Stevee

Hmmm... well, the heat transfer could be inefficient due to friction, resistance and whatnot, would it still violate the the 2nd law of thermodynamics if only a portion of the energy would get transferred while the rest gets transformed or leaves the system?

I see. Interesting. So they already tried many things. Although, One thing I don't understand about the Brownian ratched. "Although at first sight the Brownian ratchet seems to extract useful work from Brownian motion, Feynman demonstrated that if the entire device is at the same temperature, the ratchet will not rotate continuously in one direction but will move randomly back and forth, and therefore will not produce any useful work." Random motion can be used to generate alternating electric current, no?

As a discaimer, I realize that this theoratical device is both highly inefficient and is an engineering nightmare, still, I'm curious as to whether - it could, theoratically, work as intended, and if so, - is there anything like it in existence/has it been tried. Please exuse my english. From what I know (Which is not much, hence why I came here to discuss this), gases on a molecular level behave like small ping-pong balls, bouncing around all the time: http://lab.concord.org/embeddable.html#interactives/sam/phase-change/2-two-types-of-gases.json On large scale, the force inflicted on a sufrace by gas (In a peaceful environment) is more or less constant (the law of big numbers?). But what if we only have minimal quantities of gas molecules in a container? On a small time scale with highly sensitive equipment, we should be able to measure some fluctuation of pressure (Correct?). This fluctuation should also be present when dealing with larger quantitie of gas, but is insignificant compared to the average pressure. Here is where the machine comes in: http://i.imgur.com/pUtAhFE.png A box with pockets of air inside, separated by airtight metal plates. The minimal fluctuation in pressurse should slightly push and pull the plates back and fourth, this movement could then be converted to electricity, which could be used to say, heat another boxfull of air. If I understand correctly, the temperature inside the machine's box would become cooler as a result of the energy extraction. The end result would be that without use of an external force/influence, heat gets transferred from one box to another. Now I didn't go into any calculations yet, but fluctuation in pressure for each plate is probably way too small to make a real difference. Still, I would expect there to be at least a bare minimum of heat transfer taking place and that perhaps by improving it's design (Nano-tech would come in handy (Would we have to worry about the Casimir effect?)), the machine could become more efficient. So, what are your thoughts? Is there anything important I left out? Is this idea worth pursuing or is it unfeasible? Perhaps its already somehow used and I just don't know about it?