Go Back   Pro/Forums > ProCooling Technical Discussions > General Liquid/Water Cooling Discussion
Password
Register FAQ Members List Calendar JavaChat Mark Forums Read

General Liquid/Water Cooling Discussion For discussion about Full Cooling System kits, or general cooling topics. Keep specific cooling items like pumps, radiators, etc... in their specific forums.

Reply
Thread Tools
Unread 06-19-2003, 02:44 AM   #51
Gooserider
Cooling Savant
 
Gooserider's Avatar
 
Join Date: Jun 2003
Location: North Billerica, MA, USA
Posts: 451
Default Doing the half pipe & sides vs. tops

Got another opinion in the top vs.side cooling debate. See 3rotor's response to me here:
http://forums.procooling.com/vbb/sho...&threadid=7010

While he expressed preference for top cooling, he didn't seem to feel it made a big difference either way. I'm thinking it is probably a case of which approach is easier for the setup you are using.

Quote:
Wouldn't it be better to rip the pipe in half and then solder it on. So there is a more direct path for the heat to take.
Probably would be better in theory to do a split pipe, but I don't think the hassles would be worth the gains.

1. Drives don't generate THAT much heat (Seagate Cheetah 15K U160 SCSI's are only rated for 18 watts max) I doubt the efficiency difference would amount to a significant difference, possibly not even a measurable one.

2. Soldering on a split pipe means LONG seams, and lots of fussy surface prep and very exact fitting of the pipe to avoid the many chances for leaks. Soldering on an intact tube will be leak free if you don't burn a hole in it (difficult) and it's no big deal if you miss a few spots.

3. An intact tube means an easy hookup for tubing, just leave an end sticking out and slide the tube over it. No barbs, no fuss. You'll need do some fancy transition to go from a split tube to a round fitting, and you can't put a tube on a semi-circle very well.

4. Unless you use a much bigger diameter tube, you will end up with alot more flow restriction than you get with an intact tube.

5. As a minor cost issue, soldering an intact tube to a plate is very material efficient. I don't see a way to do a split tube without wasting alot of material...

I could go on, but I think you'll agree that overall the benefit doesn't look worth it.

Gooserider
__________________
Designing system, will have Tyan S2468UGN Dual Athlon MOBO, SCSI HDDS, other goodies. Will run LINUX only. Want to have silent running, minimal fans, and water cooled. Probably not OC'c
Gooserider is offline   Reply With Quote
Unread 06-24-2003, 08:45 PM   #52
sevisehda
Cooling Savant
 
Join Date: May 2003
Location: NY
Posts: 234
Default

I'm just about finished with my plans for a block and have 1 remaining question. Since the heat we're talking about here is very little, aproximately 20watts is it necessary to have a pair of side blocks or could one get away with having only a single sideblock.

The block would be a strip of Cu same profile as the HD, the thickness is still under consideration. I'll either do a rotor style block or a cross drilled style.

The chassis of a HD is a chunk of aluminum which is a rather good cunductor of heat. So if I lap the side of the drive merely to remove the paint and insure its true then apply the copper block, the drive should stay cool. Since the drive doesn't put out that much heat there shouldn't be a huge variance across the drive.

Does this sound doable or does the drive really need a pair of coolers?
sevisehda is offline   Reply With Quote
Unread 06-25-2003, 09:42 AM   #53
bigben2k
Responsible for 2%
of all the posts here.
 
bigben2k's Avatar
 
Join Date: May 2002
Location: Texas, U.S.A.
Posts: 8,302
Default

If you look at it from a perspective of thermal resistance, you would have a higher resistance with a single block.

Is it relevant?

Well, most drives will work well without water cooling, which brings up the question: why watercool it at all?

It's not necessary, and just like the rest of our watercooling ventures, it's not worth the money (not cost efficient).

So if you only want to cool only one side, go for it.
bigben2k is offline   Reply With Quote
Unread 06-25-2003, 04:07 PM   #54
sevisehda
Cooling Savant
 
Join Date: May 2003
Location: NY
Posts: 234
Default

Well I've got a pair of 7.2k rpm drives. I could always put a fan on them but that would involve putting a fan on them and virtually doubling the fans in my system. I stacked them once and they got rediculously hot in no time at all. I'd like to save room by stacking them in this case so watercooling would be the best way to keep em cold without adding fans or spacing them a good distance apart.
sevisehda is offline   Reply With Quote
Unread 06-25-2003, 04:13 PM   #55
bigben2k
Responsible for 2%
of all the posts here.
 
bigben2k's Avatar
 
Join Date: May 2002
Location: Texas, U.S.A.
Posts: 8,302
Default

Then stick to my design: I'm making a block (2 blocks) for two drives, with a large enough gap between them so that I can put a ramsink on any components that wouldn't sink any heat to the sides. Any little fan should do the trick.

It's too bad that it has to be this way, but I'm not putting another waterblock, to cool the rest of the drive.
bigben2k is offline   Reply With Quote
Unread 06-25-2003, 06:49 PM   #56
BladeRunner
Cooling Savant
 
BladeRunner's Avatar
 
Join Date: Jul 2001
Location: Chesterfield Uk
Posts: 459
Default

I always intended to mount the drives to the HDD block using a copper plate on the sides too but never got around to it. I was also going to isolate them in a double glazed box to totally soundproof them as they can still be heard if listened to close up, but the Seagate IV's in single platter are so quiet it really isn't a requirement. They are truly the only item that makes a sound in my PC when not using DVD/ Printer etc, (unless you include the slight component buzz of the NF7-S).

From my experience cooling two drives via the top plate works just fine, we are not talking CPU point heat levels, (at least not with 7200rpm drives), and heat will always migrate to a cooler surface as long as there is some type of transfer medium. My opinion is don't complicate it or over engineer, just make a block that works well in your system. I choose a slim flat block simply because I could cool two drives with one block. Oh and one other tip, do not hard mount your drives as doing this with mine makes them 10 times more noisy, (especially in seek). If the seek noise is annoying try just sitting the drive on some foam for a test if you currently have the drives hard mounted, and see if it makes a difference in your system.

I'm going to make a new HDD block soon to complete the migration to 8mm festo for all the blocks. Not important but a slight correction in that my current HDD block has 6mm festo push fits and is constructed using two 2mm plates for the top and bottom and the main part is 6mm (not 10mm), making it 10mm thick overall.
__________________
Zero Fan Zone
BladeRunner is offline   Reply With Quote
Unread 06-25-2003, 08:22 PM   #57
[-Stash-]
Cooling Neophyte
 
[-Stash-]'s Avatar
 
Join Date: Jun 2003
Location: Kent, England
Posts: 24

Quote:
Originally posted by BladeRunner
(unless you include the slight component buzz of the NF7-S)
Bah, I dream of that level of quiet. Not got the money, time, but most importantly, the dedication to do anything close to what you've done. That adn I seem to be moving house every 3-6 months for the past 2 years
__________________
[-Stash-]
----------------------------------
XP-M 2500+ (2.5GHz @200FSB), DFI Infinity NFII Ultra
2*512 Ballistix 3200, 6800GT 410/1100
WD740GD, 2*120 SATA Maxtor 6Y120M0, NEC ND2500A (external)
Eheim 1250, DD Maze3, Heater Core, 1/2" Tygon
[-Stash-] is offline   Reply With Quote
Unread 06-26-2003, 03:45 AM   #58
sevisehda
Cooling Savant
 
Join Date: May 2003
Location: NY
Posts: 234
Default

BladeRunner is my new personal Hero. Since I've soften you up what different diameter tubing do you have in your system right now?
sevisehda is offline   Reply With Quote
Unread 06-26-2003, 05:11 AM   #59
BladeRunner
Cooling Savant
 
BladeRunner's Avatar
 
Join Date: Jul 2001
Location: Chesterfield Uk
Posts: 459
Default

I'm not going to fill this topic up with my images, so go HERE to see some current images of the PC. I will have some build details on the current blocks when I get around to it but my work has been mad again lately

The Tube sizes are Tygon R-3603 for inflow & outflow, (12.5mm ID), via reamed out 5/8 barbs (13mm ID). To get the Tygon on the barbs requires about 15secs immersion in boiling water and a fair bit of force, once on though ugly hose clamps are not required as there is no way it will pull off. The inflow enters directly in the centre of the new splitter block that has four 8mm festo push fits for outflow built in. The polyurethane tube for these is 8mm OD, 6mm ID and the for separate outputs supply the 7 other blocks in my system, (the lesser ones doubled up). There is an extra splitter that can be seen dangling and this is the temporary splitter required at the moment for the current HDD and R9700 ram blocks as they are 6mm festo, (4.5mm ID). When the R9800 pro cooler i'm working on is finished, and a new HDD block made this extra splitter won't be required.

Output from the 8mm festo blocks are all collected into another low restriction splitter, (the fixed one on the far right in the images), that takes the flow out via the 5/8 outflow barb.

The silence is great but I don't even think about it much anymore. It's only really bought home to me when I get a new graphics card and test it out before making a cooler for it. It's amazing just how noisy they sound when they are the only major sound source
__________________
Zero Fan Zone
BladeRunner is offline   Reply With Quote
Unread 06-26-2003, 12:52 PM   #60
nicozeg
Cooling Savant
 
nicozeg's Avatar
 
Join Date: Sep 2002
Location: Santiago, Chile
Posts: 403
Default

Quote:
Originally posted by BladeRunner
The silence is great but I don't even think about it much anymore. It's only really bought home to me when I get a new graphics card and test it out before making a cooler for it. It's amazing just how noisy they sound when they are the only major sound source
Or when you load a cd. That's very anoying with current fast drives and a silent system..:shrug:
nicozeg is offline   Reply With Quote
Unread 06-26-2003, 01:58 PM   #61
BladeRunner
Cooling Savant
 
BladeRunner's Avatar
 
Join Date: Jul 2001
Location: Chesterfield Uk
Posts: 459
Default

Thought I'd covered that with my mention of DVD, Printer etc in my previous posting, but yes these things make a noise, however it's a different type of noise in that it's not the continuous same pitch wine of a fan or noisy HDD, and only occurs when you use the item. Normally I will install what I need from any optical medium to HDD then use a hack to bypass the requirement of having to have the CD in the drive if need be.

If someone ever makes a silent CD/DVD I'll be first in the queue same goes for printer scanner and all other noise making peripherals
__________________
Zero Fan Zone
BladeRunner is offline   Reply With Quote
Unread 06-27-2003, 12:52 AM   #62
Gooserider
Cooling Savant
 
Gooserider's Avatar
 
Join Date: Jun 2003
Location: North Billerica, MA, USA
Posts: 451
Default I haven't checked with Seagate yet,

But until I do, and unless they say otherwise, I am going to stick with top/bottom cooling, not side cooling.

I don't buy the 'side as heatsink to the case' argument, as I've NEVER seen any mention of this by drive makers, even when they talk about the need to ensure adequate cooling air flow. (over the top and bottom of the drive BTW) I know that if I were a drive engineer, I certainly wouldn't want to bet on the case as a heat sink, since I would have NO control over what kind of case or cooling that would provide - Some cases are plastic. Some cases use rails rather than solid mounts (including some plastic rails!). Some users rubber mount their drives in an effort to reduce noise (anybody we know??? ) Even in the best case of a solid mount to a steel drive cage, it's still pretty wimpy as a heat sink. Given an environment over which I would have so little control, I would plan on some other cooling path and take anything that comes from the sides as a bonus.

For myself, given current discussion, I'm inclining towards the notion of multiple top / bottom plate coolers in a 'sandwich' configuration, to cool things down with a foam wrap around the entire HD bundle for noise suppression. I will be using the new 10 and 15K rpm U160 SCSI Seagate Cheetahs, and everything I've read about them says they are noisy and hot, but that the speed is worth it

As a side note, I did a bit of work on my old dinosaur PC last night, I added a second 4GB HD (doubled my drive space) and did a bit of silencing. I removed the sheet metal grill from the PSU fan, which helped a little bit. The big change was to cut an old mouse pad (one of the cloth covered rubber ones) into strips and jam a strip inbetween each side of the hard drives and the (slightly bent) drive cage. I also had to replace the drive screws with ones that were slightly longer. I was going to screw through the mouse pad strips, but found that didn't really work because of the difficulty of getting the holes to line up, I was afraid that I'd bang the drives up to much if I kept trying so I just let it go by getting enough of the strip in place that I had no metal - metal contact other than the screws. The result is far from silent, but even with two hard drives instead of one, it's still quieter than it was.

Gooserider

PS, Bladerunner, did you get the PM I sent you the other day about the problem I was having with your website? If you didn't let me know and I can try to do a resend.
__________________
Designing system, will have Tyan S2468UGN Dual Athlon MOBO, SCSI HDDS, other goodies. Will run LINUX only. Want to have silent running, minimal fans, and water cooled. Probably not OC'c
Gooserider is offline   Reply With Quote
Unread 06-27-2003, 05:00 AM   #63
sevisehda
Cooling Savant
 
Join Date: May 2003
Location: NY
Posts: 234
Default

Although I'm sure if someone did a test comparing side vs top/bottom cooling the sides would win for the simple reason that the drive has an Al chassis so the whole thing is a kind of heat sink. So if you cooled the sides you would cool the drive. You'd also be cooling it via conduction which is by far the best.

Cooling by the top the heat has to travle through either some kind of sealent or air then through a thin sheet of Al or steel this is far less efficent. The top and bottom are also normally much more uneven than the sides normally.

However if you take into account a HD puts out less than 20w of heat than you don't need the best solution in the situation. So in reality it probobly doesn't matter for a 7200 rpm drive. I'm as of yet unsure how I'm going to cool my HDs. It will depend on A) cost of materials, B) how I want to position my drives.
sevisehda is offline   Reply With Quote
Unread 06-27-2003, 08:42 PM   #64
BladeRunner
Cooling Savant
 
BladeRunner's Avatar
 
Join Date: Jul 2001
Location: Chesterfield Uk
Posts: 459
Default

What sevisehda says is most likely true, but what is being missed is that HDD heat, (at least the HDD's I've had), its more a heat soak over time issue, not a direct heat that needs to be sinked away like a chip. Most HDD's will be cooled effectively just using a fan blowing directly on them, If like me you don't want fans they can get overly hot over time with no airflow. A waterblock will work pretty well whether top or side mounted, and If as I said you want to avoid the vibration noise you won't be hard mounting them so the top cooling block could have side plates to retain, which was my original intention. I think Top cooling is preferred mainly because its more economical in that two drives can be cooled with one simple block, even if it is a little less effecient


Gooserider

Sorry M8 but I do get a shit load of e-mail and have to work through it when I can , Thanks for the bug report, but I can't find anything wrong? Are you still getting problems as I didn't totally understand your mail as there are no "next page buttons" anyhow it seems to work fine my end so if you are still getting issue what browser are you using?
__________________
Zero Fan Zone
BladeRunner is offline   Reply With Quote
Unread 06-28-2003, 02:12 AM   #65
Gooserider
Cooling Savant
 
Gooserider's Avatar
 
Join Date: Jun 2003
Location: North Billerica, MA, USA
Posts: 451
Default

I don't have solid data to back it up, but I still think the top is better, though it probably doesn't make much difference.
Quote:
sevisehda: Cooling by the top the heat has to travle through either some kind of sealent or air then through a thin sheet of Al or steel this is far less efficent. The top and bottom are also normally much more uneven than the sides normally.
Most of the drives I've dealt with had a really flat cover (I agree the controller side is irregular), and while some have had smooth sides, several have had bosses that the drive mounting screws went into, but otherwise were recessed. A drive like that wouldn't have much more direct contact area than a CPU.

The cover is normally Al, and is thinner so as to offer less conduction resistance. The primary heat produced by a drive is supposed to be air friction on the spinning platters, so IMHO the most effective cooling tactic would be to chill the largest single flat surface that contacts the air. You could almost think of it as a backwards CPU - hot air hitting the Al top cooler, dumping it's heat which is then carried away by the copper block.

However I think Bladerunner has the right idea in that it doesn't matter all that much which area you cool.

Quote:
BladeRunner Sorry M8 but I do get a shit load of e-mail and have to work through it when I can ,
No prob, just didn't know if it had gotten lost in the shuffle or not.
Quote:
Thanks for the bug report, but I can't find anything wrong? Are you still getting problems
I tried it just before starting this reply, and in the one case I tested, I'm still having the same problem. The work-around is a bit of a nusiance but isn't that bad, so normally I've just been using it rather than doing the problem sequence.
Quote:
as I didn't totally understand your mail as there are no "next page buttons"
My mistake, I was working from memory and got that part wrong. Here is the sequence I used which produced the problem.
1. Start @ your home page. Click on 'Projects' page button (from left side menu)
2. Scroll down to Enermax PSU project, click on page 1 button.
3. Read page one, click on 'Page jump' "2" in the block at the end of the page.
4. End up on home page (NOT Page 2!)
Interestingly enough, the link shown in the browser window is "http://www.zerofanzone.co.uk/?request=liquidenermax550w&page=2"
which certainly seems to suggest the correct page, but I still got the home page.
OTOH, if I click on each page number in the block on the project page, I get the correct page. However I just noticed (funny what doing back to back tests will show you) that the link is different:
"http://www.zerofanzone.co.uk/projects.asp?request=liquidenermax550w&page=2"
I don't speak fluent URLish, but could this be the problem?
Quote:
anyhow it seems to work fine my end so if you are still getting issue what browser are you using?
I run OPERA 6.03 (Linux, ad supported, version) on Linux RH 7.2 Usually I open lots of pages in seperate windows which is one of the reasons I love Opera so much.

Hope this is useful, If it is, buy me a beer some time (When I do commercial beer, I drink Guinness, though it's a tad on the light side )
ART
__________________
Designing system, will have Tyan S2468UGN Dual Athlon MOBO, SCSI HDDS, other goodies. Will run LINUX only. Want to have silent running, minimal fans, and water cooled. Probably not OC'c
Gooserider is offline   Reply With Quote
Unread 06-28-2003, 06:41 AM   #66
BladeRunner
Cooling Savant
 
BladeRunner's Avatar
 
Join Date: Jul 2001
Location: Chesterfield Uk
Posts: 459
Default

Gooserider

Thanks for the detail but try as I might I can't get any errors using IE, maybe it's an Opera or Linux thing, but I don't use either. I'll dirrect my webmaster to this topic to see if he has any ideas. Have you tried Ctrl + F5 on those pages, (if it's the same function in Linux / opera), in case the cached page is overruling the newer updated one? probably not just an idea off the top of my head.

Worth noting I never get
"http://www.zerofanzone.co.uk/?request=liquidenermax550w&page=2"

Which ever way I access page 2 it always shows the correct page as
"http://www.zerofanzone.co.uk/projects.asp?request=liquidenermax550w&page=2"

I don't speak fluent URLish, either but trying to learn :shrug: ....... it all seems too much like school homework to me, and would much rather be designing / making new blocks

Another thing you could try, (again it may not make any difference), is going to the site with .com rather than .co.uk

www.zerofanzone.com
__________________
Zero Fan Zone
BladeRunner is offline   Reply With Quote
Unread 07-28-2003, 01:51 AM   #67
Gooserider
Cooling Savant
 
Gooserider's Avatar
 
Join Date: Jun 2003
Location: North Billerica, MA, USA
Posts: 451
Default Still haven't checked w/ Seagate,

But I am carrying on with my top / bottom cooler design. I have finished my first one, which like BR's design is intentended to do two drives at once. Unlike BR's block however, mine is on the ugly side.

I made it by getting a hunk of 1/8"(3mm) copper plate and cutting it to approx the size of a hard drive, and then milling the edges to make plates that are exactly the size of a 3.5" drive (about 4" x 6", (101 x 120mm)) I then got some copper tubing 1/4" ID, 3/8" OD and made as long a cooling loop with it as I could readily bend.

This turned out to be sort of a flattened round 'W' shape. The tube starts at the side of one end, goes to the other, u-turns comes back down the middle to the first end, does a second u-turn, back to the far end, does a third u-turn and comes back to the first end and exits out the other side of the same end. (So there are 4 passes of tubing running the length of the block)
Thus both the connection points are on the same end of the block, one on each side. Depending on how I mount the block the plumbing will hook up on either the same end as the electricals, or the opposite end. The loop is just a little smaller than the blocks, so it sits about 1/8" in from the block edges. I haven't actually measured how much tube goes into the loop, but it looks like around 2 to 2.5 feet.

I took two of the plates that I had made earlier, and sandwiched the loop in between them. Between squeezing the sandwich in my vice, and judicious application of a BMFH (Big _____ ______ Hammer ) I was able to get 90-95% contact between the plates and the tube. I took things apart, polished all the bits up with a bit of emery cloth, coated everything with plumbers solder flux and put the sandwich back together. I fired up my propane blow torch, and was able to solder all three parts together with plumbers lead free solder. It doesn't look great, but it is all copper on the inside, and today I tested it to 90 PSI with no problems. Next step will be to try doing some flow tests to see how much restriction I end up with.

The block I made was two sided, for two drives, but I could also use an aluminum block on one side (so the solder wouldn't stick to it) and make single sided plates to do individual drives. Or I could make a super cooling sandwich with any number of drives using double side plates between each drive, with a single plate on top and bottom. I also have considerable flexibility in mounting - I can stack the drives and plates in a 3.5" bay, use a stock sled to mount in a 5.25" bay, or do a sound proofing foam bundle that I could also stack almost anywhere.

I did have some difficulties. My first effort at making the loop with a tubing bender came out really lopsided, and I couldn't get it to squish down enough to fit on the block. So I went back to the drawing board and got a big block of wood which I laid the loop out on, then routed out to make a template that I could force the tube into as I bent it. This worked better, but the tube tended to kink and distort as I bent it, and this led to needing the hammer to flatten it when I was done. It is useable but not as nice as I would like.

What I would like to do is create some better jig fixturing so that I could bend the tubing more uniformly, and assemble the blocks better. But I can't really justify the cost of doing this just for the few blocks I'll be making for myself. Anyone have some ideas about how many of these I could sell if I were to make them for others? And how much I could get for them? This isn't a 'For Sale' message, as I know such are reserved for the FS/FT forums, but more an inquiry as to how much interest I would get if I was to put some of these blocks up for sale there.

Gooserider
__________________
Designing system, will have Tyan S2468UGN Dual Athlon MOBO, SCSI HDDS, other goodies. Will run LINUX only. Want to have silent running, minimal fans, and water cooled. Probably not OC'c
Gooserider is offline   Reply With Quote
Unread 01-26-2004, 03:38 PM   #68
Arcturius
Cooling Savant
 
Join Date: Aug 2001
Location: Chippewa Falls, Wisconsin
Posts: 294

I don't see why there is any argument over top/bottom vs. sides...maybe it's because I run hotter disks than everybody else.

As BladeRunner said, the heat generated from drives is not much, and is more of a 'heat soak' built up over time. Even though, it is still obvious that the sides are the ideal place to cool the drive.

1. Look at high-end, high-heat SCSI disks--they have fins BUILT INTO the aluminum chassis. Examples: the six-platter(!) IBM 36Z15s that I run. Or the Fujitsu MAM/MAN/MAP/MAS, Maxtor/Quantum Atlas XX, WD Raptor (SATA=SCSI lite), etc.
2. Also note that drive lids are made of STEEL, and sometimes is several layers thick, withouth TIMs. Not to mention, drive lids tend to be very smooth, with obviously no attempt made to increase surface area via fins or ripples.
3. What about the massive insulating PLASTIC STICKERS that are on top of the drives?

Obviously the engineers had no intentions to dissipate significant amounts of heat through the lid--they expected it to leave through the main aluminum chassis, which contacts both the spindle motor AND the heated gas inside the drive.
If heat was to be discharged through the lid, it would be aluminum, and the labels would be smaller or nonexistent, i.e. etched into the lid (or stencilled on like Quantum used to do). Also, there would be at least some attempt made to stamp little fins in to increase the surface area.

As far as the controller board goes, there is little sense in cooling it. They are almost always insulated from the main body of the drive by a nonconductive plastic sheet, and a think sheet of foam, except where a few screws attach them to the drive, and where control electrodes pass into the body. In the worst case, very few components on the board will generate enough heat to put themselves at a high temp. Most of those are spec'd to run at even higher temps. Of those that remain (probably none), they can be cooled with a tiny passive heatsink.


Kindly disregard:
1. Plexi cases. They are a very recent development, and are almost certainly NOT considered when drives are engineered. (especially SCSI drives, which usually end up in servers)
2. Drive rails. More common, but are most commonly used with high CFM fans in a server for hot-swap applications, so the entire drive gets airflow, and heat dissipation through the rails is not a concern. If you are thinking of the plastic rails in 5.25" bays, please remember that those are intended for optical/tape/front expansion, not hard drives. Also, nearly all 3.5"->5.25" bay-conversion kits for HDDs include a fan.
__________________
Everything aircooled (I honestly don't know why you guys let me stay a member...)
Always folding:
WXP Pro:
KD7R + Palomino 1800+
W2K3:
4x P3 Xeon 500 512KB (4 clients)
Linux boxes:
none Had to put them to use doing *real* work.
Arcturius is offline   Reply With Quote
Unread 01-26-2004, 07:50 PM   #69
BladeRunner
Cooling Savant
 
BladeRunner's Avatar
 
Join Date: Jul 2001
Location: Chesterfield Uk
Posts: 459
Default

Like almost everything in the real world it's never as clear cut in every situation. Drive designs tend not to have any proper thought on cooling, as the makers don't see it as an issue. If it runs at 50C and lasts 5 years instead of 10 then so what....

The thing is they have to follow a size form factor, so I guess cooling fins at the top or bottom would be harder to incorporate in the design.

Things to remember, cooling the top will generally give you more surface area in contact with the block than the sides. Natural heat paths always favour rise, (upwards), but heat will tend to migrate to a cooler surface, assuming a reasonable thermal interface.

I have a Centrino laptop that runs pretty cool with the fan off, and in doing some disk ghosting of the 2.5" drives in my main PC I've been using another spare drive in the laptop. It's a faster drive but gets extremely hot in the laptop case. Just sitting a 1U P4 copper heat sink on the plastic case of the laptop above the HDD area helps soak away some of the heat. The copper 1U P4 sink gets quite warm and that's with an air gap, and at least 3mm of plastic to pass through.

Imo, better isn't always going to be the sides in every situation, it depends on many other factors, my Cuda IV's did tend to get hottest on the top plate, (probably due to heat rise more than anything else). I just can't see it matters where you cool HDD's as long as you cool enough of it's area, (excepting high performance SCSI HDD's may require a more specific design thought on cooling). I think it's much better to design it to fit your specific needs shape and size wise, while keeping in mind contact over a lager surface area should keep the complete drive at a lower even temp. Also important is that the design doesn't leak.......... that would kill your drive faster than any heat related issues
__________________
Zero Fan Zone
BladeRunner is offline   Reply With Quote
Unread 01-27-2004, 09:46 AM   #70
MMZ_TimeLord
Cooling Savant
 
MMZ_TimeLord's Avatar
 
Join Date: Apr 2003
Location: San Mateo, CA, USA, Earth
Posts: 433
Default

I put the primary cooling on the HDD motor surface on the bottom... seems to keep the drives just over room temp without a problem... I have one upside down ... then the block... then the top drive.

Here's a picture of the block... had to mill a slight recess in the edges to get it to sit down on the HDD motor area.

__________________
MMZ>TimeLord

"Oooooooooh... that's gonna leave a mark!"
MMZ_TimeLord is offline   Reply With Quote
Unread 01-27-2004, 11:10 AM   #71
bigben2k
Responsible for 2%
of all the posts here.
 
bigben2k's Avatar
 
Join Date: May 2002
Location: Texas, U.S.A.
Posts: 8,302
Default

Nice work.


The debate stems from the whole purpose of water cooling an HDD: what's the point in doing it, if it isn't done right, given that the benefit is already marginal?

I think it's pretty clear that side water cooling, along with a small fan and heatsink(s) for top components, is the way to go. That's the conclusion I came to, when I came up with the design.
bigben2k is offline   Reply With Quote
Unread 01-27-2004, 11:42 AM   #72
BladeRunner
Cooling Savant
 
BladeRunner's Avatar
 
Join Date: Jul 2001
Location: Chesterfield Uk
Posts: 459
Default

Maybe, but define "right", what I'm saying is you just cannot for every situation, or every HDD design. My drives in my system were getting hottest on top, mostly I guess because there is no forced airflow, and the fact heat rises. It's also debatable that the benifits are marginal, I've since seen 58C max without the waterblock when stressing two drives in a 30C ambient. This may not harm them, but is far too hot for my liking.
__________________
Zero Fan Zone
BladeRunner is offline   Reply With Quote
Unread 01-29-2004, 05:14 PM   #73
bigben2k
Responsible for 2%
of all the posts here.
 
bigben2k's Avatar
 
Join Date: May 2002
Location: Texas, U.S.A.
Posts: 8,302
Default

Well, given that an HDD will, most assuredly be attached by the side, to a large steel plate, IMO, it's more than likely that if the drive was designed to have any kind of cooling, beyond airflow, it would be through the sides. What's also obvious is that it's not a complete solution.

We've debated this often. I believe that the heat that comes from the spindle motor will be dissipated into the HDD case, leaving the electronics to be cooled by another mean. We've also mentionned that cooling the top side of an HDD wouldn't work nearly as well, to cool the spindle motor, because of everything "in between".

Side cooling sucks, and I wish there was a better way, because I'm not looking forward to making a new mount for the larger drive(s).

As for the temp differences, we could go on and on about how the temp is measured, but I just don't care to get into that debate, unless someone's got some really good insights into it.
bigben2k is offline   Reply With Quote
Unread 01-29-2004, 07:05 PM   #74
Butcher
Thermophile
 
Join Date: May 2001
Location: UK
Posts: 1,064
Default

Any HDD made since the mid-late 90s has an onboard temp sensor. You just need a program that can read the SMART data to read it.
http://www.seagate.com/support/disc/.../en/smart.html is an online version, you can find various tools about the net to read it without using a net connection. Note, most don't work for drives on a raid controller, however the better one's do work on SCSI disks.
__________________
Once upon a time, in a land far far away...
Butcher is offline   Reply With Quote
Unread 01-29-2004, 08:04 PM   #75
Butcher
Thermophile
 
Join Date: May 2001
Location: UK
Posts: 1,064
Default

Further to the above I'm yet to find a program to reliably read both scsi and ide disk temps, so I'm going to write my own this weekend.
__________________
Once upon a time, in a land far far away...
Butcher is offline   Reply With Quote
Reply


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

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Forum Jump


All times are GMT -5. The time now is 03:55 AM.


Powered by vBulletin® Version 3.7.4
Copyright ©2000 - 2024, Jelsoft Enterprises Ltd.
(C) 2005 ProCooling.com
If we in some way offend you, insult you or your people, screw your mom, beat up your dad, or poop on your porch... we're sorry... we were probably really drunk...
Oh and dont steal our content bitches! Don't give us a reason to pee in your open car window this summer...