Please create an account to participate in the Slashdot moderation system

 


Forgot your password?
Close
typodupeerror
×
Hardware

Mega Heat Sinks 60

Posted by CmdrTaco from the something-that-is-literally-cool dept.
An anonymous reader submitted a stream of obscenities, but amidst them was a link to this page which is a sweet heat sink- in fact, it actually was able to accumulate a quarter inch of frost. Could be useful with Intel's track record of releasing chips that can double as microwave ovens.
This discussion has been archived. No new comments can be posted.

Mega Heat Sinks More

Mega Heat Sinks

Comments Filter:

  • Many Peltier heat sinks (Score:1)

    by Anonymous Coward
    Do a search for "Peltier heat sink" on Google or Metacrawler and you'll find several sources. They've been around for a while.

    Of course, to get frost on the outside of such a device you'd have to use one which you can plug in backwards -- so the outside would chill while it heats up the processor chip...

  • Kryotech? (Score:1)

    by avm ( 660 )
    You're possibly thinking of Kryotech [kryotech.com]. They retail overclocked Alpha and AMD-powered boxes, with their patented refrigeration system built into the case.

  • wow!! (Score:1)

    by gavinhall ( 33 )
    Posted by unixmj:

    I could really use one of these for my dtv-box. Its called Sagem and you could fry eggs on the mpeg2 chip in it, got a p133 fan in it now - but still it overheats now and then. Perhaps Sagem is really a Intel company??

  • Lame (Score:1)

    by mholve ( 1101 )
    Frosty heat sinks? Great fuckin' idea...

    Obviously a ploy by Intel aliens to get you to buy more Pentiums when your current one burns out from moisture.

    Microwave? Uh, sure... Try "convectional" instead. It would sound like you knew what you were talking about...

  • I can't imagine how ANY heat sink could produce this effect unless it had some kind of a refridgeration unit in it. Furthermore, his comments about the one wire connecting to one part and another wire connecting to another is fairly counter intuitive to normal power mechanisms.
  • Er... s/microwave/conventional/i


    ?

  • Similarly, though not quite as strong, 1.2 GHz is a harmonic for the 2.4 GHz microwaves and can have the same effects. A lot of the cellphone EMF scare has its roots here.
  • A chip operating at a certain frequency radiates at such frequency, plus every odd multiple of the base frequency. Of course the radiation is low, but it is there. Usually shielding will block most of it, although no method of shielding is perfect.

  • Where I work (for a major defence contractor) we use similar devices for cooling things where there's no possibility for airflow (i.e. the displays in an aircraft cockpit).

    They're really not all that new technology, as I recall. The concept is simple, based somewhat on the idea of a Peltier Junction.

    A Peltier Junction is formed when two dissimilar metals are physically bonded. If you put a current across this, heat is transferred from one side to the other. Thermoelectric coolers use a similar idea, but use massive arrays of specialised semiconductor diodes to achieve the same task much more efficiently.

    We get ours from ITI-ferrotec [ferrotec-america.com] and they have a pretty good web site explaining the basic concepts, as well as how to use a thermoelectric cooler in certain applications. If you let one run uncontrolled, it could easily develop frost like the web page indicates. It's typical to see TEC's with a temperature differential above 40 degrees C.

    Most applications where you want to avoid frosting your sensitive electronics require a controlled feedback loop where you can vary the current to the device based on its temperature. Just power cycling them (as with a standard thermostat) is fairly detrimental to their life-span, so in high-reliability applications (like the displays in aircraft cockpits) we have to use a specially programmed microcontroler driving a darlington power transistor to control the current, combined with a thermocouple sensor to monitor the temperature on both sides of the device.

    Even more interesting is that if you reverse the power to the TEC, the heat flow will also reverse. Great for when you have to start up your F-16 on a really cold morning and the LCD has frozen up.

  • I have heard that it was actually liquid nitrogen. I have seen pictures, and it was very cool looking (yes the pun was intended) :-)
  • cordless phones have been approved for the 2.4 GHz range. I have heard claims of 7700ft range in them!!!
  • I have always wondered about the possibility of using a small fridge w/ a bowl of silica gel to absorb all the moisture. Anyone else had this idea?
  • ANY high-intensity radiation can cook food. Remember, microwave ovens produce 750-1000 Watts of microwave power.

    The point of using microwaves, of course, is that such wavelengths are easily absorbed by water molecules, which tend to be prevalent in food. But that doesn't mean that any microwave radiation will cook food. Quantity, not just quality, counts.

    Kythe
    (Remove "x"'s from

  • So anything that cools itself down to any extent can remove the heat from any device.
  • I thought I got in ahead of the crowd.
    Arrgh!
  • Velox? The only Velox I know of makes rim strips fo bicycle tires. :-)
  • What's the deal with everyone yelling about Intel? I personally have a PII 400, it ran at about 100F when running at 400mhz and I now overclock it to 600mhz and it doesn't get over 120F, and it is rated to well over 150F, if memory serves. Have those who are talking about Intel chips being 'microwaves' actually have any experience with them, or is it just fun ripping on big companies. Remember, also that Intel is supporting linux and an Intel/Linux platform is probably the BEST (value) server out there nowadays, save possibly dec alpha.
  • Didn't say anything about a 150mhz bus, I overclocked using the bus / multiplier ratio and setting that to 6 [ 6 * 100 = 600 ].
  • Like I said, it is the bus / clock multiplier ratio that is being increased, and that is on the motherboard, not the chip. Maybe I just got a God-sent chip, but I haven't had a problem yet. All I know is I upped the multiplier, it ran faster and a little (but not much) hotter.
  • There's more to CmdrTaco's glib comment than meets the eye - a letter [newscientist.com] to this week's New Scientist [newscientist.com] points out that as Microwave ovens use a frequency of about 2.4 Gigahertz, chips in the not-too distant future will be pumping out (albeit small quantaties of) raditation that could cook your food, even if you do have a good cooling system... Personally, though I'd be more worried by stray radiation from the oven buggering about with my computer...

  • Erm - I wasn't seriously suggesting tat you could cook with it, but that we may have problems with interferance...
  • Frost on a heat sink is obviously a joke. I haven't been able to see the link yet, but if it's true it can't be a passive heat sink. You can only get as cool as the ambient temperature obviously. At 20degC ambient and 60% humidity, DEW doesn't even form until the temperature is 10degC, approximately.

    Some folks mentioned Peltier heat sinks. I have never heard it by that name, but it sounds like a thermoelectric cooler [google.com]. Put the current one way, it's a cooler, switch the direction of the current and you have a heater. We use it all the time to cool detectors used in space applications.

    Check out http://www.melcor.com/teccover.htm [melcor.com] if you really want one.

    I think it's pretty bad that you would ever need active cooling for a CPU. That really has to hurt reliability.

    ~afniv
    "Man könnte froh sein, wenn die Luft so rein wäre wie das Bier"
    "We could be happy if the air was as pure as the beer"
  • forget a better heat sink for my processor, i need one for my banshee, atleast my processor won't fry eggs
  • The problem with taking a chip down to liquid nitrogen temperatures is that the circuitry will crack due to the fact that the different materials used to build the chip (silicon, silicon dioxide, aluminum or copper) have different coefficients of thermal expansion. I'm told that freon temperatures still work, though.


    This might have a tolerable cost if you made the liquid nitrogen on the fly. This isn't difficult.

  • This weubsite [nuclearwinter.com] has some interesting writeups of a project to refrigerate a CPU. It has some Very Good Ideas like an alarm to sense if the temperature is suddenly skyrocketing, and if so, to underclock the CPU, and another mechanism to keep the temperature from going below 60 degrees Farenheit (that's around 15 degrees Celsius).

    I personally am using sandwich bags filled with water to cool my PS/2 Model P75 486. It has 24MB of memory in the form of 2MB SIMMs <shudder>. Its 486DX/33mHz also has a habit of overheating, which makes these interesting displays like ``??? 111'' when the parity checker catches the error, or if it doesn't, I get ``data modification'', which is IBM speak for random bit changings. That causes neat things like diskette copyings to suddenly fail, DHCP requests to come back with funny IP addresses, and the mouse pointer to suddenly morph into a strange shape. These are all things that have happened before I was watercooling.

    Cheers,
    Joshua.

  • Couldn't get through to the reference page (Slashdotted already?), but I can't help but wonder if this could be some kind of Pelter Junction device. For those who've never seen one (a local electronics shop here has one on its demo display), a Peltier Junction transfers heat from one side to the other when a current flows through it. It's the first thing I could think of, small enough to be built into a heat sink, that could actually make something cooler than the ambient temperature.
  • I bought a great stack-type heatsink at a local surplus place. The thing looks like the cylinder on a lawnmower, but its amazing! I mounted a case fan next to it, so it blows the heat up (tower case) into the power supply (I of course reversed the ps fan). My K62 now runs at 42 deg Celcius (before, it ran at 50-55).
  • I just thought I'd remind some folks that a heat
    sinks job is to help take heat FROM the processor, not to go frosty and cool it off... a sink which dissipated heat fast enough to freeze doesn't by itself make any sense...
  • If you're going to be actively cooling your cpu, why not go the affordable route. Liquid nitrogen... at around $0.12 a liter, maybe a liter an hour... only $2.88/day... $20.16 dollars a weak... **ONLY** $1048.32/year!!

    FROST GUARANTEED!!
  • The trouble with frosty heatsinks is: I go outta town for the weekend, the power to my apartment dies. My UPS is only good for about 15 minutes, so in a half an hour, this thing stops cooling my CPU. Then the frost melts. This is a bad thing(tm). A much better solution, having not actually gottten to this slooooooww site, would be the Kryotech system. Check out Kryotech [kryotech.com]. They cool the CPU down to -40C with no frost. They actually have little heaters in the heatsink to prevent that, but still cools the chip!
  • well, when someone figures out that cold air is good and the moisture is the bad part, then we'll get somewhere. why not make peltier card slot fans ? i have fans in my servers that push/pull air thru the card slots. why not put this peltier stuff around the tunnels they use ? would make the machine blow dry cold air into the machine, more effectively cooling the entire system. eh ?
  • card slot fans. they push/pull air through card slots in the back/bottom of the case. someone figure out how to wrap peltier stuff around the opening of those, and control the moisture, then you've got a good effective cooling system pushing dry, cold air, into the entire system.
    seems alot safer than putting condensation on my cpu and still alot more effective than just a normal fan. eh ?
  • Hmmm, OK so that's how it's done. Interesting.

    One question tho. Is having a cold surface like that on a chip ( questions of the moisture asside, at the moment ) actually going to cool it any better? I thought the idea of the heat sink was to transfer heat FROM the cpu, and dissipate it. Now, I'm not totally convinced that putting a colder surface on the CPU is really going to be able to cool it down any better ( you're still going to want your heat transfer and dissipation properties of the heatsink ). And yes, I realise that the object in question was definately a heat sink. The question being is the cold contact surface going to make it perform better than any other heatsink?

    THEN we get to the question of whether or not you'd want something that produces moisture like that near our computers. I dont' know about the rest of you, but my answer to that is a big fat NO! :)

Slashdot Top Deals

"Conversion, fastidious Goddess, loves blood better than brick, and feasts most subtly on the human will." -- Virginia Woolf, "Mrs. Dalloway"

Close