Re: NEC-LIST: calculating antenna bandwidth

While an antenna is sort of like a tuned circuit, with a bandwidth, I don't
think the bandwidth is a function of stored/lost energy, as in a RLC
circuit with a Q.

Bandwidth is also a function of the source/driving impedance. Say my
transmitter were a pure current source with infinite compliance. The
antenna's radiated power will be, by and large over a limited frequency,
proportional to the drive current(less any losses in surroundings and
elements, of course). The reactive "power" circulating between antenna and
transmitter will aggravate resistive loss in the feedline (or non-ideal
Tx), but this is a problem that power distribution engineers face every day
when feeding generally inductive industrial loads like motors.

An antenna, driven from a resistive source (i.e. a 50 ohm transmitter),
might, in fact, be an unmatched load, and reflect some power back, to be
absorbed in the source, but is that really a "bandwidth" issue, or more one
of impedance matching. With an antenna tuner made up of fairly low loss Ls
and Cs, acting as an impedance transformer, the apparent bandwidth (in
terms of radiated power vs frequency, for constant voltage (!) on the
idealized source) might be very narrow.

But, the bandwidth of the antenna, in terms of power radiated vs power
applied to the feed point, will probably be quite wide (although, the
"bandwidth" (calculated this way), in a particular direction, might change
dramatically).

For a simple antenna, you could probably use one of the approximations for
feedpoint impedance based on the sine and cosine integrals. Run NEC for a
few points to "fit the curve" (i.e. find the parameters), and then
interpolate. Another approach, based on the observation that simple
antennas have Smith chart plots that are smooth curves, is to just run the
calculation at a few points, fit a spline or other useful curve to the
curve, in Smith Chart coordinates, and then look for the intercepts with
the appropriate VSWR circle.

For what it's worth, you can make an extremely broadband antenna (in the
relatively constant feedpoint impedance sense), with low actual loss, with
an appropriate geometry... Here, the apparent "Q" would be low, but there
isn't any loss.
Such an antenna might not meet some other criteria (simplicity,
size). Equiangular spirals or other frequency independent configurations
spring to mind. Horns and biconical antennas are also possibilities.

>Is there an easy way of calculating antenna bandwidth in NEC2
>
>
>Currently I am using my Mathematica interface and the FindMinimum function to
>numerically find the point that the antenna port S11 goes to the desired level
>(-1, -3, -10 dB). It does this by repeatedly calling NEC to calculate the
>port Z for different frequencies. It is not very bad. For my simple loaded
>dipoles I can find the upper and lower band edge in 10 to 20 seconds.
>
>It seems to me that it should be possible to calculate the bandwidth from just
>the fields at resonance.
>
>One way would be to evaluate (Stored Energy/ Lost Energy) in the field. I can
>calculate the near field, but it seems that I need to integrate through the
>volume of all space. I suspect that this could be done by repeated calls to
>NEC to calculate the Near field at different R, but that seems almost as bad
>as my numeric FindRoot algorithm.
>
>Perhaps I need only the antenna currents.... or the currents and the segment
>charge.
>
>I am sure I can figure this out, but Has the Wheel already been invented?
>
>
>What is a good way to get antenna bandwidth out of NEC2 ?
>
>
>Wayne Shanks
>--
>The NEC-List mailing list <nec-list_at_gweep.ca>
>http://www.gweep.ca/mailman/listinfo.cgi/nec-list

Jim Lux
Spacecraft Telecommunications Equipment Section
Jet Propulsion Laboratory
4800 Oak Grove Road, Mail Stop 161-213
Pasadena CA 91109

818/354-2075, fax 818/393-6875

-- 
The NEC-List mailing list <nec-list_at_gweep.ca>
http://www.gweep.ca/mailman/listinfo.cgi/nec-list