----- Original Message -----
From: Alan Melia <[email protected]>
To: LF-Group <[email protected]>
Sent: Saturday, June 26, 2004 12:19 AM
Subject: LF: Earthing systems
Hi all, I offer below, culled from the LWCA message board, a description
of
an "improved" eathing system for LF. To me this does not make sense, but
maybe I am missing something...any thoughts anywhere?? I can see that the
small inductance of the very short (45feet..14m) "radial" could "impede"
the
passage of current collected at the remote ground spike, but I do not
really
see what is happening. We may have to wait for the article to hear the
full
measurement detail. I am worried that he has measured the soil
conductivity
at 60Hz ....this has little relation to ground loss effects at 136 to
180kHz....though maybe if it drops at 60Hz it is somewhat better at 136 ??
.....
Dear Alan, LF Group,
I tried some experiments along these lines a year or so ago, and can confirm
that the length of the wire connecting to the earth rods does indeed make a
lot of difference to the distribution of current. For a wire near ground
level connected to a ground spike at the far end, I reckoned that the
inductance was of the order of 1uH/m, and radials of the order of 10s of
metres long therefore have inductive reactance of the order of 10s of ohms -
comparable with the resistance of the ground rod - soil interface. The
result of this is that if you have numerous ground rods at different
distances from the feed point, the vast majority of the ground current
flows through the few rods nearest the feed point, since the return path for
the antenna current through the ground has lower overall impedance than the
path through the longer radials.
This is one of the reasons why adding to the ground system does not make
much difference to the loss resistance of the antenna once you get past a
certain number of ground rods/radials - the more distant ground rods carry
very little current, and so make very little difference to the current paths
and losses. "Ed G"'s capacitors are one way to even up the impedance in the
different ground paths (if you assume 1uH/m, they come out about right at
180kHz) - I tried two other ways. One was to have a central connecting point
for all the ground rods under the centre of the antenna, connected via a
"bus" wire back to the (inverted L) feed point . The 12 ground rods were in
a rough circle, so current had to flow the same length of wire to get to any
ground rod. The other way was to add inductance in series with the shorter
ground connections to give them roughly the same impedance as the long ones.
(Note that in either case, the ground connection at the antenna is somewhat
above "earth" potential at RF, so an isolating transformer or common-mode
choke in the TX feed is needed to stop most of the current flowing in the
transmitter earth connection). Both techniques were succesful in producing
roughly equal current in all the ground rods, but the bad news was that, in
spite of trying several configurations, only about 10% reduction in antenna
loss was achieved (i.e. about 0.5dB improvement). I went back to the
original ground system of 6 rods within a couple of metres of the antenna
feed, because by now the garden was festooned in wires with ground rods
poking up everywhere! Looking in the literature, similar schemes for
producing even ground current distribution used to be popular for VLF
antennas in the 1920s and 30s, some with elaborate systems of "balancing
coils" and over-ground distribution wiring, and "multiple tuned" antennas,
but these seem to have been replaced later by very extensive buried radial
systems, which gave better performance with fewer maintenance problems.
As discussed on the reflector some time ago, and based on this and other
experiments, I came to the view that for typical amateur antennas, the
majority of the loss occurs as dielectric loss in the ground due to the
electric field of the antenna, instead of in the ground return path
resistance, as seems to be true of the larger professional LF and VLF
antennas. In order to reduce this significantly, there would have to be a
large enough number of radials to form an electrostatic screen between the
antenna and the lossy soil underneath. Maybe 134 radials is enough to
achieve this, but I think they would also have to be quite long in order to
intercept most of the electric field around the antenna. For Eu amateurs, I
think a more effective approach would be to increase the amount of top
loading capacitance, so reducing antenna voltage and dielectric loss - but
of course, the Lowfer part 15 regulations place severe limits on the amount
of wire in the antenna.
Cheers, Jim Moritz
73 de M0BMU
|
