Lols' Quick'nDirty HDD Cooler

[lolito_fr]

I wasn't going to bother posting this, but it turns out to be surprisingly effective, so…

Since realizing that my Seagate 'cuda IV would be a lot quieter if I decoupled it from the case, I also discovered that it was running a lot hotter. As it was now installed in a spare CD bay, with DIY silent-blocs, it was no longer able to transfer any heat away through direct metal-metal contact. In fact, at one time it got to 52°, so I had to leave the cover off that slot to allow fresh air in, but this was also letting (some) sound out
So I thought maybe it would benefit from some water cooling

My spec was that the design should be:

- Cheap and simple to make (KISS)
- Have minimal flow resistance
- Mountable in the CD bay, with silent-blocs or foam, etc
- Able to keep one or possibly two hard drives below say 35°C

OK, so it isn't up to MMZ standards , but it only took an afternoon to put together and most importantly, it works! (drive temp now maxes out to 3-4 °C above water temp)

[diablito2k]

cool man. how long did it take to make them ?

[pippin88]

Neat. The way you did the barb ends especially.

What's with the springs? Used for bending?

[iakovl]

very cool man.

[AntiBling]

Nice, thanks for sharing. Hope you wont mind if I copy that sometime in the future!

[lolito_fr]

hahaha copy away!
No patents on it

The springs (inside the tubes) are just there to eek out an extra couple of °C/W, by creating some turbulance. I realise now that they're probably just killing my flow for nothing, seeing as the drive is cooled a lot better than expected :shrug:

[bigben2k]

Yep, that's the most effective design, for the effort. At 3-4 degrees above water temp, it makes one wonder if making up anything else more complex is actually worth the effort.

[killernoodle]

I approve.


I like the fit, does it screw drectly into a 5 1/4" bay?

[lolito_fr]

Yep, it is quite effective for such a simple design. Especially considering that it's probably seeing less than a quarter of my total loop flow (about 1GPM), most of it is going through the other CPU outlet which will eventually have a GPU block. Also I didn't even use thermal compound or tape between the drive and the copper, so there's probably a bit of room for improvement

It screws in the 5 1/4" bay, but via some little standoffs made of 1/4" tubing to isolate drive vibration from case). Works quite well

[Gooserider]

Nice job, simple and effective. Just out of curiosity, what are the specs on the tubing sizes and plate thickness?

Gooserider

[lolito_fr]

Thanks

Something I forgot to mention (probably the most important bit!) was that I ground a small flat on the tube (about 3mm wide) to help mate it with the plate.

Tube is 10/12mm, plate is 1.5mm. Its pretty thin, and I did have some "bendage" which I wasn't that happy about. Should have used at least 2mm with hindsight.

[kronchev]

you do realize youre going to have to start selling them

ill take one

[lolito_fr]

LMAO kronchev!

Hmmm. Ok, whats it worth!?

[MadHacker]

Originally Posted by lolito_fr

Thanks

Something I forgot to mention (probably the most important bit!) was that I ground a small flat on the tube (about 3mm wide) to help mate it with the plate.

Tube is 10/12mm, plate is 1.5mm. Its pretty thin, and I did have some "bendage" which I wasn't that happy about. Should have used at least 2mm with hindsight.

I have been thinking of making almost exactly the same thing you did...
After I saw how efficient yours is...
But I was thinking of cutting the pipe length ways and opening it up a bit and soldering that to the plate instead. But that might be excessive or unnecessary.

[lolito_fr]

Yeah, I though about that as well MadHacker. Just seemed like too much effort though
Might be tricky getting a nice straight cut, hence a good seal, if you only have basic tools. Also be a lot more difficult to connect the ends.

[FizzledFiend]

so your getting these temps with cooling just the sides of the drive? hows that work?

[Becks]

Thats real slick, I was getting ready to do the same thing over spring break... I bought some 1/2" square AL cause I figured that would make better contact on the sides... got some 1/2 brass barbs that just fit in the ends, figured with some jbweld it would work well.

Tho then when I started to think about it I don't wnat to use AL and can't find square copper anywhere around here.

Anyways that looks real nice, and I'm going to worry less... my goal was withing 5C of watertemp and i was thiking that was unrealistic but good to see its not

Nice job.

[Moparchris]

awesome man, i was thinking of doing something similar but with cases around the whole thing that will fit in the large bays with a bit of this sound proofing foam stuff i have in it

[msv]

Originally Posted by FizzledFiend

so your getting these temps with cooling just the sides of the drive? hows that work?

There´s been some discussion on this forum considering what to cool on a drive.
A disk doesn´t need much cooling in the first place, and all the warm parts (spindle motor, electronics) are mounted on the same frame on the disk. That very frame happens to be one of the halves on the sides of the disk, so cooling the sides cools the entire disk, since the heat is conducted to the sides of the disk. Take a careful look at a disk and it becomes pretty obvious, it´s simpler than describe it, actually, (English isn´t my native language).
regards
Mikael S.

[Pureh20]

nice work, like the polished copper, im thinking of doing what you did but on the tops, can you think of any problems in doing this as my drive doesnt have any holes for expanding heat (seagate barracude 7200.7 plus, 8mb cache)?

[Gooserider]

Pureh20:
nice work, like the polished copper, im thinking of doing what you did but on the tops, can you think of any problems in doing this as my drive doesnt have any holes for expanding heat (seagate barracude 7200.7 plus, 8mb cache)?

I tried doing the top plate number with 1/8" plate, and 3/8" OD copper tubing, I actually made several, some with a plate on each side and the tube in the center, and some with the plate just on one side. My idea was to stack a bunch of drives with alternating plates, so that each drive would be cooled top and bottom. I had a total of 4 passes of tubing, starting at one end corner, go to the other, U-turn, come back to the first end, U-turn, go back, make a 3rd U-turn, and come out the same end as I started, but on the opposite edge.

Great in theory, not so great in practice. Despite using a tubing bender initially, then trying a custom die, I found that it was next to impossible to bend the tube tightly enough without kinking and distorting it. The end result is a set of blocks that look OK, but have highly restrictive flows, on the order of 1-2 LPM Since drives don't put out much heat, this will probably be OK, but I would have liked to have done better.

I can't reccomend the technique the way I did it. Perhaps if one used a simpler curve (only one U-turn with a larger radius?) and / or larger tube (which has it's own problems) it would work better, but I think Lolito's design is better since he can use larger tubes with less problem, and his design avoids the bending and kinking that I had.

Gooserider