Has anyone tried passive water cooling?

The alternative to direct air cooling

Moderators: NeilBlanchard, Ralf Hutter, sthayashi, Lawrence Lee

Post Reply
Tom Brown
Posts: 246
Joined: Sun Aug 03, 2003 11:16 pm
Location: Regina, SK, Canada
Contact:

Has anyone tried passive water cooling?

Post by Tom Brown » Mon Aug 18, 2003 9:03 am

I've been thinking about trying passive water cooling. A water block, rad over the front case inlet, and a reservoir could be configured to allow the water to circulate using heat driven convection.

Has anyone tried this? I would imagine the CPU would need to be under clocked and the water block would need a substantial passive radiator as well as the water flow as the cooling efficiency would be low. Still, it might be enough to allow for totally passive CPU cooling.

Does such a water block exist?


-- Tom

Zhentar
Patron of SPCR
Posts: 918
Joined: Thu Mar 27, 2003 4:00 pm
Location: Madison, WI, USA

Post by Zhentar » Mon Aug 18, 2003 12:37 pm

you'd certainly have to create such a block yourself. I don't think water has enough convection to make this work, but I could be wrong.

Marvin
Posts: 64
Joined: Mon Nov 18, 2002 11:54 pm
Location: Tampere, Finland

Post by Marvin » Tue Aug 19, 2003 12:12 am

Heres some other threads on this topic:

Totally passive watercooling? Anyone tried it?

PC watercooling with Passive radiator Article

And lastly something that gives lots of thoughts

"PC Water Cooling with a Passive Radiator"
@overclockers.com

Jones_Rush
Posts: 6
Joined: Tue Aug 19, 2003 1:44 am

Post by Jones_Rush » Tue Aug 19, 2003 1:54 am

I have been running a completely passive cooling system for two years now (for CPU, GPU, PSU, MOBO). Everything runs just fine (the computer is on 24/7). Only thing is, I'm limited to an Athlon 1.33ghz CPU, max.

But, there are a lot of very quiet CPU fans today, and I guess that if I'll decide to use one with a faster CPU, the noise added won't be too noticeable, since the GPU/PSU/MOBO will still remain passively cooled.

CoolGav
Posts: 398
Joined: Wed Jul 23, 2003 4:26 am
Location: €ngland
Contact:

Post by CoolGav » Tue Aug 19, 2003 6:36 am

I recently upgraded my T-bird 1.33 to an XP2400+ (T-bred) which is about the same power. The die was 1/3 the size, but otherwise cooling it is the same...

Jones_Rush
Posts: 6
Joined: Tue Aug 19, 2003 1:44 am

Post by Jones_Rush » Tue Aug 19, 2003 6:48 am

I recently upgraded my T-bird 1.33 to an XP2400+ (T-bred) which is about the same power. The die was 1/3 the size, but otherwise cooling it is the same...
CoolGav,
I don't understand, are you implying that cooling an XP2400+ is the same as cooling a T-bird 1.33ghz ?. How could this be ?, the XP2400+ runs 400mhz faster than the 1.33ghz. Isn't there more heat generated by that ?.

Rusty075
SPCR Reviewer
Posts: 4000
Joined: Sun Aug 11, 2002 3:26 pm
Location: Phoenix, AZ
Contact:

Post by Rusty075 » Tue Aug 19, 2003 6:54 am

A t-bird 1.33Ghz produces 70Watts of heat max.

An XP-2400+ T-bred B produces 65.3 Watts max.

Heck even a 3000 Barton produces less heat than your 1.33, it tops out at 68.3 watts

Jones_Rush
Posts: 6
Joined: Tue Aug 19, 2003 1:44 am

Post by Jones_Rush » Tue Aug 19, 2003 8:21 am

Where can I check how much Watts each CPU produce ? (shouldn't we say "consume" ?)

Tom Brown
Posts: 246
Joined: Sun Aug 03, 2003 11:16 pm
Location: Regina, SK, Canada
Contact:

Post by Tom Brown » Tue Aug 19, 2003 8:56 am

Thanks. Great information, guys. :)

Rusty075
SPCR Reviewer
Posts: 4000
Joined: Sun Aug 11, 2002 3:26 pm
Location: Phoenix, AZ
Contact:

Post by Rusty075 » Tue Aug 19, 2003 11:17 am

Processor Electrical Specifications

it's listed on the SPRC Useful Web links

And no, it's "produce" We're talking about watts as a heat measurement, not an electrical measurement here.

axle
Posts: 64
Joined: Sat Oct 26, 2002 3:06 pm
Location: Wind Gap, PA
Contact:

Post by axle » Tue Aug 19, 2003 12:11 pm

I've been passive WC'd for quite some time now, and gone through many MANY revisions (I've had some BAD ideas), but this particular Gerwin (c) (:-D) style rad has always been good to me.

<<<edit>>> Blew the margins, sorry :(

20 ft of spiral tubing I scored off ebay for $10 is good for about 5c when used with the above rad. Right now (w/o the spiral) I idle about 28c, It'll get all the way up to 31c or so under full load.


During winter, I stick a passive rad behind the walls, exposed to the chilly air... get me some good temps then :D (like, 0c)
Image
Last edited by axle on Wed Aug 20, 2003 6:43 pm, edited 1 time in total.

Jones_Rush
Posts: 6
Joined: Tue Aug 19, 2003 1:44 am

Post by Jones_Rush » Tue Aug 19, 2003 12:38 pm

Sorry Rusty075, I didn't know that.

Anyway, I got mixed up a bit. The max CPU that my passively liquid cooled system can handle is the Athlon 1.13ghz (what I'm using currently), it's not the 1.3ghz.

Anyway, it's still a GREAT NEWS, since I can use the XP2200 (1.8Ghz) CPU, which has less heat dissipation than the 1.13 CPU.

You have no idea how happy it makes me. I really like and treasure my (trully!) 0db noise, passively cooled system, but the CPU speed was starting to make me trouble for some of the applications that I use.

The fact that I can move to the 1.8Ghz CPUm which is very close to the very expensive 2.16Ghz CPU, is like a spit in the face to anyone who told me I won't survive much longer with my silent computer. Some even called me a "silent dinosaur"!, figure that!.

My PC DOES NOT look as ugly as axle's :-). Btw, axle, which CPU are you using ?. Nice temps for a passive system.

Bat
Posts: 279
Joined: Sun Jan 05, 2003 12:09 pm
Location: U.K.

Post by Bat » Tue Aug 19, 2003 3:03 pm

I am very impressed!

Jones_Rush, how did you cool the power supply? Not many water-cooled PSUs about.

A word of caution: the modern faster processor might generate heat at a similar rate to the old one, but the die size is smaller, the power per unit area greater, so it's more of a challenge. Don't just assume that it's certain to work.

If there's no pump and you're relying on convection to move water through the system, you'll get best results if the radiator is much higher than the items that heat the water. That way the loop will have hot water on one side and cold (therefore denser) water on the other.

I've seen graphs showing how the performance of water blocks varies with flow rate, and how the flow depends on the pressure difference across the block, and I would never have considered trying a pumpless system if I hadn't seen this thread. What sort of water blocks are they? Series or parallel?

If it's a "(truly!) 0db noise, passively cooled system" then presumably you've done something to the hard drive as well, right?

Axle, you say passive... do you mean pumpless as well as fanless?

Bluefront
*Lifetime Patron*
Posts: 5316
Joined: Sat Jan 18, 2003 2:19 pm
Location: St Louis (county) Missouri USA

Post by Bluefront » Tue Aug 19, 2003 4:17 pm

I also thought the original idea here was a WC system without a pump.
Well I'm probably the only person around here who had an automobile that was water cooled....but no water pump. The term is thermo-syphon.
It could work on a PC with some mods.

As has been stated the radiator would have to be higher than the CPU. Coolant from the top of the water-block on the CPU would rise when heated, and enter the top of the radiator, fall through the radiator, out a hose and enter the bottom of the CPU waterblock.....then the cycle would continue.

My car required a fan to cool the radiator, but no pump at all. I think model "T" fords used this system also. My car was an East German DKW. This system is not practical on modern high-performance autos.

And I think it would require larger diameter hoses than I usually see on WC computers. But it sounds interesting....who wants to try one?

Jones_Rush
Posts: 6
Joined: Tue Aug 19, 2003 1:44 am

Post by Jones_Rush » Tue Aug 19, 2003 5:00 pm

deleted
Last edited by Jones_Rush on Wed Aug 20, 2003 6:12 am, edited 1 time in total.

Jones_Rush
Posts: 6
Joined: Tue Aug 19, 2003 1:44 am

Post by Jones_Rush » Tue Aug 19, 2003 5:01 pm

Jones_Rush, how did you cool the power supply? Not many water-cooled PSUs about.
I bought a complete water cooled case, the water cooled PSU came with the case.
A word of caution: the modern faster processor might generate heat at a similar rate to the old one, but the die size is smaller, the power per unit area greater, so it's more of a challenge. Don't just assume that it's certain to work.
I didn't thought about that... Do you think there is any chance that the XP22000 won't be harder to cool than the T-bird 1.13Ghz ?.
If there's no pump and you're relying on convection to move water through the system, you'll get best results if the radiator is much higher than the items that heat the water. That way the loop will have hot water on one side and cold (therefore denser) water on the other.
Yep, that's the way it works in my case. Sort of. It uses the entire side of the case as a radiator.
If it's a "(truly!) 0db noise, passively cooled system" then presumably you've done something to the hard drive as well, right?
Yes. I bought a single platter 40GB Seagate Barracuda IV, which is very quiet, and installed it inside an acoustic case. It's impossible to hear it at idle. That's the only part which is not truly 0db silent though (and the monitor too. I can hear the 100hz bazz, just a little bit)

axle
Posts: 64
Joined: Sat Oct 26, 2002 3:06 pm
Location: Wind Gap, PA
Contact:

Post by axle » Wed Aug 20, 2003 6:41 pm

:D Hey, may not be pretty, but running a 1700+ @ 2.3 w/ under 35c load comps for the bad fung shui.

Pumpless- I don't see the point. I can't hear my pump over the soft buzzing of my BEAU-TI-FUL Philips 107p, which I can't hear above my single WD HD in a smartDrive. The system is silent; I keep swearing at the breaker box every time I walk down the stairs- I think it's flipped and my comp is off.

I'll try pumpless when they come up with a solid state 200GB =-}

Guys, die size is pretty irrivelent when it comes to CPU WCing. Because we use water blocks, the heat dissipates into the block, which is cooled. Think about it, this is why direct die is so unsucessful (with AMDs; P4s have heatspreaders which help a might. But what's a heatspreader? A laser cut Intel heatsink).

Post Reply