"J. McLaughlin" wrote

1. What is (are) the conventional name for this antenna? I have not
been able to find it in any of my references.

Commonly called a "ring-stub" element. It was one of the earliest methods
of adding v-pol to an omni h-pol antenna of the "ring" design, popular
before CP was authorized by the FCC.

2. When I model one bay of the antenna I find that almost all of the
radiation is vertically polarized. This is not at all surprising in
view of the symmetry of the horizontal element [the O].
What might I be doing wrong?

There is a considerable h-pol component from the portion of the element in
the horizontal plane. But as h-pol and v-pol do not have the same radiation
center from this design AND the horizontal part of the radiator generates a
v-pol component for elevation angles not in the horizontal plane, the axis
of net maximum v-pol radiation in a ring-stub element is not located in the
vertical plane.

The PDF slide show (Paper 10) at http://rfry.org includes NEC2 pattern
studies of this and three other types of FM broadcast transmit elements,
showing the differences in their free-space surface patterns without the
effects of the tower.

Speculation: I may not have the dimensions correct though the antenna I
modeled is resonant at 94 MHz. It is possible that venders are equating
omni-directional-in-the-horizontal-plane with circularly polarized.

I have an EZNEC file that I will send to anyone who wishes it.

The impetus for these questions is a determination of an appropriate
offset of these antennas from a large diameter tower.

Paper 6 at http://rfry.org shows some of the affects of two nearby tower
structures on sidemounted FM broadcast transmit elements for one set of
conditions. Unfortunately the patterns cannot be applied to any other
conditions.

The best approach is to have the patterns measured by the antenna OEM, using
one or more mounting configurations preferred or possible in the final
installation.

R. Fry