jhc@alice.UUCP (JHCondon) (05/05/85)
>Path: ..!dec-viking!wasser_1 (John A. Wasser @ DEC Engineering Network) > > RE: The "two ellipsoid and sphere" perpetual motion machine: >> >> ...it is clear that all the light from f1 hits >> either one or the other of the two ellipsoids, and that all the light >> from f2 hits one of the ellipsoids (thereby going to f1) or the sphere >> (thereby going back to f2). No rays hit the sphere then an ellipsoid >> or vice versa. >> > The fact that all radiation from f1 travels to f2 but some of > the radiation from f1 bounces back to f2 does not mean that > f2 will have a higher temperature than f1 (and therefore there > is no temperature difference to drive the PMM). > > Consider that if you were at f1 and looked in any direction > you would see black-body radiation that depended on the > temperature of f2. If f2 is colder than you, you would > loose heat to it by radiation. If f2 is hotter than you > you would gain heat from it. If f2 is the same temperature > as you, you would neither loose nor gain. There is an appeal to fuzzy thinking here. Just what is "radiation that depended on the temperature of f2"? There is spectral distribution which depends on the temperature of the emitter, and there is flux (units of energy/(area * time)) which depends on such thing as the size of the emitter and distance from it. I agree that the spectral distribution of the radiation seen a f1 is that the temperature of f2, but power is transferred by the flux. To illustrate the difference, it should be possible to combine a bunch of mirco-wave transmitters so that the output have the distribution of 1K radiation, but the flux would be high (~able to punch holes in razor blades). Suppose that the solid angle of the sphere is 4*pi*alpha when standing at f2. Also suppose that both objects at f1 and f2 are identical and at the same temperature so they both emit power P uniformly in all directions. The power received at f1 is then (1 - alpha) * P, and the power received at f2 is then (1 + alpha) * P. This appears to be power transfer at zero temperature difference of amount alpha * P. In order to put this PMM machine discussion to rest the answer should address (and find error in) the above argument, not change the problem to single ellipsoid with different size bulbs, or depend on some non theorem. > I hope this puts an end to this version of the PMM. > Let's see some NEW ideas. > > -John A. Wasser sorry about that.