I just think I had a light come on... or I'm deluding myself but it seems so simple I can't believe I didn't think of this before. Well honestly I think I heard some one mention before if they were considering the force on the antenna in the frustum when they were doing their simulations.
Here is what I am contemplating might be going on in the frustum that could possibly cause thrust.
I apear to be discussing here a TE mode - possibly TE01 or something similar. First we can divide the frustum with the right mode into 4 parts (center-magnetic_max, 1/4 lambda-electric_max and 3/4 lambda-electric_max). [We could possibly include the plates also as
radial magnetic maximums.] The bottom plate should have only a
radial magnetic field (1st magnetic maximum). A quarter wavelength up and the
transverse electric field should be maximum (1st electric maximum). 1/4 wavelength more up or 1/2 lambda (center of frustum) we have the 2nd
radial magnetic maximum. At 3/4 lambda near the top of the frustum we have the 2nd transverse electric maximum and last at the top we should have the 3rd
radial magnetic maximum.
The idea is that the ideal placement of the antenna to induce energy into the cavity is where the electric maximums are (and that the antenna is in the shape of that electric field {transverse or circular around the axis.}). At this precise location the entire
current in the cavity appears to be moving to make light that impedes the current in the antenna, so in other words, all the current in the cavity appears to be in phase at this location. This is a time retarded effect. In other words, there should be no force induced in the antenna at this location but it is the optimal location to input energy into the cavity.
In the 1st, 2nd, or 3rd magnetic maximum the time retarded appearance of the cavity
current is such that it appears to be out of phase above and below. Here, the antenna has difficulty inputting energy into the cavity but this is the optimal location for thrust. That is, if the current accelerates in the antenna near the magnetic maximum, the top of the cavity appears to have current moving in the same direction, while below the antenna the current in the cavity appears to be moving in the opposite direction. That is the antenna is repulsed from the bottom and attracted to the top or the opposite depending on phase relations.
Also by q(v x B) we see a force on the current loop.Optimally to maximize this thrust it would take 2 antenna. One located at the 1st electric maximum and the other antenna located in one of the magnetic maximums.
However, it is likely possible if the antenna is slightly out of place and resides in both the electric, and magnetic field locations that the antenna can both store energy in the cavity while also seeing current in the cavity as slightly out of phase. The idea is to, maximize the current in the cavity while also maximizing how out of phase the antenna sees the "time retarded" current.
I suspect this is similar to a phased array but
the fields are larger and the idea is the antenna in the magnetic maximum it is taking energy in the cavity and compressing it below/above the antenna. It depends on the phase relation so the force is reversible if with a single antenna, you have it slightly above or below the electric maximum. If the antenna in the magnetic field does take energy and compress it above or below the antenna, this may cause a thermal imbalance, but I am not sure what that means exactly.
Oh if 2 antennas are used they are 90 degrees out of phase just like a phased array but one resides in electric maximum while the other resides in magnetic maximum. The force on the antenna in the magnetic maximum should be purely relativistic, or that is the force from the magnetic field is relativistic in nature.
Another fascinating thing to note is the antenna pushing on the light in the magnetic field induces its own electric field. The magnetic part of the standing wave inside the cavity has two superimposed electric fields that cancel each other out so that only the B field exists at this location. When the antenna in the B field accelerates its current it cancels part of the superimposed electric fields so that they no longer cancel. As a result an electric field appears inside the cavity but this electric field is 1/4 cycle out of phase with the cavities normal electric field.