A very interesting set of experiments Jim.
I also found that long copper rods driven into the found as deep as you
can get them worked best for 73 and 136kHz and that radials in a
suburban environment were not much good at all.
However, I did some experiments on 73kHz at Amberley museum some years
ago. This QTH was once a chalk quarry and I found it impossible to get a
ground stake earth in chalk to work. I finished up using a single
counterpoise 300m long just lying on the ground. To get it to work I
used a loading coil just like the one used for the antenna itself, which
was also about 300m long strung across the a valley created by the
mining activity. Both coils and the matching transformer were adjusted
for maximum antenna current.
Interestingly, a Tesla coil group set up a series of serious Tesla coils
at the museum for a demonstration and they had make a buried earth mat
arrangement for the large coils. The largest one, which gave spark
discharge over a metre long was driven by a RF arrangement at 70kHz. I
checked the signal using the TS850 with a short antenna in my vehicle
located around 80m distant from the coil - and heard nothing. This
surprised me - I was expecting a very large signal.
Regards
Peter, G3LDO
Over the last few days I have been doing some experiments with ground
systems with my LF/MF antenna. The purpose was partly to see whether
it is possible to reduce the overall loss resistance significantly,
but also to test out some convenient types of ground for antennas for
portable operation, in particular radial wires. The antenna is
currently a top-loaded vertical about 10m high, with about 80m in
total of horizontal wires in a narrow, asymmetrical "Y" shape about
50m long, giving a capacitance at 136k of around 500pF, and around
580pF at at 500k due to the increased effect of the distributed
inductance of the wires. As a reference, I used the normal ground
system, consisting of 6 x 1m ground rods, distributed within a couple
of metres of the base of the antenna tuners.
The measuring technique was to initially set the normal series loading
coils to resonance at 136kHz and 502khz with the normal ground system,
and then measure the antenna resistance between the "cold" end of the
loading coil and ground with an RF bridge. The total loss resistance
of the antenna (including the loading coils and the normal ground
system) was 57ohms at 136k and 24ohms at 502k. Then an alternative
ground system was connected instead of the normal ground, the bridge
re-tuned to measure the new resistance, and any change in reactance
determined from the change in resonant frequency. The reactance
measurement is only approximate, but good enough for these purposes.
The bridge equipment was all battery operated, and sat on a wooden
table, so had minimal effect on the impedance of the ground system.
When I set up a similar antenna with similar ground rods in a field a
while back, the loss resistance of the antenna was only about 8 ohms
in total, so the component of the loss resistance due to the normal
ground system is probably less than 8 ohms. The much higher total loss
resistance is due mostly to the environmental factors affecting the
antenna.
I first tried combinations of insulated radial wires laid on the
ground, including 4 x 20m radials, 4 x 40m radials and 8 x 20m
radials. It did not seem to matter much how the radials were laid out,
whether underneath the top load wires or in completely different
directions. Some had bends or doubled back - there is not really room
at my QTH for 40m radials! Some were laid on the grass, others along
the concrete driveway that runs down one side of the plot. None of
this made more than a few ohms difference to the loss resistance,
provided the wires were reasonably well spaced apart. All
configurations had significant capacitive reactance compared to the
normal ground rods:
4 x 20m radials, 136k: R = 80ohm , X = -j466ohms
4x 20m radials, 502k: R = 28ohm, X = -j113ohms
4x 40m radials, 136k: R = 66ohms, X = -j248ohms
4 x 40m radials, 502k, R = 25ohms, X = -j61ohms
8 x 20m radials, 136k: R = 62ohms, X = -j283ohms
8 x 20m radials, 502k: R = 23ohms, X = -j76ohms
So it would seem that having the total of 160m of wire as 8 shorter
radials is slightly better than 4 longer ones as far a loss goes. I
also tried elevating the radials above ground about 0.3m on lengths of
cane:
8 x 20m radials, elevated, 136k: R = 80ohms, X = -j622ohms
8 x 20m radials, elevated, 502k: R = 23ohms, X = -j158ohms
The loss resistance at 502kHz is slightly reduced, but the extra
capacitive reactance is now getting very high, especially at 136k. I
tried adding a "metallic" ground connection by adding a 200mm long
steel tent peg to the end of each radial, and driving it into the ground:
8 x 20m radial + ground spike, 136k: R = 80ohms, X = 0
8 x 20m radial + ground spike, 502k: R = 69ohms, X = -j11ohms
So this is effective at reducing the reactance, but does not help loss
resistance; in fact, it leads to a drastic increase in loss at 502k.
Instead of insulated radials, I tried a 15m x 0.6m strip of wire mesh
(a roll of chicken wire that was lying around), held down flat on the
grass with bricks:
15m x 0.6m wire mesh, 136k: R = 82ohms, X = -j102ohms
15m x 0.6m wire mesh, 502k: R = 31ohms, X = -j9ohms
So in spite of having a large area of metal in reasonable contact
with the ground, there is still relatively high resistance, and some
capacitive reactance. Since the insulated radials in general resulted
in lower loss resistance, I tried insulating the wire mesh from the
ground using polythene sheet:
15m x 0.6m wire mesh, insulated, 136k: R = 65ohms, X = -j320ohms
15m x 0.6m wire mesh, insulated, 502k: R = 26ohms, X = -j63ohms
Since the wire mesh would be quite a convenient earth system to use
for an antenna over a paved area, or on rocky ground, I tried laying
it on the concrete driveway:
15m x 0.6m wire mesh, on concrete, 136k: R = 57 ohms, X = -j320ohms
15m x 0.6m wire mesh, on concrete, 502k: R = 23 ohms, X = -j65ohms
The loss resistance is now as low as the normal ground rod system,
although the reactance is higher.
Looking at these results, the ground rods win for the fixed station
antenna. The ground here is a permanently wet clay soil, which is
probably quite good for ground rods. They give practically the lowest
loss resistance (a couple of the other ground systems give a
marginally lower loss resistance at 502k, but in practice the actual
resonant frequency was increased due to the additional reactance, and
the loss resistance of the antenna decreases at higher frequency). The
reactance is also lower. Most of the alternative ground systems
effectively behave as capacitive coupling to ground, with significant
capacitive reactance as well as some additional loss resistance. For
the fixed station, if the ground system has a high reactance, most of
the ground current will flow through the lower impedance path via TX
chassis/mains earth, instead of the ground system. This is possibly
why adding ground radials to an existing system often has no effect.
One way round this would be to have an auxilliary tuning inductor
connected between the ATU ground, and the ground radial sytem. You
could tune the inductor for maximum RF current in the radials, or
minimum current via the TX ground.
For portable antennas, it is often difficult to drive long ground rods
into the earth. It seems that short ground spikes, like the tent pegs,
or metallic conductors laid on the surface of the ground, do not make
a very good ground connection. The insulated radial wire systems
generally give the lower loss resistances for the antenna system, and
are quite easy to set up. The effect of the unwanted ground system
reactance for a portable station with no other earth connection is
that the ground terminal of the ATU is not at ground potential, and so
probably also the cases of all the station equipment. For 500kHz
stations with fairly low power (say <1A antenna current or so), where
reactance is quite low, this probably does not matter much so long as
the equipment (and operator) are reasonably well insulated from
ground. For a QRO 136kHz station, it is likely to be a different story
- with 1kW I can get about 4A antenna current - if I was using the 8 x
20m radials, the 283ohms reactance would mean the equipment ground
would have more than 1kV of RF relative to ground! The capacitance of
the insulated radials works out very roughly to 28pF per metre of
wire, and one wants to make the total capacitance large enough to keep
the voltage down to a reasonable level. More and longer radials would
clearly be desirable, which are not really feasible in my garden, but
might be quite easy at a /P location.
The insulated wire mesh ground required less area than the radial
wires, and had low loss resistance with similar reactance, so could be
quite promising, especially where the ground is already insulated; on
concrete or tarmac for example.
Hopefully some time this summer, I will be able to borrow a field
again, and do some further experiments with more space available.
Cheers, Jim Moritz
73 de M0BMU
|
