Scratch Build – In Progress 7th.GeN - Rebooted ^^

water cool all the drives, it would be cool. no pun intended.

 

Have had it in my thoughts, but havent found a decent way to implement watercooling for this amount of drives 0.o

Anyone know how much heat that can be dissipated from aluminium? Some heatspreaders in front and in the back should have some effect on the overall, if waterblocks would be implemented on the back of them, the heat should be transfered as the plates would be colder then the coolant? Or am i way off here?

 

Something I made up in paint, you could use custom blocks for the HDD's...

 

Oh JESUS not again...

*Subb'd*

 

If only it would be that easy =/

Most of the heat from harddrives are located on the top of them, just cooling the sides of the disks wouldnt be sufficient, would need a single waterblock in between each pair of drives, guess that would effectively be around 8 blocks or so... hmm... actually not all that bad 0.o

 

Something like this?

 

I didnt think you could become even more insane! 8xHHD waterblocks...

 

Thats actually a nice idea ^^

Would be nice if it could be implemented on a drive row of lets say, 4 drives or so aswell,

Ill see what i can figure out 0.o

How much heat in wattage does a disk produce anyways?

 

I don't think heat dissipation rate varies from metal to metal, it's surface area that affects that. Aluminium stores heat effectively (heat capacity), that's why it rocks for passive cooling. But for water cooling I think you'd be better with copper all the way as it's the better conductor.

And I have to echo others on this thread. YOU IS A MENTAL...AND THAT IS AWESOME

 

less than 5-7W each, usually.

This myth that something "stores" heat (=better cooling!!!!) is gonna kill me one day or another . There's nothing important aside from surface area, thickness of the metal and conductivity. if the former two are equal, the only important factor is conductivity: no matter if its passive cooling or whatever. However the conclusion is right: go with copper if possible and if you like the aestethics. Also, ist better not to mix different metals in the same waterloop due to possible galvanic corrosion.
(Physics student here)

 

So... Mr.Physics student, how much heat can be transfered trough a waterblock to a larger surface area? How much heat would a passive block be able to remove if you take standpoint in a room with low circulation and normal room temp of around 20c, the weight of copper is quite different from that of alu aswell. Would a passive cooler on the outside of the case have any affect when its ran along with two 480 radiators? atleast the heat wont be stored fully inside the case since half of the frontplate is on the outside

I know from benchmarks that each 480 radiator should be able to dissipate around 740w of heat each, with 2000rpm fans at 24db, so thats a total of 1480w on the two in the top module, wich should be more then sufficient to cool a cpu, chipset and some 5970's ^^

Dont see why producer's doesnt utilize the large surface areas around a case for cooling if it might have a effect on the overall cooling 0.o Though, large quantities of alu would put a heftier pricetag i guess.

 

@Odin. Dude, sorry to nitpick but I'm not suggesting that a greater heat capacity necessarily = better cooling. I'm saying that aluminium has a greater heat capacity than copper. Copper is by far the better conductor and therefore the better material for use in a computer heatsink given the constant heat source (CPU,GPU, etc.). Alas it is also, unfortunately, far more expensive than aluminium or we'd all be packing our systems full of it

Having said that silver is an even better thermal conductor than copper.
...
Grevaeg, how rich are you feeling

 

Still waiting for industrial diamond heatsinks.....

 

Well, the factors are many, though i would say that a passive radiator like the one you designed would not have much effect on the cooling of the whole system. Its possible that you would gain 0.5-1 degree Celsius but that's it. Compared to a 480 radiator it has a way lower surface area. What you could do is:

- first try to get a higher surface area, say machining the cover with channels (is this the right word?). If well designed, you could easily triple the surface area while lowering the weight, and keeping somewhat the same design.

- lower the thickness of the copper in contact with the waterblock. This will speed up the heat transaction process.

- Get as much as possible surface contact between the external cover and the internal waterblock, which will probably have to be massive to cover as much as possible of the external radiator. If possible, put the water in direct contact with the esternal copper cover. If its aluminium (less cost, weight and performance) then put a copper layer in between: you dont wanna mix metals. Also, the mass of the copper/alu and the relative heat capacity will only affect how fast you reach the stable temperature, not what this stable temperature is. The more mass=the slower temperature will change, equilibrium temp remains the same (given same surface area etc.)

the formula for heat transfer from a sheet of a material is:

K*surface area*deltaT/thickness​

where K is a constant, for alu is around 205 W/(m*Kelvin), for silver 415 W/(m*Kelvin). For copper I dont have the numbers here but its in the middle.
So, for example, an aluminium piece 1cm thick, 0,02m^2 will potentially be able to transfer 205*0.02*15/0.01= 3075 Joules of heat per second (=3075Watts), if one side of the surface is kept at 15K (=15°C) above the other one.
Then of course you have to consider how many W the water will be able to give to the aluminium plate, and how many the air will be able to absorb (which is much lower given K for air is 0,02, compared to the 205 of alu. For water its 0,6.). You could use diamond . Its K is between 1000 and 3000, depending on the quality XD.

Hopes this helps, however on Wikipedia you can find everything regarding this subject. And dont be afraid to ask.

sorry if i seemed rude, didnt intend to. The thing about passive cooling makes me angry every time.

 

Carbon nanotubes and graphene Cheesecake

 

Well... there is a budget limit at 2000$ for the casebuild with watercooling, no hardware present, as long as its within those frames, i really just care about the final product and how it will perform ^^

 

Watercooled drives?

So... been working a bit with some basic models again, please give me some input if a design like this could work at all ^^

The general idea would be to have waterflow on the top and bottom of the harddrive mount, while the disks heat up the sideplates they are mounted on, not sure how much heat that can be removed from a disk trough the sides though, but hey, if its around 8-15w a disk, this should be sufficient?

 

Hmmm.... just noticed, the waterflow holes in the top will be blocking for sata and power cables, guess ill have to move them a bit ^^

 

I would believe that the middle two HDD's will have a harder time dissipating heat that way.

Then again, how much heat does a HDD produce?

For me, one single 80mm 1000RPM fan is enough to keep 4 7200RPM HDD's running at barely 7 deg above ambient. It's like that in my admittedly stock OEM case.

 

Thats kinda why i havent been watercooling drives earlier aswell.... its so much easier to just place a low rpm fan on them, if i were to watercool the drives, then it would most likely be for the design challenge, and nothing else 0.o