New Form Factor

Discussion in 'General Hardware' started by scriptasylum, May 3, 2003.

  1. scriptasylum

    scriptasylum Moderator

    Messages:
    832
    Location:
    Des Moines,IA
    I was just looking at heat sink reviews and realized the heat issue is not gonna ever get better unless something different is done. Even with the shrinking of the manufacturing processes, heat output is still climbing as chips get faster.

    So, here's my suggestion: Create a new form factor that mounts the CPU (and any other heat producing devices) directly to the computer case. Then, the case (or at least that side) could be made from a thicker aluminum/copper. Tada! back to passive cooling! You could even use small fins if you want, esp near the CPU core mounting area.

    You'd even be able to use this as a room heater. :)
     
  2. RazerBack

    RazerBack The First Rebirth

    Messages:
    159
    Tiny Bubbles

    I'm waiting for that, my computer is pretty quick, and should last me for a couple more years. By then different standards should be out and more mainstream. I guess I'll upgrade then.
     
  3. dubstar

    dubstar format c:

    Messages:
    1,357
    Location:
    Southern California
    that tiny bubbles thing is cool, nice article.




    (link found in this article *onions found in space *)
     
  4. scriptasylum

    scriptasylum Moderator

    Messages:
    832
    Location:
    Des Moines,IA
    That "tiny bubbles" thing is cool, but I still think the idea of retooling the current ATX mobo standard would help cool things off. I made a quick drawing of what I am talking about.

    The picture was created from a top-looking-down perspective.
     
  5. Zedric

    Zedric NTFS Guru Folding Team

    Messages:
    4,006
    Location:
    Sweden
    This is all in all a good idea. The ATX needs redoing to take care of the heat.

    But there is a problem with your particular solution. The CPU sits on the back of the mobo which means that you must be able to attach components on both sides of the mobo. This isn't normally done and the reason is probably monetary. The boards are probably wave soldered on the underside (liquid solder washed over the board). If you have a plastic socket there it will melt or at least get clogged. Putting it on afterwards means manual soldering on the top side.

    The solution isn't trivial, but keep spawning! :)
     
  6. blinden

    blinden OSNN Senior Addict

    Messages:
    268
    Location:
    Pittsburgh, Pa, USA
    not to mention having a case that you can get burnt on...

    bigger isn't always better as far as Heatsinks are concerned.
     
  7. silky62678

    silky62678 Guest

    Would the cpu to the aluminum case be cooler than a heatsink/fan?
     
  8. lmi91

    lmi91 Guest

    might also be dangerous if something fell or bumped that area. And it would have to be a copper side to be efficient at transporting heat, so $$$$, and you'd probably still need a fan. but instead it would be on the outside of the case
     
  9. LeeJend

    LeeJend Moderator

    Messages:
    5,291
    Location:
    Fort Worth, TX
    Congratulations. You just reinvented the laptop.

    The laptop does exactly what you descibed. Sinks the heat out the bottom of the case. This is not a good thing. Check the warnings in the high laptop manuals. They say cool stuff like do not place on good wood tables, they damage the finish. And just to really heat up the concept there is a story floating around (think I saw a copy in here and at The Regsiter) about the guy who had his laptop in his "Guess where" lap for several hours. Silly place to put a laptop! Seems the poor fool severely burned his pecker and sued the laptop manufacturer. I think he won the case. But the really good news is his pecker healed. Well, maybe not good news that means he will possibly be breeding...

    Also, the new "faster" Intel laptops do neat stuff like slow the CPU clock down if they are getting too hot.

    Oh yeah, back to the subject at hand. Passive coolers (that's what your idea is called) are severely limited by those nasty old laws of physics. The thiness of the plate you are mounting to and amount of natural convection cooling across the plate surface suck for removing heat Now if you wanted a 1/4 inch thick block of gold for a case cover that will help, but you still can't move enough air across the surface to get the heat out without fins and a high flow rate (i.e. noisy) fan.

    For a reality check, the 2-3 gig processors are dissipating about 65 watts. Stuck your finger on a light bulb lately? And the light bulb is 2 inches in diameter,not the size of a fingernail and a chunk of that 60 watts is emitting as visible photons not heat.

    Liquid cooling or chilled air is how we get the big heta out of electronics. Both have major problems (condensation, leaks, cost weight, power usage...).
     
  10. Taurus

    Taurus hardware monkey

    Messages:
    3,206
    Location:
    Sacramento, CA
    agreed!

    i think this 'use the case as the heatsink' idea isn't one of the best. you would have to keep the case in a well-ventilated area just to dissipate it and nothing could be up against that 'hot side'. not to mention the costs of aluminum cases.

    i can't really think of a feasible and reasonable solution to the future heat problem. increasing the core area would sure help. :huh:

    one good solution for now is if motherboards start reading the cpu's thermal diode and output the voltage to the cpu fan accordingly.
     
  11. Zedric

    Zedric NTFS Guru Folding Team

    Messages:
    4,006
    Location:
    Sweden
    Hush makes case-cooled cases (umm....). To bad they're not powerfull enough to cool a P4 or AMD XP. They can cool 1GHz VIA C3 though.
    http://www.hushtechnologies.net/
     
  12. Indomidable

    Indomidable Guest

    I find the case cooler very dangerous! I totaly agree that liquid cooling is the next step. There are dangers to these forms like using a pilter to cool with or using a water block cooler you can get condensation, and that (could/eventualy would) destroy your chip. If you look at the slot systems they were trying to fix that cooling problem but their preformance sucked because of increasing the distance from the componants. I'd say larger chip's would increase the ability of heat dissipation. But, you'd run into the slot problem with increased distance reduces speed.
     
  13. scriptasylum

    scriptasylum Moderator

    Messages:
    832
    Location:
    Des Moines,IA
    Ok, I see everyone's point.

    Zedric: I know that the socket would be on the other side, but I have seen components mounted on both side of a PCB before. I am sure you are right in saying it would increase costs. Even if the cost was $30 higher for a case like this? Don't most of us spend at least that much on a HSF anyway?

    Blinden: But bigger is better... at least when talking about more surface area for a HS anyway.

    silky62678: Not sure, haven't tried it. But since the case is already exposed to ambient room air and since a mid/full tower side panel is so big, it makes sense I would think.

    lmi91: If the aluminum or copper was mounted sturdily enough, then I wouldn't think that would be a big issue, but then, I'm not an engineer. :) I don't know about you, but rarely does anything bump hard enough against my PC to cause any kind of damage like that. I thought aluminum might be sufficient instead of copper since the surface area is so big. Maybe a copper contact plate and the rest of the side aluminum?

    LeeJend: Yeah, I suppose there are similarities to a laptop. BUT, most people do not put mid/full towers on their lap. :blink:
    I already figured the side plate would have to be thicker than it is now (even 1/4 to 1/2 inch I figure and aluminum is cheap), but the sheer size of the panel would be sufficient. Even use shallow fins to help the surface area... might actually look cool too. :)
    BTW: After running benchmarks a bunch of times, I can touch my HSF bare-handed and it isn't hot enough to burn. And that coming from a small HSF compared to a much larger case.
    A light bulb, as you used for an example, is severly limited in surface area as well as heatsink material: glass isn't a very efficient HS. I'm sure we would use copper or aluminum, but it would be hard to see the light through those materials, plus heat isn't as much as a concern. Sure, it might shorten the lifespan of a bulb, but they only cost a buck or two to replace. Once CPUs reach that price point, I won't care about heat as much I'm sure. :)

    taurus: Ok, maybe this isn't the best ideas, but it seems like we are re-inventing the wheel (albeit with improvements) with these liquid ideas. I was trying to come up with something different. Well, not really different because my idea sorta comes from how car audio amplifiers work. They produce tons of heat, esp those big ones. At a typical efficiency of 55% (ish), a 1,000 RMS watt (not a class D) amp draws about 1800 watts of power from a battery. That means there is 800 watts (thats right kiddies, not a "mere" 65 or 75 watts) of power dissipated on the heatsink. Granted, that heat is generated from many different devices, but the point is the case is the heatsink and it does a good job, provided the amp is not installed in a severly airflow-challenged location.

    Zedric : That's what I'm talking about! Sure, it is limited right now, but thats because it's new. I like that idea and looks bad@$$ too. :) I am just going a bit further than hushtech is going and not as drastically retooling the case.

    Indomidable: I guess I don't follow. My design does not use peltier, or water cooling. Nor does it involve increasing the size of the chip.


    Indomidable gives me an idea though: why not water cool the panel? Extrude the aluminum panel/fins with holes along the inside for liquid to travel through, and only where it is needed. The water would stay outside the case and not have any chance of leaking on your expensive hardware.

    Thanks everyone for their thoughts (and criticisms).
     
  14. canadian_divx

    canadian_divx Canadian_divx

    i rember seeing something like a mobo cooler and it looked something like this
     
  15. Taurus

    Taurus hardware monkey

    Messages:
    3,206
    Location:
    Sacramento, CA
    looks like a radiator for a water-cooling system.
     
  16. canadian_divx

    canadian_divx Canadian_divx

    kinda but it is basicaly a huge one that is just a brick of copper or what ever. i am trying to find where i saw it but yea that is the basic of it
     
  17. Zedric

    Zedric NTFS Guru Folding Team

    Messages:
    4,006
    Location:
    Sweden
    Not that it would cool anything besides the points where the board is screwed to the block...
     
  18. Taurus

    Taurus hardware monkey

    Messages:
    3,206
    Location:
    Sacramento, CA
    unless there was refridgerant flowing through it and a fan blowing onto the board...

    but that's a whole 'nother realm. ;)
     
  19. GoNz0

    GoNz0 NTFS Stoner

    Messages:
    2,781
    Location:
    the year 2525



    i stuck my finger on a CPU and powerd the system up.. was a duron 900 can you say OUCH.. the funny thing is both the lads i worked with watched me and tried it themsleves... tw@'s
     
  20. scriptasylum

    scriptasylum Moderator

    Messages:
    832
    Location:
    Des Moines,IA
    You stuck your finger on the CPU core?!?! You are braver than I. :)