Variant "Homebrew" Hard Drive Enclosure

[Fëanor]

http://www.silentpcreview.com/files/ima ... index.html

My idea starts with this SPCR member's concept of gel-packs cushioning the drive inside an enclosure which I will then suspend in the case.

Here's what I'm thinking. I want to use some heat-sealable anti-static bags. I was thinking of making my own gel-packs using a vacuum sealer.

Everyone else is using those insulative gel packs meant for storing thermal energy rather than conducting it, and they are still still getting decent temperatures. I was thinking I would have even better results if I filled gel-packs with a thermally conductive compound such as Arctic Silver 5. I would probably vacuum-seal a second bag over the first one just because I'm not sure how tough these bags are.

What do you guys think?

[jhhoffma]

The problem is that thermal paste/compound is usually only conductive is thin layers. If you used a massive amount of it, it's not going to behave the same. Thermal paste works mainly because it removes the air trapped between the chip and the heatsink, not because it's conductive. I saw an article where someone used different household materials to see if any worked like thermal paste. He tried glue, lithium grease, and Miracle Whip or ranch dressing or something. Basically all worked the same (within a couple degrees) of the thermal paste, but they dried out over time.

If you use a big bag of Arctic Silver, not to mention being prohibitively expensive, it will still have some air trapped in it as those vacuum sealers really don't go that low pressure-wise, and if you did go too low, you might evaporate some of the ester oil, causing the mix to separate.

Practically, water is the best method of transferring heat (which is why we use it in water cooling setups, for cars and computers), but it's just not safe to use to encapsulate a hard drive.

[BillTodd]

http://www.silentpcreview.com/files/ima ... index.html

My idea starts with this SPCR member's concept of gel-packs cushioning the drive inside an enclosure which I will then suspend in the case.

Here's what I'm thinking. I want to use some heat-sealable anti-static bags. I was thinking of making my own gel-packs using a vacuum sealer.

Everyone else is using those insulative gel packs meant for storing thermal energy rather than conducting it, and they are still still getting decent temperatures. I was thinking I would have even better results if I filled gel-packs with a thermally conductive compound such as Arctic Silver 5. I would probably vacuum-seal a second bag over the first one just because I'm not sure how tough these bags are.

What do you guys think?

The gel pack has two jobs; one is to suspend the drive in a compliant, well-damped material (actually, it also adds mass that helps to reduce vibration as well). The second job, is to fill the gap between case and drive with something that conducts heat better than an air space (pretty much anything conducts heat better than air )

A bag filled with heat-sink compound would conduct heat considerably better than a water based gel (water isn't a great heat conductor*), but it would also be far less compliant (i.e. stiffer), so transmitting more vibration (hence noise) to the case. You may find that what you gain in cooling maybe lost in noise (if you see what i mean )

Whatever you use in your bag, if you use a 'vacuum' sealer, make sure you leave enough room in the bag for expansion (you really don't want the bag to burst!)



*water has a thermal conductivity of only 0.6 W/m K, compare that to 385 W/m K for copper or 419 for silver (Artic silver seems to have a conductivity of about 9 W/m K - if I have converted their 350,000Wm2'C [0.001" layer] value correctly and Servisol's nonsilicone thermal compound 0.9 W/m K ).

What water does have, is a huge heat capacity, which is why it is used in cooling systems because it can store, and remove, a large amount of heat for a given mass of coolant. (The gel used in packs is too thick to 'flow' at the temperatures of a HDD, so the heat transfer is all by conduction, rather than by convection, typical of water or air cooling systems)

[Fëanor]

So do you guys think it would be better if I used a non-conductive water cooling fluid like Fluid XP+ Ultra or something? (cost for a 32-oz bottle is roughly the same as a couple 12g tubes of AS5)

I wonder what density/viscosity would be best to use. Liquids transmit sound much better than solids, but solids transmit vibration much better than liquids.

I imagine I want something a little thicker than just the liquid coolant, so what could I add to thicken it a little (without making it conductive or corrosive)?

[jhhoffma]

Liquids transmit sound much better than solids, but solids transmit vibration much better than liquids.

Sound is vibration.

I would stick with the proven method of the gel packs. I don't know of any instances where anyone's killed a drive because of the heat. If they have, it was because they were enclosing a drive that shouldn't be enclosed in the first place.

[BillTodd]

a non-conductive water cooling fluid like Fluid XP+ Ultra or something?

Hmmm, water, pure water, is non-conductive. (I wonder what magically substance they add to the water? - probably a ph balancer of some description)

If you want to play with different viscosities just use wall-paper paste (methyl-cellulose) mixed to different consistancies.

[Fëanor]

Liquids transmit sound much better than solids, but solids transmit vibration much better than liquids.

Sound is vibration.

True. I'm no physics expert but let me give this a shot.

A solid will directly transmit vibration in the same direction and amount it is applied to the solid (ie if you hold a brick against your ear and I vibrate it, each and every shake will transmit to your ear and you'll feel it)

A liquid will disperse the vibration equally throughout the liquid. If you put your ear in the water and I vibrate the water, you will feel some weak water currents but you definitely won't feel the full force of the shaking.

Hmmm, water, pure water, is non-conductive. (I wonder what magically substance they add to the water? - probably a ph balancer of some description)

Once again, I'm no physics expert. But if you leak water on your computer components you are screwed. Non-conductive cooling fluids are designed so that a leak will not short out your system.

Also water is very corrosive whereas the cooling fluid is not.

Anyway it sounds like the general consensus is that it does not matter. Maybe I'll quit wasting my time and go with the standard gel packs (but maybe I'll experiment, we'll see)

[coylh]

With the right arrangement of several powerful magnets, you could "levitate" the drive in mid air.

Oh, wait...

[BillTodd]

But if you leak water on your computer components you are screwed.

Oh yes:), the water starts off non-conducting, but as contaminants disolve into it, it starts to conduct so, if the machine is powered, eletrolosys will rapidly destroy the board!

Non-conductive cooling fluids are designed so that a leak will not short out your system.

Yes, it appears that Fluid XP is such a liquid I.e it is something more exotic than a water additive, probably glycol based.

[Fëanor]

With the right arrangement of several powerful magnets, you could "levitate" the drive in mid air.

Oh, wait...

Too risky. The last time I was playing with really powerful magnets I managed to hurt myself. The magnets were stuck together so strongly that when I finally pried them apart one with a sharp piece of wood, I stabbed my hand

Oh yes:), the water starts off non-conducting, but as contaminants disolve into it, it starts to conduct so, if the machine is powered, eletrolosys will rapidly destroy the board!

Ok, that makes sense. Sad thing is, now that you say it like that, I remember I actually did a science project in middle school about electrolytes and how water doesn't conduct electricity until you dissolve an electrolyte in there.

So does this mean I was supposed to remember stuff from school and apply it to my life? Ah, now I get it