Dear LF Group,
The 137kHz ferrite rod RX antenna I have been experimenting with in the last
few weeks has proved to have more than adequate sensitivity, but the narrow
<1kHz bandwidth is a problem. It is necessary to re-tune when changing
frequencies within the band, also the power-supply type ferrite used seems
to have a high temperature coefficient, which causes considerable centre
frequency drift with changes in ambient temperature. Since it is usually
desireable to have the antenna some distance from the receiver, and remote
tuning is a nuisance, a wider bandwidth of about 5kHz would be useful.
The bandwidth could be increased by adding a damping resistor to reduce the
rod antenna Q , but unfortunately this also reduces the SNR. Approximately,
the signal voltage is proportional to the Q, but the thermal noise voltage
produced by the loss resistance is proportional to sqrt(Q), so the signal
level decreases more quickly than the noise level as the Q is reduced. For
my rod, the desired bandwidth increase would result in about 7dB increase in
noise floor, which would be marginal. What is needed is a loading resistor
with reduced thermal noise (and without the need for a liquid helium
supply...). As DF6NM suggested, one way to do that is to use a preamplifier
with a feedback network that defines its input resistance to provide the
loading.
The preamp circuit in the attachment achieves this by using a shunt feedback
resistor around an amplifier with a well-defined inverting voltage gain to
provide the load resistance. The equivalent input resistance is R(shunt) /
(1-A); in this case gain A is -10 and R(shunt) is 100kohms, giving Rin about
9k. However, the noise voltage at the input caused by R(shunt) is also
reduced by about 1/(1-A), and is about 10dB less than if a 9k resistor was
connected directly across the input. The same principle is used in the
lower-noise types of "modamp" gain block ICs, and also in my "bandpass loop"
50ohm preamp - it gives quite a big reduction in resistor noise provided
voltage gain >>1.
Compared to the original ferrite rod antenna circuit, the result is that the
bandwidth is now increased to about 5kHz, and a preset tuning adjustment is
adequate. The output level is a few dB higher than the original unity gain
buffer circuit. The internal noise level of the rod + preamp combination is
increased by a few dB, but is still well below the external band noise, even
with the quiet band conditions existing yesterday morning. So the idea
certainly works. Some circuit improvements are possible though, I think,
i.e. reduce the current consumption for /P operation, the gain is currently
too high for a sensitive receiver, etc. The basic idea is quite adaptable to
different gain, impedance levels, frequency and so on.
Cheers, Jim Moritz
73 de M0BMU
Ferrite_wide.png
Description: PNG image
|
