Jim,

Do you think this approach would be worth trying at the surface station of a 
cave radio link?

Regards,
John G3PAI

*********** REPLY SEPARATOR  ***********

On 17/07/2005 at 10:27 James Moritz wrote:

Dear LF Group,

Since the new Rugby Loran renders all except local 136k amateur signals
inaudible at my QTH, I have been trying some experiments to see what can
be done to improve the situation. I have used loop antennas to null the
noise from Lessay coming from a rougly SW direction for a long time, but
the direction to Rugby is roughly NW, so using a loop it is only possible
to null one Loran at a time - plus, with a null in a NW direction, the
other null is SE, which reduces many European signals too.

What is required is an antenna with two nulls which can be "steered" so
that the nulls have an arbitrary difference in angle between them, rather
than being fixed at 180degrees. A unidirectional cardioid pattern with a
single null can be obtained by summing equal signal levels from a loop
antenna and a vertical, as used in D/F. If the signal from the vertical is
attenuated, you get a pattern intermediate between a cardioid and a
figure-of-eight, i.e. an asymetrical "8" with one lobe smaller than the
other, and the null angles skewed away from 180 degrees, by an amount
which depends on the relative signal levels from the two antennas.

To implement this, I used one of the "bandpass loops" (see G3YXM's
"features" page). In principle, the TX vertical antenna could be used, but
at my QTH this picks up a lot of mains noise, plus it has a narrow
bandwidth which would cause the amplitude and phase to vary a lot across
the band, so I built a seperate "bandpass vertical", with a double-tuned
input similar to the loop, and a bandwidth of about 15kHz. Each antenna
has a preamp, and the signals are combined using an RC amplitude/phase
adjusting network not unlike a "noise canceller".

To set this up, I aligned the loop so its maximum direction was mid-way
between the bearings to Lessay and Rugby - about E-W at my QTH. Then I
adjusted the signal levels from each antenna individually so that the
Loran noise was about the same. Then both signals were added together ,
and the phase control adjusted for minimum Loran. Then amplitude, phase
and the position of the antenna were iteratively tweaked for the best
results.

The results are better than I expected - see the attached SpecLab spectra
- the audible Loran chatter is virtually eliminated , and only very slight
tweaking is required from one end of the band to the other to maintain the
null. As a bonus, it also nulls the "Watford Whistle" power line carrier
on about 136.645kHz, which is always annoying here. I was able to copy
DL1DQZ, DK7SU and F6BWO this morning, which with nulling disabled were
completely inaudible. It proved important to de-tune the TX vertical to
prevent it "sucking out" signal from the RX vertical - this is much more
pronounced than with the RX loop antenna. I think the antenna pattern
should give good coverage of Europe and Scandinavia and beyond, although
not very good for the western UK and GI/EI, or North America. But with the
very narrow band modes used for transatlantic DX, the Loran noise is less
of a problem. So once I get the current lash-up sorted out, with some T/R
switching, I should be back in business again :-)

Cheers, Jim Moritz
73 de M0BMU