When I built the SVDA antennas for Clipperton/Midway/Kanton I used the
following procedure:
Erect a parasitic element with a length close to a reflector for the desired
frequency, just a little longer to allow for trimming (easier to trim than
lengthen).
Erect the driven element of approximately the correct length, again a little
longer to allow for trimming and isolated from the feed line.
Trim the driven element for resonance (0 ohms reactance) at the desired
frequency.
Lower the parasitic element.
Erect a sample element about 100' away from the driven element and isolate
it from it's feed line.
Starting 500 KHz below the target frequency I observed the AGC level at 100
KHz intervals to 500 KHz above the target frequency for normalized signal
levels.
Then I re-installed the reflector and repeated the frequency sweep to
determine at what frequency maximum gain occurred. When I had the frequency
of maximum gain I then adjusted the length of the reflector to move the
frequency of maximum gain to where I wanted it and repeated the sweep as a
check.
Then I removed the reflector, shortened it until it was almost as shorter of
the driven element as it used to be longer (as a reflector) and installed it
as a director.
Repeated the frequency sweep to determine where maximum gain was, adjusted
the length of the director element as required and repeat the sweep
as a check.
For the first array I then reversed the elements and repeated the sweeps to
determine where maximum front to back was.
After the construction was fairly well complete, I started over with the
sweeps to ensure that nothing has changed and I had a repeatable design.
Front to back was a curiousity, not a concern. When you are on an island
DXpedition you really want maximum forward gain and don't care about front
to back. I did find that front to back was not that great - to the point I
was wondering where the gain was coming from. I finally modeled the design
in EZNEC and found the signal was squished down and that (to me) explained
where the gain was coming from.
I found that the arrays could not be scaled from one band to another, each
band array required a fresh start. I concluded that was because of the
proximity of the lower end of the elements to ground. This distance was
always about 4' above ground. The idea was to get the tips as high as
possible and still be able to reach the tip of the director so I could
insert the strap that converted it to a reflector.
Because the measured gain was always higher with the reflector/driven
element configuration that was the way the antennas were used.
In operation we found the antennas worked as well off the back as the front.
Something about a salt water ground plane and being in a rare location, so
we never did change from the reflector/driven element configuration.
The feed point impedance was measured at about 40 ohms. After a hundred
feet or so of coax was added the SWR at our desired operating frequency was
1:1 at the amplifier end.
de Paul, W8AEF
ZF2JI/ZF2TA FO8DX/FO8PLA 8Q7AA ZX0A VU7RG TX5C A52PP
-----Original Message-----
From: David Gilbert
Sent: Friday, May 18, 2012 10:26 AM
To: towertalk@contesting.com
Subject: Re: [TowerTalk] 2 el yagi question
>From what I understand of the theory, a director is capable of slightly
(on the order of a tenth of a db) higher gain, but the tuning (spacing
and director length) is considerably more critical in order to achieve
that. I'd have to check, but it might even be that case that in order
to get it optimized you'd need to adjust the driven element length as
well. A reflector, on the other hand, gives it's best gain at wider
spacing and generally has a broader SWR curve, and also results in a
higher feedpont impedance for the driven element. I have no doubt that
the combination of those effects could make a reflector give higher gain
in practice.
Unless space (or boom length on a yagi) is absolutely critical, I'd vote
for the reflector.
73,
Dave AB7E
_______________________________________________
_______________________________________________
TowerTalk mailing list
TowerTalk@contesting.com
http://lists.contesting.com/mailman/listinfo/towertalk
|