Flowrate...

[garasaki]

90 degree bends need to be place before and after the heat xchanger and the waterblock to enhance trubulent flow, would you agree with that? Also turbulent flow will enhance heat transfer?

As mentioned, turbulence after the HX and WB wouldn't be of benefit. Also, I don't believe turbulence in the HX will improve heat transfer (much anyway). Turbulence improves heat transfer in waterblocks, but from my calculations, the CPU temp difference between turbulent flow and laminar flow is very small. I believe this is why increasing fluid flow thru the WB results in small temperature reductions (which is the subject of this post...), when compared to, say, increasing the airflow thru the rads. When trying to maximize your system, I have noticed that the HT between the waterblock and water is not where you should concentrate your efforts. Changes in that area seem to have negligable effects.

 

[parapapa?]

BillA -
Your scotch is on the shipping dock - you should see it sometime early next week.

I'm still trying to figure out why it takes so much water to cool a 70 watt cpu. Perhaps you can tell me where my error is?

70 watts x 1 calorie/4.2 watts is a little less than 17 calories.

17 calories will raise 17 grams of water 1 degree C every second.

Move 17 grams of water through the block each second and away go 70 watts.

17 grams/sec * 60 sec/minute is about 1000 grams/minute. That's a kilogram of water. Kilogram is 2.2 lbs so we're talking about .27 ~ .28 gallons /minute. If the wb is 100% efficient, it'll raise the water 1 degree and dump all the heat.

If the copper/water interface is 20% efficient, that is 1 calorie of every 5 gets through to the water, then the cpu will rise 5 degrees for every degree rise in the water. So is that where the problem is? The waterblock is failing to completely transfer the heat to the water and hence it heats up?

 

[vandersl]

Parapapa?,

You were coming close there at the end.

First, you are correct - even at low flow rates, the increase in temperature from inlet to outlet in the WB is very small. See my earlier post for a few examples.

The problem is not the mass of water required to remove the heat. The problem is with the efficiency of the heat transfer between the block and the water.

First, all heat generated by the CPU must transfer to the water, unless it it lost to other sources (air, MB, etc). To transfer the heat to the water, it must be 'pushed' into the water from the block. The only way to push heat is with a temperature difference. So, the block must be warmer than the water for the heat to flow.

How much warmer the block needs to be depends on the efficiency of the transfer to the water. If the water is not moving, all heat transfer would need to be through conduction. This is not very efficient with water (though better than air). To improve heat transfer, the idea is to move heat just a little way into the water next to the block, then move away that bit of water and replace it with another bit (simplfied example). This type of heat transfer is called convection (convey = move).

The faster you move the water, the more efficient this process becomes. However, since water next to a surface tends to 'stick' there without moving (friction et al), just moving water through the pipe slowly sometimes isn't enough. The water in the middle of the pipe might be moving, but the water velocity decreases as you come closer to the pipe walls, and eventually approaches zero - back to conduction again.

This is where turbulence comes in - if you can get fluid from the center of the channel to 'mix' with fluid toward the walls, you end up with more efficient heat removal. Turbulence increases heat transfer significantly over slower 'laminar' flow.

Turbulent flow occurs 'naturally' in a pipe when the fluid velocity exceeds a certain point, which is dependent on a lot of factors. Also, turbulence isn't an on/off thing - you can have more or less of it. Moving faster will result in more.

So, in short )) moving water through the block faster improves heat transfer between the block and the water, which reduces the temperature differential between the block and water required to move an amount of heat. It is not intended to reduce the temperature increase in the water as it travels through the block.

Note: as with everything, the improvement in heat transfer has diminishing returns.

Hope this helps.

 

[BillA]

This post deleted by its "derogatory and insulting" author so that it will not be (so) necessary
"to defend the little guys against the tyranny of the know-it-alls."

 

[vandersl]

all described quite improperly, of course
but I suspect you understand

Nope, not even close. Either you already received that Scotch parapapa? sent you already, or my reading comprehension skills have dropped off markedly towards the end of the day

Either way, have a good one

 

[BillA]

This post deleted by its "derogatory and insulting" author so that it will not be (so) necessary
"to defend the little guys against the tyranny of the know-it-alls."

 

[Sirpent]

Well, turbulence is complicated. But for ideal liquid and laminar flow, how heat flux in a system with a waterblockand and a radiator depends on the flow?

 

[GreenmanWD-40]

I'll answer these out of order.



Absolutely agree. Turbulent flow enhances heat transfer. Well documented, fits the theory. Got that one right.



Absolutely disagree. Especially with the 'after' the exchanger and waterblock, where introducing turbulence would have no real benefit.

You can get turbulent flow in a straight tube with smooth walls if the water flows fast enough (usually within the range of fluid velocities used in WC). This is the result of fluid viscosity and velocity differential between the fluid at the tube walls and the center of the channel. I hesitate to bring up 'fluid friction' again in this discussion, but there it is

Spikes, swirls, 'tubulators' and other flow channel designs to induce turbulence may not be needed in many cases (but they look cool, and sell more blocks).

I think the reason most blocks use inlets at 90 degrees to the CPU surface is that its hard to get clearance around the socket for the barbs if they are parallel to the CPU surface.

Injecting the coolant from above the block surface is a reasonable way to induce turbulence as well (a la the swiftech block), without restricting flow needlessly.

Wait a second though,

I ment to say turbulent flow inside the waterblock and the heat exchanger, laminar everywhere else. Something like 90 degree bends on the inlets and straight hose barbs on outlets. Right? I think a test was don't on overclockers, and it showed about a 2 degree difference, but don't know where right off hand.

Now, For every heat exchanger there is a limit on how much heat can be removed verses the flowrate? Is there any way possible to predict this number? Or is it more complicated?

BTW, SOMEONE I need conformation on the mounting hole on a P4 motherboard. What I have is 59.44mm and 76.2mm? Anyone?

GreenmanWD-40

 

[GreenmanWD-40]

I'll answer these out of order.



Absolutely agree. Turbulent flow enhances heat transfer. Well documented, fits the theory. Got that one right.



Absolutely disagree. Especially with the 'after' the exchanger and waterblock, where introducing turbulence would have no real benefit.

You can get turbulent flow in a straight tube with smooth walls if the water flows fast enough (usually within the range of fluid velocities used in WC). This is the result of fluid viscosity and velocity differential between the fluid at the tube walls and the center of the channel. I hesitate to bring up 'fluid friction' again in this discussion, but there it is

Spikes, swirls, 'tubulators' and other flow channel designs to induce turbulence may not be needed in many cases (but they look cool, and sell more blocks).

I think the reason most blocks use inlets at 90 degrees to the CPU surface is that its hard to get clearance around the socket for the barbs if they are parallel to the CPU surface.

Injecting the coolant from above the block surface is a reasonable way to induce turbulence as well (a la the swiftech block), without restricting flow needlessly.

What if you used copper transport tubes and wanted them to take some heat?? Just joking.

GreenmanWD-40

 

[GreenmanWD-40]

Parapapa?,

You were coming close there at the end.

First, you are correct - even at low flow rates, the increase in temperature from inlet to outlet in the WB is very small. See my earlier post for a few examples.

The problem is not the mass of water required to remove the heat. The problem is with the efficiency of the heat transfer between the block and the water.

First, all heat generated by the CPU must transfer to the water, unless it it lost to other sources (air, MB, etc). To transfer the heat to the water, it must be 'pushed' into the water from the block. The only way to push heat is with a temperature difference. So, the block must be warmer than the water for the heat to flow.

How much warmer the block needs to be depends on the efficiency of the transfer to the water. If the water is not moving, all heat transfer would need to be through conduction. This is not very efficient with water (though better than air). To improve heat transfer, the idea is to move heat just a little way into the water next to the block, then move away that bit of water and replace it with another bit (simplfied example). This type of heat transfer is called convection (convey = move).

The faster you move the water, the more efficient this process becomes. However, since water next to a surface tends to 'stick' there without moving (friction et al), just moving water through the pipe slowly sometimes isn't enough. The water in the middle of the pipe might be moving, but the water velocity decreases as you come closer to the pipe walls, and eventually approaches zero - back to conduction again.

This is where turbulence comes in - if you can get fluid from the center of the channel to 'mix' with fluid toward the walls, you end up with more efficient heat removal. Turbulence increases heat transfer significantly over slower 'laminar' flow.

Turbulent flow occurs 'naturally' in a pipe when the fluid velocity exceeds a certain point, which is dependent on a lot of factors. Also, turbulence isn't an on/off thing - you can have more or less of it. Moving faster will result in more.

So, in short )) moving water through the block faster improves heat transfer between the block and the water, which reduces the temperature differential between the block and water required to move an amount of heat. It is not intended to reduce the temperature increase in the water as it travels through the block.

Note: as with everything, the improvement in heat transfer has diminishing returns.

Hope this helps.

Isn't this where decreasing the surface tension comes in? for example Water Wetter? Wouldn't that help?

 

[h2sammo]

Greenman

Isn't this where decreasing the surface tension comes in? for example Water Wetter? Wouldn't that help?

Decreasing surface tension is obviously beneficial when it comes to reducing friction. However, i wonder how much water wetter would accomplosh this. Water wetter is some sort of an alcohol. When it disrupts the H-bonds of water molecules, it breaks (lowers) the surface tension of water.

When this is accomplished, H2O will take a hit in terms of polarity and will "cling" less to the tube walls (which are also polar).

However, this is a 2 edged sword. Decreasing surface tension is very well be beneficial for reducing friction, but it decreases the heat capacity of H2O. Water's attribute is that it has its molecules tightly held together by H-bonds (very strong intermolecular bonds). When these interconnects are disturbed, water will behave poorly when absorbing heat loads.

A very good solution to getting rid of much of the friction between water and tubes (or anything else for that matter) is coating the vinyl tubes with a few layers of polymers (PSS/PDAD, etc), and coat the last polymer layer with some nonpolar "sealant" molecule layer. Then, water will flow with much less friction thourgh the tubing (close to frictionless), the only friction would be the water molecules bumping into the last nonpolar molecule coating (which will not be flat at molecular level, like nothing else would be flat at molecular level). However, there will be no attarctin between water and that layer, just them touching each other.

The very well known "meniscus" shape of the water flowing though a tube will be replaced by a flat shaped end of the water column. This method is used in making different kinds of chromatographic columns elute the samples faster.

Hope this helped.

 

[GreenmanWD-40]

Greenman




Decreasing surface tension is obviously beneficial when it comes to reducing friction. However, i wonder how much water wetter would accomplosh this. Water wetter is some sort of an alcohol. When it disrupts the H-bonds of water molecules, it breaks (lowers) the surface tension of water.

When this is accomplished, H2O will take a hit in terms of polarity and will "cling" less to the tube walls (which are also polar).

However, this is a 2 edged sword. Decreasing surface tension is very well be beneficial for reducing friction, but it decreases the heat capacity of H2O. Water's attribute is that it has its molecules tightly held together by H-bonds (very strong intermolecular bonds). When these interconnects are disturbed, water will behave poorly when absorbing heat loads.

A very good solution to getting rid of much of the friction between water and tubes (or anything else for that matter) is coating the vinyl tubes with a few layers of polymers (PSS/PDAD, etc), and coat the last polymer layer with some nonpolar "sealant" molecule layer. Then, water will flow with much less friction thourgh the tubing (close to frictionless), the only friction would be the water molecules bumping into the last nonpolar molecule coating (which will not be flat at molecular level, like nothing else would be flat at molecular level). However, there will be no attarctin between water and that layer, just them touching each other.

The very well known "meniscus" shape of the water flowing though a tube will be replaced by a flat shaped end of the water column. This method is used in making different kinds of chromatographic columns elute the samples faster.

Hope this helped.

Yes Sure did.

 

[h2sammo]

simpler idea

Well, consider my last post just an intelectual exercise. A much more practical way to deal with H2O's affinity for polar plastics (vinyl) is to use a nonpolar tube, like polyethylene.
Polyethylene is a nonpolar polymer, and there will be no dipole affinity between water and the tube walls. I wonder how much that would increase flow, as compared to a vinyl tube for example.

 

[GreenmanWD-40]

Re: simpler idea

Well, consider my last post just an intelectual exercise. A much more practical way to deal with H2O's affinity for polar plastics (vinyl) is to use a nonpolar tube, like polyethylene.
Polyethylene is a nonpolar polymer, and there will be no dipole affinity between water and the tube walls. I wonder how much that would increase flow, as compared to a vinyl tube for example.

What about Tygon?

 

[h2sammo]

Tygon

Yeah, nice site. Tygon

There are several tubes they say are "non-wetting". I suppose that means they faciitate flow of water containing fluids. Tygon would be a good place to get tubing from. I wonder if they sell anything but bulk.



Just an aside... Are you thinking what I'm thinking?...no fans, no radiator...no noise, not cold enough to worry about condensation.

 

[garasaki]

Uh, that walmart most likely does have fans and a rad...and my guess is its pretty loud! Would be worth a shot though...if you have 140 bucks laying around

 

[GreenmanWD-40]

Re: Tygon

Yeah, nice site. Tygon

There are several tubes they say are "non-wetting". I suppose that means they faciitate flow of water containing fluids. Tygon would be a good place to get tubing from. I wonder if they sell anything but bulk.



Just an aside... Are you thinking what I'm thinking?...no fans, no radiator...no noise, not cold enough to worry about condensation.

Good one, LOL

A couple of months ago I bought some, from mcmaster. 1/2 inch and 10 feet long. They sell it by the foot.

Could you drink from the water cooler if it cooled the cpu, but would you want to??

lol

 

[h2sammo]

Tygon response on nonpolar plastic tubes

Direct quote:

Hi Claudiu!

Our Tygon(R) 2075 and Tygon(R) 2275 would be nonpolar formulations and have
extremely low water absorption (hydrophobic = antagonistic to water) since
they are flexible, non-plasticized PVC.

Our proprietary Tygon 2000 series products are more flexible than standard
non-plasticized PVC products and can be slid onto barbed plastic or metal
fittings more easily.

The Tygon B-44-3 and Tygon R-3603 formulations are polar, since they contain
plasticizers and slightly absorb water.

You may contact the following authorized Tygon tubing distributor in your
area:

Southern Industrial Supply / St Petersburg, FL
Phone: 727-323-1300
Fax: 727-323-6905
Contact: Dean Burnette

Your interest in our products is appreciated!

Regards,
Linda

End quote.


hope this specific information helps

 

[Robbie]

Just before I started watercooling I noticed a lot people always said that more water flow means better cooling.

But if water moves too quickly won't it not have enough time to cool?


I cut my hose lenghts in half to fit the setup into my case. Just to check what effect this would have I left everything outside so the only difference was the hose lenght. My temperature went up!

I'm guessing the smaller hose lenght reduced the time it needed to get from one part of the setup to another. Basically, the speed increased.

I think it's because the water spent a longer amount of time AWAY from the cpu.
Rob

 

[PYROMANIAC]

whats a good standard GPH for just a cpu block?