Pro/Forums - View Single Post

greenman100 · 08-16-2004, 09:36 PM

Watercooling Myths Exposed

Written by Greenman100 (Tim Elmore) edited by members at ProCooling.

Disclaimer:
The following is the work of a team of technically minded individuals who worked together to make the facts as correct as possible. However, one should not believe everything they read on the internet blindly, and if you have evidence that contradicts the information presented here, fell free to voice your thoughts.

Myths about Water and Flow

Myth: Water must slow down to fully absorb heat.
Reality: In a closed loop, a given water molecule actually spends the same amount of time in the radiator no matter how fast it is moving, as long as the water is indeed moving. If this is a difficult concept to understand, think about a racecar on a track. If the track is 1 mile (5280 ft) long and the car is driving at 60 mph, the car will spend about 1 second in a 100 ft stretch. Think of the 100 ft stretch as the radiator. Now, if the speed is doubled, the car only spends ½ a second in the 100ft section, but it passes through that same section twice a minute, so it spends a total of 1 second in the 100ft section per minute. If this is unclear, please post.

Myth: Order of components makes a significant impact on temps. (e.g. radiator must be before CPU)
Reality: Order of components makes a difference of less than .5C in most watercooling systems. The physics:

pump---->radiator---->CPU---->pump
pump---->CPU---->radiator---->pump

There is only one difference, and that is the position of the pump in the loop, be ir before it after the CPU.

Assuming the pump dumps about 50w of heat into the water, and flow rate is 1GPM (very reasonable):

Water has a thermal capacity of 4186J/Kg-C at 22C, and a density of about 1g/mL

With a flowrate of 1GPM, that’s ~3.75LitersPM.

3.75LPM/60 seconds= 0.0625Liters or Kilograms through the waterblocks per second.

4186*0.0625=261.625W/C

So that's 1C warmer for every 261W

But only 50 watts of heat are present, so: 50/261.625=.19C

So, there is a .19C difference in water temperature between the inlet and outlet of the pump. This does not mean the water is only .19C warmer than air, that is an entirely different calculation.

And that’s with 50 watts. If you’re running a smaller pump, like the D4, you're looking at 15w or so.

So, do what allows for the simplest tubing runs, tubing length/kinking will have a greater impact on temps.

Myth: Pump power consumption makes a significant impact on temps
Reality: It is difficult to know exactly how much heat a pump dumps into water, but a good rule of thumb is the following: An inline pump generally dumps 70-90% of its heat into the water. A submerged pump dumps 100% of its heat into the water. A dual 120mm radiator with a decent fan is good for about .03C/W, with a decent pair of 120mm fans. That is, the water temps will rise 1C for every 33 watts in the water. So, if your pump dumps 33 watts into the water, water temps will rise 1C. Therefore, the difference between a Mag3 at 40w and an Iwaki WMD-30 at 90 watts is fairly insignificant; 32w into water versus 72w, so about 1.1C. Note that the performance of a waterblock will improve with diminishing returns as the pressure increases.

Myth: A pump's flowrate is the only consideration to make when choosing a pump
Reality: A pump's maximum head pressure is just as, if not more important. Waterblocks are relatively restrictive, and many aquarium pumps are not made for that kind of restriction. In order to estimate one's flowrate, Calculate all pressure drops, then overlay on top of the pump's P/Q curve. In other words, it's not easy, but consider head pressure, too.

Myth: A t-line must be at the top of the system or water will leak out when you take the top off.
Reality: If the water were to leak out, air would have to replace the water. Since the entire rest of the system is sealed, air can’t get in to replace the water. Thus, no water will leak out of an opened t-line, unless there is a leak in the top of your loop

Myths About Materials/Coolants

Myth: Aluminum absorbs/dissipates heat faster than copper
Reality: All thermal properties of copper are better than aluminum. The only advantage to aluminum is that is lighter. So, if one were given a pound of copper and a pound of aluminum, you could make a better performing heatsink with the aluminum. However, this is not directly applicable to watercooling, as weight is very rarely a factor.

Myth: Gold is the best thermal conductor.
Reality: Diamond (6-50w/cm-k, dependent on purity) is actually best, but too cost prohibitive. Silver (4.173w/cm-k ) is second, and copper (3.937w/cm-k) is a close third. Gold’s thermal conductivity is 2.913w/cm-k. It’s primary use is its electrical conductivity, and resistance to corrosion. The reason for this is that gold is largely chemically inert, and thus will withstand many harsh environments. Graphite has many different thermal conductivity values, ranging from .6-10 w/cm-k, dependent on the formulation.
Sources: http://www.ai.mit.edu/people/tk/tks/tcon.html http://www.pyrographite.com/pyrogr.htm

Myth: Mercury would make a good coolant.
Reality: Mercury is way too environmentally unfriendly to be used in a non-industrial cooling system. In fact, engineers studied using mercury to cool a power plant. If mercury is too unsafe for workers working around a controlled nuclear reaction, it probably has no business in the home PC. Speaking of the home PC, if you ever did have a leak, liquid mercury is a very good electrical conductor, and would leave you with a dead motherboard rather quickly. Mercury vapor is poisonous.

Myth: Antifreeze improves the thermal properties of water
Reality: Antifreeze is actually makes the thermal properties of water worse, and is 18 times thicker at room temperature, resulting in more backpressure and slightly lower flow. However, in a multiple metal loop (like aluminum and copper) antifreeze or some other anticorrosive is needed to keep metals from galvanic ally reacting, resulting in corrosion. Antifreeze is good at raising the boiling point of water and lowering the freezing point. Unless you are boiling your CPU (not recommended) or running below ambient (see condensation myth) then you are not in need of those two characteristics.

Myth: Bleach is a good coolant additive.
Reality: Bleach is actually very corrosive, as indicated by both its datasheets and ionic makeup. The chemical formula for bleach is NaOCl, Which is ionic and in water becomes Na+ and OCl-. Ions in solution will lead to galvanic corrosion, as previously noted, as well as bleach's natural tendency to corrode.
Source: http://www.fact-index.com/s/so/sodium_hypochlorite.html

Myths About Temperatures and Measurement

Myth: On-board temeprature sensors are accurate.
Reality: You shouldn't even ask what someone else's temps are. This is one of the biggest problems i see in forums. YOUR temps are not even accurate relative to your own system, much less someone else’s. Let me elaborate. Just because your board reads 41C does not mean that if the temps drop 5C in real life the board will read 36C. Motherboards are not "absolutely accurate", nor are they "relatively" accurate. This is even worse with in-socket monitoring while watercooling. It is quite possible for the area around the socket to heat up to 50-60C, and influence the temperature measurements of the in-socket thermistor.

Myth: Digidoc5/ThermalTake/Vantec temperature monitors are accurate
Reality: While generally not as easily influenced by other factors, generally these sensors are off by more than +-3C. that means if your sensor is reading 3C high, and your friends is reading 3C low, you could be off by 6C on your comparison. More accurate sensors are available, the questions is whether that is important to you. The point is, do not believe onboard/digidoc temps. You shouldn't even tell other people what they are, they are so inaccurate. Tell them your overclock. There is very little inaccuracy in mhz measurements.

Myth: Condensation will form when watercooling.
Reality: The water in a watercooling cannot get any lower than the temperature of the air around the watercooling system/flowing through the radiator. In order to make th water colder than the surrounding air, energy must be applied either in the form of a compressor system or a thermoelectric device, both of which will cool the water, sometimes below ambient. The temperature point at which water condenses is a function of both temperature and humidity.

Myths About Waterblocks

Myth: If a block is shiny, it will conduct heat better.
Reality: In reality the flatness of the block is what counts. Flatness ensures maximal contact with the CPU and minimal TIM necessary to bridge the gap between. Shiny just looks nice. It is also of note that hand lapping a flat machined base will only make the base worse, if the base was fairly flat from the factory.

Myth: You can tell how a waterblock will perform just by looking at it.
Reality: The only way to know for sure is to run a highly calibrated test with a well thought out method.

Thanks to:
Cathar for the racetrack analogy, and several corrections and comments.
Crimedog for doing a search on OCers for me when I was unable to.
nikhsub1, Etacovda, Razor6, AngryAlpaca, and DrMemory for corrections and comments.
All of those not names previously who have given me an education in the field of watercooling in general.

Add any i missed, and correct me where I'm wrong

I'm thinking of submitting to pH and OCers

08-16-2004, 09:36 PM	#1
greenman100 Cooling Savant Join Date: Jun 2004 Location: Florida Posts: 414	Water Cooling Myths Watercooling Myths Exposed Written by Greenman100 (Tim Elmore) edited by members at ProCooling. Disclaimer: The following is the work of a team of technically minded individuals who worked together to make the facts as correct as possible. However, one should not believe everything they read on the internet blindly, and if you have evidence that contradicts the information presented here, fell free to voice your thoughts. Myths about Water and Flow Myth: Water must slow down to fully absorb heat. Reality: In a closed loop, a given water molecule actually spends the same amount of time in the radiator no matter how fast it is moving, as long as the water is indeed moving. If this is a difficult concept to understand, think about a racecar on a track. If the track is 1 mile (5280 ft) long and the car is driving at 60 mph, the car will spend about 1 second in a 100 ft stretch. Think of the 100 ft stretch as the radiator. Now, if the speed is doubled, the car only spends ½ a second in the 100ft section, but it passes through that same section twice a minute, so it spends a total of 1 second in the 100ft section per minute. If this is unclear, please post. Myth: Order of components makes a significant impact on temps. (e.g. radiator must be before CPU) Reality: Order of components makes a difference of less than .5C in most watercooling systems. The physics: pump---->radiator---->CPU---->pump pump---->CPU---->radiator---->pump There is only one difference, and that is the position of the pump in the loop, be ir before it after the CPU. Assuming the pump dumps about 50w of heat into the water, and flow rate is 1GPM (very reasonable): Water has a thermal capacity of 4186J/Kg-C at 22C, and a density of about 1g/mL With a flowrate of 1GPM, that’s ~3.75LitersPM. 3.75LPM/60 seconds= 0.0625Liters or Kilograms through the waterblocks per second. 41860.0625=261.625W/C So that's 1C warmer for every 261W But only 50 watts of heat are present, so: 50/261.625=.19C So, there is a .19C difference in water temperature between the inlet and outlet of the pump. This does not mean the water is only .19C warmer than air, that is an entirely different calculation. And that’s with 50 watts. If you’re running a smaller pump, like the D4, you're looking at 15w or so. So, do what allows for the simplest tubing runs, tubing length/kinking will have a greater impact on temps. Myth:* Pump power consumption makes a significant impact on temps Reality: It is difficult to know exactly how much heat a pump dumps into water, but a good rule of thumb is the following: An inline pump generally dumps 70-90% of its heat into the water. A submerged pump dumps 100% of its heat into the water. A dual 120mm radiator with a decent fan is good for about .03C/W, with a decent pair of 120mm fans. That is, the water temps will rise 1C for every 33 watts in the water. So, if your pump dumps 33 watts into the water, water temps will rise 1C. Therefore, the difference between a Mag3 at 40w and an Iwaki WMD-30 at 90 watts is fairly insignificant; 32w into water versus 72w, so about 1.1C. Note that the performance of a waterblock will improve with diminishing returns as the pressure increases. Myth: A pump's flowrate is the only consideration to make when choosing a pump Reality: A pump's maximum head pressure is just as, if not more important. Waterblocks are relatively restrictive, and many aquarium pumps are not made for that kind of restriction. In order to estimate one's flowrate, Calculate all pressure drops, then overlay on top of the pump's P/Q curve. In other words, it's not easy, but consider head pressure, too. Myth: A t-line must be at the top of the system or water will leak out when you take the top off. Reality: If the water were to leak out, air would have to replace the water. Since the entire rest of the system is sealed, air can’t get in to replace the water. Thus, no water will leak out of an opened t-line, unless there is a leak in the top of your loop Myths About Materials/Coolants Myth: Aluminum absorbs/dissipates heat faster than copper Reality: All thermal properties of copper are better than aluminum. The only advantage to aluminum is that is lighter. So, if one were given a pound of copper and a pound of aluminum, you could make a better performing heatsink with the aluminum. However, this is not directly applicable to watercooling, as weight is very rarely a factor. Myth: Gold is the best thermal conductor. Reality: Diamond (6-50w/cm-k, dependent on purity) is actually best, but too cost prohibitive. Silver (4.173w/cm-k ) is second, and copper (3.937w/cm-k) is a close third. Gold’s thermal conductivity is 2.913w/cm-k. It’s primary use is its electrical conductivity, and resistance to corrosion. The reason for this is that gold is largely chemically inert, and thus will withstand many harsh environments. Graphite has many different thermal conductivity values, ranging from .6-10 w/cm-k, dependent on the formulation. Sources: http://www.ai.mit.edu/people/tk/tks/tcon.html http://www.pyrographite.com/pyrogr.htm Myth: Mercury would make a good coolant. Reality: Mercury is way too environmentally unfriendly to be used in a non-industrial cooling system. In fact, engineers studied using mercury to cool a power plant. If mercury is too unsafe for workers working around a controlled nuclear reaction, it probably has no business in the home PC. Speaking of the home PC, if you ever did have a leak, liquid mercury is a very good electrical conductor, and would leave you with a dead motherboard rather quickly. Mercury vapor is poisonous. Myth: Antifreeze improves the thermal properties of water Reality: Antifreeze is actually makes the thermal properties of water worse, and is 18 times thicker at room temperature, resulting in more backpressure and slightly lower flow. However, in a multiple metal loop (like aluminum and copper) antifreeze or some other anticorrosive is needed to keep metals from galvanic ally reacting, resulting in corrosion. Antifreeze is good at raising the boiling point of water and lowering the freezing point. Unless you are boiling your CPU (not recommended) or running below ambient (see condensation myth) then you are not in need of those two characteristics. Myth: Bleach is a good coolant additive. Reality: Bleach is actually very corrosive, as indicated by both its datasheets and ionic makeup. The chemical formula for bleach is NaOCl, Which is ionic and in water becomes Na+ and OCl-. Ions in solution will lead to galvanic corrosion, as previously noted, as well as bleach's natural tendency to corrode. Source: http://www.fact-index.com/s/so/sodium_hypochlorite.html Myths About Temperatures and Measurement Myth: On-board temeprature sensors are accurate. Reality: You shouldn't even ask what someone else's temps are. This is one of the biggest problems i see in forums. YOUR temps are not even accurate relative to your own system, much less someone else’s. Let me elaborate. Just because your board reads 41C does not mean that if the temps drop 5C in real life the board will read 36C. Motherboards are not "absolutely accurate", nor are they "relatively" accurate. This is even worse with in-socket monitoring while watercooling. It is quite possible for the area around the socket to heat up to 50-60C, and influence the temperature measurements of the in-socket thermistor. Myth: Digidoc5/ThermalTake/Vantec temperature monitors are accurate Reality: While generally not as easily influenced by other factors, generally these sensors are off by more than +-3C. that means if your sensor is reading 3C high, and your friends is reading 3C low, you could be off by 6C on your comparison. More accurate sensors are available, the questions is whether that is important to you. The point is, do not believe onboard/digidoc temps. You shouldn't even tell other people what they are, they are so inaccurate. Tell them your overclock. There is very little inaccuracy in mhz measurements. Myth: Condensation will form when watercooling. Reality: The water in a watercooling cannot get any lower than the temperature of the air around the watercooling system/flowing through the radiator. In order to make th water colder than the surrounding air, energy must be applied either in the form of a compressor system or a thermoelectric device, both of which will cool the water, sometimes below ambient. The temperature point at which water condenses is a function of both temperature and humidity. Myths About Waterblocks Myth: If a block is shiny, it will conduct heat better. Reality: In reality the flatness of the block is what counts. Flatness ensures maximal contact with the CPU and minimal TIM necessary to bridge the gap between. Shiny just looks nice. It is also of note that hand lapping a flat machined base will only make the base worse, if the base was fairly flat from the factory. Myth: You can tell how a waterblock will perform just by looking at it. Reality: The only way to know for sure is to run a highly calibrated test with a well thought out method. Thanks to: Cathar for the racetrack analogy, and several corrections and comments. Crimedog for doing a search on OCers for me when I was unable to. nikhsub1, Etacovda, Razor6, AngryAlpaca, and DrMemory for corrections and comments. All of those not names previously who have given me an education in the field of watercooling in general. Add any i missed, and correct me where I'm wrong I'm thinking of submitting to pH and OCers __________________ "Employ your time in improving yourself by other men's writings, so that you shall gain easily what others have labored hard for." --Socrates "greenman100 = obnoxious ass hole"-gazorp Last edited by greenman100; 08-17-2004 at 09:13 PM.