I see that they have a 4.8/5 rating on eham.net. I will definitely look
into them.

From the picture, it looks like one of the poles has a SO239 connector,
but the others don't. Are
the two other poles passive?

The Arrow is an "open-sleeve" J-pole design. The other poles aren't
actively fed, but make up a critical part of the antenna design.

A standard (traditional) J-pole is an extended zepp. It consists of a
half-wave end-fed radiator, connected to a transmission-line matching
section (nominally a quarter-wavelength). The matching section is
usually made up of an extension to the radiator, and a second element
located an inch or two away.

The matching section is usually shorted at the bottom, and fed at a
tap point an inch or two above the short... a point at which the high
impedance of the end-fed half-wave radiator has been transformed down
by the transmission line to something close to 50 ohms.

An "open sleeve" J-pole (which Cebik calls a "nontraditional" J-pole
on his web site) is a bit different. It's somewhat shorter (the long
arm is less than 3/4 wavelength), the short arm is longer than 1/4
wavelength, and it's fed differently (at the bottom, with one arm
being grounded and the other being insulated from ground and fed
directly from the feedline). Cebik's website has a very nice
discussion of it.

The Arrow OSJ is actually two open-sleeve dipoles, in parallel. The
longest arm and the shortest arm are grounded, and the medium-length
arm is connected to the hot side of the feedline.

On 2 meters, the long and medium arms form an OSJ and are active...
the short arm is simply a grounded stub which carries very little
current and doesn't radiate much.

On 70 cm, the medium and short arms form an OSJ and carry most of the
current. The long arm probably carries a small amount of current, but
since it's a couple of wavelengths long its radiation resistance is
high and its current flow and radiation are only a small percentage of
the total.

What's the radiation pattern?

Very similar to any other half-wavelength radiator... 0 dBd, plus or
minus a few fidgets.

The Arrow OSJ is not actively decoupled from the feedline (there's no
balun) and thus there may be a small amount of RF current flowing on
the outside of the feedline, and some amount of disturbance of the
pattern. The amount of current on the feedline will depend on the
details of the individual installation (e.g. the length of the
feedline). I don't think the effect will be significant in most
installations; reflections from the ground, building structures, etc.
are likely to alter the pattern just as much or even more.

You may, if you choose, mount the Arrow on an insulated mast, and
include some sort of choke in the feedline (e.g. a small loop of a few
turns in the coax, or some ferrites) in order to reduce the effect of
feedline radiation.

I think that the Arrow is a good performer and a good value. You can
even make 'em yourself if you wish - Arrow sells the parts
individually and (last time I looked) had complete construction
dimensions and plans somewhere on their web site.

--
Dave Platt AE6EO
Friends of Jade Warrior home page: http://www.radagast.org/jade-warrior
I do _not_ wish to receive unsolicited commercial email, and I will
boycott any company which has the gall to send me such ads!