sides however, both for the pump and for the radiator.

Afaik Alphacool has or had radiators and pump blocks with edge ports and G1/4 which would allow building something like this right away.

 

[Thehack]

IMO "shorter" tubing is not the way to go, especially in SFF.

The users should be able to cut tubes to length, to fit a particular build perfectly. Using G1/4 threads throughout should be enough to address maybe half the points from the first post. Fittings are easily available, and more importantly threaded fittings can be combined to get around layout limitations. It wouldn't be a "true" AIO, more like a water cooling kit, but who cares.

I did something similar with a BeQuiet AIO, but with larger clearances, and with a 240mm radiator.
BQs have ports on the flat sides however, both for the pump and for the radiator.

Afaik Alphacool has or had radiators and pump blocks with edge ports and G1/4 which would allow building something like this right away.

We can already do what you described, grab any pump block and an aftermarket radiator.

An AIO from the factory will be filled correctly, quick and easy to install. Not everyone want to cut their own tube, run their own leak test, and deal with a bad fill since you have no resevoir to eat the bubbles.

 

[alexs]

Never mind, I should have read the whole thread before replying. Completely misunderstood the point.

As a custom product, my only question would be, why not custom heatpipes.
AIO needs to bring a lot to offset the ease of installation and simplicity of a large C-shaped cooler.

 

[Thehack]

Never mind, I should have read the whole thread before replying. Completely misunderstood the point.

As a custom product, my only question would be, why not custom heatpipes.
AIO needs to bring a lot to offset the ease of installation and simplicity of a large C-shaped cooler.

Heatpipes can't be easily bent and moved to a number of motherboard layouts. A 120mm should be more powerful than low profile cooler up to 83mm clearance.

An AIO is not difficult. Install the backplate, attach the cpu block, and mount the radiator. The purpose is to have performance that exceeds a C type cooler at this size.

 

[W4RR10R]

A 120mm should be more powerful than low profile cooler up to 83mm clearance.

I'm kind of skeptical of that, I'd have to do some work but I'm fairly certain that there is little to difference in specific cooling capacity (cooling/volume) of a a water cooling loop vs an air cooler. Heat pipes are fairly efficient and even offer some cooling on there own due to the latent heat of vaporization, energy lost in the phase change from liquid to gas. The main thermodynamic advantage of a water cooler is increased surface area, usually because you can move the cooler away from the cpu to a less space restricted area of the case. You could make an air cooler that placed its fin stack at the top of the case (where say a 240mm rad would go) and make it the same dimensions 240x120x30 mm with 2 120mm fans mounted to it and it would cool just as well if not better than any water cooling loop (assuming the contact between the IHS and cold plate are equal between the two). The issue is that unlike hose, heat pipes aren't flexible so you would have to make a different cooler for every case and motherboard combo that you wanted to support, so an AIO Loop or custom loop are a better option for compatibility.

The Noctua L12s (70mm) and C14s (115mm), would likely out perform an AIO in that small of a space.

 

[Thehack]

I'm kind of skeptical of that, I'd have to do some work but I'm fairly certain that there is little to difference in specific cooling capacity (cooling/volume) of a a water cooling loop vs an air cooler. Heat pipes are fairly efficient and even offer some cooling on there own due to the latent heat of vaporization, energy lost in the phase change from liquid to gas. The main thermodynamic advantage of a water cooler is increased surface area, usually because you can move the cooler away from the cpu to a less space restricted area of the case. You could make an air cooler that placed its fin stack at the top of the case (where say a 240mm rad would go) and make it the same dimensions 240x120x30 mm with 2 120mm fans mounted to it and it would cool just as well if not better than any water cooling loop (assuming the contact between the IHS and cold plate are equal between the two). The issue is that unlike hose, heat pipes aren't flexible so you would have to make a different cooler for every case and motherboard combo that you wanted to support, so an AIO Loop or custom loop are a better option for compatibility.

A radiator itself has a significantly denser fin stack. As you've said it yourself, surface area is the main driver of cooling capacity. Nowadays with highly pressure optimized fan design, you can get more performance.

Other factors to consider:

1. Water has high specific heat capacity, so it reduces the highs usage spike and cools off during the low.

2. For larger water cooling systems, idle time is used to cool the water which is like a heat tank.

3. Microchannels are very effective at heat transfer. This may be a bigger effect than we think, but I haven't seen any evidence.

4. It is much easier to manufacture a large but slim radiator. As air travels through the fin stack it heats up, reducing its effectiveness to cool the fins. A large 360 radiator has significantly more fresh L/min than a D15. A wide but shallow cooler as you've mentioned would be a nightmare to manufacture and package. It also be difficult to mount safely.

A 120mm AIO, if built to about 95mm clearance height, will pretty much decimate any air cooler of a similar clearance.

I believe it is also an effect of better fan tech out pacing cooler tech. Noctua just did a press release for their dense fin 120mm cooler, so there are still gains to be had in performance/volume on the side of air coolers.

Finally, as I've stated, being able to slightly move the location of the cooler is a HUGE benefit even if the performance/volume is the same. This is sff we're talking about and we design by the mm. A solid heatsink design you have to design more clearance around it to fit different motherboards.

 

[W4RR10R]

A radiator itself has a significantly denser fin stack. As you've said it yourself, surface area is the main driver of cooling capacity. Nowadays with highly pressure optimized fan design, you can get more performance.

Other factors to consider:

1. Water has high specific heat capacity, so it reduces the highs usage spike and cools off during the low.

2. For larger water cooling systems, idle time is used to cool the water which is like a heat tank.

3. Microchannels are very effective at heat transfer. This may be a bigger effect than we think, but I haven't seen any evidence.

4. It is much easier to manufacture a large but slim radiator. As air travels through the fin stack it heats up, reducing its effectiveness to cool the fins. A large 360 radiator has significantly more fresh L/min than a D15. A wide but shallow cooler as you've mentioned would be a nightmare to manufacture and package. It also be difficult to mount safely.

A 120mm AIO, if built to about 95mm clearance height, will pretty much decimate any air cooler of a similar clearance.

I believe it is also an effect of better fan tech out pacing cooler tech. Noctua just did a press release for their dense fin 120mm cooler, so there are still gains to be had in performance/volume on the side of air coolers.

Finally, as I've stated, being able to slightly move the location of the cooler is a HUGE benefit even if the performance/volume is the same. This is sff we're talking about and we design by the mm. A solid heatsink design you have to design more clearance around it to fit different motherboards.

1. Water is the working fluid in heat pipes used for the temp range of computers.

2. You can make the fins denser in an air cooler (much like noctua just did).

3. Air coolers are very simple (compared to radiators) to manufacture on a large scale, the fins are stamp cut Alu sheet and the heat pipes are bent and fixed to the cold plate and pressed through the fins.

4. Most importantly I'm not suggesting that this is a bad idea, merely throwing my 2 cents that I don't believe there are significant gains to be had in use cases this small.

 

[Thehack]

1. Water is the working fluid in heat pipes used for the temp range of computers.

2. You can make the fins denser in an air cooler (much like noctua just did).

3. Air coolers are very simple (compared to radiators) to manufacture on a large scale, the fins are stamp cut Alu sheet and the heat pipes are bent and fixed to the cold plate and pressed through the fins.

4. Most importantly I'm not suggesting that this is a bad idea, merely throwing my 2 cents that I don't believe there are significant gains to be had in use cases this small.

1. Water is the working fluid, and it has a high heat capacity. This means it takes longer to reach saturation (but also longer to completely cool it off) and is incredibly useful as a feature.

2. Exactly. This is sff so performance/volume is king. Denser fins=better performance in equivalent sized radiator (heatpipe included). I'd be interested in a dense fin top down cooler and seeing what they can do.

3. We're talking specifically about designing the heatsink like you've mentioned, one where the heatsink stack extends ridiculously long with two fans.

4. I highly disagree. I have yet to find any testing where a 100mm top down cooler can bet a 120mm AIO.

A C14S comes close but it is 15mm taller so significantly bigger. A 120mm AIO seems to be equal to it.

 

[Necere]

I just wanted to chime in here on a couple of points:

1. Water is the working fluid in heat pipes used for the temp range of computers.

True, but the actual quantity of liquid in heatpipes is insignificant compared to what's in a water loop. Most of the heatpipe is empty (or rather, filled with gas), with only a small amount of working fluid. The working fluid boils and vaporizes at the hot end, travels to the cold end, and turns back to a liquid state, where it's wicked by capillary action (sintered powder or other materal) back to the hot end. @Thehack's point was that there's enough water in a water loop that it takes a while for it to heat up, providing a kind of buffer to your CPU temperature. You'll see this in action where it can take 10-20 minutes or longer (depending on exactly how much liquid is in the loop) for temperatures to stabilize during a load test, whereas with an air cooler temps will hit their peak pretty much instantly.

2. You can make the fins denser in an air cooler (much like noctua just did).

You can, but there's a trade-off between surface area and airflow restriction when you increase fin density, such that you need higher CFM/static pressure fans to make use of the increased surface area. That means more noise, of course, so you have to decide what's important.

Radiators have an advantage in that the fins can be made significantly thinner than air coolers, allowing them to have a higher surface area to restriction ratio. If you've handled a rad before, you'll know that the fins are tissue-paper thin and fragile because of it. In comparison, an air cooler's fins are typically much thicker (though still <1mm), so you can't pack as many into the same space.

Another advantage radiators have is a more even distribution of heat throughout the rad than air coolers can manage, resulting in more effective use of the available surface area.

 

[Thehack]

I just wanted to chime in here on a couple of points:

True, but the actual quantity of liquid in heatpipes is insignificant compared to what's in a water loop. Most of the heatpipe is empty (or rather, filled with gas), with only a small amount of working fluid. The working fluid boils and vaporizes at the hot end, travels to the cold end, and turns back to a liquid state, where it's wicked by capillary action (sintered powder or other materal) back to the hot end. @Thehack's point was that there's enough water in a water loop that it takes a while for it to heat up, providing a kind of buffer to your CPU temperature. You'll see this in action where it can take 10-20 minutes or longer (depending on exactly how much liquid is in the loop) for temperatures to stabilize during a load test, whereas with an air cooler temps will hit their peak pretty much instantly.

You can, but there's a trade-off between surface area and airflow restriction when you increase fin density, such that you need higher CFM/static pressure fans to make use of the increased surface area. That means more noise, of course, so you have to decide what's important.

Radiators have an advantage in that the fins can be made significantly thinner than air coolers, allowing them to have a higher surface area to restriction ratio. If you've handled a rad before, you'll know that the fins are tissue-paper thin and fragile because of it. In comparison, an air cooler's fins are typically much thicker (though still <1mm), so you can't pack as many into the same space.

Another advantage radiators have is a more even distribution of heat throughout the rad than air coolers can manage, resulting in more effective use of the available surface area.

Thanks for the additional clarification. Your explanation is very helpful and is more insightful.

There are few tests that really compare performance/noise (one of the few is gamer nexus but they primarily deal with liquid based cooling). Do you have any anecdote of how a C14S (single fan) compare in performance to a good quality AIO? We can account for using our aftermarket fans as well, since it is a popular thing to do.

You haven't chimed in on the idea of this yourself. I think NCase M1 may benefit from something like this. Do you have any thoughts on it?

 

[alexs]

Radiators have an advantage in that the fins can be made significantly thinner than air coolers, allowing them to have a higher surface area to restriction ratio.

Do they? Imagine a 120mm radiator, but with heat pipes in place of water pipes.
Air coolers tend to be frameless and self-supporting, that's why they need to be sturdy.
But it doesn't mean an air cooler cannot be made with a frame and thin copper fins.

 

[Thehack]

Do they? Imagine a 120mm radiator, but with heat pipes in place of water pipes.
Air coolers tend to be frameless and self-supporting, that's why they need to be sturdy.
But it doesn't mean an air cooler cannot be made with a frame and thin copper fins.

It doesn't really matter if it is technically possible or not. We're asking for a modified AIO, not trying to manufacture a new type of cooler.

I'm not sure why you're so deadset on this heatpipe design to be of similar performance for volume so we that can throw away two advantages of water: able to easily move the location of heatsink and the high heat capacity.

Regardless, I think continuing on that discussion path is really off topic. Comparing the technical performance of the two type of designs is interesting but is not the point of this thread, which is to have a tailored product for our specific application.

 

[Necere]

There are few tests that really compare performance/noise (one of the few is gamer nexus but they primarily deal with liquid based cooling). Do you have any anecdote of how a C14S (single fan) compare in performance to a good quality AIO?

I can't recall seeing any comparisons offhand.

You haven't chimed in on the idea of this yourself. I think NCase M1 may benefit from something like this. Do you have any thoughts on it?

I think for the M1 specifically it may not be that useful, and actually less convenient overall. Longer tubes allow for enough slack to bring the rad (along with the fan bracket) completely out of the case for access to the motherboard for servicing/installation, whereas short tubes would prevent that. I think it would make more sense for case layouts where the rad is adjacent to the motherboard (e.g., something like the Dan A4's 120mm optional rad mount) rather than directly over it.

 

[Necere]

Do they? Imagine a 120mm radiator, but with heat pipes in place of water pipes.
Air coolers tend to be frameless and self-supporting, that's why they need to be sturdy.
But it doesn't mean an air cooler cannot be made with a frame and thin copper fins.

In principle, perhaps, but I think there may be some difficulties in brazing the fins to a heat pipe as you would to a radiator channel. I believe most if not all heat pipe heatsinks use a press-in attachment method, rather than any kind of heat bonding, and that may have to do with the unsuitability of heat pipes to that method. Heat pipes are tuned to operate in a rather narrow range, and the higher temperatures required for brazing could potentially damage or destroy the effectiveness of the heat pipe by cracking the wick or pipe. But I'm just guessing here based on what little I know about it - you'd want to research this for a more definitive answer.

 

[SaperPL]

I didn't see this thread earlier, but I'd very much like to have a 120 mm AIO with the same 90 degree knee bends as 645LT for Sentry 2.0 which would make it a less hardcore AIO build by easing the tube routing and also enabling use of the 2.5" slot over the AIO.

I hope this idea gets some traction in the future.

 

[AseDen]

Still keeping track in here.

My major concern remains air in the pump in 8/10 use cases.
If we can get some solid numbers on a reasonable solution, I'd be happy to bring it up internally, though

 

[riba2233]

I also vote for 120mm AIO with angled connectors on the radiator, it would be blessing for console cases. Also it would allow for thicker rads, like 34-38 mm.

 

[AseDen]

I also vote for 120mm AIO with angled connectors on the radiator, it would be blessing for console cases. Also it would allow for thicker rads, like 34-38 mm.

For normal rads we use either 27mm or 38mm, so I suppose that would just manage to squeeze in to those specs.

Will be keeping further track of this thread.
Any estimates on how many/which cases would need this support? (apart from the Sentry 2.0)

 

[riba2233]

For normal rads we use either 27mm or 38mm, so I suppose that would just manage to squeeze in to those specs.

Will be keeping further track of this thread.
Any estimates on how many/which cases would need this support? (apart from the Sentry 2.0)

My P-ATX could fit 38mm in that case. But even for 27mm it would be great for all console cases, sentry, salvo s402, I'm not sure if sligers console or conswole support aio. Also sandwich cases that support 120mm aio would also benefit from not having to bend tubes as much since there is not much space below the psu.