Slashdot is powered by your submissions, so send in your scoop

 


Forgot your password?
Close
typodupeerror
×
Hardware Technology

New 3D CPU Water Cooling Method 239

Posted by timothy from the mmm-water-cooling dept.
captain igor writes "According to this story on Wired News, a new company launched by researchers from Stanford has come up with a way to layer a silicon network of tiny tubes on top of a microprocessor. The system then uses a solid-state motor (no moving parts!) to pipe cold water through the silicon network. According to the article, this system can handle 1000 watts (yes, a kilowatt) per square centimeter."
This discussion has been archived. No new comments can be posted.

New 3D CPU Water Cooling Method More

New 3D CPU Water Cooling Method

Comments Filter:

  • Oh Great.... (Score:5, Funny)

    by GeneralEmergency ( 240687 ) on Wednesday October 08, 2003 @02:13PM (#7165254) Journal


    Now my PDA can wee-wee in my pocket.

  • Am I the only one? (Score:2, Insightful)

    by PrintError ( 708568 )
    I'm still wary about pumping water into my computer. What if one of those microthin pipes were to burst? Then you'd get a microsized stream of water shorting out your not-so-micro-priced processor. I'll stick to windtunnels and heatsinks... maybe a heatpipe or two.
    • I am sure they could use a nice nonconductive liquid like alcohol, though it would reduce its ability to handle watage.
      • It can't be nonconductive; the pump works on electrolytes. I imagine that it is basically a small version of the "caterpillar drive" used by red october in the book of the same name.

        It's an ion drive, but for liquid. For that to work, you need ions, and thereby conductivity.

        All this from a quick skim of the article, so add salt (:-)) to taste.

    • So use transformer oil, like used in big power transformers. Or something similar. There must be plenty of chemicals out there with high levels of thermal conductivity but are fairly inert.

      We use isopropyl alcohol here at work to clean our circuit boards, and one day one of my colleagues accidentally powered up his board while it was still in the bath being washed. It powered up fine, with no difference to being in open air. The conclusion is that there are liquids that can be used for this purpose, withou

  • Pump with no moving parts? (Score:4, Interesting)

    by Keck ( 7446 ) on Wednesday October 08, 2003 @02:15PM (#7165263) Homepage
    Forget about the cooling, tell me more about that pump! /me googles electrokinesis ..

    apparantly it uses osmotic pressure to drive it, how cool is that?
  • Reminds me of "The Diamond Age" by Neal Stephenson. There was a bit in there where a future super computer was hooked up to a pipe delivering ice. As the batch runs, ice is sucked through the computer at amazing speeds, passed straight through the CPU. and boiling water is delivered (in large quantites) at the bottom. Not what I would like for me home PC :-) Still, should make water cooled systems much more efficient.

  • How long... (Score:4, Funny)

    by YuppieScum ( 1096 ) on Wednesday October 08, 2003 @02:19PM (#7165283) Journal
    ...before someone at IBM notices their use of 'MicroChannel'?

  • It's not a motor (Score:3, Insightful)

    by signe ( 64498 ) on Wednesday October 08, 2003 @02:19PM (#7165284) Homepage

    It's not a solid state motor. I dare say, there's no such thing. By definition, a motor turns, therefore it has moving parts. In fact, the word "motor" appears nowhere in the article, so I'm not sure where the submitter dreamt that up.

    It's a solid state pump that moves an electrolyte through it using osmotic pressure.

    -Todd

    • Re:It's not a motor (Score:4, Informative)

      by CaseyB ( 1105 ) on Wednesday October 08, 2003 @02:31PM (#7165367)
      By definition, a motor turns, therefore it has moving parts.

      No. A motor is by definition "one that imparts motion". This device certainly qualifies.

      • Would you call this [uwex.edu] a motor than?
    • "By definition, a motor turns"

      Linear induction motors don't turn.

  • Quality Journalism: (Score:3, Insightful)

    by pavon ( 30274 ) on Wednesday October 08, 2003 @02:23PM (#7165308)
    This is pretty cool, and I thank the poster and slashdot editors for putting up the story. I just thought that this was funny:

    Apple, Intel, DARPA and Cooligy did not respond to requests for comment.

    Well that includes just about everyone mentioned in the article, so exactly where did the information come from? I see, I'm reading a posting about an article about another article about information gleaned from a website. Oh, well at least they told me :)
  • If we put in a miniature steam turbine we can generate power to charge laptop batteries and perhaps add a steam whistle to the sound system. Actually, since I can't get any sound out of my laptop a steam whistle would be a nice addition!

    • Re:Power from waste heat (Score:5, Interesting)

      by Baron_Yam ( 643147 ) on Wednesday October 08, 2003 @02:34PM (#7165385)

      Actually, I wonder what the theoretical limit is on converting waste heat back into electricity in a laptop... would it be worth the extra weight? Even if it's NOT worth the extra weight, it might be fun to do it just because it can be done.

      Off the top of my head, though, I'm not aware of any laptop-scale device for generating power from a heat source.

      • The laws of thermodynamics say that you need a large temperature difference to efficiently generate power from heat. That's why power plants use steam or gasses heated to hundreds or thousands of degrees to run their turbines.

        A semiconductor chip won't work if the temperature gets too much above 100 C. Therefore, by definition, you would be stuck with a very low power generation efficiency.

        Your efficiency is limited by trying to extract work between the (say) 100 C CPU and the 25 C ambient environment.

      • Maximuim Theoretical Efficiency, due to the second law of thermodynamics, is (Thot-Tcold)/Thot, where the temperatures are measured in Kelvins. So for boiling exit water (100C = 373K) and room temp (20C = 293K), so about 20%

        The engineering problem is getting Thot to be the microprocessor temperature, not the exiting cooling water temperature - this would give you much better efficiency, but at the possible cost of cooling power.

      • Peltier's cell. (can exchange voltage into temp difference and opposite) But you'd need a decent cooling device (radiator? Water cooler?) on the other end. You can't generate energy from heat itself. You can generate it from temperature difference (and thus heat flow) though. Plus this would impair heat dissipation, rather undesired with processors. You surely wouldn't be able to build a perpetuum mobile like cooling system (powered only by CPU heat, the hotter the CPU, the cooler theother side because wate
        • I've always wondered how these work to generate electricity from heat flow...

          I assume a tiny amount of flowing heat is absorbed and converted into electricity? If not, then where does the energy come from?

          What happens when you sandwich these cells? Wouldn't you get electricity from each cell, until the heat is completely absorbed?

          And if that is the case, then wouldn't a sufficently well designed array of these allow you to reabsorb nearly all waste heat and recycle it back into electricity to be re-use
          • No! It's heat TRANSFER that gives the electricity. So if you put the cell on top of dry ice block and put a pot of boiling water on top of it, you have 100C on one side and some -30 on the other, and the cell provides some electricity because some heat flows from the pot to the block, water cools, ice evaporates. If you stack two of these, you get some 30C between them. Combined they will provide stronger current, but because the temperature difference is lower, the voltage will be lower. (or something like
    • I'd rather have a mini pipe organ. With a USB plug!

  • Cool Suit (Score:5, Interesting)

    by binaryDigit ( 557647 ) on Wednesday October 08, 2003 @02:24PM (#7165314)
    /. is acting weird, so someone will probably have posted a similar idea, but ...

    If you could figure out a way to sew this into material, then you could have some really "cool" (literally) clothing. I'm sure people like the Army would be very interested in a suit or body armour that offered effective cooling, esp in the desert where a system with a motor could be undesireable. I know it would be sweet to get a set of motorcycle leathers with something like this built in (those Texas summers get a bit toasty).

    • Re:Cool Suit (Score:3, Interesting)

      by blibbleblobble ( 526872 )
      "If you could figure out a way to sew this into material, then you could have some really "cool" (literally) clothing."

      It was demonstrated on british television a few years ago (more than 5 years), being used by firefighters, who could carry a refrigeration unit on their back, and walk through flaming buildings without getting hot (tubes built into the clothing)

    • I think a more immediate use would be in Internal Combustion engines. A whole lot of heat is produced by engine (almost 40 percent of total fuel consumption) , this method would be ideal in those situations if the 1000 watts/cm^2 is correct. Engines would become a whole lot smaller and efficient.
      • This does not change the amount of heat produced, but provides a more efficient way to move the heat away from the heat source. As such, it would have the possibility to lower the size of the radiator, not the engine. I haven't RTFA, so I can't say if such a system would be sensible for an auto cooling system.
        • Ofcourse this would affect the radiator mainly.This will in no way by itself make engines better.But I was thinking that this would allow you to use newer kind of materials for the engine, allow you to put in larger/smaller cylinders, allow you to make the cylinders in odd shapes etc. etc.
          This indirectly would make the engines run better and smaller. This technology was developed for Chips but I think this should be portable to car engines too (the only difference from the current water cooling tech i can
          • -(maybe someone could change the cooling fluid?).

            Suzuki GSX-1100R/750R, circa 1986 through ?? a few years ago. Used the same lubricating oil that circulates through the engine as a coolant (had a massive oil cooler, ran the oil all through the engine and actually had streams of oil directed on the hot spots.) Theoretically the engine was more efficient because it could run a little hotter than a water cooled engine without concern for a boil-over or friction induced ultra hot spots.

            Pretty much the same
    • They already make cool suits for race car drivers. They pump cold water through tubes in the suit. This technology could miniaturize the pump but would add greatly to the complexity of the suit if the pumps are built in. Plus you'll still need a water chiller and power for it. Not much of a problem if you have power and minimal cargo room like on a motorcycle. More of a problem if you're on foot.
      • Consider evaporative cooling, particularly on a motorcycle rider because the wearer is moving fairly fast and is exposed to the constant stream of wind.

        At that point all you need is a water reserve and a simple mechanism to pump is slowly onto the evaporative material, insuring that it gets distributed to all the places needing cooling.
        • Evaporative cooling won't work in humid weather, but it's great for dry heat. They make neck bandanas filled will water absorbent gel. They'll hold enough water to keep your head cool with evaporative cooling for a pretty long time. Another trick if you have a vented nylon jacket is to wear a wet t-shirt under your gear.
    • Nike makes something like this for football players. They wear a special vest. When they come to the sidelines between series they are hooked up to water tubes which ciruclates cool water across their torso.
  • Cooling with water reminds me of primitive automobiles and their liquid cooling . Most of the energy of an auto is wasted HEAT.

    Something new needed in chip technology. Moores law is about to END.

    How about opical-electronic computer chips. Lets reduce heat ! These chips already exist ,just google(optielectronics) but most are used for high speed networking. Optical buses are achievable in my opinion right now. Logic gates are another story.

  • stove top boiling water experiment (Score:3, Insightful)

    by goombah99 ( 560566 ) on Wednesday October 08, 2003 @02:26PM (#7165325)
    lets do some order of magnitude, spheical cow type estimates using simple everyday experience in the kitchen.

    A typical stove top burner is order of magnitude 1000 watts spread out over around 500 sq cm: so were talking order of magnitude less than 10 watts per sq cm.

    if I take a teaspoon of water an put it on a sq cm of stove top and it boils in far less than a second. really almost instantly so its probably like less than a tenth of a second.

    Thus if this thing is going to not explode, the flow rate required to avoid boiling at 1000 watts /sq cm is going to be on the order of hundreds to thousands of teaspoons per second.

    If I take a tiny swizzel straw and try to suck through it I cannot suck 1000 teaspoos per second. Since my ability to suck is probably within an order of magnitude of the cavitation pressure for atmospheric pressure water a pump trying to flow this stuff through an equally small crossecttion may not be able to sustain such a flow rate. And any on-chip pump is probably going to have a simmilar crossection for its fluid intake port. (off -chip is another matter)

    unless this thing is actually flowing the water based on the steam pressure itself, I'm skeptical that this can meet the claimed specs.

    but I assume these people aren't fools. Perhaps the science reporter slipped a few digits.

    • Sounds to me like the system is closed, so it may have high enough pressure and low enough temperature to remain liquid, or very low quality steam. All conjecture of course, but if the coolant changed to steam I would fully expect the device to explode.

    • Re:stove top boiling water experiment (Score:5, Informative)

      by Thagg ( 9904 ) <thadbeier@gmail.com> on Wednesday October 08, 2003 @02:49PM (#7165497) Journal
      Say water goes in at 30 degrees C and comes out at 50 degrees C. According to the spectacular Google calculator [google.com], 1000 watts is 239 calories per second, and it takes 1 calorie to increase the temperature of 1 cc of water 1 degree C, so you'd have to move 239/20 or about 11 cubic centimeters of water through the cooler every second assuming a delta-v of 20 degrees C. Doesn't sound unattainable.

      thad

      • I concur (Score:5, Insightful)

        by goombah99 ( 560566 ) on Wednesday October 08, 2003 @03:01PM (#7165598)
        Yep you're answer is better than my initial post. I agree with your math. (4.8Joules/gm-C)

        my post erred because the reason the water boils is not the heat flux but the stored heat in the stove top coil. The transient delivery of this stored heat vastly exceeds the rate of power delivered to the stove and thus the water boils fast. but this would not be sustained.

        I withdraw my original answer.

    • Re:stove top boiling water experiment (Score:5, Informative)

      by rcw-home ( 122017 ) on Wednesday October 08, 2003 @03:01PM (#7165601)
      This is a lot easier if we stick to metric units.

      The factor they always leave out is how much of a temperature rise one can tolerate at the heat sink. Let's assume that the incoming water will be no higher than 40C and the CPU can become no hotter than 60C - that's 20C rise.

      1 kilowatt is 1000 joules per second, or 238 gram calories per second. Conveniently, a gram calorie is the energy needed to raise a gram of water one degree celcius. For water, one gram is also one milliliter. So, a single gram of water will be raised 238 degrees C in one second. We don't want it to be raised more than 20C, so we need to exchange water at a rate of 238/20 = 11.9 mL/sec.

      Heat sinks aren't perfect - the outgoing water will always be colder than the CPU. Let's pretend that this sink is 50% efficient (the CPU rises to a temperature, relative to the incoming water, of twice that of the outgoing water). Ergo, we need 23.8 mL/sec.

      How is this a problem?

      • 1 kilowatt is 1000 joules per second, or 238 gram calories per second. Conveniently, a gram calorie is the energy needed to raise a gram of water one degree celcius. For water, one gram is also one milliliter. So, a single gram of water will be raised 238 degrees C in one second. We don't want it to be raised more than 20C, so we need to exchange water at a rate of 238/20 = 11.9 mL/sec.

        You're ignoring the convective heat transfer coefficient for water.

        The heat transfer rate is a product of the temparatu

    • OK, let's see here, first off, most stove top burners are closer to 2000 to 3000 Watts, not 1000.

      http://www.consumersearch.com/www/kitchen/range s /c omparisonchart.html

      But regardless, your analogy is much too much work, let's just figure out how much water you can boil each second with 1KW of power.

      This page:

      http://www.infinitepower.org/calc_watts.htm

      Says you can evaporate 0.0001172 gallons each second. According to Google, this works out to:

      http://www.google.ca/search?q=0.0001172+gallon&i e= UTF
    • That is a pretty good off the cuff Spherical Cow analogy. I would suspect that the network of silicon tubes uses something akin to counter current flow [passporttoknowledge.com] to achive higher rates of cooling.

      Also, with a small network of tubes the relative surface area of the water to the heat would be higher than a teaspoon on a stove. While this probably means that the water would vaporize more quickly this might not be a bad thing. There was (is?) a company that produced PC cases that contained a compressor and supercooled l

    • Heat Transfer (Score:3, Informative)

      by nuggz ( 69912 )
      They claim the potential to move 1kW through this surface, but they don't mention the conditions.

      If you make it really cold on one side, and really hot on the other this could happen by itself.

      Think of your cooler, it doens't leak heat much on a cold day, but on a hot day it will warm up much quicker.
      Change your temperature difference, the heat flow rate will change.

      On your boiling water, take steady state water evaporation vs energy input. Your 1kW Burner isn't going to be boiling thousands of teaspoons

  • Next step: in the processor (Score:5, Interesting)

    by hcetSJ ( 672210 ) on Wednesday October 08, 2003 @02:28PM (#7165349)
    I wonder how long until we have nanotubes running all through the processor. There's a professor at my school [cornell.edu] doing research on 3-D photolithography, which would allow much more complex structures to be built out of crystalline silicon. This sounds like a good application.

  • I'm sure it will be sealed with non-water coolant. (Score:3, Insightful)

    by antimith ( 683310 ) on Wednesday October 08, 2003 @02:32PM (#7165375)
    Considering the consequences of a little algea or whatever in tubes so small, I'm sure they'll provide the coolant(likely non-water) and perhaps even an on board Closed coolant system.

    Considering the size of 3rd party coolants shown on site's like Tweak3d.net [tweak3d.net] I wouldn't be suprised at all if the setups didn't look like some of ThermalTakes [thermaltake.com] larger models.
    If most of the tubing is kept in the in-die, and the motor is solid state (not sure what size we're talking about) then I'd envision something that would leak about as mutch as an air cooled system. hehe.

  • Don't you mean... (Score:3, Funny)

    by product byproduct ( 628318 ) on Wednesday October 08, 2003 @02:33PM (#7165382)
    1000 watts (yes, 0.9765625 kibiwatts) ?

  • Seymour would approve of the focus (Score:3, Interesting)

    by Baldrson ( 78598 ) on Wednesday October 08, 2003 @02:37PM (#7165407) Homepage Journal
    When the first supercomputer was built on Seymour Cray's farm by 34 guys with 1 PhD among them (a junior programmer) [geocities.com] the key technology turned out to have been refrigeration devised by a kid from the Amana Colonies. Seymour spent his career fighting heat as he strove to get path-lengths between components smaller thereby driving up power density as the cube of his system's scale and speed. He most certainly would have approved of the focus if not the approach taken by the Stanford team.
  • How much force would it take to burst a pipe? I would think that would be instant death for your cpu... imagine THAT for a blue screen of death -- "Sorry, your CPU has drowned. Go buy another one!"

  • Screw 3D CPU's (Score:2, Insightful)

    by SoTuA ( 683507 )
    Put it in the main CPU, not just the GPU. That way we can get rid of the screaming banshees/cooling fans in our towers. (So you can leave your favorite p2p running overnight without the whirrrrrrrring)
  • Why not do a more radical case-mod? Build an aquarium and put the motherboard at the bottom. Remove all fans and put good heatsinks on. Have a longer SATA cable to the HD (and other peripherals) and put it outside of the aquarium. Fill the aquarium with a suitable clear liquid that don't affect the MB or electronic signals. Glycol?

    Now I would have a passive glycol-cooled computer!

    I wonder if this kind of cooling would work. I'm not sure how to calculate such things... Which is the best cooling fluid?

    Ah y
    • Ah yes, and if anyone know what the colored stuff in lava-lamps is, please tell me. Pouring some of that stuff in could make the case-mod very interesting...

      wax?
  • This is an old technology... you can create a tremendous amount of attractive-looking vapor by emitting a bunch of hot air in the right direction. Although there's a short-lived burst of heat and light (and sound and fury), pretty soon, everybody is cool to the technology.
  • Wow! Not only can I run Java on the
    processor - it can brew it too!

    I can't wait to get my NeverEmpty
    coffee cup on ThinkGeek!

  • While I think this is a great step forward in terms of cooling, I'm afraid that if this gets widely implemented, it'll end up making processors less efficient in terms of heat management.

    A modern processor creates around 60-70 watts of heat. Heat management is currently a problem in many systems, and the fact that the processors convert so much power into heat creates a need for larger power supplies. Now, if this was to be widely implemented, the incentive for reducing the heat waste in the processors wou
  • Oddly enough, I received the following seminar announcement this morning (hope Dr. Santiago doesn't mind):

    Electrokinetic Microfluidic Systems
    Prof. Juan G. Santiago, Stanford
    October 14, 3:00 pm

    Abstract

    Electrokinetics involves the interaction of solid surfaces, ionic solutions, and electric fields. Electric fields can be used to generate bulk fluid motion (electroosmosis) and to separate charged species (electrophoresis). Microfabrication technology has ena

  • Didn't the computer industry take over watts yet? Shouldn't that be 1024 watts?

    or

    Insisitive clods! They should say 3412 BTUs!
  • I'm serious. I've heard that pure distilled water is not conductive at all. Some big transformers have used this cooling method.

    Of course, you get any impurities in there and PFFFFT!

  • this system can handle 1000 watts (yes, a kilowatt) per square centimeter.

    I'm pretty sure a kilowatt is actually 1,024 watts.

    Though, I suppose that's depending on who is measuring it. Not to mention that some electrical systems cannot handle larger wattages, or do so through Logical Wattage Access (LWA).

    ;)

  • Liquid cooling? (Score:2, Interesting)

    by Adumbratus ( 711714 )
    Someone will have to doublecheck this for me, but I seem to recall distilled water as being very likely to damage any system due to simple chemistry. Last I recall from high school chem was that impurity concentrations will travel from highest to lowest, and that if the metal involved has enough of a charge to it, it'll just leach out into the water resulting in local pitting of the metal (and eventual failure of the surrounding structure).

    Other then that, isn't it more of a matter of finding the right li
  • Heatpipes have been around and used in laptops and GPU's for a while.

    Yeah, the heat dispersal is extrodinary. But, I am certain when the end-user product arrives it won't be flawless. Perhaps not as functional.

    Heatpipes for every day of the week:
    http://www.google.com/search?hl=en&ie=utf-8 &oe=utf -8&q=heat+pipe+cpu

Slashdot Top Deals

Pascal is not a high-level language. -- Steven Feiner

Close