Hi Alan, LF Group,
I found out why the pot setting was so critical. I measued the value for
best notch and it was about 330 ohms ! I have replaced the 25k with a 1k pot
and it is much better behaved. I have also added a dual gang 500pF airspaced
tuning cap in parallel with the existing caps (22n and 4n7) and that now
brings it down onto DCF39 nicely.

I set up the antenna this afternoon and have been experimenting with "real"
signals.
I can now reduce DCF39 from -54dBm without the filter in circuit, to -112
dBm with the filter. When a burst of modulation occurs the level rises by
10dB suggesting the frequency shift is fairly small compared to the size of
the notch.

I have continued to use the filter with the 12 : 1 toroidal transformer to
achieve matching, so I still have an overall reduction in signal from the
antenna due to the turns ratio. I generated some signals that could be
picked up by the antenna, and found that a signal at 136.562 kHz is reduced
by 23dB and one at 137.700 kHz is reduced by 27dB.

But the main thing is the filter does the job. I can set the SLM to 3.1kHz
band width and not be totally swamped by DCF39. So now I can leave Argo
running and see what is happening right across the band. For example I was
able to receive DJ9IE sending ordinary morse on 136.590 by spotting the line
appearing on the screen. And someone's sending some QRSS40 (?) on 137.800
and just now (23:15)I've got some more long dots or dashes around 137.800.
This is good. Before I had to tune around the band with the 80Hz filter to
find things and probably missed lots of stuff.

However getting rid of the worst interfering station does mean that one can
hear other flaws in the set-up. I can now hear a broadcast station very
faintly, and I guess it must be resonsible for some patterning on the
screen.

73
Hugh M0WYE