Diamond CPUs (not a joke)

This is an interesting one... These guys can grow big diamonds. Diamonds have exceptional thermal conductivity... So, it is theoretically possible to use diamonds to create diamond CPUs that will endure operation at much higher temperatures than current CPUs will.

The link:<A HREF="http://www.wired.com/wired/archive/11.09/diamond.html?pg=1&topic=&topic_set=" target="_new">The New Diamond Age</A>!!

<i>So, Intel: See? Prescott should have been made out of diamond wafers!</i>
Quote:
Today's speedy microprocessors run hot - at upwards of 200 degrees Fahrenheit. In fact, they can't go much faster without failing. Diamond microchips, on the other hand, could handle much higher temperatures, allowing them to run at speeds that would liquefy ordinary silicon. But manufacturers have been loath even to consider using the precious material, because it has never been possible to produce large diamond wafers affordably.

<i><font color=red>You never change the existing reality by fighting it. Instead, create a new model that makes the old one obsolete</font color=red> - Buckminster Fuller </i>
38 answers Last reply
More about diamond cpus joke
  1. cool, uhh or should it be hot :P

    _____________
    whompiedompie
  2. Yeah, this one is hot!!

    But I've got a cool one here: <A HREF="http://www.infoworld.com/article/04/02/11/HNeden_1.html" target="_new">Via's CPU line</A> now has 1Ghz processors dissipating some 7W. This is less than 10 times what a typical Intel/AMD CPU dissipates (much less than prescott, heh)... Now <b><font color=blue>that's cool!</b></font color=blue>

    Unfortunately, though, it's so cool that its performance levels will probably make you think the whole damned processor is frozen... (or at least, it'll not perform well...)

    <i><font color=red>You never change the existing reality by fighting it. Instead, create a new model that makes the old one obsolete</font color=red> - Buckminster Fuller </i>
  3. Though diamond can handle excessive temperature of 3500 degrees, making it is a poor conductor of heat, and electricity. It has also no semiconducting properties.
  4. Quote:
    But I've got a cool one here: Via's CPU line now has 1Ghz processors dissipating some 7W. This is less than 10 times what a typical Intel/AMD CPU dissipates (much less than prescott, heh)... Now that's cool!

    Intel's Pentium-M at 1.1 GHz also only dissipates 12 Watt. Ok that's nearly double but it has a 1 MB cache compared to the VIA's 64 kB cache. Still a nice achievement that's very useful for really small laptops of course, and the Pentium-M low-voltage probably costs a lot more.
  5. Quote:
    Though diamond can handle excessive temperature of 3500 degrees, making it is a poor conductor of heat, and electricity. It has also no semiconducting properties.

    I've heard the opposite: <A HREF="http://www.eetimes.com/at/hpm/news/OEG20030822S0005" target="_new">NTT verifies diamond semiconductor operation at 81 GHz</A>.
  6. in 5 years:

    -No you should not overclock that Amd 45000+ At more than 34ghz , because you will go over 430C.
    -But Amd say that their chip support 400C
    -Yea but you will shorten the life of your cpu
    -Hoh i see.....heuu should i put a fan on my stock heatsink then?


    Athlon 2700xp+ (oc: 3200xp+ with 200fsb) , Radeon 9800pro (oc: 410/370) , 512mb pc3200 (3-3-3-2), Asus A7N8X-X
  7. all very interesting - I'm just glad that the DeBeers pseudobusiness is on the verge of extinction.

    If a tree falls on coop, but noone is there to hear it - do less people rejoice?
  8. Quote:
    -Hoh i see.....heuu should i put a fan on my stock heatsink then?

    Wouldn't that heatsink have to be titanium or something? :lol:
  9. Quote:
    Wouldn't that heatsink have to be titanium or something?

    uhh, hardly any metal melts at such a "low" temp..

    _____________
    whompiedompie
  10. Just saw that the other day on TV. The guy made a Diamond wafer they called it. Mini dvd what it looked like.
  11. Quote:
    uhh, hardly any metal melts at such a "low" temp..

    Well, I looked it up and aluminium melts at 660 degrees. But it already starts to become easily deformable at 470 degrees. Either way it's not healthy to have such metal at high temperatures for a very long time. You're literally burning it and it slowly sublimates.

    Which reminds me... what would they do with the interconnections? If it's aluminium or copper it definitely won't withstand these temperatures for long. These 'wires' are so tiny that any heavy temperature fluctuation could just vaporize, break or deform them.

    Hmm, nanotubes...
  12. They can add positively or negatively charge boron to the diamond as it's being made... that was the key.

    <font color=red> If you design software that is fool-proof, only a fool will want to use it. </font color=red>
  13. perhaps it will not ever reach such tempuratures to worry about it

    at what temp would you think the interconnets could tollerate safely, a few hundred anyhow...what ever style of heat disapator is used it would likely be able to keep it all cool enough.

    that is a rather funny out look on the future THGC conversations.

    <b><A HREF="http://www.1112.net/lastpage.html" target="_new">Finally...the LAST PAGE</A></b>
    ASUS P4S8X-P4 2.4B - 2x512M DDR333 - ATI 9500Pro - WD80G HD(8M) - SAMSUNG SV0844D 8G HD - LG 16X DVD - Yamaha F1 CDRW
  14. i think at 430c your desk would combust. and your case would be too hot to touch, let alone put a plastic media into it. i dont think 430c is what they had in mind, i bet they were thinking more along the lines of 100c, thats still pretty damn hot.

    wpdclan.com cs game server - 69.12.5.119:27015
    now featuring (optional) cheating death!
  15. wow, i just can't imagine the electric bill that will follow. ;)

    A fine day!
  16. Quote:
    Though diamond can handle excessive temperature of 3500 degrees, making it is a poor conductor of heat, and electricity. It has also no semiconducting properties.



    diamond is pure carbon

    -------
    <A HREF="http://www.albinoblacksheep.com/flash/you.html" target="_new">please dont click here! </A>
  17. Diamond is a semiconductor, and an EXCELLENT conductor of heat (much better than even copper).

    Some day I'll be rich and famous for inventing a device that allows you to stab people in the face over the internet.
  18. Quote:
    i think at 430c your desk would combust. and your case would be too hot to touch, let alone put a plastic media into it. i dont think 430c is what they had in mind, i bet they were thinking more along the lines of 100c, thats still pretty damn hot.

    These chips would probably only be a few square millimeter in size. It's like having a cigarette lighter in your case, nothing to really worry about if you can keep the other components shielded. Even if it's bigger than that, much still depends on the cooling method, and it might even work passively with a small heatsink. Free convection improves linearly with temperature difference. So the hotter the chip is allowed to be, the more effective it becomes at actually loosing its heat.

    A bit similar, my girlfriend will be experimenting with sonoluminescence this semester. She will produce temperatures close to a million degrees in water. That's no problem at all since it's focussed on just a minuscule gas bubble, and the surrounding water won't even boil!

    So, whenever we talk about temperatures it's very important to mention at what scale it happens. Actually it's more useful to talk about the amount of Watts dissipated.
  19. At higher temperature diamonds behavioural properties become like those of silicon, cant remember what the exact temperatures are, (and too lazy to go back and look it up).

    "" Society is to blame ""
  20. thats what i figured...it'd be so small heat build up would not be a problem for other eliments inside the case.

    Quote:
    So the hotter the chip is allowed to be, the more effective it becomes at actually loosing its heat.

    i didn't realy think about it that way before but that makes sence too...sorta like driving a car 100Km/h (60mph) and taking you foot off the gas you will slow down 50Km/h(30mph)to 50km/h(30mph)in say 10 seconds, but if you were doing 150km/h(90mph)it would only take 5 seconds to drop down to 100km/h(60mph)

    the faster or hotter something is the steapper the curve is back to nutral


    <b><A HREF="http://www.1112.net/lastpage.html" target="_new">Finally...the LAST PAGE</A></b>
    ASUS P4S8X-P4 2.4B - 2x512M DDR333 - ATI 9500Pro - WD80G HD(8M) - SAMSUNG SV0844D 8G HD - LG 16X DVD - Yamaha F1 CDRW
  21. Maybe too late???
    <A HREF="http://msnbc.msn.com/id/4242493/" target="_new"> Photonics? </A>

    The loving are the daring!
  22. I've read about that before. But that's not the only way to make diamonds, in fact it's probably not the most practicle for making wafers!

    A LONG time ago someone figured out how to make a diamond film on a plastic lense. This idea is probably already in use for making scratch resistant sun glasses at the high end of that business.

    The problem is, the film is only microscopicly thin. Not good for making stones, but PERFECT for making wafers (just build up several layers of the diamond film on a substrate).

    <font color=blue>Only a place as big as the internet could be home to a hero as big as Crashman!</font color=blue>
    <font color=red>Only a place as big as the internet could be home to an ego as large as Crashman's!</font color=red>
  23. ooo wow realy? thats seems more promising

    <b><A HREF="http://www.1112.net/lastpage.html" target="_new">Finally...the LAST PAGE</A></b>
    ASUS P4S8X-P4 2.4B - 2x512M DDR333 - ATI 9500Pro - WD80G HD(8M) - SAMSUNG SV0844D 8G HD - LG 16X DVD - Yamaha F1 CDRW
  24. O no!! One of my good friends works with sonoluminescence and u just had to mention it hehe. Where she working at (please don't say RPI...)
    They using water? I thought it'd be more like acetone....
    Anyway, best of luck to her getting her experiment underway, that shite gets annoyingly touchy so u have to be real precise for it to work...

    The one and only "Monstrous BULLgarian!"
  25. Quote:

    Though diamond can handle excessive temperature of 3500 degrees, making it is a poor conductor of heat, and electricity. It has also no semiconducting properties.

    Although you probably already read this in the topic, your statement is false. I would like to point to you why as your statement contains an implicit inconsistency (is that english? whatever)

    Diamond can handle such excessive temperatures because it is a great conductor of heat. Although the crystal structure is very stable, it would blow up in your face if it was not such a great heat conductor.


    BigMac

    <A HREF="http://www.p3int.com/product_center_NWO_The_Story.asp" target="_new">New World Order</A>
  26. Yes, and actually, thermal dissipation increases with temperature according to the fourth power. This means a great deal: Something at 0C dissipates 3.5 times less heat than something at 100C. If you double temperature, you multiply heat dissipation by 16!!! This is from infrared/electromagnetic emissions alone - it doesn't account for conduction.

    As for temperatures in Diamonds, not only do they endure 100+Cs, but the important part is that they <b>conduct heat</b> extremely well. So, if you put some sort of heatsink on top of them, they can dissipate 100+W or much more and still operate at low temperatures. This is what's highly interesting...

    <i><font color=red>You never change the existing reality by fighting it. Instead, create a new model that makes the old one obsolete</font color=red> - Buckminster Fuller </i>
  27. Quote:

    i didn't realy think about it that way before but that makes sence too...sorta like driving a car 100Km/h (60mph) and taking you foot off the gas you will slow down 50Km/h(30mph)to 50km/h(30mph)in say 10 seconds, but if you were doing 150km/h(90mph)it would only take 5 seconds to drop down to 100km/h(60mph)

    Hey, this is spot on :smile: ! Your analogy is abolutely correct. Thermal conduction and temperature variations are direct functions of the temperature the individual elements are in; and they're usually very powerful correlations.

    Take black body radiation, for instance: it is that glowing coal effect. It's infrared and electromagnetic radiation combined, and all objects emit it to a lesser or greater degree. Black body radiation emission power is related to, as I said, the fourth power of temperature. Double temperature, and you 16x the power dissipation.

    Thermal conduction curves relating to two different objects that are in direct contact follow similar patterns (and exponential ones at that). Thermal conduction is important because it controls the equilibrium state of all these things - if you can make more heat flow, then the final temperature will probably be lower because heat is not allowed to build up.


    <i><font color=red>You never change the existing reality by fighting it. Instead, create a new model that makes the old one obsolete</font color=red> - Buckminster Fuller </i>
  28. Indeed, you're right, Crashman. Thin layer coating by sputtering in extremely low pressures, for instance, can create a layer of Diamond-like Carbon (it's called just that, DLC) of anywhere between 5000 to 50000 Angstroms or more. This should be good enough for making a processor... I've studied the deposition process a bit...

    If I'm not mistaken, one major japanese automobile industry pioneered diamond-like carbon layer-deposition and used it widely in its engines; increased efficiency and durability were reported. Unfortunately, I don't know if that's exactly what they use for lenses; DLC is very much black and shiny. But!! You said sun glasses; this I can believe. And... wow. What an idea.

    Actually, once you know how admirable DLC is... it would be very cool to have DLC-coated sunglasses. Wow. It's probably highest-end. Diamond-like black-shiny coated carbon. Absurdly scratch resistant. Extremely expensive.

    As for this particular article that was mentioned in the first post, well... it refers to bigger gemstones. But DLC is very, very interesting too...

    <i><font color=red>You never change the existing reality by fighting it. Instead, create a new model that makes the old one obsolete</font color=red> - Buckminster Fuller </i>
  29. Quote:
    One of my good friends works with sonoluminescence and u just had to mention it hehe. Where she working at (please don't say RPI...)

    Nice coincidence. Seems to be a <i>hot</i> topic with possibly some future for nuclear fusion. She's studying at Ghent University and it's a project to prepare for Masters.
    Quote:
    They using water? I thought it'd be more like acetone....

    Just distilled water.
    Quote:
    Anyway, best of luck to her getting her experiment underway, that shite gets annoyingly touchy so u have to be real precise for it to work...

    Thanks, I'll let her know! :wink:
  30. Quote:
    Yes, and actually, thermal dissipation increases with temperature according to the fourth power.

    That's only the emmissive component of heat transfer. The chip would actually have to glow brightly. Then you can get a tan while working! :lol:
  31. Yes, well, I think I mentioned in my post that this is black body radiation, which is only a part of the picture... I didn't make it too clear though, sorry.

    Heat conduction is a very important one too, but it is more sophisticated and depends on temperature a lot, too - at least linearly. The hotter, the faster dissipation goes, even proportionally.

    <i><font color=red>You never change the existing reality by fighting it. Instead, create a new model that makes the old one obsolete</font color=red> - Buckminster Fuller </i>
  32. Found out another fact: diamond <b>burns</b> at a little more than 4,000 degrees! Of course that's not a realistic temperature for chips...
  33. I saw a video of a super heated diamond being dropped into a dish of liquid oxygen. It turned straight into carbon dioxide.

    Some day I'll be rich and famous for inventing a device that allows you to stab people in the face over the internet.
  34. Just a few comments:

    The thermal emissivity of a black body would require something to transmit to - a heatsink clamped down on top would remove that transmissible gap (microscopically you may still benefit, not if you polished the bottom of the heatsink though!)

    Also talking about temperature and proportionality is sure to confuse some people. Given that Celsius and Fahrenheit can confuse people enough, it's worth pointing out that 0 deg C is not zero - it's 273.16 K. So, as you implied before, going from 0 to 100 deg C is actually going from 273 to 373 K, a 37% increase.

    Also, this talk has mostly centred on temperature. However, power dissipation is more important. Hence why the toleration of higher temperatures is important - the chip can be run faster without significantly improving the cooling as the increased temperature improves the rate of heat transfer through the heatsink and convection into the air. This comfortably improves power dissipation without increasing cooler complexity and cost.

    Finally, a colleague recently finished a doctorate in thermodynamics studying phase change cooling - but not in the normal way you may imagine. He studied the extra cooling achieved using the turbulence generated when a bubble formed from boiling a liquid moves away from the surface. The turbulence acts to improve flow over the surface, bringing more mixing into the liquid and improving the conduction from the solid to the coolant. This is in addition to the normal benefits of phase-change cooling. The link to the thread is that the most common liquid is water - cheap! So cooling systems that can be designed to operate above 100 deg C actually can have a significant improvement in power dissipation using a readily available liquid - like a compressor cooler but at practical temperatures (well within the operating range - no worries about condensation on surrounding components) and prices!

    Rambled a bit but hope it explains things a bit more for the less thermodynamically knowledgable readers.

    Peter
  35. Quote:
    The thermal emissivity of a black body would require something to transmit to - a heatsink clamped down on top would remove that transmissible gap (microscopically you may still benefit, not if you polished the bottom of the heatsink though!)

    If you put a heatsink on top of your processor, then I can see that black body radiation will not be very relevant, because you'd have a microscopic contact. But I didn't mean the processor itself; I was thinking about the Processor+Heatsink combination, on top of which the fan resides. The heatsink itself is unlikely to melt; therefore, the main effects that dissipate heat off the heatsink are: (1) thermal conduction to the flowing air and (2) black body radiation. So, while the actual processor die might not have a lot of black body emissions, the heatsink certainly has. Black body radiation <i>cannot be avoided</i> at all.
    Quote:
    Also talking about temperature and proportionality is sure to confuse some people. Given that Celsius and Fahrenheit can confuse people enough, it's worth pointing out that 0 deg C is not zero - it's 273.16 K. So, as you implied before, going from 0 to 100 deg C is actually going from 273 to 373 K, a 37% increase.

    This is why I said that going from 0C to 100C means a 250% (x3.5) increase in black body radiation. If anyone is truly interested, I did this math:

    increase = (373^4)/(273^4) = 3.5

    ...Exactly because black body emission is proportional to the fourth power of temperature, but in the absolute - Kelvin - scale.

    About this phase change cooling: it's very interesting, because with phase change, you can transport huge amounts of heat <i>virtually without any change in temperatures</i>. This is unusual, because heat buildup usually means temperature increase. Therefore, it really brings out what you said: temperature is what actually matters, not heat. If you can get rid of heat, temperatures won't rise.

    That's what makes phase change cooling so great.


    <i><font color=red>You never change the existing reality by fighting it. Instead, create a new model that makes the old one obsolete</font color=red> - Buckminster Fuller </i>
  36. Good point about the heatsink and prcessor being considered as one black body, the only problem is that the heatsink will be at a significantly lower temperature than the die, reducing the benefit (though it's still worth considering).

    Also, for the person who mentioned aluminium melting earlier, a lot of car engines are made using it now. Admittedly they need cooling, but a well designed heatsink and fan combo used within its correct operating range would not be a problem.

    I didn't mean to seem to disagree with your posts, more wanted to guide the less thermodynamically able readers to appreciate your messages. When I learnt thermo I found it quite straight forward but some of my cleverer course mates seemed to struggle, so I thought that this thread could do with a bit of explanation!
  37. Hmmm... I'm not an engineer or anything, but isn't there an issue with the conductive material (copper or whatever) handling the electricity melting at some point? Wouldn't manufacturers have to deal with heat threshold much less than what a diamond can withstand regardless? How close are current processors to that threshold?

    Also, what sort of manufacturing process would be needed to work with diamond wafers or DLC? I'd imagine it's more difficult than working with silicon.
  38. Absolutely, that's OK.

    Actually, I'm about to graduate in physics, but I still can't seem to like thermodynamics. I understand it alright, but... Like it?... Not really... :eek:

    <i><font color=red>You never change the existing reality by fighting it. Instead, create a new model that makes the old one obsolete</font color=red> - Buckminster Fuller </i>
Ask a new question

Read More

CPUs Diamond