I was referred to this science paper:
Maxwell's Demon, rectifiers, and the Second Law: Computer simulation of Smoluchowski's trapdoor P.A.Skordos and W.H.Zurek Dec. 1991
http://arxiv.org/abs/cond-mat/0402665 pdf version:
http://arxiv.org/PS_cache/cond- mat/pdf/0402/0402665.pdf
The figures (illustrations) did not open for me.
I do not agree with the paper's conclusion that 2LoT is supported.
To simulate cooling of the trapdoor to see if the heat difference would
turn it into a pump the computer removed velocity from the door and returned an equivalent amount of energy to the particles. The simulation should be inspected to see if friction does the same thing. The solid pivots of the door, perhaps a hinge, have zero degrees of freedom and friction, due to partial stickiness, so the door should slow down at any temperature. The particles can afford a small transit burden.
In one run type all the particles were started on the far wall of the trapping side of the door and the door was closed. The receiving chamber should have kept its particles indefinitely. (another run type had all the particles started on the far wall of the loosing side).
Aloha, Charlie
