We can feed a vertical off-set by extending the vertical element and shortening the radials to raise the feed-point impedance while maintaining resonance.
The radials must be electrically isolated from a metallic mast (if used), and a high quality 1:1 current balun to isolate the coax shield from the antenna system, I'm using a Balun Designs 1115t.
Elevation Plot:
SWR Plot:
In this case making use of a 40 ft / 12 meter Spiderbeam pole mounted 10 ft / 3 m above ground on the 2nd level deck at the back of the house.
Modeling this using bare 12 AWG wire, the radial length to achieve resonance was 23 ft / 7 m. At least part of the reason my radials ended up shorter will be because I used plastic coated wire, and the vertical wire is attached to a fiberglass pole, but there may be something else having an effect..
However after trimming the radials for resonance the analyzer showed a perfect match:
The analyzer hung off the balun with a 6" jumper, since the SWR didn't change after connecting the feed-line is an indication the balun is doing its job.
The SWR response turned out to be more broad-banded than predicted according to the IC-7300, it covers the whole 40m band only rising to 1.5:1 at 7.3 MHz (feed with 70 ft of LMR-400):
Photo of the antenna:
The longer radiator moves the current maximum slightly higher above surrounding clutter.
One radial runs out near the end corner of the house, one out to a tree in the back yard, and one along the under side of the deck, this also allowed mounting the balun out of the weather. With the shorter radials I was easily able to maintain ideal symmetry.
Compared to a 1/4 wave ground plane model same height above average around with four radials, the 1/3 wave vertical peak gain is 0.2 dB less but one degree lower at 22 degrees.
Some or more of that 0.2 dB might be gained back with the higher current maximum in the real world environment with ground clutter, and broader 1:1 SWR vs 1.3:1 or higher with a ~36 ohm 1/4 wave. Either way a negligible trade off (if any) for a much more convenient radial configuration and 50 ohm feed. Field strength readings would be the next step.
Noise floor is not too bad at S4 to S5, compared to S3 to S4 on my 130 ft doublet at 55 ft height.
We'll see how it works out :-)
Model file: GP-40m-40ft.