Howard K2HK wrote:
>Original Message Follows
>...I was able to find one _fixed_ setting of the tuner for which the
>SWR on the coax (and therefore the loss) was tolerably low across
>_all_ HF bands 80 through 10 meters, except 30 meters where I don't
>transmit much power anyway. Since then I've left the tuner fixed.
>Now QSYing is much faster....
>73 de Chuck, W1HIS
>
>This is becoming even more interesting !! I am looking forward to
>seeing your method. I wonder if it would apply to a shunt fed tower.
>Substituting a fixed value for a variable capacitor for example. I
>hope we hear some more of this.
>73, de Howard..K2HK
Here's an outline of my method. There's not much to it.
First, my antenna is a doublet 30 meters (102 feet) long, like a G5RV.
My tuners have been CLC "T" networks; these are highpass filters.
I start by adjusting them for a perfect match on the lowest band (80
m in my case). Thus, they are relatively transparent on the higher
bands. The lowestband match places two constraints on the three
component (two C's and one L) values; one degree of freedom remains.
(One does not have unlimited freedom here; tuner losses must be
considered.)
The length of the parallelwire line between the antenna and the
balun can be varied; this is another degree of freedom. A change of
just a few feet matters much more than you might think. After
changing this length you must readjust the tuner.
The length of coax between the balun and the tuner can be varied;
this is a third degree of freedom. Here again, you'd be surprised by
how much a fewfoot change matters. Again, after changing this
length you must readjust the tuner.
By varying these three free parameters you can achieve a perfect
match on a second band (e.g., 40 m) and can minimize but not
necessarily zero out the SWR on a third band (e.g., 20 m).
The first time I tried this method, I was lucky. I got perfect
matches at my target frequencies of 3.7 and 7.1 MHz (this much was
not luck), and a high but tolerable SWR of 4:1 on 20 m. Thirty
meters was terrible but my transmitter, coax, etc., could handle it
because I transmit only 200 W on this band. The SWR was 2.0 or less
across 17, 15, 12, and 10 m. I could transmit legallimit power on
every band (1500 W, except 200 W on 30 m) but I burned out my balun
(a commercial product rated at 5 kW) transmitting less than the legal
limit on 20 m.
When I tried the method with a different tuner, a different 4:1
balun, and a different commonmode choking arrangement, my luck was
not as good. I got perfect matches at my target frequencies of 3.7
and (in this case) 14.1 MHz (again, this much was not luck), but a
higher, less tolerable, 6:1 SWR on 40 m. Thirty meters was still
terrible; but my transmitter etc. could still handle it at the lower
power limit. The SWR was still 2.0 or less on 17, 15, 12, and 10 m.
I could transmit legallimit power all bands including 40 m (the
tough one now) because my new balun (homebrew, and good for 30 kW
with SWR = 1:1) was bigger than the old one, and my coax is Heliax
LDF550.
Additional degrees of freedom are available but I have yet to exploit
them because they require outdoor work in addition to turning knobs
and changing coax jumpers indoors. The length of my doublet (so far
fixed at 30 m) and the characteristic impedance of my openwire line
are the leading candidates to be varied; but, in principle, there are
infinitely many more, e.g., the ratio of the balun transformer, and
all sorts of transmissionline stubs and matching sections.
If I had accurate measurements of the feedpoint impedance of my
doublet antenna at all relevant frequencies, then I would let a
computer program do the simultaneous multivariable optimization.
Microsoft Excel could probably do it. The relevant transmissionline
and lumpedelementnetwork formulas are not too complicated, and I
have Excel macros for most of them already. Unfortunately, so far
all of my several attempts at antenna impedance measurement have
failed. Strong signals from nearby 10 and 50kW AM broadcast
stations have clobbered the detectors of the simple impedance bridges
that I've tried using, and the noise bridge that I've tried had
insufficient range and precision. I tried the conjugatematching
trick with a couple of antennatuners, but the accuracy I got seemed
inadequate.
Right now I see this antenna impedance measurement problem as the
Gordian knot, the key to achieving the Holy Grail of an allband,
notune, fullpower, efficient, antennafeedline system. (Forgive my
metaphors; I'm frustrated.)
73 de Chuck, W1HIS
