Wimpie wrote:

So you will end up with a high Z transmission line with significantly
lower propagation speed. A difficulty is: where is the return
conductor? You can assume the return conductor about 0.125 lambda
away when the influence of the ferrite is not that large. Very close
to the feedpoint, the impedance of the ferrite loaded line is strongly
dependent on the distance from the feedpoint.
. . .

The second conductor of the transmission line is the ground plane.

When you would take a long thick ferrite tube, the field will even not
come out of the ferrite because of the attenuation in combination with
the low EM wave propagation speed inside the ferrite. Ferrite
absorbing tiles also make use of both the epsilon.r and permeability.

So finally I did not answer your question about how to determine the
feedpoint impedance. Probably you will need full 3D EM software that
can handle 3D structures of different mu and eps. As far as I know,
all momentum based EM software cannot handle this.

I don't believe that's necessary. As you say, the field doesn't escape
the ferrite. This means that the E field between transmission line
conductors (the whip and ground plane) is zero. Therefore, the impedance
of the line (E/H) is infinite. The loss of the ferrite is adequate to
suppress any reflections from the end of the line. A reflectionless,
infinite impedance line will have an infinite input impedance. This is,
as it should be, the same result I got from a somewhat different
perspective.

Roy Lewallen, W7EL