Watercooling HDDs
 

There is a discussion at bit-tech concerning watercooling hard drives.
My participation has been limited to PFTA ("pulled from the ass") conjecture and limited personal experience...and seemingly, so is everyone else's.

So I thought I'd seek some info here, too.

I've never seen a HDD waterblock subjected to anything like the rigorous examination that the CPU/GPU blocks receive, indeed, there does not seem to be any consensus over what part of a drive is suitable to cool.
Some blocks just attach to the chassis side, some plate the top or bottom.

Anyone here have any substantive data on the efficacy of watercooling a hard drive- does one method benefit more than the other and is this a subject worth investigating at all?

Re: Watercooling HDDs
 

HDDs casings are designed to pass heat out to the sides, using the chassis effectively as a heatsink. Hence sidemounting tends to be the recommended solution.

Plating the bottom shouldn't be necessary - it's not like you're overclocking anything on the drive, or giving it additional voltage... so the circuitry is always in spec.

Most common way to look at it - MTBF of drive vs how long you're likely to keep that drive in service. All you're likely to do is increase the MTBF.

Re: Watercooling HDDs
 

personally i think it's a waste of resources (money, time, etc etc),
i have two 80mm fans blowing on mine (i have five) at the lowest possible rpm and they are inaudible ( tested with every other fan off ), keeping them cool enough,
hence, not worth the effort for more.

agree with Marci, lateral has larger contact area to conduct heat from.
however it would be interesting to see the use of block with a thermal pad, like the koolance :
http://www.koolance.com/shop/product...roducts_id=177
... versus a lateral block.

Re: Watercooling HDDs
 

How do side mounted blocks deal with this?

Re: Watercooling HDDs
 

They dont. My take is that the PCB chips are what gets hottest, as your quote suggests as well. Since all PCB's are different, there is no way to affix a HD WB to it. Airflow is IMO better than H20 for hard drives because of this.

Re: Watercooling HDDs
 

Don't tell this to the guys at koolance. Link:

http://www.koolance.com/shop/product...roducts_id=231

Re: Watercooling HDDs
 

Well, in the interest of science I'm volunteering to take one for the team.
Kismet struck and this just showed up in the [H]classifieds....
http://img.photobucket.com/albums/v7...uaDrive_01.jpg
Not sure about a testing protocol, open to suggestions.
Prolly have in hand just before the holiday, so nothing substantive will happen for a while.

Re: Watercooling HDDs
 

That is cool. Although, I'd never water cool my HD's.

Re: Watercooling HDDs
 

They would be suitable for 15K rpm drives, in a fanless system. I think that even 10K rpm drives can be properly cooled the passive way with big heatsinks. So the testing protocol would include some red hot 15K drive, like an Atlas 15K II maybe.

Re: Watercooling HDDs
 

I think he was referring to the placement of temp probes.

There's no easy way to answer that, since there are multiple heat sources, and they are not the same from one manufacturer to another.

Off the top of my head, I'd say use an IR imager, but that's cost prohibitive. Otherwise you'd take a temp measurement on the side of the HDD (where it's mounted), but that won't give you a complete picture.

Follow Marci; it's all designed to go through the sides and convect to air.

Re: Watercooling HDDs
 

Yep i recall a discussion with the Maxtor guy on Storagereview who explained that the heat path in HDDs was meant to pass through sides (and certainly not the top part). He said that the air friction on the platters, and the engine itself were usually the most hot elements. So either you put a probe on the bottom of the spindle (which is usually accessible) or you put it on a side (on the bottom half, of course)
But i insist, there's no use testing that with a 'common' ATA drive which needs no active cooling anyway, or even a Raptor which does fine with a good passive cooling box.

Re: Watercooling HDDs
 

I used one of the Koolance HD waterblocks in my early days (it was their first Water cooled case, a total PoS). All I succeeded in doing was killing my drive. I can't imagine a setup where you would ever need to cool your HD with a waterblock. It's trivial to provide more than adequate air cooling for HD's in almost any case and even keep that cooling below audible levels if your are concerned about sound. In my experience water cooling HD's is a waste of time, money, space, and adds unnecessary complexity to the system.

Re: Watercooling HDDs
 

watercooled hard drives are the only way for a quiet system. If you dont have them in enclosures then either:

1) you dont care for quiet
2) you think your system is quiet but in actual fact its noisy as hell and there's something else in your system thats even louder than the noisy hard drives
3) you have 1 weeny laptop drive
4) your case is made out of 10 inch thick rubber with steel plating

quiet PC = 1 fan maximum
quiet pc = hot internal case temps
hot internal case temps = fried HD very very quickly unless its watercooled

Re: Watercooling HDDs
 

ferdb, you totally miss the point. You certainly have no fanless system, and no 15K rpm drives in it... The HDD cooling in *real* quiet systems is a real problem. One solution is heavy heastinks well designed for passive cooling, the other one is watercooling. Both solutions are comparably expensive. Questioning the goal here (watercooling HDDs) is irrelevant.

Re: Watercooling HDDs
 

Actually I found 15K drives to be quieter than my 10K drives. I have a Fujitsu 15K and a Seagate 15K and they both seem quieter (or in a audio band thats outside of my hearing range) than the 10K which have a clearly audible tone.

Re: Watercooling HDDs
 

Now I'm a little bit concerned. I made a waterblock for my 2 HDDs pretty much like Bladerunner, although mine looks like crap comparing to his...
Anyway the way it's mounted it's cooling the top of the 2 disks, not the sides. Since I'm building a new case for me and I'll probably make at least 1 more block, for my GPU, I might just as well create another one for the HDDs if indeed the best cooling will be on their sides. Can I consider it to be true or can I stick with the top cooling?? :uhh:

______
TaTs

Re: Watercooling HDDs
 

How is your setup working?
How does it compare to air-only cooling?

Re: Watercooling HDDs
 

clocker: actually a more relevant question would be "which HDD do you have, and is the internal drive temperature acceptable ?", as air cooling is not an option for many people ;). Most drives have an internal probe and can report temperatures through diagnostic tools (which are quite different between SCSI and ATA)
TaTs: well, which HDD do you have ? Not all HDDs have the same cooling requirements.

I had exactly the same experience, until i found the WD Raptor to be quieter than the SCSI 10K.6 Seagate i have, maybe closer to the 15K.3 noise level (i didn't test them side by side). I think it's related to fluid bearings, more than spindle speed...
BTW all these require active cooling :P

Re: Watercooling HDDs
 

Heh... my SCSI 15k's can be heard in the next room... and definitely require active cooling when there's 8 of em in a stack...

Re: Watercooling HDDs
 

I have 1 Seagate 80GB and 1 Maxtor 80GB. I can't tell the model right now since I'm not at home. Usually the Seagate runs more or less 10ºC hotter than the Maxtor, I'm getting the temps from HDDTemp.
With air I used to get temperatures around 45ºC for the Seagate and 33ºC for the Maxtor. Now with water I'm getting 33~35ºC in the Seagate and 28~30ºC in the Maxtor, yes temperature raised a little for it.

I'm planning on getting a new PC next year, and definetly I'm going to replace the disks, I'm not sure for what yet but it will be a RAID 0 with something like 2x200GB. That's why I'm concerned with temperatures, and if I should worry on getting a new block for them. I don't care about ease to move or replace the drives, my main worry is with silence...
So with this in consideration should I go for a solution like this block clocker posted or what?

Thx.
_______________
TaTs

Re: Watercooling HDDs
 

I have 1 15.3 and one 10.7 and the 15.3 is quieter by a good margin over the 10.7, but the 10.7 is still VERY quiet (and its 300GB!)

The Fujitsu MAS 36GB 15K I have is louder than both... but also faster so I'll accept that ;)

for active cooling, I have a 80mm fan controlled by my CF633 based on the drive cage temp.

Re: Watercooling HDDs
 

Marci: the 'noisy' generation of 15K drives are prehistoric :P I remember having those 15K IBM drives in a server, noisy as hell (they sound like a jet engine), and hot (temps were over 70°C) ! But now 15K drives are quite silent, thanks to FDB, and quieter even than 10K or 7.5K drives, since the noise frequency is at the limits of human hearing bandwidth. Even the seeks are toned down (that's because of the lesser travel of the actuator, platters are 2.5" instead of 3.5"). The Seagate 15K.3 is awesome for that, only equalled by the Raptor which beats it in the price department.

TaTs, from your drive capacities (and temperatures) i conclude you have only ATA drives, those are 7500 rpm max, and nowhere near as hot as 15K SCSI drives. Actually most ATA drives don't need any active cooling, heat transfer through the sides into the PC case frame is enough. IMHO you will better go with a silent enclosure like this one:
http://silentmaxx.de/pics/sm-hddaemmung_gross.jpg
which is excellent and quite enough for 7.5K drives (i use it for raptors, which do very fine in there). No need to blow air over them, they're designed for passive cooling. Just make sure there's room around them so convection can do its work.

Re: Watercooling HDDs
 

I've got a pair of Atlas 15Ks in a SilentStar watercooled silencing enclosure and it makes a HUGE difference. Not dead silent, but to the point where the pump vibration and other ambient noise drowns out the whine.