[extropy-chat] Casimir Torque Project

[Hal Finney]

Scerir writes:
> But also constantly (and strong enough). It is - perhaps - 
> easy to create a *temporary* asymmetry.
>
> ----    ----    ----    ----          fixed multi-plate
>   < ----    ----    ----    ----      sliding multi-plate
>
> (You can also imagine round multiplates, one fixed and
> the other spinning).
>
> Due to the "lateral" Casimir force, the sliding multi-plate
> moves to the left, because the plates of the sliding multi-plate 
> and of the fixed multi-plate tend to overlap. During the
> process the sum of field energy and mechanical energy is
> (supposed to be) constant. 
>
> ----    ----    ----    ----    fixed multi-plate
> ----    ----    ----    ----    sliding multi-plate

I thought of something similar earlier.  You could unroll Adrian's design
and get a linear accelerator.  What you do is put parallelogram shaped
"blockers" between the two rows in the first diagram, so that the bottom
moving plates only "see" the plates above them to the left.  This is
supposed to pull them in that direction.  Then when they get to the
position in the second diagram, the bottom plates are completely covered
by the blockers and so they don't see the forces that would normally
pin them there.  Then when they move a little farther (by momentum,
or by careful arrangement of the blockers) they will again only "see"
a force pulling them to the left.  Presto, perpetual motion again, only
this time we create a linear rather than rotary accelerator with no input.

I am confident that this won't work either, because with the blockers
in place the plates can no longer move in such a way as to minimize
potential energy, hence there will be no force.  Or else, there will be
a minimum in the PE, as in Scerir's second diagram (with no blockers),
and then the plates will just go to that position and stay there.

Hal