Why Doesn't a Fractal Heatsink exist?

[Twigsoffury]

...you can have pressure without compression?

Yes, turbulence totally aids thermal transfer. But it also causes heat in its own right. And at a certain scale, micro effects tend to overcome macro effects...

read through that article i posted it sorta explains it better then i can, maybe you'll get more out of it then i do.

lots of these questions seem like things you'd need R&D to figure out.

[mDust]

The problem I still see with this design, is the turbulence. Yeah, some turbulence is good for heat transfer, but you need to balance that with flow restriction (look at an EK waterblock, they figured out the balance long ago). The head pressure of most PC water cooling pumps simply isn't high enough to force water through a series of tubes that change direction dozens of times each. You would need a large, noisy, high-pressure industrial water pump to do so. The pressures needed for decent flow would likely not be held by the standard fittings and tubing (at least between the fractal cooler and pump). One way to remedy this, would be to have many, many shorter fractal tubes per device so the total 'area of flow' is at least a few times that of your water tubing. You could also round out any right or sharp angles...the simple change in flow direction will create more than enough turbulence for the desired performance increase while not over-burdening the pump too much.

Also, you've managed to increase thermal transfer from the water to the radiator by several times but you haven't increased the efficiency of thermal transfer from the radiator to the air at all. As you work from the heat source, to the waterblock, to the water, to the radiator, to the air, each segment must be able to absorb heat at a rate greater than or equal to the previous segment or else there is no point to any efficiency improvements anywhere in the entire system. You need huge air fins, insane airflow, ridiculously low air temps, or some radical redesign of that component as well.

[Twigsoffury]

.

if cork screw pumps can force liquid concrete up 2,600 feet to the top of the burj dubai, i bet they can also force water through a foot or two long tube.

its how dams get water back up through the damn to the res lake

And super chargers get insane pressures.

I bet if you shrunk those designs down they'd be effective enough to push water through at reasonably acceptable noise levels.

the traditional blade inside a water pump looks kind of pussy after seeing corkscrew style pump blades. TBH.



And looks like you guys would need to get used to threaded fittings.

Have no idea why they already aren't using threaded fittings in PC cooling systems.

kind of obvious that they are going to leak regardless of the pressure when you look at the fitting from a face value.

Also, you've managed to increase thermal transfer from the water to the radiator by several times but you haven't increased the efficiency of thermal transfer from the radiator to the air at all. As you work from the heat source, to the waterblock, to the water, to the radiator, to the air, each segment must be able to absorb heat at a rate greater than or equal to the previous segment or else there is no point to any efficiency improvements anywhere in the entire system. You need huge air fins, insane airflow, ridiculously low air temps, or some radical redesign of that component as well.

I sense someone just skimmed over my posts and then hit "Respond" and typed out some s@#t to attempt to make me feel bad.

[dr.walrus]

What we're getting at is this: I don't want that anywhere near my computer!

[Twigsoffury]

What we're getting at is this: I don't want that anywhere near my computer!

well I'll take the plunge first with a old pentium 4 system.

now for the 50 grand for building it and fabricating everything.

obama is still handing out grants and free money for stupid purposes right?

[blaze15301]

well I'll take the plunge first with a old pentium 4 system.

now for the 50 grand for building it and fabricating everything.

obama is still handing out grants and free money for stupid purposes right?

Why yes he is. Withe all the damn money we are saving from cancelling the shuttle program and all so more dead beats can stay living off the system.

[Oneslowz28]

I have some photos from CES of a pretty radical water block design from Cooler Master. I need to see if its OK for me to post them. While its not a fractal, the surface area was more than tripled over traditional water blocks.

[Fuganater]

Way too much thought has been put into this thread...

Just stating the obvious for everyone.

[mDust]

if cork screw pumps can force liquid concrete up 2,600 feet to the top of the burj dubai, i bet they can also force water through a foot or two long tube.

Have no idea why they already aren't using threaded fittings in PC cooling systems.

I sense someone just skimmed over my posts and then hit "Respond" and typed out some s@#t to attempt to make me feel bad.

1) Exactly, you need an industrial pump...like I said.
2) Cost and difficulty of modification.
3) I've read everything you've posted and I've never skimmed over a post anywhere and then responded. I've also never attempted to make you feel bad. Like you said earlier, you need people to point out flaws so you can improve your idea. Stop feeling bad and make this damn thing work.

All I was saying there is that if your fractal cooler is absorbing 40x more heat from the water than a normal radiator, then the fins on the side need to be able to dump all of that heat (and more) into the air...which they are not going to do.
Let's say you leave the fins how they are:
The cpu will dump x watts/sec into the water block (a tiny fraction will dissipate into the air along the entire journey to the radiator, but it depends on temp differentials and other complicated variables, so we'll just ignore that), if the block cannot absorb that much heat, the cpu overheats. If it can, the block dumps x amount of heat into the water...you'd have to do something really terrible to the water in order for it to not be able to absorb that heat, so it's assumed that the water dumps x amount of heat into your fractal cooler. You've designed it so it can absorb much, much more heat than necessary...but when the heat starts to dump into the air fins on the side to dissipate into the air, the fins can only dissipate .5x/s. What happens is like a traffic jam. The temp difference between the fins and the interior of your fractal cooler becomes nil, and so does heat transfer. As the temp of the fractal cooler rises, it stops cooling the water effectively, which stops cooling the waterblock effectively, which stops cooling the cpu effectively. Even if the air fins can keep up and effectively dissipate the required amount of heat to cool the CPU, if it can't dissipate as much as your fractal cooler can absorb, then there is no point to improving any other part of the system.
Get designing and surprise me with something new and unexpected.

Way too much thought has been put into this thread...

There is no such thing as 'too much thought', it's rampant 'too little thought' that is dangerous...which describes most people nowadays.