Running Passive in ambient 30C temps
 

Hello,

I setting up my first watercooling rig. My computer is in my bed room and am considering running the setup passively. My concern is that ambient temps where i live average 30C. Most threads i read recommend placing the rad horizontally, some distance from the surface and (maybe) a chimney.

Will a thermochill 120.3 be enough for just the CPU(Prescott)?
If not wat would you recommend? (other than fans :D)

Please share your summer experiences w passive setups. Thank You.

Bing

Hi duh_bing. I'm a big fan of passive radiators.

I can't comment on the Thermochill, but I've been using an 8 foot length of 3/4" diameter copper tube as a radiator for about a year now on an undervolted, stock speed Athlon XP 2400+. My goal is silence, not performance. The copper tube is laying on the floor in back of the computer. My CPU temps run about 20 degrees C above ambient under load as reported by motherboard sensors. The motherboard sensors' results are probably not accurate, but do seem to follow the temperature changes in the room. Higher summer room temps mean higher CPU temps.

My first try at a passive radiator was a variation of the copper tubing mounted in a chimney. The chimney didn't seem to make any difference in temperature, not even when I added a fan to the chimney.

Cheers, Joe

Hi Joe,
Thanx for you input. found ur tower over at spcr. looks very interesting! giving me new ideas!! :D mind if i print out for reference? :D

Do you have a rough idea of wat kind of thermal output you have from ur undervolted Althon? Cos from data i found, the 3.2E can reach 100++W of heat output. My concern is that not nuff heat can be dissipated.

Was lookin at this(http://www.overclockers.com/tips1093/) too and seems like quite a bit of space is needed to get a good passive setup...

Feel free to print it out. That design never quite lived up to my expectations, but I did get a lot of practice soldering it together. :) My inspiration for the Tower of Cooling Power was the article you linked at overclockers.com. I thought the added fins and chimney would improve the cooling.

I'm not sure of the thermal output of the Athlon, but the undervolting made about a 4 degree CPU temperature reduction. If I remember right, I went from the stock Vcore of 1.65 volts to either 1.50 or 1.55 volts. Check your BIOS to see if you can do it on your motherboard.

I've been considering adding another 10 foot length of copper tubing to the loop. Right now, the 8 foot length is cut into two 4 foot pieces, joined with elbows on one end, so it takes up a space of about 4 inches x 4 feet behind a desk.

We're getting close to heating season in Wisconsin and my tubing is running near the baseboard hot water register since I moved it a couple of months ago. I need to see how the temperatures change when the heat goes on. I've been wondering if more tubing = cooler temps.

Joe

Waiting for LHG to jump in here, he's had some experience running passive rads over in Sydney, and I assure you from my time there, in summer ambients get a fair bit more than 30degC. :D

His reply in the oc aus forum
erm... so if all else const... the 120.3 shd perform better?
1m above ground in a corner of the room, dun get much airflow in that area thoits the 3.2e, have the rbx, a pair of ddc and the 120.3 ready, just need to get some tubing. Asking around to get a rough idea of wat to expect.

I think you will get better temps, since there is a larger area for convective heat transfer to occur ya?

hmmm... gonna get my tubings later...then start testing w the 120.3... and cal how much copper pipe i can squeeze into my room ;) hee...

Joe,
Can you do a measurement of the ambient temp and liquid temp with the same thermometer? Then i can calculate the average heat transfer per meter/feet of copper pipe. mean to put in a "please" hee...

bing

Not to thread crap, but you've chosen the absolute worst platform for passive cooling. If you're committed to Intel, a p-M rig would do a much better job.

I just got in to work, but I'll try it when I get home.

heh... you see... i had my rig b4 i got my water cooling :)
And here comes the challenge to go extreme!!! 2km of copper pipe!!! :D hehe... naw... too expensive... will go broke at 50cents/ft (~US$0.35)

Thanx dude! :dome:

I've often thought that copper pipe with swaged aluminum fins (like in a baseboard heat register) might make a good passive radiator, although IMHO it'd work better with longitudinal fins (rather than perpendicular) to assist air-thermosyphon - went searching for an example and found this:
http://www.tex-fin.com/images/longit.jpg
which, sadly, probably costs a lot - but offers the possibility of a non-aluminum reserator-like-thing...

I checked some temperatures tonight. Ambient temp and water temp were checked with my Micronta Indoor-Outdoor digital thermometer. CPU temps were logged with Speedfan . The computer was warmed up with a 30 minute game of Sven Coop, then left running at idle for 30 minutes.
I started another game of Sven, started logging with Speedfan, and measured water temperature in the resevoir every 2 minutes for the next 40 minutes until the computer blue screened. :mad: No joy after a restart, BIOS was stuck on 'Detecting IDE Drives'. I've got a Zalman passive heatsink on the Northbridge chip that seemed too hot, so I let everything cool for a few minutes and it's working fine now.

Start of test
Ambient 25.0C
CPU 40.0C
Water 31.0C

End of test
Ambient 25.1C
CPU 45.5C
Water 33.0C

I hope this helps, duh_bing. I'm not sure of the accuracy of the onboard CPU temp sensor or the accuracy of the digital thermometer, so view the results with some suspicion. Let us know what you find out when you do the math. :)

Here's a chart of temperatures/time:

gee... THANX!!! :D risking your computer for a whimp of mine *muacks* hheeh

Will work on the numbers.. get back to you once they are done ;)

Hi I run a Barton 2500 (65Watt) and a Hydor L30

I cant run it passive vertikaly however in horrisontaly there is no problem for me, the water temp is 38.3°C after 3 h with burn in (amb. =21,3°C) running = my CPU = 51°C.
There is a new radiator 3*3*12cm over at www.vcore.dk, that one should do the trick. (or use a Landrover-71 radiator).

If I where you I'l mount one papst 12cm and run it @ 5 volt, seal of the rest of the area. When I run my (3) papst @ 7 volt I cant hear them (and I have a very sensitive ear) throw the radiator, then my water temp is 3 °C above abm. (@12 volt delta T = 2°C)

Unclear to me exactly what radiator you are running horizontally - but I had another thought: why not use a long and narrow heatercore horizontally? For instance, the fedco 2-766 is 14.2"x3.25" (1.625 thick). Put that horizontally along the "dead" side of your PC case (the one that doesn't open up) a few inches up from the floor. It's then easy to build a "chimney" the same size as the rad, using the side of the PC as one wall. A chimney will provide some "stacking pressure" that'll help move air through the rad.

BTW, the 2-766 goes in an 84-96 AMC eagle and is $40 at partsamerica.com (currently free shipping) - and it's a single pass rad, which, for the silent-running crowd, means you can use a slightly lower powered pump and possibly reduce noise a tad

Just a thought, of course...

joesgarage11:
the numbers i cal from your data doesn't make sense to me... it looks like your convective heat transfer coeff is ~30W/m2K that is way above the ~6W/m2K i cal for free convection. the energy equation -> Ein + Egen - Eout = Estored
look at the water loop system as a control volume
Ein - CPU (60W from AMD spec sheet)
Egen - pump (assume negligible compared to CPU)
Eout - Rad
Estor - Energy that is stored by the fluid with increase in temp. should be 0 at steady state. I suspect that your sys have not reached equilibrium when you did your measurements and thus a skewed result set.

________________________________________________

Just set up a passive loop with a 120.3, 6m of garden hose and a small res... took 2hrs to reach ~60CPU ~45H2O ~29Ambient at full load (running dual prime95, an mpeg convertor software and rise of nations) might still go up but then i had a breeze :D

The rad was place vertically against an open window. Will repeat the experiment horizontally tomorrow night. need to sleep its 6am in the morning here... hee... leaving the sys to run to see if temps will get any higher w the noon sun.
_______________________________________________

mwolfman:
as mentioned above, i will try the horizontal orientation once i get my 1 day data. I think i shd get the same sort of results as yours. Can't expect the winds to favour me all the time ya? ;)
_______________________________________________

will post some pics of the temp setup when i wake up... good morning! :D zzzzZZZ

hmm, that could actually be portable pretty easily too, with some hinges on both the rad and the chimney...

Not sure you'd even need hinges - a tower case would just be, say, three and a half inches wider. If it was sheet metal and painted in a matching color it might not even get noticed...

"Gigabyte GA-7N400 Pro2, Barton 2500@2300MHz, TurboBlock Rev. 1.0, HL120.3, Hydor L30II, X800 XL, 1 GB Cosair 3200 (6-2-2-2) "

________________________________________________

"Just set up a passive loop with a 120.3, 6m of garden hose and a small res... took 2hrs to reach ~60CPU ~45H2O ~29Ambient at full load (running dual prime95, an mpeg convertor software and rise of nations) might still go up but then i had a breeze :D"
Way to hot, what happens if there are just a little bit of sun on a black radiator...

mwolfman:
as mentioned above, i will try the horizontal orientation once i get my 1 day data. I think i shd get the same sort of results as yours. Can't expect the winds to favour me all the time ya? ;)
_______________________________________________

Test the rad. with one fan (block the rest of the area on that side, an run it @5volt
the result should be arund 5-8 °C above amb temperature.

Over view of test setup
http://www.photoi.org/share/upload/1...7/overview.jpg

Close in shot of mobo set up
http://www.photoi.org/share/upload/1.../mobo_shot.jpg

Results, using mobo thermo sensor since using no probes on hand. Ambient is 28C to 30C in left and middle block(night), Ambient max 33 for right block.
http://www.photoi.org/share/upload/1..._4_posting.png

A pic of the wind.
http://www.photoi.org/share/upload/534712048/wind.jpg

Or go for BIG radiator, 3*3*12cm
http://www.vcore.dk/images/2210-1-p.jpg ~$120
http://www.vcore.dk/images/2209-1-p.jpg ~$100

hehe.. i've decided to go ghetto and make my own pipe rad if i need more cooling power. But those are cool rads tho ;)

Pipe-radiators aren’t that efficient, the area is way to low compared to the weight/space.

Actually, i think there is something to them that is overlooked a little. They hold loads of liquid too. The additional liquid will take it much longer for the temps of the water to rise high enough. In my setup it took approx 5 hrs to reach equilibrium and that is much longer then the time i play games. If i increase the volume of liquid, in terms of cost effectiveness(for my purposes) it shd be better.

Then the best rad for you may be no rad at all, but just a large storage container with a sealed lid. It can radiate a bit of heat, but if you only do quick bursts of PC usage (as in hours, not 24 x 7), then 20L to 50L takes a long time to heat up.

60-15=45
45-30=15
h/V=(15*4180)/((pumppower+CPUpower)*3600), asuming 100 Watt CPU and 27 Watt pump, that gives 0,137 h/l
5h gives 36,5 liter

Results for horizontal orientation
http://www.photoi.org/share/upload/1...17_results.png

Heres a copy of the results for the vertical orientation
http://www.photoi.org/share/upload/1..._4_posting.png

From the results of the tests. The vertical orientation seem to yield better results in terms of consistency. This is prob due to air moving in and out of the house, even if it is minimal. The using the rad in horizontal orientation to "move" air by heating it doesn't seem to provide any obvious benefits(prob due to tight spacing of the rad). Will dig out my 120mm fans to fix on the rad for another set of results(prob in 2 days?).

Bing

Disclaimer: Results are based on one test each, and most prob not conclusive. Getting results over a longer period will prob provide even more data. But i'm getting lazy :D

:p will have to find a place to stuff such a big res tho, hardly have any space left for walkin! :p *ok ok... its a guys room ;)* On the serious note, I'd prob keep the rad And have as much water as i can (makes temp rise slower ya?)

Thats heat capacity / energy generated per hour?