I haven't modeled a rototiller yet but I suspect it will have the same problem as Circularly Polarized beams have (at CB frequency).
The ground increases the gain of the horizontal component dramatically more than the vertical component making the signal more elliptical than circular..
The axial ratio quoted was 3 dB in free space
To my old eyes those half circle shaped elements are not flat, that is they are really sections of a helix.
Approximate diameter of the helix 135 inches if it is a 1/2 wave long half circle
That looks like a hard antenna to model let alone build, then try making identical ones to stack or phase.
Is this antenna reversible ? Left handed to right handed.
It sounds like it's not.
Don't you think that is caused/controlled by proper spacing and phasing of the stacked elements rather than the geometry of the elements?
Circularly Polarized antennas don't always scale well especially omnidirectional ones.
Wow! that's from 1986 they don't separate slanted linear from circular.polarization they lump it all together as mixed. They also accept as fact that they are unable to produce a true CP omni. 27 years is a long time in technology. Computers had barely impacted antenna design in 1986.
I'm sure it was relevant to 1986 technology but it's a bit dated to just trust verbatim 1986 minus another 27 years was 1959 just saying
If the transmitter and the receiver both use the same directional rotation of CP then multipath is almost non-existent because reflected signals are reversed in direction and attenuated about -20 dB. But could you explain how the receiver having a linear polarized antenna helps multipath?
DualAntennas: I'd agree with shockwave rescaling this to 11 meter and making clones to phase together is a very ambitious I would look to another design. If you're really wanting an omni CP antenna that you can try in a phased array I'd suggest instead the cycloid dipole at least you can correct the axial ratio problem.
It has simple straight sections and 90 degree angles so it's easier to model and build/clone.
I've already scaled one to 27.1 MHz like this one
in free space. Notice as shock stated total gain is less than a dipole and the vertical and horizontal component gain is really low at -
I've been interested in a four element square array like a super scanner +1 with 4 cycloid dipole elements instead of 3 standard dipoles and steered by phasing instead of reflection or direction, all 4 elements hot. 
The south and east elements are fed 90 degrees delayed separation is ~ 10 ft.
left hand vs right handed is intrinsic in the geometry.
