I remember modeling the Imax with the capacitor for the first time. The model worked well with the capacitor right where it is on the Imax, and the capacitance as measured from an actual Imax capacitor.
As I recall, position was very important, as moving it to far in either direction on the model would drastically reduce the phase shift effect of the capacitor. I presume that much like coils, which are current devices, have far more of an effect when near a current node, the capacitor, a voltage device, seems to need to be near a voltage node for best effect.
Another thing I noticed with said models is the feed point impedance was very close to a half wavelength antenna, and so I figured that part of the goal was to use the same matching system as they used on the a99.
The phase change effect, and the feed point impedance are why I have, on multiple occasions, called the Imax an extended electrical half wavelength antenna in the past.
Like in the .pdf bob posted, I have modeled EDZ antennas with this setup, and said model had more gain than a standard EDZ antenna, and didn't have the two minor lobes near the main lobe.
The DB