Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

Refrigerating a sealed room?

  1. Jul 28, 2010 #1
    Seems like a pretty basic thermodynamics question, but i can't think of a way how or get a clear reason why not.

    Let's take a heat-insulated room with a usable energy source provided in it (electric wires sticking out of a wall).

    Can this room be refrigerated as a whole?

    One thing is, any refrigeration scheme i've seen requires a heat sink besides a power source, so one half of the room would have to be made hotter to get the other half colder, and since whatever way is used would add more heat to what it moves, the overall temperature would increase until the power runs out. First law of thermodynamics.

    On the other hand, why can all the excess of heat/entropy be generated on the other side, where the energy source is? Energy can be use to order things, and ordering things reduce entropy. There are ways to slow atoms down directly on small scale.

    So, is it a fundamental or technical problem?
  2. jcsd
  3. Jul 28, 2010 #2


    User Avatar
    Science Advisor
    Homework Helper
    Gold Member

    It's a fundamental problem. To cool the room, you need to remove entropy as well as energy, and the heat insulation prevents the entropy from leaving.
  4. Jul 28, 2010 #3
    If it were theoretically possible to do what your suggest, room air conditioners would not have to be hung out windows to get rid of extra heat...

    But if you had a water supply coming into the room instead of electricity with water colder than the room, and allowed the water to leave the room, you of course could cool the room.
  5. Jul 28, 2010 #4
    Thank you for clarifying.

    Yesterday i would have agreed and not asked the question above, but today i've been to one official's office, which is quite like the room described - heat insulated and with windows that don't open (after all the usual problem here is -30C, not +40C). And there on a cabinet stood a small AC unit, the kind you stick into a window leaf, blowing cold air on the clerk in question. The thing had the hot end in the same room it was cooling, yet there was a stream of cold air, and the rest of the room was only marginally hotter than outside.

    That got me thinking.
  6. Jul 28, 2010 #5


    User Avatar
    Staff Emeritus
    Science Advisor
    Gold Member

    Sure, the fellow sitting right in front of the cool air flow will feel cooler. HOWEVER, the room as a whole will be hotter.
  7. Jul 28, 2010 #6
    I don't think that was a standard air conditioner. If it were... dumb, de dumb, dumb.
    Not only would it NOT cool the room, it would heat it up. Like said before, a person directly in front of it might expierience a cool air flow, but the entire room would actually get hotter.

    Assuming the "not dumb" factor, the unit might have been an indoor swamp cooler as oppossed to a regular window mounted air conditioner.
    Note that, even with an indoor swamp cooler, you MUST crack a window open, preferably in the same room.

    Even still, a simple fan will cool someone in front of it, yet will ADD heat to the rest of insulated room.
  8. Jul 29, 2010 #7

    Andrew Mason

    User Avatar
    Science Advisor
    Homework Helper

    You should go back to his office and find out how it worked. All we can say is that it has to be putting out more heat than it absorbs (ie. if the hot reservoir is the room, it heats the room more than it cools).

  9. Jul 30, 2010 #8


    User Avatar

    Staff: Mentor

    Ugh, unfortunately, the "not dumb" factor is not a safe assumption.

    A few years ago, my company was hired to help my alma matar (!) [hangs head in shame] deal with an overheating data center. The college had been growing and as a result they had been packing their primary data center with new servers to handle the email and network storage of new students/profs/classes. When I went to survey, I found they had been doing their own cooling upgrades in an attempt to keep ahead of the rising heat load. The installed systems, in order of installation, were:

    1. Standard Computer Room Air Conditioning (CRAC) units. These have air handlers in the data center that blow cold air under a raised floor, where it percolates up through perforated floor tiles. Air is returned through the top of the unit, directly from the room. The heat is rejected to remote (outside) condensing units like what you have sitting outside your house. Standard, straightforward.

    2. Operating on a similar theory, they added a large residential grade split system to dump cold air into the floor plenum and pull warm air from the ceiling plenum. The ceiling also had perforated ("eggcrate") tiles to allow the hot air rising off the servers up into the ceiling. Again, a remote condensing unit rejected the heat to outside. I believe it had 17.5 kW (5 tons) of cooling capacity.

    This wasn't sufficient so you know where this is going.....

    3. For hotspots that were cropping up in the room, they added a handful of spot coolers. These are completely self contained units - no separate unit that sits outside for heat rejection. They use a duct that pulls a little bit of air from the room to cool the integrated condenser, then you are supposed to run that duct to a window to dump the waste heat outside. Instead, they ran the heat rejection duct up to the ceiling. I think there may have been three of these units, producing 3.5 kW of cooling each, while requiring 1.5 kW of input each, for 3.5+1.5=4 kW of heat rejection each.

    So what they "gained" here is instead of letting the fan of the system in #2 pull warm air up into and through the ceiling plenum as it was already doing, they added 4.5 kW of "pumps" to pump it up there. The net result is the combination of 2 and 3 was 17.5-4.5=13 kW of cooling. Adding #3 decreased the cooling to the data center. And they were genuinely confused about why they were losing control of the temperature even while adding more A/C units.

Share this great discussion with others via Reddit, Google+, Twitter, or Facebook