Harddrive water-cooling

the creature · 01-12-2002, 05:23 AM

I would like to water-cool my hard-disk's and cd-roms, and to get them soundlees.

What is the best aproche? have anyone done this before?

BladeRunner · 01-12-2002, 08:00 AM

I recently made my system fanless. Unless you have particularly noisy HDD's then the PSU fan will be an equal noise source.

I've water cooled both and bought a couple of the new Seagate Barracuda IV HDD's with the FDB (silent fluid drive motor), to help with the silence goal. I have an open case, no fans in a silent room, I can only hear the 2 HDD's if I get within a foot of the PC case and then only just. they really are near silent.

CDroms / DVD roms are another matter and silencing these is not so easy. Considering they only make a noise when used they do matter much imo.

I should have a new site up in a week or so detailing the build.

GigaFrog · 01-12-2002, 04:20 PM

<B>My God</B>, that's awesome.
I will have a spare Maze1 once I'll have put my Maze1C+pelt in, so I think I will use it for my PSU.
I was wondering if the wires make any problems in the regulation of the PSU. They certainly add inductance to the leads of the transistors, but it may not be a big deal. Have you notice any more fluctuations of your voltages after this awesome mod?

You could certainly market these HDD coolers.
Does cooling a drive make it soundless? how?
I would consider cooling them because they heat up, particularly my SCSI drives. It sometimes leads to HDD crashes in the summer. They do get very hot. I use an air cooler on the drives, but they collect a lot of dust and everyonce in a while, they will need clean up.
Can you put the HDD cooler in sandwitch, thus cooling 2 drives?

BladeRunner · 01-12-2002, 05:50 PM

Voltage fluctuations was one of the first things I checked after heat on initial start up. I was pretty amazed the PC booted after all the de-soldering, re-soldering & handling the PSU had during the build. I can't say I've noticed worse fluctuations but I was looking for big voltage discrepancies at the time. I will log the fluctuations from now on, thanks for bringing it up. Just for interest what ways are there to combat inductance of the transistors leads other than the obvious, make them shorter. The PSU has been used for everything Gaming, IE, etc since the PSU has been liquid cooled and the PC has been very stable overclocked. I'll run CPU burn for a bit and see what I get. It's a 550watt Enermax running no fans, and minimal hardware so it's hardly stressed.

Liquid cooling a HDD won't quieten it significantly but if it doesn't run hot it could be boxed in. I was thinking of a double glazed Perspex box, but to be honest these Seagate's really are so quite it might not be required for me. The HDD block will take one underneath and one on top and I still have to make the side brackets and isolation fixings to prevent vibration noise.

GigaFrog · 01-12-2002, 08:46 PM

That's a professional soldering job, very nice on the transistors.

You can reduce the inductance of wires by making them bigger in diameter (aside of making them shorter as you said) but it seems that you are fine.

This is definitively a mod that I will plane for soon.

You might actually get more stable voltages, because the transistors will stay cooler.

futRtrubL · 01-12-2002, 10:29 PM

If you have noisy drives then water cooling them will allow you to pad them with foam or something to acoustically insulate them. I'm planning that on my next system.

Edward

BladeRunner · 01-13-2002, 08:37 AM

Ok did some tests.

in a 12 hour period of various usage including gaming and CPU burn internet and idle these are the highs and lows.

Enermax liquid cooled (with extended transistor wiring)

Core 0 high 1.89 low 1.84 difference .05

Core 1 high 1.94 low 1.89 difference .05

+3.30 high 3.31 low 3.22 difference .09

+5.00 high 5.09 low 4.96 difference .13

+12.00high 12.59low 12.40 difference .19

I am now running an unmodified fan cooled 300Watt Aopen PSU to compare, (It's only been running 4 hours though but here are the voltages with no gaming)

Core 0 high 1.90 low 1.84 difference .06

Core 1 high 1.95 low 1.90 difference .05

+3.30 high 3.30 low 3.20 difference .10

+5.00 high 5.00 low 4.87 difference .13

+12.00high 12.65low 12.46 difference .19

It must be mentioned that both these PSU's show more fluctuations on my newer MSI mobo (nforce), than they did on an Abit KT7-A Raid, (Kt133a), so some of it could be the mobo or monitoring IC. The Enermax however still seems generally better even with extended transistor wires.

gmat · 01-13-2002, 11:10 AM

Hey BladeRunner,
About the HDDs: are these SCSI ?
About the PSU:
(how) did you insulate mosfets backplate from holding screws ? Also are those insulating leafs coming from original PSU or did U get them somewhere else ?
And do you have any leakage current going through ? i mean between waterblock surface and the ground. This could lead to corrosion in the cooling system.

Joe · 01-13-2002, 02:50 PM

Blade man you never cease to amaze me

Nice looking setup!

Brad · 01-13-2002, 03:56 PM

una's hdd cooler showed you don't need too much of an area, he used a normal cpu waterblock.

of course, for 2x hdds, you'd need a block with side bards, like the dden gpu cooler. I wonder how well one of those would work

BladeRunner · 01-13-2002, 07:35 PM

Joe

I never cease to amaze myself, today I re-routed the remoted pump to tidy it up, and get it a bit further away as it could be heard humming slightly. Got it all plumbed in running just testing for leaks and decided I'd check the input neck clamp on the pump was tight. Turned it the wrong way. My god! an Eheim 1060 can create some output with no flow restriction, I got drenched.

Anyway the goal is achieved silence. It may not be 100% absolute silence but as the only thing I can hear if I strain when sitting at my PC is the monitor I consider that silence. It's been a fun project with many different parts, the PSU cooling success being perhaps the most satisfying. In truth it's still in testing as it could yet fail or not cope with the summer but for now it's working fine, also no hardware has been killed during or by the project build.

gmat

The drives are IDE 7200 ata 100

The mosfets are plastic packaged except the back plate but the screw hole and an area around the hole is also plastic so the blot does not come in electrical contact with the back plate. All I have do is transfer the mosfets as they were on two heatsinks (that were live, no apparent reason for the liveness), to a water block that is not live keeping all the isolating mica shims as they were on the sinks. These mica isolators have good thermal properties and excellent electrical isolation. can be bought from Maplin in the UK and probably Radio Shack in the USA.

I wasn't sure if this lack of a "liveness" to the sinks would be a problem but in asking around it didn't appear to be. I'd researched this PUS cooling for 6 months to ensure it had a good chance of success and would be safe. The water block is earthed so would trip the RCD if there were any leakage.

As said it's working fine but still very much in beta testing. The silence is wonderful.

Here is how the block is mounted which also helps cool the transformer a bit. It's a copper and brass "table" that fits snugly over the transformer casing. the water block is then bolted to it. pipe unions exit the side of the PSU to ensure any pipe connection drips are outside the PSU and it's spun the other way up to help the natural heat rise of the PCB.

redleader · 01-13-2002, 10:21 PM

Hey Blade, did you make that HD block? If not any idea where I could get one?

BladeRunner · 01-14-2002, 12:58 PM

Yes I made it, Rotor makes some for sale, can't remember his url now I can find it if you need, but search for him on hardocp forums, or search for water cooling HDD topics you should find something.

Brad · 01-14-2002, 06:03 PM

http://www.angelfire.com/ms/codemasters/index.html

01-12-2002, 05:23 AM	#1
the creature Cooling Savant Join Date: Jun 2001 Location: Norway Posts: 108	Harddrive water-cooling I would like to water-cool my hard-disk's and cd-roms, and to get them soundlees. What is the best aproche? have anyone done this before?

01-12-2002, 08:00 AM	#2
BladeRunner Cooling Savant Join Date: Jul 2001 Location: Chesterfield Uk Posts: 459	I recently made my system fanless. Unless you have particularly noisy HDD's then the PSU fan will be an equal noise source. I've water cooled both and bought a couple of the new Seagate Barracuda IV HDD's with the FDB (silent fluid drive motor), to help with the silence goal. I have an open case, no fans in a silent room, I can only hear the 2 HDD's if I get within a foot of the PC case and then only just. they really are near silent. CDroms / DVD roms are another matter and silencing these is not so easy. Considering they only make a noise when used they do matter much imo. I should have a new site up in a week or so detailing the build. __________________ Zero Fan Zone

01-12-2002, 04:20 PM	#3
GigaFrog Cooling Neophyte Join Date: Nov 2001 Location: NorthWest (French & US) Posts: 88	<B>My God</B>, that's awesome. I will have a spare Maze1 once I'll have put my Maze1C+pelt in, so I think I will use it for my PSU. I was wondering if the wires make any problems in the regulation of the PSU. They certainly add inductance to the leads of the transistors, but it may not be a big deal. Have you notice any more fluctuations of your voltages after this awesome mod? You could certainly market these HDD coolers. Does cooling a drive make it soundless? how? I would consider cooling them because they heat up, particularly my SCSI drives. It sometimes leads to HDD crashes in the summer. They do get very hot. I use an air cooler on the drives, but they collect a lot of dust and everyonce in a while, they will need clean up. Can you put the HDD cooler in sandwitch, thus cooling 2 drives? __________________ Got water? Water cooled for 3 years. Safe with thermal switch. Now silent with fan thermostat. Last edited by GigaFrog; 01-12-2002 at 04:23 PM.

01-12-2002, 05:50 PM	#4
BladeRunner Cooling Savant Join Date: Jul 2001 Location: Chesterfield Uk Posts: 459	Voltage fluctuations was one of the first things I checked after heat on initial start up. I was pretty amazed the PC booted after all the de-soldering, re-soldering & handling the PSU had during the build. I can't say I've noticed worse fluctuations but I was looking for big voltage discrepancies at the time. I will log the fluctuations from now on, thanks for bringing it up. Just for interest what ways are there to combat inductance of the transistors leads other than the obvious, make them shorter. The PSU has been used for everything Gaming, IE, etc since the PSU has been liquid cooled and the PC has been very stable overclocked. I'll run CPU burn for a bit and see what I get. It's a 550watt Enermax running no fans, and minimal hardware so it's hardly stressed. Liquid cooling a HDD won't quieten it significantly but if it doesn't run hot it could be boxed in. I was thinking of a double glazed Perspex box, but to be honest these Seagate's really are so quite it might not be required for me. The HDD block will take one underneath and one on top and I still have to make the side brackets and isolation fixings to prevent vibration noise. __________________ Zero Fan Zone

01-12-2002, 08:46 PM	#5
GigaFrog Cooling Neophyte Join Date: Nov 2001 Location: NorthWest (French & US) Posts: 88	That's a professional soldering job, very nice on the transistors. You can reduce the inductance of wires by making them bigger in diameter (aside of making them shorter as you said) but it seems that you are fine. This is definitively a mod that I will plane for soon. You might actually get more stable voltages, because the transistors will stay cooler. __________________ Got water? Water cooled for 3 years. Safe with thermal switch. Now silent with fan thermostat.

01-12-2002, 10:29 PM	#6
futRtrubL Cooling Savant Join Date: Jul 2001 Location: Kingston, Jamaica Posts: 204	If you have noisy drives then water cooling them will allow you to pad them with foam or something to acoustically insulate them. I'm planning that on my next system. Edward

01-13-2002, 08:37 AM	#7
BladeRunner Cooling Savant Join Date: Jul 2001 Location: Chesterfield Uk Posts: 459	Ok did some tests. in a 12 hour period of various usage including gaming and CPU burn internet and idle these are the highs and lows. Enermax liquid cooled (with extended transistor wiring) Core 0 high 1.89 low 1.84 difference .05 Core 1 high 1.94 low 1.89 difference .05 +3.30 high 3.31 low 3.22 difference .09 +5.00 high 5.09 low 4.96 difference .13 +12.00high 12.59low 12.40 difference .19 I am now running an unmodified fan cooled 300Watt Aopen PSU to compare, (It's only been running 4 hours though but here are the voltages with no gaming) Core 0 high 1.90 low 1.84 difference .06 Core 1 high 1.95 low 1.90 difference .05 +3.30 high 3.30 low 3.20 difference .10 +5.00 high 5.00 low 4.87 difference .13 +12.00high 12.65low 12.46 difference .19 It must be mentioned that both these PSU's show more fluctuations on my newer MSI mobo (nforce), than they did on an Abit KT7-A Raid, (Kt133a), so some of it could be the mobo or monitoring IC. The Enermax however still seems generally better even with extended transistor wires. __________________ Zero Fan Zone Last edited by BladeRunner; 01-13-2002 at 09:20 AM.

01-13-2002, 11:10 AM	#8
gmat Thermophile Join Date: Mar 2001 Location: France Posts: 1,221	Hey BladeRunner, About the HDDs: are these SCSI ? About the PSU: (how) did you insulate mosfets backplate from holding screws ? Also are those insulating leafs coming from original PSU or did U get them somewhere else ? And do you have any leakage current going through ? i mean between waterblock surface and the ground. This could lead to corrosion in the cooling system.

01-13-2002, 02:50 PM	#9
Joe The Pro/Life Support System Join Date: Dec 1999 Location: Denver, CO Posts: 4,041	Blade man you never cease to amaze me Nice looking setup! __________________ Joe - I only take this hat off for one thing... ProCooling archive curator and dusty skeleton.

01-13-2002, 03:56 PM	#10
Brad Thermophile Join Date: Oct 2001 Location: Nuu Zeeelin Posts: 3,175	una's hdd cooler showed you don't need too much of an area, he used a normal cpu waterblock. of course, for 2x hdds, you'd need a block with side bards, like the dden gpu cooler. I wonder how well one of those would work

01-13-2002, 07:35 PM	#11
BladeRunner Cooling Savant Join Date: Jul 2001 Location: Chesterfield Uk Posts: 459	Joe I never cease to amaze myself, today I re-routed the remoted pump to tidy it up, and get it a bit further away as it could be heard humming slightly. Got it all plumbed in running just testing for leaks and decided I'd check the input neck clamp on the pump was tight. Turned it the wrong way. My god! an Eheim 1060 can create some output with no flow restriction, I got drenched. Anyway the goal is achieved silence. It may not be 100% absolute silence but as the only thing I can hear if I strain when sitting at my PC is the monitor I consider that silence. It's been a fun project with many different parts, the PSU cooling success being perhaps the most satisfying. In truth it's still in testing as it could yet fail or not cope with the summer but for now it's working fine, also no hardware has been killed during or by the project build. gmat The drives are IDE 7200 ata 100 The mosfets are plastic packaged except the back plate but the screw hole and an area around the hole is also plastic so the blot does not come in electrical contact with the back plate. All I have do is transfer the mosfets as they were on two heatsinks (that were live, no apparent reason for the liveness), to a water block that is not live keeping all the isolating mica shims as they were on the sinks. These mica isolators have good thermal properties and excellent electrical isolation. can be bought from Maplin in the UK and probably Radio Shack in the USA. I wasn't sure if this lack of a "liveness" to the sinks would be a problem but in asking around it didn't appear to be. I'd researched this PUS cooling for 6 months to ensure it had a good chance of success and would be safe. The water block is earthed so would trip the RCD if there were any leakage. As said it's working fine but still very much in beta testing. The silence is wonderful. Here is how the block is mounted which also helps cool the transformer a bit. It's a copper and brass "table" that fits snugly over the transformer casing. the water block is then bolted to it. pipe unions exit the side of the PSU to ensure any pipe connection drips are outside the PSU and it's spun the other way up to help the natural heat rise of the PCB. __________________ Zero Fan Zone

01-13-2002, 10:21 PM	#12
redleader Thermophile Join Date: Jun 2001 Location: The deserts of Tucson, Az Posts: 1,264	Hey Blade, did you make that HD block? If not any idea where I could get one?

01-14-2002, 12:58 PM	#13
BladeRunner Cooling Savant Join Date: Jul 2001 Location: Chesterfield Uk Posts: 459	Yes I made it, Rotor makes some for sale, can't remember his url now I can find it if you need, but search for him on hardocp forums, or search for water cooling HDD topics you should find something. __________________ Zero Fan Zone

01-14-2002, 06:03 PM	#14
Brad Thermophile Join Date: Oct 2001 Location: Nuu Zeeelin Posts: 3,175	http://www.angelfire.com/ms/codemasters/index.html

Currently Active Users Viewing This Thread: 1 (0 members and 1 guests)