> Being an analytical guy I modeled the antenna system in both
> Ansoft Designer and an antenna modeling program. In Ansoft
> Designer the antennas were modeled as a series RLC load with
> 50 ohms at resonance and a Q of 15. Transmission lines were
> modeled with VF equal to that of RG-58A/U coax. In the
> antenna modeling program the antennas were modeled as 1/4
> verticals with transmission lines having VF=0, antenna
> spacing of 1 meter, and perfect ground (assuming auto metal).
> The Motorola note suggests a minimum spacing of 3 feet
> between antennas.
I think one minor detail you may have overlooked is that the diplex document
specifies the use of base-loaded antennas, most likely of the design that
Motorola was manufacturing at the time. They may have been shunt-fed (cold
side of the base loading coil grounded for static/noise supression, coil
tapped, possibly with a series cap from the tap to coax center). My guess
is that if you could model the actual Motorola antenna (if you knew how it
was constructed) you would find that it appears to have a much higher
feedpoing Z at the "wrong" frequency as compared to modelling it as a true
linear end-fed quarter-wave element, thus a half-wave cable from the T to
the opposing antenna would be the way to go in this diplex arrangement.
Also note that the Motorola doc says that an increase in VSWR as seen by the
radio up to 2:1 is acceptable, i.e. it would be overly-optimistic to assume
you're going to get total decoupling of the two halves of the antenna
system.
--- Jeff WN3A
