Project Orion related fusion powerplant:

Carve out a huge spherical region underground. Then plate the boundary (to rock
or whatnot) with some very good heat transmitter material and a traditional
Promethean heat exchanger. Keep the interior at a vacuum so that you get most
of the energy as gammas and don't have to clean up radioactive lithium or
whatever is being used and deal with enormous pressures.

Then drop a pulse object into the chamber and detonate it. This pulse object
is basically a hydrogen bomb, but it should be relatively small as well as
clean (have a large fusion component compared to fission component). Its
detonation will create a bunch of radioactive isotopes (from the fission
component) as well as EMF (mostly gamma and x-rays, but also visible and IR) 
and various elemental particles (mainly n and p).

You'll need some very good pumps to get the gaseous waste out of the chamber in
time for the next bomb. Then repeat.

This probably wouldn't be well received by various environmental groups.
Nevertheless, ignoring various engineering problems, it should be possible to
use this example to show that fusion power is possible in theory and with
Q > 1.

--

As an reductionist extreme, consider Project Orion. Replace the fission pulse
units with fission-fusion units and a comparatively larger blast shield. It'll
still work, as far as I can see.

--

Questions: Could we use a "driver fluid", say of Na, and deal with the
pressure in a similar way to how this is done in hydroelectric power plants?
The rapid pulsing probably would be problematic to handle for the turbines,
and the driver fluid would get increasingly more polluted with time. As an
advantage you wouldn't have to use inefficient heat capture anymore.

Also, seeing a plot of the lowest possible power (tons of TNT) versus
fission/fusion ratio of output energy would be interesting. If it's in the
megaton range, this thing falls apart as anything but a gedankenexperiment.

However, India apparently created and detonated a <1KT thermonuclear device.
(http://www.flonnet.com/fl1625/16250810.htm)

-

Other ideas: Could one charge a metallic DF with magnets around and get energy
from the Lorentz force in reverse? In this case the reactor would consist
of a spherical "main chamber" with multiple horizontal cylinders (length axis 
is x axis) surrounded with ring magnets, and with the entire thing filled with a
DF gas or liquid.

(A 3d configuration would probably be more efficient; here one of the vertical
 cylinders would have to fit the dual use of an injector as well)

[Added Aug 10 2003: Maybe it would be possible to create a continuous analog
to such a fusion power plant in a similar way to how Zubrin's nuclear salt
water rocket is a continuous analog to Project Orion? Our plant could then
be classified as an Inertial Confinement Fusion approach]