Who has watercooling with no fan, no pump, and no case fan?

The alternative to direct air cooling

Moderators: NeilBlanchard, Ralf Hutter, sthayashi, Lawrence Lee

Visitor
Posts: 8
Joined: Sun Apr 29, 2007 9:06 am

Who has watercooling with no fan, no pump, and no case fan?

Post by Visitor » Sun Apr 29, 2007 2:46 pm

Have any of you made such a thing? I imagine it'd take a large external radiator?

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

Post by Bluefront » Sun Apr 29, 2007 3:32 pm

Er......without a pump, the radiator will do nothing. But there is an alternative. Hook your water-cooled system directly to the cold water line. Rather than a radiator, route the water to the front lawn......double-duty water cooling. Don't laugh too hard......this sort of cooling was used for air conditioner systems a few years ago. If you're not on a water meter, and don't mind the waste......it'll work.

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

Post by Rusty075 » Sun Apr 29, 2007 4:35 pm

Bluefront wrote:But there is an alternative. Hook your water-cooled system directly to the cold water line. Rather than a radiator, route the water to the front lawn......double-duty water cooling. Don't laugh too hard......this sort of cooling was used for air conditioner systems a few years ago. If you're not on a water meter, and don't mind the waste......it'll work.
A couple of problems with that:
-Condensation. The cold water supply is often below the dew point, so your lines and waterblocks will all be dripping with condensation. Generally not a good recipe for the inside of your PC.
-It's often illegal. Single-pass cooling loops are illegal in most places nowadays, with good reason. Granted, no one is likely to be calling the EPA on you, but its still a pretty dumb way to solve your computer cooling problems.


There is at least one guy that I know of who did build a pumpless, fanless, watercooling rig. Not exactly a practical solution, but interesting.

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

Post by Bluefront » Sun Apr 29, 2007 5:06 pm

That system I proposed would work ok in the summer, since the water is not cold enough to condensate. And it's perfectly ok to water your lawn, or garden, or the trees. Pretty dumb way to cool a computer however, just like water inside a computer is a dumb idea no matter what. The usual result is fairly obvious. :lol:

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

Post by Rusty075 » Sun Apr 29, 2007 5:44 pm

Actually exactly the opposite: it would work better in the winter when both the room temp and the dew point are lower. Anyone who's ever had a running toilet could testify to that.

There's nothing dumb about watercooling a computer per se, but using a system that dumps thousands of gallons of water a day onto your yard whether it needs it or not is rather silly. Compared to some of the complex fan-cooled schemes people come up with, throwing a couple of waterblocks and a pump into a case, and sitting a radiator behind it is a very simple and rational way to get to levels of quiet that are as good or better than aircooling. Not without its own set of issues, of course, but catastrophic failures that result in dead parts are about as rare as fan failures are.

CyberDog
Posts: 206
Joined: Tue Jan 30, 2007 5:54 am
Location: Kuopio, Finland

Post by CyberDog » Mon Apr 30, 2007 12:39 am

I have seen fully passive WC system. It worked but it was on PIII Celeron :D So I think that modern PC:s are just too hot.

peteamer
*Lifetime Patron*
Posts: 1740
Joined: Sun Dec 21, 2003 11:24 am
Location: 'Sunny' Cornwall U.K.

Post by peteamer » Mon Apr 30, 2007 10:59 am

My XP2400+ Thoroughbred 'B' and NVidia 6600 thingamee, both overclocked ~10% and running higher than stock voltages, were cooled in the same loop with no pump and no fan.....

No probs even if Folding 24/7.....

Fairly obvious result!!!... :roll:




Behold my leat 'clickable' skils:
Image



Pete 'Down wid it' Amer

nici
Posts: 3011
Joined: Thu Dec 16, 2004 8:49 am
Location: Suomi Finland Perkele

Post by nici » Mon Apr 30, 2007 12:01 pm

How do you get the water running in that thing? Or does it work because the water heats up? If that's the case then it should work better on a hotter system? im a bit slow in the head this week.

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

Post by Rusty075 » Mon Apr 30, 2007 12:21 pm

Yeah nici, its the old idea of "hot water rises".

I think you were using the wrong waterblocks Pete. I almost got a rig like that working...using a .5m WACC, and an old-style "straight through" Zalman VGA block as the waterblock strapped to an XP2100. The water would flow, albeit very, very slowly, but not fast enough to keep the CPU from overheating. I toyed with the idea of adding a TEC to the mix...that way you could keep the CPU cool, and let the hot side of the TEC be 80 or 90°'s to keep the water moving.

But then I bored, or distracted, or something, and never went back to it again.

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

Post by Bluefront » Mon Apr 30, 2007 1:00 pm

The Model "T" ford used this water-cooling technique......no water pump, but a big radiator. Tended to run hot if you weren't moving. By hot I mean 100C+. With a low power CPU no doubt it would work.....not so much by water circulation as by simple radiation. Why have a water cooling system without a pump anyway. Fins on a standard heatsink would probably be more effective. Guess what happens when you cut a hole in your roof. Now guess what happens when you have water, hoses, clamps, joints, radiators, water blocks, pump seals, etc....all crammed into your (up to this point) dry computer. :lol:

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

Post by Rusty075 » Mon Apr 30, 2007 1:20 pm

Yes, but as we all know fins attract dust, and dust is evil...with a nice big external radiator you can just vacuum the dust off once a month. :wink:

peteamer
*Lifetime Patron*
Posts: 1740
Joined: Sun Dec 21, 2003 11:24 am
Location: 'Sunny' Cornwall U.K.

Post by peteamer » Mon Apr 30, 2007 1:46 pm

Rusty075 wrote:I think you were using the wrong waterblocks Pete.
Why? :?

It was a WACC CPU block and a Gainward(?) GPU block.


Pete

peteamer
*Lifetime Patron*
Posts: 1740
Joined: Sun Dec 21, 2003 11:24 am
Location: 'Sunny' Cornwall U.K.

Post by peteamer » Mon Apr 30, 2007 2:02 pm

Bluefront wrote: With a low power CPU no doubt it would work.....
Thoroughbred 'B' is not low in terms of heat output.

Low heat output chips would probably hamper the effect.

Bluefront wrote:Why have a water cooling system without a pump anyway.
Sorry, I thought this was Silent PC Review..... :roll:

Bluefront wrote:Fins on a standard heatsink would probably be more effective.
How on earth do you reach that conclusion?

Bluefront wrote:all crammed into your (up to this point) dry computer.
Do you read what other people write?

Rusty quoted a .5M WACC and I quoted a 1M WACC.....


You're obviously 'out of your depth' in this forum... :lol:

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

Post by Bluefront » Mon Apr 30, 2007 4:16 pm

I don't swim very well.....and my floors have carpets and/or hardwood. So I try to avoid water incidents at all costs. Sorry but I'd have to build a water-cooled computer with passive water circulation, before I'd ever believe it was possible. Every water pump I ever heard is too noisy for me.

thetoad30
*Lifetime Patron*
Posts: 368
Joined: Fri Jul 23, 2004 11:33 am
Location: King of Prussia, PA

Post by thetoad30 » Mon Apr 30, 2007 8:45 pm

Not sure exactly what you mean... but it sort of sounds like you are "all wet" ;)

I was (and still am I'd say) a complete newb to watercooling. I purchased for my first set a Storm R2, Maze4GPU, MasterKleer 7/16" tubing, a PA160, and a MCW655 pump.

I used to run the pump full blast and had Nexus fans in a P180 case running full. I couldn't hear the computer because of the room I was in and the resulting ambient noise.

However, when I moved, I then started to notice my computer. So I upgraded to a PA120.3, and added three Nexus fans to that, then took all system fans down to 7V.

I did run a 3.4 GHz Prescott (a 550, one of the hottest CPU's you could buy at the time) on the PA160. It ran fine. I am now running the PA120.3 on my C2D E6700. The pump is at the lowest setting, which is inaudible with the P180. I have all fans still at 7V and cannot hear the thing turn on unless I put my ear to the front panel fan grills. On top of that, it never gets hotter than 55C on TAT fully loaded with ATITool pumping out all the wattage I can muster from my X800GTO^2 pipelined at 16 and screaming at 600/575. This, at one point, I calculated as running near 100W or more of power just to keep the thing running.

Water cooling can be just as or even quieter than air cooling. The only thing quieter is passive cooling, which, as it goes right now, is not very viable in a portable or realistic everyday setup. You'd have to really modify things to make it work.

Oh, and I've never had a water cooling accident either. No drips, no major problems. Only did some maintenance that I wanted to do and that involved flushing the system because the pipes became cloudy. Other than that, nothing bad has happened. I even transport the computer back and forth up to 8 times or more a year with no breakdown of components or anything.

Just my .02 on water cooling and it's effectiveness.

PS: Take swim lessons... one day they could save your life (and ears) :) ;)

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

Post by Rusty075 » Mon Apr 30, 2007 9:13 pm

peteamer wrote:Why? :?

It was a WACC CPU block and a Gainward(?) GPU block.
Sorry, my fault for not typing as much as I was thinking.

Here's what I was thinking:

For these passive loops to be be effective you have to have virtually no flow restrictions, since you're talking about a pumping force that basically has zero head pressure. And, ideally, you want the two parts of the system that have the ability to accelerate the water to be able to function to their maximum ability. On the cooling end the WACC already does that pretty well, assuming that you have it oriented vertically with the "hot" end up. it's the waterblock side that is tougher to get working. Regular waterblocks are all full of narrow nozzles, and twists and turns. The restrictions don't give the hotter, less dense water an easy way to rise up and out of the block. It doesn't matter how effective the block is at moving the heat out of the CPU if the hot water can't get out of the block. That's what I meant by "you're using the wrong waterblock"

That's why I thought the old Zalman VGA blocks might be on the right track. They're basically restriction free, and if you mount them and the board vertically you get a nice straight shot for the water to flow through. Too bad they're so poor at actually moving heat out of the CPU.



And don't be too hard on Bf...lots of people feel the same way about watercooling that he does. And lots of people think its pretty silly to fill homemade computer cases full of automobile intake filters and cobbled together ductwork too. :wink: So we're all even. Different sides of the same coin.

Visitor
Posts: 8
Joined: Sun Apr 29, 2007 9:06 am

Post by Visitor » Tue May 01, 2007 7:06 am

peteamer wrote:My XP2400+ Thoroughbred 'B' and NVidia 6600 thingamee, both overclocked ~10% and running higher than stock voltages, were cooled in the same loop with no pump and no fan.....

Fairly obvious result!!!... :roll:
Peter, I'm not sure I understand. Did it work or did it overheat? I looked up some old posts of yours that say it worked perfectly.

---
The waterblock sure seems to be critical:

The guy in Rusty's link achieved 300L/h with a DIY waterblock...

Same setup: 20L/h using narrower inlet-diameter Danger Den Maze a commercial waterblock
Last edited by Visitor on Thu May 03, 2007 3:52 pm, edited 5 times in total.

nici
Posts: 3011
Joined: Thu Dec 16, 2004 8:49 am
Location: Suomi Finland Perkele

Post by nici » Tue May 01, 2007 7:19 am

It looks like a Reserator would be ideal for this thing... Just make an inlet at the top.

peteamer
*Lifetime Patron*
Posts: 1740
Joined: Sun Dec 21, 2003 11:24 am
Location: 'Sunny' Cornwall U.K.

Post by peteamer » Tue May 01, 2007 10:07 am

Russ, Oh... now I see :D shall we start another debate on flow rate or do you think it's been done to death (with no concrete conclusions) already? :lol:


Hi Visitor, excuse my... 'humour' :D . Yeah it worked fine once you got the water circulating.
Normally let it sit in the BIOS set-up for a while to generate a little heat and start the process off. Though that was rarely a problem as my machine is on 24/7... with the occasional reboot not causing any worries.

Block design would be important if not critical. Both my blocks are low resistance.

My CPU block is cylindrical and originally had side by side inlet/outlet... a little 'adjusting' of the cradle :twisted: and it's now top/bottom.

I will/must get round to ordering some 1/2" barbs so I can use 12mm ID tubing instead of the 8mm I've currently got. That should help the flow quite a bit... but as the system worked that upgrade was stored in the 'To-Do' pile... :lol:

When I get round to WCing this new set-up I'll run it as is and then upgrade to larger bore tubing as an experiment, tubing size is a hotly debated subject, don't think I've seen it discussed for a passive system though. :D. I'm sure there are other sado's like me who would be interested in the results too. (Note to self: 'Get A Life!')

I can't give you any idea of temps as was because I'm convinced the board talked crap on that score. Suffice to say the auto shut down in the BIOS didn't.

nici wrote:It looks like a Reserator would be ideal for this thing...
Pretty much what a WACC is, though it's narrower and longer and with vertical grooves (Fluting ?) on the inside to increase surface area.
nici wrote:Just make an inlet at the top.
I've said that before...
And I'm sure some have talked about an internal pipe extension to send the incoming water to the top. Makes much more sense than to just mix the water up...

IF I recall correctly my WACC had a temp difference of 15-20C on the external surface. Will try to remember to test this also when I rebuild.



Regards
Pete
P.S. Fair few words but not as many as in my head. (Done a Rusty075 :D ) Feel free to laugh or ask questions... Even the air-cooled Heathens... :lol:

res0r9lm
Posts: 22
Joined: Tue Mar 27, 2007 12:11 pm

Post by res0r9lm » Tue May 01, 2007 10:57 am

you might be able to use mercury

peteamer
*Lifetime Patron*
Posts: 1740
Joined: Sun Dec 21, 2003 11:24 am
Location: 'Sunny' Cornwall U.K.

Post by peteamer » Tue May 01, 2007 11:18 am

Well that depends..... is it quiet?.....

Visitor
Posts: 8
Joined: Sun Apr 29, 2007 9:06 am

Post by Visitor » Tue May 01, 2007 1:43 pm

Too bad the WACC water block is no longer sold. I have no idea which of the available water blocks have low enough resistance but adequate heat transfer...

This link shows how good CPU-Water heat transfer is, but says nothing about flow resistance, which is probably positively correllated with heat transfer... Pressing the liquid through many small channels improves heat transfer (large water-metal contact area) but ruins flow rate.

jaganath
Posts: 5085
Joined: Tue Sep 20, 2005 6:55 am
Location: UK

Post by jaganath » Tue May 01, 2007 2:33 pm

Pressing the liquid through many small channels improves heat transfer (large water-metal contact area) but ruins flow rate.
just FYI, this is also true for air-cooling.

peteamer
*Lifetime Patron*
Posts: 1740
Joined: Sun Dec 21, 2003 11:24 am
Location: 'Sunny' Cornwall U.K.

Post by peteamer » Tue May 01, 2007 10:16 pm

Visitor wrote:Too bad the WACC water block is no longer sold.
http://www.overclock.co.uk/product.php? ... 7&view=rel

:D


Pete

CyberDog
Posts: 206
Joined: Tue Jan 30, 2007 5:54 am
Location: Kuopio, Finland

Post by CyberDog » Wed May 02, 2007 12:24 am

How about the blocks http://www.swiftnets.com/products/mcw6000.asp
They have really low flow restriction. Those are discontinued products but maybe there are some similar on the market.

unimatrix0
Posts: 88
Joined: Sat Jul 15, 2006 2:07 pm

Post by unimatrix0 » Wed May 02, 2007 9:32 pm

The D-Tek Fuzion is low-restriction. I wonder how it compares with DangerDen's maze cpu block, which I think was one of the least restrictive blocks around.

HammerSandwich
*Lifetime Patron*
Posts: 1288
Joined: Sat Oct 25, 2003 3:21 pm
Location: 15143, USA
Contact:

Post by HammerSandwich » Sun May 06, 2007 12:28 pm

peteamer wrote:I will/must get round to ordering some 1/2" barbs so I can use 12mm ID tubing instead of the 8mm I've currently got. That should help the flow quite a bit...
My gut says not, but I haven't done any calculations. Because the pressure drop increases with the square of flow, a pumpless system must have minimal flow resistance. OTOH, every reduction helps, and a passive setup is operating in the flat part of the pressure drop vs. flow curve. Hmmm... Now I'm curious about the results.

Elixer
Posts: 520
Joined: Wed Jun 02, 2004 10:31 am
Location: Las Cruces, NM
Contact:

Post by Elixer » Sun May 06, 2007 1:23 pm

Bluefront wrote:Er......without a pump, the radiator will do nothing. But there is an alternative. Hook your water-cooled system directly to the cold water line. Rather than a radiator, route the water to the front lawn......double-duty water cooling. Don't laugh too hard......this sort of cooling was used for air conditioner systems a few years ago. If you're not on a water meter, and don't mind the waste......it'll work.
Here's a thought for those in the dry states during the summer - What about using this idea with a swamp cooler? IE - having the water go through the computer before it gets to the swamp cooler. Granted, swamp coolers don't use a lot of water (at once), but with the high heat capacity of water I could see this working.
Rusty075 wrote:its the old idea of "hot water rises".
Also, my understanding of water is that it is not like air. Hot water doesn't rise like air does. The effect is neglidgeable I would say in pumpless system. Water is simply a good conductor of heat. Therefore my belief is that in a pumpless system, low resistance blocks do not matter. You simply want as large of tubing as possible and as short of a connection as possible to the radiator.

nici
Posts: 3011
Joined: Thu Dec 16, 2004 8:49 am
Location: Suomi Finland Perkele

Post by nici » Sun May 06, 2007 1:53 pm

Water gets less dense as the temperature rises, so it will move upwards. The change is not big, but it's there. here's some tables. http://www.simetric.co.uk/si_water.htm

mnanderson
Posts: 2
Joined: Sun Jun 04, 2006 10:42 am

Ultimate Linux Box 2005

Post by mnanderson » Tue May 22, 2007 6:13 pm

All the silence that money can buy.

http://www.linuxjournal.com/article/8292

Post Reply