Content-Transfer-Encoding: 8bit
Message-ID: <5.1.0.14.0.20020304114102.00aa6498@gemini.herts.ac.uk>
Date: Mon, 04 Mar 2002 12:40:54 +0000
To: rsgb_lf_group@blacksheep.org
From: "James Moritz" <j.r.moritz@herts.ac.uk>
Subject: LF: Noise Cancelling RX antennas
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii; format=flowed
Precedence: bulk
Reply-To: rsgb_lf_group@blacksheep.org
Sender: <majordom@post.thorcom.com>

Dear LF Group,

As I mentioned a few weeks ago, I have been having problems with local 
mains-generated QRM from an unknown source which made it impossible to copy 
the weaker signals on the band. Over the weekend I tried out a noise 
cancelling system which seems to more or less solve the problem.

I first tried using various antennas I already had as "signal" and "noise" 
antennas, with the HF type phasing circuits that have been around for a 
while. I found that it was possible to null the noise, but the bandwidth of 
the null was at best a few 100 Hz; changing frequency or re-tuning the 
antenna meant complete re-adjustment of the phasing network was required, 
which was not easy because the noise level varies rapidly and  makes it 
hard to tell if the null has been found. So the system was not very practical.

Thinking about it, the problem compared to HF is that the LF antennas and 
phasing networks are relatively very narrow band, and so phase and 
amplitude of the signals changes rapidly with frequency. In order to get a 
null, the noise signals from the 2 antennas must be equal in amplitude and 
180degrees out of phase when the signals are combined, therefore a null is 
obtained only over a narrow bandwidth. I decided the solution would be to 
use two identical, matched, loop antennas with fairly wide bandwidth, in 
combination with differential phase shift networks (similar to those used 
for the LO in a phasing SSB exciter), so that the whole system has a 
reasonably constant relative phase over a considerable bandwidth.

I made 2 tuned loop antennas with 2m sides from 15mm copper water pipe. 
These have low noise pre-amps and a bandwidth of about 30kHz, and by 
themselves have more than adequate sensitivity to hear the external noise 
in a quiet band. One is located at the end of my garden nearest the noise 
source, while the "signal" antenna is positioned at the opposite end of the 
garden. Both loops are oriented to null the Loran signal form Lessay, which 
at the same time gives quite good Europe/NA coverage at my QTH. The signals 
are each passed through RC phase shift networks tuned by a dual gang pot, 
arranged so that as the phase shift of one increases, the other decreases. 
The phase shift networks are buffered to prevent loading effects. The 
relative phase can be varied over a range of about +/-120degrees, and a 
transformer and switch gives a further 0/180degree shift means any phase 
shift can be nulled. The "noise" antenna signal also passes through a 
variable attenuator, and the two signals are combined in a transformer and 
fed to the RX.

In operation, once the null is found it seems to be quite stable, and 
covers the whole amateur band without adjustment. While operating, there 
was no noticeable interaction with the TX vertical, although one of the 
loops is directly under it (the TX ant is de-tuned on receive). Without 
noise cancelling, the normal mains noise is about 10-15dB above the natural 
band noise, with peaks up to 30 - 40 dB. After nulling, the normal mains 
noise is inaudible, while the peaks can still be heard but are obviously 
much weaker. The system is still in the breadboard phase, but is certainly 
a huge improvement over the situation before - it is entertaining to switch 
the noise antenna on and off and hear the change in noise level!

Cheers, Jim Moritz
73 de M0BMU