As Doug mentioned, it's not so easy to get 20% (20MHz centered at
98MHz) bandwidth and gain and "omnidirectionality". Of course, it
can't be truely omni and have gain: best you can hope for is omni in
the horizontal plane.

That said, the next issue is "best all around". What exactly do you
mean by that? How much gain is "good"?

If I were doing it and needed it to cover the whole 20MHz, I'd stack
however many full-wave vertical doublets I could manage and feed them
with a harness to keep the feeds in-phase. I'd separate them
vertically by enough to keep the mutual coupling moderately low and
perhaps experiment with that (or model it) to optimize the phase of
the mutual coupling. I'd probably make the doublets with fairly large
diameter tubing, or use a "bowtie" arrangement, to keep bandwidth
high.

If I was concerned with only one frequency, or a narrow band, I'd use
a coaxial collinear because it's relatively easy to build and mount
and feed and get working properly (IF you understand how it works).
That's assuming I wanted to use at least four elements; fewer than
that and it's probably not worth the effort.

But also beware that the vertical collinears will give you vertical
polarization and that may not be what you want!

Cheers,
Tom

(Patrick) wrote in message . com...
Hello all,

I would like to build the best all around omni-directional antenna,
that would have a good gain, in the FM broadcast band.

Any thoughts would be commended.

Sincerely,

Patrick Cambre