I was frustrated that the ARRL design never worked right, and quite a
few years ago thought through the problem pretty carefully. Since
then, I've been able to build them "at will" since I know what's
needed now. I've posted this in the past, and you can probably find
it in the Google archives. Keys, as with essentially any antenna, are
to arrange elements to get the pattern you want and then feed them
properly, and decouple from other conductors (esp. the feedline). In
a coaxial collinear, the feed is tied closely to the elements. If you
feed between each adjacent element pair with in-phase equal-amplitude
voltages, even though the elements are shorter than 1/2 wave, the
currents in the elements will be very nearly in-phase (though of
differing magnitudes). It's fairly easy to simulate in EZNEC to see
what happens. The "end-fed" half-waves show high feedpoint impedance,
but the coax puts them all in parallel, and what I see with typical
ten element antennas is a net ROUGHLY 150 ohms with some reactance,
easily matched with an "L" network to 50 ohms.

Though I prefer using foam Teflon dielectric line (which has about
0.82 VF), the coaxial collinear DOES work with solid polyethylene line
with 0.66 VF. I prefer the foam Teflon because it's really nice to
solder, more than anything else, and an 1100' roll was very cheap...

Others have posted (on web sites) similar findings, with designs
similar to what I've used. You will find people who tell you you
can't build one, but I know better: I've simulated them and built
them, and the performance as-built matches what I expect from the
simulation.

Cheers,
Tom

"Dave VanHorn" wrote in message ...
can anyone point me to an article that has the details right, for vhf/uhf
use.

i don't understand how the velocity factor of the coax used in the elements,
works in this antenna.