Topband: Strange behaviour of radial systems

[Kees Nijdam]

I have a north-south switchable 21m high phased  pair for 80m with 1/8 lambda spacing. It is optimised for 3520 kHz, the feed -Christman-system was made according to ON4UNs book and software. Measurements were made with a General radio bridge 1606B, oscillator and receiver. The system has a good F/B ratio in the CW dx part of the band (checked with a signal soure at several wave lenghts distance) and the SWR is almost 1, in both directions. With increasing frequency the directivity disappears and the system becomes more and more omni directional, however with acceptable SWR (maximum=2)
Vertical nr 1 was my old vertical with 50 radials (24x12, the rest between 16 and 20m).
Vertical nr 2 was made a few years ago and has 30 radials (24x12,5m and 6x20m). I stopped adding radials at the point where both verticals had the same feed point impedance.
The radials are made from isolated wire, not connected to each other and about 5 cm under the grass. The longer radials partly run along the boundary of my property.

I planned to use one of the radial systems also for an inverted L for 160m. So, I measured the feed point impedance of both verticals on 1825 kHz. Keeping all other wires in the neighbourhood far from resonance.
To my ashtonishment, vertical 1: 28-j328 Ohm and vertical 2: 12-j320 Ohm both on 1825 kHz.
The vertical with the smallest radial system has the lowest earth resistance (around 5 Ohms, estimating 7 Ohms radiation resistance for a 21m high vertical).
The oldest vertical (nr1) with 50 radials seems to have more than 20 Ohms.....

I measured the earth conductivity with the 2-conductor method and I found that the soil under vertical nr 2 has better conductivity (maximum found: 25 mS/m) compared to vertical nr
 1 (maximum: 21,5 mS/m). 

I have no explanation for this behaviour. However, a 170 year old map of my property (on a mound) shows a farm with, may be, a dung-hill at the location of vertical nr 2.

Preliminary conclusion:  do not rely on the behaviour of your 80m radial system when planning to use it for a 160m antenna.