Passive water cooling

eniacs · Post by **eniacs** » Thu Feb 26, 2004 5:27 am

Has passive water cooling been done?

I have looked about for a while and not seen any.

Anyways i have been toying with the idea of passive for a while and noticed two obvious solutions;
1; use heatpipes and enormous heatsink (zalman case etc)
2; use enormous radiator and water cooling

I found the cost of heatpipes too high at about £120 for the two i would require so i have decided on water cooling.

Will comprise of the following components;
eheim 1048 pump
maze3 cpu cooler
escort 1985-1991 1.4-1.6l radiator (this has copper fins, aluminium core and parts of brass aswell - brand new thankyou ebay...)

The radiator is 28cmx50cmx2cm thick

Is this going to be big enough? looking at the thing its fairly small (it looked larger in the picture) anyway i will try it and see if it cools it sufficiently, if not its back to ebay.

Oh the computer;
3.0g p4
1024 3200 ram
4x 120gb seagate drives (2 in raid 0 2 for storage all mounted in custom foam rack)
2 cd drives etc
ati9600xt with cheapy zalman passive cooler
zalman 350w psu
no case fans to be used

And the reason for this?
1; i want to try to be passive and leave the comp on 24/7 no reason really execpt i feel annoyed every time i have to wait 2 mins for it to boot up!
2; the noise irritates me

3; cos i want to try overclocking with watercooling (it wont be overclocked all the time tho)

So major question is:
Has anyone else tryed this? if so post information on it.

Thanks for reading.

Dr.CrackEnHore · Post by **Dr.CrackEnHore** » Thu Feb 26, 2004 8:24 am

Here is a couple of links to some really cool passive rads.

http://www.highspeedpc.com/Merchant2/me ... ovaKonvect

http://www.highspeedpc.com/Merchant2/me ... de=HTCSrad

Tad bit expensive though.

Qwertyiopisme · Post by **Qwertyiopisme** » Thu Feb 26, 2004 10:47 pm

Well, I had passive watercooling, I just used a huge 30 liter water tank used normally for wine fermentation (clear plastic). Now i have a 150x200x50mm radiator from something that is ewnough to run pasive for 2 hours on a stock 2500+, after that i turn on my enermax 120mm fan at 5v with the pot screwed down to slowest speed (I turn it on when the spu get to about 50 degrees, it then stabilises at 45 degrees at non-stop 100% cpu usage). I have a home-made water block and an eheim 1048 pump. (I got rid of the big tank becuase my parents were afraid of a leak).

zuperdee · Post by **zuperdee** » Fri Feb 27, 2004 12:36 am

I guess maybe everyone knows this already, but just thought I'd mention it anyway: if you want greater COOLING capacity in a water-cooled system, a bigger RADIATOR is what you need, not a bigger tank. I'd say if you REALLY want to be adventurous, you might want to consider using an AUTOMOTIVE radiator with your computer!!!!

(I have heard of people doing this before, actually.)

silvervarg · Post by **silvervarg** » Fri Feb 27, 2004 12:59 am

Where did you check the prices for heatpipes to get to that high cost?
Would they have been specially produced for you?

As an example the Zalman HDD cooler with 10 heatpipes cost ~$30 in shops. So the real cost for 2 mass produced heatpipes should be in the ~$6 range.
Problem is to find stuff that is suitable for your needs and that you need small quantities, so shipping becomes high compared to the cost of the heatpipes.
You might also need to produce the blocks for the CPU etc yourself, so some tools and craftmanship are neccessary.
In general the cost for the heatpipe solution should be far less than the cost for water cooling. At least if you just cool the CPU. The reason is that the initial cost for watercooling is a lot higher than the cost for heatpipe cooling.

Real passive watercooling would be without a pump, but I guess we call the semi-passive watercooling passive as well when we don't use fans.
If you go for passive watercooling your biggest problem is to get heat away from the radiator. The placement and orientation of the radiator makes a big difference.
Putting the radiator on the side of a tower chassi is NOT a good place.

Will you still have some chassi fan(s)?
Will you still have a PSU fan?
Do you expect the GPU to be passively cooled without airflow in a closed chassi?
Will you have waterblock on your NB?
Will you have waterblocks on your harddrives?
Will the harddrives be in some kind of enclosure? (If so, what will airflow inside the enclose be?)

axhind · Post by **axhind** » Fri Feb 27, 2004 1:01 am

It definitely has been done before, I even know a guy on a forum of which I'm a regular who has done it. Unfortunately for you it's a Belgian forum ( http://www.born2oc.be), so here's a pic (courtesy of homerke):

link to his site: http://www.sknet.be/homerke/born2oc/wc/

On that same forum some guy even used groundcooling! Let me explain with a drawing he made: (courtesy of Xentagon)

He just dig a hole and put in a radiator you use for heating your house

.

Watertemp was about 19° then, but he had to stop his experiment because his father didn't agree with it

.

My point: it's perfectly possible to have passive watercooling, but it will take a lot of place and material

.

eniacs · Post by **eniacs** » Fri Feb 27, 2004 1:22 am

silvarg;
I needed 2 heatpipes, of about 50cm and the sink was on the outside of the case, the prices were from two uk companies thermacore and some other one.

For the watercooling i will leave the case open for a while to cool passively the hardrives, graphics and north bridge, if all goes well i will expand to a watercooled northbridge and gpu, and mod up a cheap psu for watercooling. the current psu does have a fan but it is extremely quiet. the case is very thickly padded with cheap carpet underlay and so when closed the hardrives will not be audible. i have also experimented with hardrive enclosures (home made) i cooled 3 drives in a carpet insulated enclosure with a papst fan at full 12v (there were inlet and outlet areas!). nothing could be heard from the enclusure, but it overheated the drives to much and one began to stop working so i cancelled that idea and currently they sit in a foam block with most of the drive area in open air.

Dr.CrackEnHore;
Those are some cool radiators!

Qwertyiopisme;
If you can passively cool on that size radiator for 2 hours i dont think i will have a problem!

zuperdee;
That is exactly what i am doing, using a large automotiv radiator. Please post links to people who have done the same!

axhind;
Thanks for the link, that setup is good, but my rad is bigger so i guess seeing that mine should work really well!
Thats well funny about a guy digging up his garden to bury a house radiator! i can just imagine his mum planting petunias and finding a large radiator buried in the garden!!

Qwertyiopisme · Post by **Qwertyiopisme** » Fri Feb 27, 2004 6:38 am

I meant that I used the tank as my radiator. It doesn't have all that much surface area, but if you only run it 12 hours a day and use an average of 50% cpu usage my proc got to 35 degrees, after 12 hours of 100% (say a long 10000 frame 3d app render). i got to 37 degrees. It then cooles of during the night using the tank as a radiator. You need a radiator with a large distance between the flanges for best passive cooling, my heater from a bus cabin work very well with a fan, but not all that well passivly (up to 55degrees and still rising after 1 hour of 100% usage). Of course my proc gave off 85 watts because I had overcloked it alot and had the vcore at 1.85V, so you would probably get better results.

pdf27 · Post by **pdf27** » Fri Feb 27, 2004 6:46 am

Bladerunner (www.zerofanzone.co.uk) has a totally passive watercooled system (IIRC using a buried tank).Well worth a look, particularly if you're interested in a watercooled PSU as well.

eniacs · Post by **eniacs** » Fri Feb 27, 2004 7:26 am

Qwertyiopisme: you said:
Now i have a 150x200x50mm radiator from something that is ewnough to run pasive for 2 hours on a stock 2500+, after that i turn on my enermax 120mm fan at 5v etc...

if you can run that then i should have no problems with the size rad i have.
yes i understood what you meant about the tank, but i would need to run my comp 24/7 as thats one of the resons i want it quieter.

pdf27;
Yes i have seen the over excessive buried tank system! its good but he went to a lot of effort to get it buried where a normal radiator would do!

Thanks tho as i do intend to make the psu watercooled sometime.

silvervarg · Post by **silvervarg** » Fri Feb 27, 2004 8:00 am

Note that the XP2500+ is about 61W where the P4 3.0 is about 81.9W.
The external volume of your radiator is about 86% bigger, and hopefully that translates to roughtly 86% bigger surface area, but that depends on the construction.

That said it is not certain that you will be able to run passively with 100% load for many hors in a row. So first question is if you will do that? E.g. if you will run folding@home or something like that?

Due to the noise from your harddrives I would rather have a slow chassi fan and the chassi closed than no chassi fan and the chassi open. If you look at the noise levels you will do better with the fan.

Seal · Post by **Seal** » Fri Feb 27, 2004 1:34 pm

Hey guys! Another watercooling post wuhoo! I watercool too.

I really dont reccomend running your case without any fans in, i used to do something similar and because there was so little airflow around the processor area the capacitors swelled up and i had to stop using the motherboard. Also no/very little airflow will also probably cause your graphics card to overheat and do funny things which mine also used to do. (in my old setup i left my case side open with no exhaust fan and a small fan on my radiator at the front)... infact let me get some pictures...

****Back! with piccys!****

Thats my old setup, and it was run 24/7 exactly like you see in the picture, case side off. The fan on the rad was run on 6v (using a rheo).

Thats what happens with zero airflow around the processor area.

From another angle just for perspective.

Learning from my mistakes i learned that i need at least a little airflow running through my case at all time, the most efficient way is sucking air from the front and blowing it out the back with the case side on. I figured that since i am putting at least 1, 5v 120mm fan on the front i may as well have an internal radiator and have the air from the 120mm fan blowing over the radiator which would be enough to cool my heavily overclocked processor (oh yeah, running 100% load 24/7 too - folding@home). For good airflow i also added a 120mm exhaust fan also running at 5v.

Having my setup like that also allows me to passively cool my graphics card with a zalman heatpipe. (i havent got it yet but im getting it soon)

anyway, heres a pic of my current setup:

i figured that internal watercooling is probably the best and most efficient method for me - as i have 2 120mm fans in my computer for airflow already, i might as well use it to my advantage and put my radiator in there as well.

The setup is going to be re-done soon, mainly to tidy it up and give my watercooling a good clean.

Seal

Post by **Rusty075** » Fri Feb 27, 2004 4:43 pm

You can't really call it "passive" if you're using a pump......

Pete's Passive Prototype

Seal · Post by **Seal** » Fri Feb 27, 2004 6:37 pm

http://www.plees.f2s.com/ec/pas-cool/pa ... coffee.jpg

omg! didnt your gfx card snap in half when you did that?

Post by **Rusty075** » Fri Feb 27, 2004 8:05 pm

Yeah, that does look just a wee bit risky. Maybe he has it supported my more than just the AGP slot, I hope?

eniacs · Post by **eniacs** » Sat Feb 28, 2004 3:05 am

6 entries found for passive:

3: Existing, conducted, or experienced without active or concerted effort:
5: Electronics. Exhibiting no gain or contributing no energy: a passive circuit element.

(i skipped out the other entries)

Ok so a passive system would be a system with no moving parts of any kind, like a huge heatsink bolted on top of the cpu with no fan or anything.

So what about a heatpipe solution is this passive cos there are no powered parts, but inside the pipe the water moves, evaporates etc?

quote from rusty's site:
I watched small specks of dirt through the translucent pipes which travelled round at about 0.3m/s. I did a calculation and reckon I am getting around 300 litres per hour of water flow.

30cm per second flow!? thats a lot of flow. i wouldnt expect that the 80w of the athlon would be enough to move much water but it seems that your setup worked! when i get mine put together i'll have to try it without a pump and see how it copes!

silvervarg · Post by **silvervarg** » Sat Feb 28, 2004 4:04 am

Seal:
I really dont reccomend running your case without any fans in, i used to do something similar and because there was so little airflow around the processor area the capacitors swelled up and i had to stop using the motherboard.

That was a very interesting post.
Note that it is not the temperature of the capacitors that is the problem, it is to temperature of the FET's close to the capacitors that probably got overheated so they bring the motherboard out of spec.

Could you give us some more info on exactly what you where running?
E.g. What motherboard did you use? (looks like an Abit board)
What CPU did you use?
vCore?
FSB?
Multiplier?
Any other things that draw power from motherboard that might be extraordinary? (Like lots of RAM or powerfull cards etc)

I am in the process of doing something similar, so I hope to avoid doing the same misstakes.

Post by **Rusty075** » Sat Feb 28, 2004 5:14 am

I think I may have confused eniacs and Seal. That link isn't to my machine. It's owned by a very resourceful (and very brave) guy named Pete. To my knowledge he doesn't actually post here, but he has been active on the Silent PC mailing list.

eniacs · Post by **eniacs** » Sat Feb 28, 2004 8:39 am

seal;

(being a communications engineer) i have seen this sort of thing on caps, usually when they get either very hot or the polarity is reversed, or some times a higher voltage can cause it.
I would think that the voltage regulating FETs that silvarg has mentioned probably get very hot (considering they power the cpu with around 80amps!), and with no airflow probably caused the caps to get too hot. Most newer boards have sinks on the regulator FETs and so this wont be a problem for me or most people.

Just out of interest did the board actually stop working?

rusty;

you should get pete to post here! he is obviously very inventive! and would be an asset to the forums!

Seal · Post by **Seal** » Sat Feb 28, 2004 4:33 pm

Hi,

Basically the motherboard was an Epox and i, and a few other people on the overclockers forum put it down to the low quality capacitors used, doing a google search on the subject returned a surprising amount of results too. Epoxes are known to be a cheap make and they must have cut their budget back on things like caps.

When i was using the board i did feel the caps whilst the machine was on and they ran very hot so you could barely touch them for more than 2-3 seconds.

In that setup i ran a gentle overclock of just the ram if i recall correctly balancing the multiplier so processor speed was retained but fsb speed was increased. Voltage wise, the processor voltage stayed the same (it might have been increased very little - 0.05v or around that) and the ram voltage tuned up by 0.1-0.15v ish.

I noticed the problem when i was checking my computer and cleaning it out routinely, although my computer was still working i stopped using the motherboard in the fear of the caps blowing and taking out other devices attached to the motherboard (im not 100% sure if it would have taken out other things - maybe you know eniacs?) I also recall my computer becoming a little unstable which would make sense thinking about it, as any voltage spikes would cause disrupted signals to and from the processor etc...

in a watercooling setup where there is little airflow around the cpu and fet area, would you reccomend putting heatsinks on the fets eniacs?

eniacs · Post by **eniacs** » Sun Feb 29, 2004 3:38 am

seal;
i think the moral of the story is dont buy a cheap mobo! i used to fall for the cheap prices long ago,and i would get serious stability problems. since i started paying more for the mobos they have given me no trouble at all and i have had a happy 5 years computing!

I have never found caps to get hot as in their nature they are passive circuit parts and should not get hot. i would think that the fets heated up the area around them and caused the caps to be very hot, those fets are probably rated upto 150degrees C but the caps are probably no more than 80 degrees. If the caps did die the comp would go unusable and crash all the time because they are smoothing capacitors for the cpu's power supply. i doubt it woud take anything else out however so i would continue using the board until they pop. your choice tho.

Seal · Post by **Seal** » Sun Feb 29, 2004 3:13 pm

yeah, now i wont buy or reccomend cheap motherboards, i used to run the same setup with a very good asus motherboard and had no problems with the caps and stuff, then i switched to the epox which died, and now ive gone back to asus, one of the more reputable brands, i dont think ill ever buy an epox board again. Quality-wise you can just tell by looking at the capacitors that the asus ones are of much higher quality. and not as cheap-looking as the epox ones.

Gooserider · Post by **Gooserider** » Sat Mar 06, 2004 6:58 pm

I have had caps fail on me in other contexts, not on a PC, and I would not reccomend using a board with a suspect cap on it. An electrolytic can cap, like the ones Seal was showing in his pictures, is essentially a tin can with a plastic plug on the end where the leads go in. The inside is full of a sort of foil spiral (actually two of them) seperated by a non-conductive dielectric. This is immersed in a liquid made from something nasty, I'm not sure just what (and it may vary) with the works sealed up tight.

When a cap overheats, the liquid starts to vaporize and build up pressure, just like a spray can in a fire. When the pressure gets to high, something has to let go. On Seals board, those little three legged depressions are machined in weak points intended to let go 'safely'. On older caps without those lines, I've seen the can shoot off like a rocket (In one case hard enough to embed in the ceiling tile eight feet overhead...) Either way if the cap lets go, it can spray nasty goop around, possibly damaging other components. I've also seen tantalum 'gumdrop' caps do the same thing.

Thus, while the cap failing itself might or might not cause a problem, there is significant risk that the cap might damage other parts when it lets go.

OTOH, there isn't anything to prevent you from replacing the suspect caps - electrolytics are relatively low cost and easy to come by. Just get the same uF value and the same or higher voltage rating parts, and fire up the soldering iron...

Gooserider

Seal · Post by **Seal** » Sun Mar 07, 2004 3:11 am

Thx for the info gooserider. Ive tried to look for replacement caps but the sizes are impossible to come by. After doing a bit of research, i read that motherboard manufacturers have capacitors taylor made to specification. I was also told that i could use a capacitor of slightly higher value if i couldnt find the exact one i am after.

joecuba · Post by **joecuba** » Sun Mar 07, 2004 4:39 am

Higher value caps would be fine. These things are really impresise, I think tolerences can be as high as +/- 20%.

RS components in the UK carry a good selection of caps. You want the ones rated up to 105°C.

An electrical engineer I know beleives that caps such as these are placed close to hot components to deliberatly shorten the life of the board.

A simple mod would be to extend the wires to the caps (i.e. not have them mounted hard on the board). This could be taken further with a small seperate stripboard made for the caps, and connected to the motherboard with extension wires.

Seal · Post by **Seal** » Sun Mar 07, 2004 1:48 pm

interesting concept joecuba, thx for that webby ill check them up.

eniacs · Post by **eniacs** » Wed May 05, 2004 9:48 am

joecuba;

Caps are placed near to the hot components as they are part of the power supply and aid smoothing of the supply line. they are placed nearer to the devices so that they act quickly when the supply is dipped by devices switching. Hence why the cpu is covered in caps as the cpu needs lots of smoothing due to it switching billions of times per second.

A mod to move the caps away could cause instability due to increased supply dip times because of the increased wire lengths. Give it a go it you want it may work! But it may not...

My suggestion is to go for high temperature caps from RS.

Passive water cooling

Passive water cooling

Bladerunner...