The general consensus seems to be that attempting to use odd-1/4 wave
multiples to feed a tribander stack would be impractical (no argument there)
and would add very little if anything.
One person did firmly support the premise that it was best, with
monobanders, to follow this practice, but thought that the advantage was
not large. Another reported that his 15M stacks seemed to work
significantly better with half-wave multiple feedlines than with 5/4 WL.
Just for fun, I built a simple model in EZNEC, using NBS 3-el yagi models
stacked .75 and 1.5 WL over MiniNEC ground. Using 2 separate sources,
in-phase and with equal currents, produced the expected pattern. Examining
the source data revealed that there were very small differences in
feedpoint impedance between the two antennas, presumably related to the
different distance above ground. The deviation was ~.1 percent.
Then I "constructed" two theoretical 50-ohm transmission lines, one feeding
each antenna, and connected them together at the far end. A single current
source was placed at that point. First I tried 1/4 WL transmission lines.
The pattern was identical to the separate sources. Then I tried one .33
and one .66 WL, simulating a badly dissimilar feedline length situation.
As expected, the pattern was badly distorted and severe gain losses
resulted. Finally, I tried the .33 WL feedline length for both, and the
pattern returned precisely to nominal, just as if the two antennas were
being fed without the transmission lines.
I'm sure I haven't exhausted all the potential sources of error in this
exercise, but taking it together with the anecdotal information, I feel
fairly comfortable dispensing with the 1/4-wavelength constraint.
Of course, if the feedpoint impedance of the antennas is other than 50
ohms, whatever length of 50-ohm coax you use *will* transform that
impedance to something else at the other end. I suppose that could,
inadvertently, transform the antenna feedpoint impedances to an awkward
value for the Stackmatch to handle. To eliminate that variable, you could
use an electrical half-wave from Stackmatch to antennas.
Thanks to N8UG, K5TR, W2XX, K3ND, NQ4I, K2AV, and N5CQ for their comments.
73, Pete Smith N4ZR
n4zr@contesting.com
Sometimes a tower is just a tower
--
FAQ on WWW: http://www.contesting.com/towertalkfaq.html
Submissions: towertalk@contesting.com
Administrative requests: towertalk-REQUEST@contesting.com
Problems: owner-towertalk@contesting.com
Search: http://www.contesting.com/km9p/search.htm
|