Passive watercooling

Boli · 08-15-2003, 06:57 AM

OK here's somethig weird that may interest some of you.

Now as usual before I went out to work I set my computer up on one of its many downloading/folding mission only I forgot to turn on my pump.

Yet when I can back and realised my mistake the temps were fine... As I have burnt out my temp sensor a long time a go I rely on the old touch the componants to see if anything is burning up on me... but it all was cool. even the water in the systsem was nothing more than luke warm.

Now thinking about this later I came to the conclusion that the entire thing cooled itself by convection alone (I had at least put the fans running on my rads.)

Now considering my componants... black ice rads maze 3 it would not take a hugely powerful pump to pump things round at all. add that to my layout where the rads are at the top so the heated water would rise easily to the rads and cool there. There was no impingment needed nor was there any contrictioning in the tubing/layout to worry about.

Now consider this... would it be possible to cool a computer passivly (As I seem to have done) WITHOUT need for a pump relying on convection of the water entirely n a 24/7 timescale?

As I blew my temp sensor up yonks ago I cannot verify that my CPU wasn't burning out under the block but I think it is worth thinking about and maybe looking into.

Wondering

~ Boli

Brians256 · 08-15-2003, 10:41 AM

I can verify that this MIGHT work if you aren't overclocking or overvolting your CPU. I've done it once by accident. One time it worked, and another time, the temps just kept going up and the computer shutdown safely (due to MBM watching the CPU temp).

Part of this is that the AMD CPUs can continue to work at insane temperatures like 60C to 70C (can someone please respond with the actual max CPU temp?). Some of that is the chip that you use, too. A high-frequency chip simply won't work well unless you cool it well, since it is already running at the edge of tolerance.

Now, the reason your water may be cool is that convection cooling really isn't that efficient. The heated water may not move around as much as you'd like, so .... I wouldn't recommend this if you don't have a functioning CPU temp sensor and a really good shutdown program watching that CPU temp. People have melted stuff (vinyl hose, CPUs, etc...) by having a pump go out. So, I know it doesn't work all the time.

eRNo · 08-19-2003, 08:51 AM

85-90c, depending on the model, if I remember correctly.

trit187 · 08-19-2003, 09:00 AM

intel's spec sheets say 70 for p4's but I think that's for continous use, but I'm 95% sure that at least the xeon's are all tested at 80 when testing for multiplier, mhz and voltage
so they're actuall max running temp would have to be somewhere above that

pHaestus · 08-19-2003, 11:03 AM

I have had an AMD TBird running at 85C (according to the thermal probe epoxied under the core) for days at a time with no problems and no apparent change in overclockability. Dunno how much of its life span I compromised though

Zhentar · 08-19-2003, 11:22 AM

None, the thing is already completely worthless

Gooserider also mention he was looking into convection watercooling.

08-15-2003, 06:57 AM	#1
Boli Cooling Savant Join Date: Apr 2003 Location: Blackburn / Dundee Posts: 451	Passive watercooling OK here's somethig weird that may interest some of you. Now as usual before I went out to work I set my computer up on one of its many downloading/folding mission only I forgot to turn on my pump. Yet when I can back and realised my mistake the temps were fine... As I have burnt out my temp sensor a long time a go I rely on the old touch the componants to see if anything is burning up on me... but it all was cool. even the water in the systsem was nothing more than luke warm. Now thinking about this later I came to the conclusion that the entire thing cooled itself by convection alone (I had at least put the fans running on my rads.) Now considering my componants... black ice rads maze 3 it would not take a hugely powerful pump to pump things round at all. add that to my layout where the rads are at the top so the heated water would rise easily to the rads and cool there. There was no impingment needed nor was there any contrictioning in the tubing/layout to worry about. Now consider this... would it be possible to cool a computer passivly (As I seem to have done) WITHOUT need for a pump relying on convection of the water entirely n a 24/7 timescale? As I blew my temp sensor up yonks ago I cannot verify that my CPU wasn't burning out under the block but I think it is worth thinking about and maybe looking into. Wondering ~ Boli __________________ 1800+ @ 2247 (214x10.5) - STABLE, 512MB PC3700 TwinX Cosair RAM, NF7-S v2.0, GeForce3 Ti200 Parallel BIM, 120.1 Thermochill, Eheim 1048, Maze 3, Maze4 GPU, "Z" chipset, 1/2" tubing, PC-70: 5x120mm & 9x80mm fans. Internet Server & second machine (folding 24/7): 512MB DDR RAM, XP2000+

08-15-2003, 10:41 AM	#2
Brians256 Pro/Staff Join Date: Oct 2001 Location: Klamath Falls, OR Posts: 1,439	I can verify that this MIGHT work if you aren't overclocking or overvolting your CPU. I've done it once by accident. One time it worked, and another time, the temps just kept going up and the computer shutdown safely (due to MBM watching the CPU temp). Part of this is that the AMD CPUs can continue to work at insane temperatures like 60C to 70C (can someone please respond with the actual max CPU temp?). Some of that is the chip that you use, too. A high-frequency chip simply won't work well unless you cool it well, since it is already running at the edge of tolerance. Now, the reason your water may be cool is that convection cooling really isn't that efficient. The heated water may not move around as much as you'd like, so .... I wouldn't recommend this if you don't have a functioning CPU temp sensor and a really good shutdown program watching that CPU temp. People have melted stuff (vinyl hose, CPUs, etc...) by having a pump go out. So, I know it doesn't work all the time.

08-19-2003, 08:51 AM	#3
eRNo Cooling Neophyte Join Date: Apr 2003 Location: sweden Posts: 6	85-90c, depending on the model, if I remember correctly.

08-19-2003, 09:00 AM	#4
trit187 Cooling Savant Join Date: May 2003 Location: shanghai, china Posts: 200	intel's spec sheets say 70 for p4's but I think that's for continous use, but I'm 95% sure that at least the xeon's are all tested at 80 when testing for multiplier, mhz and voltage so they're actuall max running temp would have to be somewhere above that

08-19-2003, 11:03 AM	#5
pHaestus Big Player Making Big Money Join Date: Aug 2001 Location: irc.lostgeek.com #procooling.com Posts: 4,782	I have had an AMD TBird running at 85C (according to the thermal probe epoxied under the core) for days at a time with no problems and no apparent change in overclockability. Dunno how much of its life span I compromised though

08-19-2003, 11:22 AM	#6
Zhentar Cooling Savant Join Date: Apr 2003 Location: Madison, WI Posts: 225	None, the thing is already completely worthless Gooserider also mention he was looking into convection watercooling.

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