Improving cooling efficiency

[Voodoo Rufus]

I am running water cooled, and I want to run it cooler than I am. My specs are on my sig, and here's a few more details.

Temps:

Ambient: 20C
Load under Seti: 38C

I know my components are better than that. My system has an inlet hose going from the rad to the reservoir that takes a rather sharp 90 degree turn that bends the tubing when the water gets warm. I think a 90 degree elbow would help. I'm using Radio Shack thermal grease, but I will eventually change that to Arctic Silver. Then I am running 50/50 antifreeze and distilled water. Water Wetter and distilled should help that area. Any other suggestions for improving the cooling efficiency? I'm stuck at 1600 MHz and I'd like to squeeze some more out.

[RangerJoe34]

umm, have you tried better fans on your radiator...or sticking the entire radiator in an ice chest, thatd cool ALOT better.....yea fix that bend in the hose....but your temps are fine if you are trying to overclock more

[Voodoo Rufus]

What I am trying to do is to minimize noise while having the greatest cooling power. I can't hear my Panaflo at 5.8V. My computer is near silent, except for the PS and the hard drive. My computer is fully encloesd also, and not very convenient to place in an ice chest. I'll fix that hose over Christmas break.

[Christoph]

90 degree angles are too sharp. You'd get better results with a gentler angle change, ie 2 45 degree elbows. Also, 1/2" ID tubing will help, if you can do that. If not, just try to get the most gradual change in angle readily possible.

[Voodoo Rufus]

I am using 1/2" ID hose all around. I'll look for those 45 degree elbows or 90s . Whatever I can find at the hardware store I'll use.

[JetMech]

Originally posted by Voodoo Rufus
I am using 1/2" ID hose all around. I'll look for those 45 degree elbows or 90s . Whatever I can find at the hardware store I'll use. Sometimes it is possible to eliminate a kinking problem by installing a longer hose and allowing it to form a loop. If you have the room this would be an inexpensive solution and would increase the amount of surface area in the cooling system. The latter mentioned affect is only beneficial if you have really good internal air movement otherwise any heat dump inside the case would be counterproductive. That is one of the advantages of super case airflow. It gives more design flexibility.

[Sonny]

I'm assuming that the fan on the radiator is blowing out so were is the supply of cool air coming from? Do you really need the antifreeze & if you do isn't a 50/50 mix too much? Just a thought or two;)

[Voodoo Rufus]

I now know a 50/50 mix is excessive. I've read a 80/20 H2O-antifreeze would be good enough. I'll try to get my hands on some water wetter sometime.

The fan on the radiator is blowing out and the cool air is coming from the front 120mm case vent, which I have an aluminum filter on.

[Sonny]

You'll probably better temps with less or no anti-freeze in your set-up. Since you do not have a peltier/TEC/scary piece of hardware:eek: there is really no need for it. Surfactants will be a better choice to mix with water & they do not have minute particles that can destroy the pump like anti-freeze.

[KLowD9x]

Does that pump interfere with anything? and what kind of pump is it?

[Voodoo Rufus]

Eheim 1048. I don't think it interferes with anything. It's fairly free-flowing except for the hose kink and the fluid I use.

[Voodoo Rufus]

Which is better for decreasing system temps? A water chiller with a couple 172 Pelts or a 172 Pelt on the proc? Also, wouldn't a water chiller only be able to get the water to 0C unless one uses antifreeze to keep the water fluid?

[Christoph]

Originally posted by Voodoo Rufus
Which is better for decreasing system temps? A water chiller with a couple 172 Pelts or a 172 Pelt on the proc? Also, wouldn't a water chiller only be able to get the water to 0C unless one uses antifreeze to keep the water fluid?

Huh? What do you mean "get the water to OC?"

[Voodoo Rufus]

0 degress Celcius

[Christoph]

I get it. If you want the ° sign, use alt+0176.

[Warlord2]

omg thats almost the same resavor I made gj:D

anyways

you want about 5% water wetter and 95% distilled water that should decrease temps 2-3 degress

also try going from radiator to cpu that may get you a nother degree cooler

AS2 will also get you maybe a degree cooler so that would be 4 degrees you can nock off quite easly

waterflow should not be a problem but if you add more waterblocks you will need to have no crimping

your right with ambient of 20c you should be getting about 5-6c cooler thin that

the only other thing you could do is add a nother radiator/fan and nock a nother couple of degrees off

[Voodoo Rufus]

Do you think by doing these mods it could make my proc able to run at 1733? 13x the next multiplier up for my. I can't take the bus speed up above 137 because of the USB issue with the mobo.

[Warlord2]

you can never say for sure how far you can overclock a cpu but it is posible to go to 1.7+ watercooling

I would do just the 5% water wetter and 95% distilled and see where that gets you and thin get some AS2 on there

those are easy mods and I would try those first

btw

see if it will post at 1.7 with 1.85v and see what your temps are like

[FrozenInHI]

looking at that setup, i see no real reservoir, if that pipe tee is your "reservoir" then you're not really gonna get great performance, as the water never really cools down much, it's moving in a tight circle really fast. you might try making a bigger reservoir and see how that helps, more water makes more cooling effort and thermal absorbsion. and you don't want the antifreeze, ur gonna ruin that awesome pump that way. use a 90/10 water to water wetter ratio, and don't use dyes, they cause corrosion in aluminum blocks, which leads to clogs, which leads to reduced flow, etc......

[Voodoo Rufus]

My whole system I think takes a little less than a quart of fluid. Why would the antifreeze wreck the pump? It's not going to corrode. It may be a little more work for the pump but the same goes for a lot of water head. Is there any reservoir that might fit in the space of 4 extra 3.5" drive bays? I really don't have much room left after that T-joint I put in there.

[Voodoo Rufus]

I took the voltage up to 1.85V and 13x133 mult., but no go. At 1733 it wouldn't even post. I think the limit of my proc is around 1650, since when I took the bus up to 12x137, it would post but wouldn't get into windows. I was hoping for a little more out of the proc with H2O and the stepping. Oh well.

[FrozenInHI]

My whole system I think takes a little less than a quart of fluid. Why would the antifreeze wreck the pump? It's not going to corrode. It may be a little more work for the pump but the same goes for a lot of water head. Is there any reservoir that might fit in the space of 4 extra 3.5" drive bays? I really don't have much room left after that T-joint I put in there.



If you go over to overclockershideout, they have a reservoir there that will fit in there, it fits into my lian li pc61's lower hdd holder with extra room to spare, pumps 145gpm, holds just over a liter of liquid, and is anodized with a fill port that needs a rubber gasket to seal properly, mine came with it, but it fell off somewhere along the line and i didn't notice! Doh! only small leakage though, just light weeping from the top of it, and it's not over any electrical parts. That's what id do, i'd get a reservoir so that you could at least have a little liquid running around there.

antifreeze is ethylene glycol, this solution is mildly corrosive to metals-hence a car's radiator getting corroded inside, water pump burning out, etc... Imagine what it's doing to that little plastic pump with the nylon or rubber impeller? I think you might get the picture now, but in case you don't, it will rot the impeller, and it will eat away at the metal over time. plus it's actually inhibiting your cooling efforts, it's not made for a short circuit coolant loop, it's made for longer travel time and transfer area, two of which none of us really have on any of our watercooling setups, if you think about a car engine, which is what this stuff primarily cools these days, it travels a long way and at a much slower rate compared to the size of the system. Ever felt how sticky your hands are after you handle it? that's a deposit of glycol, similar to sugar, it is supposed to coat the interior surfaces of the cooling system in a car (metal and heavy duty rubber, not poly/nylon) to prevent premature fungus growth.

[Warlord2]

looking at that setup, i see no real reservoir, if that pipe tee is your "reservoir" then you're not really gonna get great performance, as the water never really cools down much, it's moving in a tight circle really fast. you might try making a bigger reservoir and see how that helps, more water makes more cooling effort and thermal absorbsion. and you don't want the antifreeze, ur gonna ruin that awesome pump that way. use a 90/10 water to water wetter ratio, and don't use dyes, they cause corrosion in aluminum blocks, which leads to clogs, which leads to reduced flow, etc......

WRONG

you can have as little of water as you want and it will run at the same temps as with 2 gallons of water just will take longer to get to those times.

almost ALL the heat is disapated at the radiator
I dont think you will get to 1.7 unless you do a voltage mod and thin maybe you can get there
I would get a new mobo and get your fsb up

[Voodoo Rufus]

I believe the impeller in the pumpo is a sort of ceramic. It is very hard and definitely not metal. I don't know how the glycol will affect it but I will be changing out the fluid sometime next month.

I think that Warlord is right about the reservoir problem. The more water you have the longer it will take for it to reach a stable temperature. In an aluminum reservoir some heat will be lost through the walls, and therefore be dumped in my case, which then increases the case temp. Then the heat still goes through the radiator making it less effective.

I'll probably be stuck with this motherboard for a while until I can afford to sustain my computing hobby on my own. Student budget. But hey, Christmas is coming.

[Sonny]

Maybe the reason your not getting 1.7GHz+ is there is not enough voltage even at maximum settings. Everyone that have hit 1.7GHz with the Thunderbirds have done a voltmod. Only the XPs can do it w/o the vmod.

[Voodoo Rufus]

My Bios gives me 1.8 and 1.85V exactly when I choose those voltages. I'm not sure if the Bios is lying to me or the CPU is regulated well. I don't really want to do a voltage mod to my motherboard as I think that going over 1.95V is crazy for Athlons. The reccomended max voltage is 1.85 for those, right?

[FrozenInHI]

WRONG

you can have as little of water as you want and it will run at the same temps as with 2 gallons of water just will take longer to get to those times.


hehe, you really need to find some information on watercooling, or for that matter, simple physics. The only reason I know for sure that what i said is true--barring that it may not apply for this guy, but more than likely does--is that i've done it both ways! The radiator does remove most of the heat from the system, but if the coolant is moving through it so fast that it can't absorb much heat, then you're losing efficiency. Now, on to physics 101 which teaches us that the larger amount of thermal capacitiveness in a coolant system, the greater the ability to remove heat from the source, and the maximum surface area contact over time allows us to remove the maximum amount of entergy(heat) from the coolant. so, putting this all into practice, and laymans terms, if you have a short circuit of coolant running in a tight circle with no real time in the heat extraction portion of the loop, then you lose efficiency.

Now, in case that's not enough for you, I'll conduct an experiment just for further proof:

My cooling system consists of a bong w/ reservoir (approx 4gal) and an in-case reservoir (about a liter) basically operating on two coolant circuits "wired in series" with one another, my flow is about 155gph. I have the ability to remove the bong from the loop, and tweak the speed of the pump in the case to higher levels (225gph) so we'll do that to show ya how it happens, i'll be back with numbers that show you that the more water you have in the system, with the maximum amount of time in the "cooling zone" which is the radiator, the better you will be able to absorb/remove heat from the sytem. It's simple physics, but i'll make up an experiment to prove it for you.

[Sonny]

Originally posted by Voodoo Rufus
My Bios gives me 1.8 and 1.85V exactly when I choose those voltages. I'm not sure if the Bios is lying to me or the CPU is regulated well. I don't really want to do a voltage mod to my motherboard as I think that going over 1.95V is crazy for Athlons. The reccomended max voltage is 1.85 for those, right? Since your O/Cing & want to hit 1.7GHz you will need the extra juice. That's just how it goes. Faster=Voltage+=HEAT. You got watercooling so what are you worried about?

[Voodoo Rufus]

I'm worried about the life span of the processor and the possibility of it burning out over a couple years. I'd like it to last around 5 years.

[Warlord2]

hehe, you really need to find some information on watercooling, or for that matter, simple physics. The only reason I know for sure that what i said is true--barring that it may not apply for this guy, but more than likely does--is that i've done it both ways! The radiator does remove most of the heat from the system, but if the coolant is moving through it so fast that it can't absorb much heat, then you're losing efficiency. Now, on to physics 101 which teaches us that the larger amount of thermal capacitiveness in a coolant system, the greater the ability to remove heat from the source, and the maximum surface area contact over time allows us to remove the maximum amount of entergy(heat) from the coolant. so, putting this all into practice, and laymans terms, if you have a short circuit of coolant running in a tight circle with no real time in the heat extraction portion of the loop, then you lose efficiency.

I yes where your going with that but maybe about 4 months ago I believe overclockers did a test to see if it does mater and they came to a conclusion that for the most part it will not mater unless you have a huge amount of water in a open container
also I dont think it would really mater if you get the water any cooler thin it is already maybe 1-2c as I can leave one of may radiators of and temps will only go up 1-2c during load. I believe the only realy way to improve temps besides bong or supercooling is for people to come out with better waterblocks out of better meterial.

[Sonny]

Originally posted by Voodoo Rufus
I'm worried about the life span of the processor and the possibility of it burning out over a couple years. I'd like it to last around 5 years. They are going to be really cheap in the next couple of months but if you really want it to last then you gotta start conditioning yourself to just:rolleyes: 1.6GHz. I would be happy if I get 1.5GHz from my chip.