LB,
Many, many thanks for your assistance! I appreciate you bringing your time
and expertise to bear on my question.
>Steve,
>
>Your situation of wanting a compensating capacitive reactance to reduce
>SWR at 3.8 MHz for a resonant 3.5 MHz dipole has several possibilities.
>For your problem, I modeled a resonant 3.5 MHz dipole at 50' up and then
>checked the impedance at 3.8 MHz, getting 94 + j155. These numbers must
>be considered ballpark, since the dtails of your precise antenna were
>lacking. However, they will get us started.
The complete description of the antenna system is this:
Three vertical dipoles suspended from the top of a 90' Rohn 45G tower. Each
is pulled out from the tower at a 45 degree angle by means of a line
attached to the dipole feedpoint and the lower legs of the dipoles are
pulled back in toward the base of the tower (sort of an invertedvee on
end). The dipoles are spaced 120 degrees apart around the tower, 60 degrees
from the tower guywires. The dipoles are #13 AWG insulated copperclad, and
are cut for resonance at 3.5 MHZ. Tower guywires are 3/16" EHS broken up
with insulators (I can furnish the guywire segment lengths if that is
helpful).
The feedlines for each of the dipoles extend back to the tower horizontally
from the dipole feedpoint, and extend back down the tower part way, to a
relay box that allows selecting one of the three dipoles to be fed  the
other two dipoles and their transmission lines are "floated", with both
center conductor and shield of the coax open. The lines from relay box to
dipoles are cut to be 153 electrical degrees at 3.5 MHZ so that the unfed
dipoles are loaded by the trnsmission line to act as reflectors.
/\
/  \
/  \
/  \
Side view (only 2 of 3 dipoles /  \
shown /  \
/  \
/______________\ feedpoint of dipole
. \  / .
. \  / .
. \  / . dacron line
. \  / .
. \  / .
. \  / .
. \  / .
________________________\/______________________
>From the relay box a single 1/2" hardline feedline continues to the shack,
along with the relay box control cable, some 160' away. Hence, my thought
to switch in and out capacitance by means of a relay instead of tuning via
a variable cap (Although, my control cable is a 6 conductor, and I only
need three conductors to switch antennas. . . I could rig up a reversable
motor and have a remotely tuned 80M antenna. . . I think that's a project
for another day)
>At the 166 degree point (about 78.8' of 0.66 VF 50ohm coax) the
>uncompensated 3.8 MHz impedance has been transformed to about 30 + j88
>ohms. At this point, a series capacitive reactance of j88 ohms will
>compensate, leaving the 30 ohms resistive.
I haven't attempted the calculation yet, but I am assuming I can determine
what j88 ohms is by the formula:
Xc = 1/2pi x f x C or, C = 1/2pi x f x Xc
What are the "units" associated with the imaginary number j88 ohms?
Simply ohms?
>
>However, switching a fixed capacitance may be more complex than you need,
>if you can obtain a hefty variable capacitor. Place it in a good
>(shielded) box with coax connectors. Isulate it well from everything (box
>and its panel) and connect the coax connectors to the plates to place it
>in sereis with the line. Now you can select the capacitance you need to
>compensate for inductive reactance anywhere in the band.
I am beginning to like this idea (with a motor drive) more and more.
>
>This 1component ATU works if the line presents the terminals with
>inductive reactance. Many hams in the past have cut their dipoles a
>little long, even at 3.5 MHz and then used 180degree or multiples thereof
>at the lowest frequency in the band (3.5) MHz to ensure that the line at
>the capacitor box always shows an inductive reactance of a magnitude that
>the capacitor can handle. Your 155166 degree line appears adequate to
>ensure this in your case.
I derived the 166 degree feedline length from relay box to antennas by
proportion: if the feedlines are 153 degrees at 3.5 MHZ then they are:
3.8/3.5 x 153 degrees (at 3.8 MHZ)
>
>Remember that the max capacitance of the capacitor you use determines the
>lowest reactance you can compensate for, while the minimum capacitance
>determines the highest reactance. Since impedances are low, you do not
>need a 7kV rated capacitor, but something at 1.5 kV seems wise.
Thanks, I was concerned with the voltages, because I knew there would be
some VSWR, but it would be as you've said, moderate to low. Can you suggest
a range of capacitance for a variable capacitor that will span the 80M band
with the antenna and transmission line parameters as given?
>
>Remember that you still have a resistive component other than 50 ohms, so
>your will still have some SWR (about 1.6 or so for the case you set up).
>However, losses in coax at 80 meters are relatively low even with SWRs
>above 2:1 (which the line will have at 3.8 MHz).
If there is any real concern about my solid state finals (Ten Tec Omi VI+
with or without the Hercules II solid state amp), I have the Ten Tec model
253 Autotuner that I can use at the shack. The feedline from relay box to
shack will be 1/2" 75 ohm hardline CATV surplus), so that is a source of
some mismatch, too. I had planned to use a multiple of a halfwave at 80M of
the hardline and make up the rest with 50 ohm RG213, but that may not be
worth the trouble if the VSWR is pretty low at the relay box. I may just
use all hardline from shack to towermounted relay box and save a few
connectors and some RG213 for another project.
>
>The equations for calculating the impedance along a transmission line are
>fairly complex and difficult to reproduce in ASCII, even for lossless
>lines. However, there is an item at my site with essentially the same
>program as in HAMCALC that includes a BASIC listing from which one can
>view at least one way of writing the equations for voltage, current (and
>phases angles) and R and jX along a lossless feedline for any Zo. It is
>possible to use the R and jX values at the shack end of the line, and with
>appropriate sign changes calculate the antenna impedance as well. If
>memory serves, Walt Maxwell's book has such equations and a Basic program
>for that direction.
I'll look at my copy of Maxwell's REFLECTIONS (should have thought of
that!). At least your comments about the complexities of the calculations
make me feel like I'm not a total dummy.
>
>Future versions of HAMCALC will contain a group of programs drawn from
>Peter Dodd's work that show a simple way to measure impedance at the shack
>end of the line with voltmeter readings and then to calculate back to the
>antenna impedance. Murph is working on this for Version 31, which I have
>reviewed in advance for him. Unfortunately, the Canadian postal workers
>strike prevents me from mailing my suggestions.
>
>I hope these notes are useful to youand to others who might have
>forgotten the old timer's way of stretching the SWR bandwidth of an
>80meter dipole with a series capacitor at the shack end of the line. It
>will also work for 160 meters, with the right dipole, capacitor, and
>feedline length range.
>
>73
>
>LB, W4RNL
Once again, thank you, LB. I am more of an electrician than an electronics
technician by trade and training (almost all of what I know about
electronics is selftaught from the amateur radio texts I've bought). I
appreciate you taking the time to give me a lesson in basic theory. I will
definitely track down the software you mentioned!
Steve Zettel KJ7CH
near Libby, MT USA

FAQ on WWW: http://www.contesting.com/towertalkfaq.html
Submissions: towertalk@contesting.com
Administrative requests: towertalkREQUEST@contesting.com
Problems: ownertowertalk@contesting.com
Search: http://www.contesting.com/km9p/search
