I wrote:
>>Interestingly, in the W9LT/K3LR-design lazy-vee parasitic array, the losses
>>in the floating feedlines of the non-driven dipoles (which are
>>open-circuited stubs) actually improve F/B ratio substantially (though at
>>some cost in gain). You can see the change dramatically if you model the
>>feedlines as transmission lines (which are lossless in NEC-2) and then
>>replace them with loads of the appropriate X and R values.
>
At 09:06 AM 1/18/99 -0600, WE9V Chad Kurszewski wrote:
>Ahh, but if you model them correctly, you won't be as impressed!!
>
>I was working on my own version of the above antenna for 80M and using
>a perfect X value (0 + j50 ohms). Antenna looked totally kick
>butt. Then, because I couldn't use a 1/8th wave open stub, I
>went for the 3/8ths wave open stub. Using TLA.EXE (by N6BV),
>and using RG-8X as the transmission line, the actual impedance is,
>
>are you ready?
>
>
>5.62 + j49.69
>
None of this is a surprise. In fact, Chad's procedure is exactly what I
did. Then I split the resistive and reactive components of the stub into 2
equal loads, placed as close as possible to the junction of the 2 halves of
the vee. This is necessary with NEC-2 because you can't use either a
single source or a single load at the juncture of 2 non-parallel wires.
There is a special trick with this antenna. Good F/B is very dependent on
the dipole elements themselves being long for the operating frequency.
Only 100 kHz below resonance, most of the F/B goes away, while best F/B
(and only slightly diminished gain) is found well above resonance.
Logically, this should mean that longer feedlines (longer than the 135
degree standard) would enable operation closer to resonance with good
pattern, by making the reflectors electrically longer. In my case, since
my linear is quite forgiving, I chose to go with the standard length and
resonate the antenna below 3500 kHz. The resulting antenna has good F/B on
CW DX, typically 2-3 S Units, as well as useful gain. The F/B holds up to
3800 KHz at least. Even nicer, on some higher-angle signals the F/B is 5-6
S units -- not a bad place to have a null for an East Coast station in DX
contests.
Of course, modeling and reality don't necessarily converge. In my case,
the bottom halves of each dipole, rather than describing a nice straight
line back to the base of the tower, go through more of a quarter circle,
because the elements aren't at quite 45 degrees from the vertical and my
tower is a little short for the job. The curved elements are almost
impossible to model well. But it does play, as contest scores demonstrate.
73, Pete Smith N4ZR
n4zr@contesting.com
Loud is.
--
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
|