> I avoid antenna + feedline lengths that terminate midway
between a high
> voltage point and a high current point on the feedline.
Such a length
> presents a highly reactive load that is very difficult to
tune to resonance
> while maintaining an efficient transfer of power, hence
the switchable
> additional 1/8 wl of feedline. When I first tried the
antenna at the odd
> 1/8 wl, I could arc over a 7 kv tuning capacitor with only
about 200 watts,
> and although I could get the swr down to unity by tapping
down the coil to
> attach the feeders, an rf ammeter in one side of the
feedliner indicated
> substantially less current than I got at that same spot,
running the same
> input power, with the additional 1/8w switched in, and
simple parallel
> tuning. Ever since, I have avoided odd 1/8 wl's like the
plague.
Hi Don,
Antenna tuners are more complex than we often credit them
for. Any particular tuner configuration has pros and cons.
There is no universal "best" network. Because a particular
configuration "acts up" in one situation, a change to make
it work with that tuner configuration will not apply
universally. So we have to use general rules to fit most
situations.
In your case I'm sure it wasn't the odd 1/8th wave that was
the problem, but rather the network configuration you had in
combination with that length. The very fact you tapped line
down on the coil to obtain a match indicates the impedance
and voltage on the line terminals at the tuner was modest,
even though it contained a reactive component. Otherwise the
tap point would have been well up the coil.
What happens with a link coupled parallel network is the
operating or loaded Q and network voltages across the
capacitor is set by the impedance loading the capacitor and
the turns ratio between the link and the capacitor. In other
words the network loaded Q and tank voltage is affected by
ratio of turns between the link and the tank coil, the
location you tap the feeder, and the overall number of
turns. When the feeder tap is moved down without changing
turns ratio between the link and the tank, we have all those
extra unneeded turns ratio doing nothing but stepping up
voltage to the capacitor. The complex impedance of the line
was likely just the correct value to maximize network Q and
cause enough voltage to arc the capacitor.
My Johnson Matchbox was fussy like that. I added a second
link tap so I could double the number of turns in the link
or use the original link position for certain load
combinations. Eventually I just got rid of the KW Matchboxes
in my system and went to T networks with good 1:1 ratio
choke baluns on the output. The T networks are not so load
impedance critical as the Matchbox.
In every single tuner manual I write I warn people NOT to
use a 1/4 wl feeder (or multiple) to feed a 1/2 wl antenna,
and suggest an odd-1/8th wave feeder on the 1/2 wl band of
the antenna. That's because overall *most* problems possible
come from the 1/4 wl feeder 1/2 wl antenna configuration.
Using the odd 1/8th wave line with 450 ohm ladder line
provides a modest impedance of about 400R + or - 400J on the
lowest band (where the antenna is a 1/2 wl dipole), about
20-100R with nearly J0 on the second harmonic, and so on.
While we can build a tuner that doesn't like those
impedances on some bands, overall they are closest to the
center of a Smith Chart on all bands and almost never a
problem with conventional networks.
Using a 1/4 wl line puts us along the zero reactance line,
but as far from Smith Chart center as possible in the high
impedance direction (highest feeder terminal voltage) on the
lowest bands. A parallel network with a link indeed might
work well with that (especially with a high link turns to
tank coil ratio), but virtually every other balun and
network combination wants to be closer to the center of the
chart like the odd 1/8th wave line likes to be.
73 Tom
_______________________________________________
Topband mailing list
Topband@contesting.com
http://lists.contesting.com/mailman/listinfo/topband
|