Topband: Topband Puzzle

[Tod -ID]

In the ARRL 160 meter Contest this year I used an antenna that I have
experimented with off and on during 2008. It has a vertical element 57 feet
high insulated from ground. The feed point is at the base of this vertical
element. At the 57 foot point a 55 foot horizontal wire is extended as a
'top hat'. Very much like an inverted L you might say.

I added an 80 meter trap in the horizontal wire -- at a point 22 feet from
the connection to the vertical element. The trap is made from a 30 uH coil
and a 60 pF capacitor in parallel. This combination provides excellent
matching for the 3500-3660 kHz part of the band and good matching for the
1815 to 1845 or so. 

I found an interesting effect when I used the system with high power [1.0 to
1.5 kW]. After a very active operating period [lots of answers to CQ's] the
power output from the amp began to drop. Eventually, the power dropped to
about 500 watts. At the 500 watt level there was some noticeable mismatch
which I presume caused the setting of the plate tuning circuit to be
incorrect and as a result the power output dropped. If I paused a bit the
power level when I resumed was close to the initial power. 


After fighting this on Friday night I removed the 80 meter trap and replaced
it with a large AirDux coil with the same inductance. I figured that maybe
the capacitor in the trap [it was not mica or vacuum or air dielectric] was
experiencing dielectric heating and that might affect the input impedance of
the antenna.

Saturday night I saw the same effect as I did on Friday night. At this point
I am speculating that the coil is heating sufficiently to cause changes. The
coils are about 3 inches long, 2 inches in diameter and use bare #15 wire.
The trap coil is wound on an epoxy type grooved form. The AirDux coil is
essentially air wound with small spacers. 

Anyone else have a similar problem historically? Anyone feel they know with
certainty that a resistance change in a coil will cause this problem? Is
there a chance that the inter-turn capacitance is changing significantly and
detuning the coil?  Any other ideas that might help understand and solve
this?

Clearly I can discard the trap and do the matching at the base, but I would
rather use the trap because of the simplicity. 

The "good" news is that the temperature is dropping and by the Stew Perry I
can expect -10 C which should give longer CQ runs before the power drops.
There is no similar problem on 80 meters.

Tod, K0TO